Surg Endosc (1997) 11: 902906

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The incidence of port-site metastases might be reduced

M A. Reymond,1 Ch. Wittekind,2 A. Jung,2 W. Hohenberger,1 Th. Kirchner,2 F. Ko ckerling1
1 2

Department of Surgery, University of Erlangen, P.O. Box 3560, 91023 Erlangen, Germany Department of Pathology, University of Erlangen, P.O. Box 3560, 91023 Erlangen, Germany

Received: 26 October 1996/Accepted: 14 February 1997

Abstract Background: Laparoscopic resection for cancer is controversial and port-site metastases are not infrequent. The mechanisms of occurrence of port-site metastases remain unclear. Animal experiments have suggested a role for carbon dioxide (CO2), but port-site metastases also occur after thoracoscopy, where no CO2 is used. The aim of this study was to define the role of CO2 in the seeding of tumor cells in the human patient. Methods: CO2, instruments, trocars, suction device, and peritoneal washing were examined during 12 staging laparoscopies for pancreatic cancer. The presence, viability, and biological significance of cells were investigated using conventional cytology, polymerase chain reaction (PCR), and restriction fragment length polymorphisms (RFLPs) to detect the presence of a mutant k-ras gene as a genetic marker of cancer cells. Results: Cytology exam of peritoneal washing, instruments, the suction device, and trocars revealed many cells. Tumor cells were detected in 6/12 peritoneal, in 4/12 trocars and 4/11 instruments washings, but not in 12 CO2 samples. The DNA content of CO2 was very lowas assessed by PCR. Mutant DNA was detected by RFLP in four out of 12 aerosols. Six aerosols did not contain any DNA. Two aerosols were borderline. Conclusions: During staging laparoscopy for pancreatic cancer in humans, CO2 contains only very low levels of free-floating tumor cells, even in the presence of massive peritoneal contamination. These results suggest that the incidence of port-site metastases might be reduced if mechanical contamination of the port sites with instruments or with the specimen can be avoided. Key words: Neoplasms, staging Laparoscopy, adverse effects Pancreatic neoplasms, surgery

abdominal malignancies via the laparoscope. Although such treatment is accepted for palliative operations, laparoscopy for curative resection remains controversial. Items of concern are questionable radicality, a lack of long-term results, and port-site metastases. The last-mentioned complication, occurring after a curative laparoscope-assisted colonic resection, was first reported in 1993 [1] by Alexander, and numerous case reports have since followed. The first experimental animal study on port-site metastases was published in 1995 [12]. Clinical conclusions are difficult to draw from the results obtained so far. The present study was undertaken to determine how many cells are present in the CO2 during a clinical staging laparoscopy for pancreatic cancer, whether these cells are viable, and whether they include tumor cells. The staging laparoscopy for pancreatic cancer is an excellent model for investigating these questions, because over 90% of such cancers show mutation on codon 12 or 13 of the k-ras gene [4], thus permitting detection of a small number of cancer cells. Materials and methods Tumors and cell lines
Samples were obtained from 12 patients who underwent a staging laparoscopy for pancreatic cancer at the universities of Erlangen, Wu rzburg, Regensburg, and Ulm between 15.1 and 8.5.96. Eight tumors could be resected. Biopsies were performed in cases 8, 10, 11, and 12. The diagnosis of pancreatic adenocarcinoma was based on histopathological analysis in all cases. Since neither the indication nor the procedure was modified by the analysis, no informed consent was required. The cell line of human pancreatic carcinoma (PC44) was a kind gift from H. Heinmo ller from the University of Regensburg; it showed a mutation on codon 12. K-ras negative control cell lines (T-ALL Jurkat) were obtained from our laboratory.

Sampling Recent technological advances now allow surgeons to treat

Correspondence to: F. Ko ckerling Twenty milliliters of peritoneal washing was sampled at the beginning of the procedure. The CO2 was delivered during all the duration of the procedure through an artificial leak into a closed Bu lau bottle, which was filled with 400 ml phosphate-buffered saline (PBS). The CO2 was allowed to

903 diffuse through the medium, so all its contents were captured in the bottle; 89.3 3.8% of CO2 was collected. The pH was maintained around 7.4 using titration with NaOH. By the end of the procedure, trocars and instruments were rinsed separately with 50 ml PBS per sample. The procedure was followed by immediate centrifugation of the samples at 400g for 10 min at 4C. The collecting system was washed and sterilized between the procedures according to surgical standards. After three cases, the analysis of a standard laparoscopic cholecystectomy excluded a possible contamination.

Cytology
After vital staining with trypan blue, the viability of the cells was assessed by conventional microscopy. An aliquot of 1.5 ml was centrifugated at 400g for 5 min at room temperature. After Giemsa staining, the cytology samples were analyzed by a senior pathologist. Cell counting was performed after fixation and adapted to the initial sample volume. Fig. 1. Polymerase chain reaction (35 cycles) of the k-ras gene, with serial dilutions in phosphate-buffered saline solution (PBS), showing a sensitivity of less than 10 cells.

Statistical analysis DNA preparation

The cell pellet was resuspended into 1 ml PBS and centrifugated at 400g for 5 min at room temperature. Then the pellet was resuspended into 380 l Higushi buffer (50 mM KCl, 20 mM Tris-HCl, pH 8.3, 2.5 mM MgCl2, 0.1 mg/ml gelatine, 0.45% (v/v) NP-40, 0.45% [v/v] Tween 20), and 20 l proteinase K (Boehringer, Mannheim, Germany), and incubated at 56C overnight with continuous shaking at 1,200 rpm. After denaturation at 95C for 5 min, the DNA was stored at 20C until analysis. Because of the presence of inhibitors the samples had to be purified using a QIAamp DNA purification kit (Qiagen, Hilden, Germany) according to the manufacturers protocol. To rule out the null hypothesis that the cellular contamination was equal in all samples, the Kruskal-Wallis test for equality of populations was used. Data were not normally distributed. Tests were two-tailed.

Results Cellular content of CO2 Figure 2 shows the characteristics of the various items. The peritoneal washing sample was the most cellular (5.3 105 4.1 105 cells), although the suction device (7.7 104 1.2 105), instruments (4.9 104 4.6 104), and trocars (2.9 104 1.8 104) also bore many cells. The CO2 contained almost no nucleated cells (3.4 12.4) but did contain carbonized material and some cell fragments. The difference between the cell contents of the different items was statistically highly significant (p 0.0001).

K-ras PCR
A nonradioactive method was used, involving PCR amplification of K-ras first exon sequences (157 pairs of bases) [19]. Oligonucleotide primers were as follows: K-ras 5: ACT GAA TAT AAA CTT GTG GTA GTT GGA CCT K-ras 3: TCA AAG AAT GGT CCT GGA CC Primers were purchased from MWG Biotech (Ebersberg, Germany); 1 l of high molecular weight DNA was amplified in a volume of 50 l containing 1 PCR reaction buffer, 4 l MgCl2 (25 mM), 1 l primers (50 pmol/l), 1 l NTP (10 mM), and 0.2 l 1U TaqDNA polymerase (Perkin Elmer). The reaction was overlaid with 50 l mineral oil (Sigma Chemical Co., St. Louis, MO). DNA was amplified with a Hybaid Omnigene 3 Thermal Cycler using following parameters: 3 cycles with 95C, 0.5 min/42C, 0.5 min/72C, 1 min 35 cycles with 95C, 0.5 min/55C, 0.5 min/72C, 1 min The PCR products were run on 2% (w/v) 0.5 TBE ethidium-bromidecontaining (1 g/ml) agarose gels, and the result was photographed on an ultraviolet transilluminator. The PCR allowed us to detect less than 10 cells in PBS, as assessed with the mutant cell line (Fig. 1). The 40-cyclessecond-PCR was performed with the PCR products of a first 40-cycles PCR, after 1:1,000 dilution. The negative controls remained negative.

Cell viability The viability of the cells found on the instruments and trocars was estimated to be over 90% with trypan blue and conventional microscopy. No cells could be seen in CO2 samples.

Biological significance of the cells Table 1 shows the results of conventional cytology for the various items: no tumor cell was detected in the retrieved CO2 samples, but many suspect cells were detected on the instruments (Fig. 3) and trocars (Fig. 4). A more precise analysis was enabled by molecular biological methods. It was possible to amplify the k-ras gene in six out of 12 aerosols (data not shown). An RFLP analysis of four of the six positive PCR products was possible and showed the presence of tumor cells (Fig. 5). The remaining two aerosols contained insufficient DNA for RFLP analysis. A second PCR excluded the presence of DNA in the other six aerosols, and thus the presence of tumor cells. Table 2 shows the results of the gene analysis for the various items.

Restriction enzyme analysis

The PCR was followed by restriction to detect either endogenous restriction polymorphisms, which can be found in the PC44 pancreas carcinoma cell line. For the codon 12, the wild gene was digested by MvaI (Boehringer Mannheim) into a main fragment of 114 pb; the mutant gene was digested into a 143-pb main fragment. For codon 13, the mutant gene was cut by HphI (New England Biolabs) at a length of 114 pb; the wild gene was not restricted. The digestions were incubated at 37C for 3 h, following the instructions of the supplier. An electrophoresis using 3% (w/v) 0.5 TBE Metaphore gel containing 1 g/ml ethidium bromide was then performed, and the result was photographed on an ultraviolet transilluminator.

904

Fig. 2. Cell content of the various items, showing significant differences between the subgroups (Kruskal-Wallis test for equality of populations, p < 0.0001).

Table 1. Results of the cytologya Patient 1 2 3 4 5 6 7 8 9 10 11 12

a

Peritoneal washing N S T S T N N T N T N N

Instruments N T S N N S N N N S N

Trocars N S S S N N N S N T N N

Suction device N S N N S N N

CO2 No cells No cells Normal No cells No cells Normal No cells No cells Normal No cells No cells No cells

Fig. 3. Patient 2: Typical cytology of the instruments, showing many suspect cells (Giemsa staining, magnification 200). Fig. 4. Patient 8: Cluster of suspect cells found on a trocar (Giemsa staining, magnification 200).

N, normal cells; S, suspect cells; T, tumor cells; , no sample.

Discussion Our results show that during a staging laparoscopy for pancreatic cancera procedure for which port-site metastases have been described [21]the risk of seeding tumor cells through CO2 remains very low. On the other hand, a high cellular contamination of the instruments and the trocars including sometimes tumor cellsis documented. These results are consistent with the clinical data available. Some experienced surgeons have performed hundred of colectomies without any port-site metastases [12] being seen. Beside two other publications [8, 13], Downey recently reported on a series of 21 port-site metastases after thoracoscopy where no CO2 is usually used [11]. Such troubling complications have also been reported after mediastinoscopy [26]. Incision site metastases occurring after fine needle biopsies are well known [9, 24, 27]. Since 1995, many experimental studies on laparoscopy and cancer have appeared. Compared with control animals (anesthesia alone), tumor implantation at trocar sites is enhanced by CO2 in the hamster [20] and the rat [7, 10, 16, 18] when cells are injected intraperitoneally, and there is a definite dose-response relationship between tumor implantation and the number of cells in the inoculum [20]. This does not

happen when a solid [6] or retroperitoneal [23] tumor model is used, which is similar to the clinical situation in early tumor stages. Interestingly, laparoscopy involving the use of air enhances intraperitoneal tumor growth more markedly than CO2 in laparoscopy [17]. The influence of helium in laparoscopy is presently being debated [10, 18]. These results suggest both a gas-independent mechanism and an inhibitory effect of CO2 at high concentrations, and this deserves further investigation. After laparotomy, intraabdominal recurrences occur preferentially in the laparotomy wound or at the anastomosis [25]. Intraabdominal recurrences after laparoscopy occur only at serosal lesions (port sites [16, 20] or liver surface [20]), which underscores the importance of an intact peritoneal barrier for preventing the implantation of free cancer cells in the peritoneal cavity [14]. Although CO2 laparoscopy enhances tumor implantation when compared with anesthesia alone, it is associated with less experimental tumor growth than laparotomy when cancer cells are injected under the renal capsule [7], or into the pancreas after manipulation of the tumor [23], or subcutaneously [3, 5] or even intraperitoneally [17]. Using a radioactive cell line in pigs, Allardyce has recently shown that although CO2 may increase wound-site

905

and relatively gas independent. Although tumor growth is enhanced through CO2 laparoscopy when compared with anesthesia alone, there is less tumor growth after CO2 laparoscopy than after conventional laparotomy, as the results obtained in five independent animal studies [3, 5, 7, 17, 23] have shown. If this finding is confirmed in the clinical setting, it might be an argument for using laparoscopy in curative surgery in selected indications. In summary, only very low levels of free-floating tumor cells can be detected in CO2 during staging laparoscopies for pancreatic cancer using CO2 at pressures between 10 and 15 mmHg. Mechanical inoculation of the abdominal wall by the surgical specimen or an instrument is significantly more likely to occur, since they carry tumor cells. This must be the case in thoracoscopy and mediastinoscopy, where no CO2 is used. Proper prevention measures including a good surgical technique and proper indications [12, 22]could reduce the incidence of port-site metastases.
Acknowledgments. This study was funded by the universities of Geneva and Erlangen. We gratefully acknowledge the help of K. Fuchs (Wu rzburg), R. Kunz (Ulm), and K. Jauch (Regensburg) for the recruitment of patients and the technical help of S. Kastl, J. Riese, and T. Brabletz.

Fig. 5. Four aerosols contained mutant DNA; two aerosols were borderline; six aerosols did not contain any DNA. U uncut PCR product; 12 restricted with MvAI (codon 12); 13 restricted with HphI (codon 13).

References
1. Alexander RJ, Jaques BC, Mitchell KG (1993) Laparoscopically assisted colectomy and wound recurrence [letter; comment]. Lancet 341: 249250 2. Allardyce R, Morreau P, Bagshaw P (1996) Tumor cell distribution following laparoscopic colectomy in a porcine model. Dis Colon Rectum 39: S47S52 3. Allendorf JD, Bessler M, Kayton ML, et al. (1995) Increased tumor establishment and growth after laparotomy vs laparoscopy in a murine model. Arch Surg 130: 649653 4. Bos JL (1989) Detection of ras oncogenes using PCR. In: Erlich H (ed) DNA amplification using the PCR. Stockton Press, New York, pp 225233 5. Bouvy ND, Marquet RL, Hamming JF, Jeekel J, Bonjer HJ (1996) Laparoscopic surgery in the rat. Beneficial effect on body weight and tumor take. Surg Endosc 10: 490494 6. Bouvy ND, Marquet RL, Jeekel H, Bonjer HJ (1996) Impact of gas(less) laparoscopy and laparotomy on peritoneal tumour growth and abdominal wall metastases. Surg Endosc 10: 1618 7. Bouvy ND, Marquet RL, Lambert SWJ, Jeekel J, Bonjer HJ (1996) Laparoscopic bowel resection in the rat: earlier restoration of IGF-1 and less tumor growth. Surg Endosc 10: 567 (Abstract) 8. Collard JM, Reymond MA (1996) Video-assisted thoracic surgery (V.A.T.S.) for cancer: risk of parietal seeding and of early local recurrence. Int Surg 81: 343346 9. Dick R, Heard BE, Hinson KEW (1974) Aspiration needle biopsy of thoracic lesions: an assessment of 227 biopsies. Br J Dis Chest 68: 8693 10. Dorrance HR, Oein K, ODwyer PJ (1996) Laparoscopy promotes tumour growth in an animal model. Surg Endosc 10: 559 (Abstract) 11. Downey RJ, McCormack P, LoCicero III J, and the Video-Assisted Thoracic Surgery Study Group (1996) Dissemination of malignancies following video-assisted thorac surgery. J Cardiovasc Thorac Surg 111: 954960 12. Franklin ME, Rosenthal D, Abrego-Medina D, et al. (1996) Prospective comparison of open vs. laparoscopic colon surgery for carcinoma. Dis Colon Rectum 39: S35S46 13. Fry WA, Sidiqqui A, Pensler JM (1995) Thoracoscopic implantation of cancer with a fatal outcome. Ann Thorac Surg 59: 4245 14. Goldstein DS, Lu ML, Hattori T, Ratliff TL, Loughlin KR, Kavoussi LR (1993) Inhibition of peritoneal tumor-cell implantation: model for laparoscopic cancer surgery. J Endourol 7: 237241 15. Hewett PJ, Thomas WM, King G, Eaton M (1996) Intraabdominal cell

Table 2. Contents of CO2: amplification and restriction fragment length polymorphisms (RFLP) Patient 1 2 3 4 5 6 7 8 9 10 11 12 PCR (k-ras gene) Present Present Present Absent Absent Present Present Absent Present Absent Absent Absent RFLPa (codons 12 & 13) Mutant Mutant Mutant Mutant

a no reaction possible in the absence of DNA; mutant presence of tumor cells.

implantation, the major variable influencing tumor cell deposition is whether or not the port is used by the surgeon [2]. Using filters, Hewett demonstrated that the movement of cells throughout the peritoneal cavity during laparoscopy is via contaminated instruments, with local contamination of the port by dispersion within water vapor remaining a possibility [15]. Using in vitro and in vivo models, Whelan showed that trocar-site recurrence is unlikely to result from aerosolization of tumor cells [28]. To explain the experimental findings, pressure gradient must be assumed to play a limited role in tumor dissemination, but only when tumor cells are already present in the peritoneal cavity. This effect seems to be dose dependent

906 movement during abdominal carbon dioxide insufflation and laparoscopy. Dis Colon Rectum 39: S62S66 Hubens G, Pauwels M, Hubens A, Vermeulen P, Van Marck E, Eyskens E (1996) The influence of pneumoperitoneum on the peritoneal implantation of free intraperitoneal colon cancer cells. Surg Endosc 10: 181 (Abstract) Jacobi CA, Ordermann J, Bo hm B, Zieren HU, Volk HD, Mu ller JM (1996) Increased tumor growth after laparotomy and laparoscopy with air versus CO2. Surg Endosc 10: 551 (Abstract) Jacobi CA, Sabat R, Bo hm B, Zieren HU, Volk HD, Mu ller JM (1996) Pneumoperitoneum with CO2 stimulates malignant tumor growth. Surg Endosc 10: 551 Jiang W, Kahn SM, Guillem JG, Lu SH, Weinstein IB (1989) Rapid detection of ras oncogenes in human tumors; applications to colon, esophageal, and gastric cancer. Oncogene 4: 923928 Jones DB, Guo LW, Reinhard MK, et al. (1995) Impact of pneumoperitoneum on trocar site implantation of colon cancer in hamster model. Dis Colon Rectum 38: 11821188 Jorgensen JO, McCall JL, Morris DL (1995) Port site seeding after laparoscopic ultrasonographic staging of pancreatic carcinoma [letter]. Surgery 117: 118119 22. Ko ckerling F, Reymond MA, Schneider C, Hohenberger W (1997) Mistakes and hazards in oncological laparoscopic surgery. Chirurg 68: 215224 23. Mutter D, Hajri A, Tassetti C, Solis-Caxaj C, Aprahamian M, Marescaux J (1996) Experimental pancreatic tumor growth and spread after laparoscopy versus laparotomy in the rat. Surg Endosc 10: 490494 (Abstract) 24. Nankhonya JM, Zakhour HD (1991) Malignant seeding of needle aspiration tract: a rare complication. Br J Dermatol 1243: 285286 25. Skipper D, Jeffrey M, Cooper AJ, Alexander P, Taylor I (1989) Enhanced growth of tumour cells in healing colonic anastomosis and laparotomy wounds. Int J Colorectal Dis 4: 172177 26. Sullivan WD, Passamonte PM (1982) Mediastinoscopy incision site metastasis: response to radiation therapy. South Med J 72: 1428 27. Voravud N, Shin DM, Dekmezian RH, Dimery I, Lee JS, Hong TW (1992) Implantation metastasis of carcinoma after percutaneous fineneedle aspiration biopsy. Chest 43: 15331540 28. Whelan RL, Sellers GJ, Allendorf JD, et al. (1996) Trocar site recurrence is unlikely to result from aerosolization of tumor cells. Dis Colon Rectum 39: S7S13

16.

17. 18. 19. 20. 21.

Surg Endosc (1997) 11: 899901

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The phagocytosis activity during conventional and laparoscopic operations in the rat
A preliminary study
C. N. Gutt, P. Heinz, W. Kaps, V. Paolucci
Department of General Surgery, Johann-Wolfgang-Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany Received: 8 December 1996/Accepted: 20 February 1997

Abstract Background: Numerous experimental and clinical investigations indicate that the mononuclear phagocyte system (MPS) has a relevant function in terms of physiological defense against tumor metastasis and bacterial infection. Consequently, a point of major interest is the influence of surgical techniques on the MPS function. Method: The model investigation examines the phagocytosis activity of the rats MPS during conventional fundoplication (group 1, n 10), laparoscopic fundoplication using a pneumoperitoneum (group 2, n 10), and gasless laparoscopic fundoplication (group 3, n 10). The MPS function is evaluated by an intravascular carbon clearance test (G. Biozzi). Results: The fastest carbon elimination half-life was found in group 3. By way of contrast, there was a significant increase of carbon half-life in group 2 (p < 0.005). Even group 1 caused less MPS depression (p < 0.1) than group 2. Conclusion: Gasless laparoscopic procedures have a favorable effect on phagocytosis activity. The CO2 pneumoperitoneum seems to be the main reason for a decreased antigen elimination in laparoscopic treatments. Key words: Laparoscopy Pneumoperitoneum Gasless laparoscopy Mononuclear phagocytes system Rat model

erative infection and metastatic tumor spreading [6, 9, 10, 11, 13]. The investigation of the mononuclear phagocyte system (MPS)s phagocytosis activity is determined by means of the Carbon-Clearance test which, itself, is carried out in a modified form, i.e., following the method as described by Biozzi et al. [3] and Lemperle et al. [7]. In the course of this test, the degree to which carbon particles are eliminated from the circulating blood by local macrophages of the MPS is determined. This phagocytosis activity depends on the kind and extent of an operative manipulation and can be judged to be a measure of interference with the organism caused by surgical interventions. Our investigation compares the phagocytosis activity during conventional operations (group 1) with that of laparoscopic surgery using a pneumoperitoneum (group 2) and gasless laparoscopic surgery without the use of a pneumoperitoneum (group 3) (Table 1). Method Laparoscopy in the rat
A surgical table for small animals (KAPS Company, Asslar, Germany) is used consisting of an operation platform and a base plate with a flexible supporting arm. In a supinated position, the narcotized animal lies on the raised operation platform which is set over the base plate. By slightly moving the operation platform via two micrometer screws with fixed laparoscope, the surgeon or his assistant can finely adjust the picture segment. The surgical instruments consist of a 4-mm arthroscope and micro-alligator forceps and microscissors (Aesculap, Tuttlingen, Germany). All instruments, as well as the suture material, can easily be inserted into the abdomen via 3-mm synthetic ports [2, 5].

The human and animal organisms answer to an injury is characterized by an early rise of the serum level for stress hormones and a drop in the immunoresponse conveyed by the cells. The stress reactions force and duration are considered to be proportional to the severity of the injury. There is a correlation between a reduced perioperative cellconveyed immunoresponse and an increased risk of postopCorrespondence to: C. N. Gutt

Experimental procedure
Before the operation starts, the rats are injected via the v. dorsalis penis with 0.1 ml/100 kg of fount india shellac-free ink (Pelikan Company,

900 Table 1. Nissen fundoplication was performed in three animal groups Group 1 Conventional laparotomy Group 2 Laparoscopy with pneumoperitoneum Group 3 Gasless laparoscopy (n 10) (n 10) (n 10)

Hannover, Germany) which, in order to obtain a better colloid stabilization, is mixed with gelatine. Periodically, 200 l blood is taken from the animals. After the hemolysis in 2 ml of a 0.1% Na2 CO3 solution, the mixtures extinction at a wavelength of 640 nm is determined as a blank value against Na2 CO3/blood (before ink injection). The extinction depends on the ink particles quantity (particle size approx. 250 )/volume and consequently represents a measure for the ink concentration in the peripheral blood. Throughout that intervention period, the inks half-life in the animals circulating blood can be determined by exploratory blood takings. The bigger the half-life, the longer the ink remains in the blood and the lower is the MPSs phagocytosis activity under the corresponding operation conditions. At first, the rats of the first group of animals (group 1, n 10) are subject to a conventional laparotomy; then a fundoplication after Nissen is made. Throughout the operation, and in defined intervals (3, 15, 30, and 40 min after the ink injection), 200 l of blood is taken from the animals retrobulbar plexus venosi and analyzed. In the second group of animals (group 2, n 10), the fundoplication is carried out under an endoscopic operation with a pneumoperitoneum of 7 mmHg. The rats belonging to the third group of animals (group 3, n 10), as well, are subject to an endoscopic fundoplication; however, a pneumoperitoneum is renounced and, instead, the abdomen is stretched out by a specially constructed wire bow. Still narcotized, all animals are killed by carbon monoxide.

Fig. 1. Average carbon elimination half-life of the experimental groups, graphically described.

Statistics
The mean carbon half-life and the standard deviation (SD) of the collected data were recorded. Carbon half-life was calculated by using simple linear regression (Pearson-regression). To calculate p values a Kruskal-Wallis test was used.

Results The lowest elimination half-life of ink particles in the laboratory animals bloodstream (t1/2 12.86 min) could be ascertained in group 3 (gasless laparoscopic). Compared to this group, the elimination half-life is significantly increased (t1/2 21.91 min; p value < 0.005) in group 2 (laparoscopic with pneumoperitoneum). In group 1 (conventional laparotomy) the ink particles elimination half-life ranged below the one of group 2 (t1/2 16.1 min; p value < 0.1) (Fig. 1).

Discussion Repeatedly, literature has described the phenomenon of the implantation of tumor cells during laparoscopies [4, 10, 13]. Why this problem occurs and how often are still unclear. Yet there are hypotheses according to which suboptimal technique and instrumentation are said to play a decisive part in that matter [1, 12]. Under certain conditions, even the pneumoperitoneum seems to support the intraperitoneal dissemination of tumor cells [8]. The mononuclear phagocyte system (MPS) represents an important component of the endogenous defense against

circulating tumor cells. A decreased perioperative, cellconveyed immunoresponse correlates not only with an increased metastatic tumor spreading risk but also with postoperative infections; the phagocytotic activity of the MPS consequently seems to be of significance when judging the etiology of implantation metastases during laparoscopic interventions. The Carbon-Clearance test is a well-established method of measuring the phagocytosic activity of the mononuclear phagocytes system (MPS). In conformity with the method described by Biozzi et al. [3] and Lemperle et al. [7], it is applied in a modified form which offers the advantage of measuring in vivo whole the extent of a hematogen antigen elimination. Since the phagocytosis activity depends on the type and extent of an operation it, therefore, can be considered a measure of the disturbance of the organism by any such interventions. Therefore, a highly standardized proceeding, as well as the choice of an anesthesia method not influencing the test procedure, is a basic condition. From a technical point of view, the rat fundoplication is a simple operation which, without any problems, can be carried out both conventionally and laparoscopically [5]. As shown by the results of this investigation, the fastest way to eliminate carbon particles from the bloodstream by local macrophages of the MPS is via gasless laparoscopic (group 3). Compared with the conventional method (group 1), the minor disturbance of the organism seems to be the minimally invasive access. The use of a pneumoperitoneum (group 2) has an even stronger influence on the antigen elimination from the circulating blood than the choice of surgical access. Although, in case of the rat, with a pneumoperitoneum of 7 mmHg there is no long-term decrease in the cardiac output, the intraperitoneal pressure seems to cause a reduced circulation speed of carbon particles in the blood. Further impacts that may prejudice the organism, apart from the circulatory influences of the CO2 pneumoperitoneum, are not excluded. The results of the present investigation show that a pneumoperitoneum in the rat is an invasive arrangement leading to a significant reduction of the phagocytosis activity of the MPS.

901

References
1. Allendorf JDF, Bessler M, Kayton ML, Whelan R, Treat MR, Nowygrod R (1995) Tumor growth after laparotomy or laparoscopy. Surg Endosc 9: 4952 2. Berguer R, Gutt CN (1994) Laparoscopic colon surgery in a rat model. A preliminary report. Surg Endosc 8: 11951197 3. Biozzi G, Benacerraf B, Halpern BN (1953) Quantitative study of the granulopectic activity of the reticulo-endothelial system. II. Br J Exp Pathol 34: 441457 4. Childers JM, Aqua KA, Surwit EA, Hallum AV, Hatch KD (1994) Abdominal-wall tumor implantation after laparoscopy for malignant conditions. Obstet Gynecol 5: 765769 5. Gutt CN, Berguer R, Stiegmann G (1993) Laparoskopie an der Ratte: Beschreibung einer neuen Technik. Zentralbl Chir 118: 631634 6. Keller SE, Weiss JM, Schleifer SJ, Miller NE, Stein M (1983) Stressinduced suppression of immunity in adrenalectomized rats. Science 221: 13011304 7. Lemperle G (1972) Der Funktionszustand des retikuloendothelialen Systems bei chirurgischen Erkrankungen. Habilitationsschrift Frankfurt/M1517, 3842

8. Mouiel J, Gugenheim J, Toouli J, Crafa F, Cursio R, Chastanet S (1995) Port-site recurrence of cancer associated with laparoscopic diagnosis and resection: the European experience. Semin Laparosc Surg 2(3): 167175 9. Pollock RE, Lotzova E (1987) Surgical-stress-related suppression of natural killer cell activity: a possible role in tumor metastasis. Nat Immunol Cell Growth Regul 6: 269278 10. Ramos JM, Gupta S, Anthone GJ, Ortega AE, Simons AJ, Beart RW Jr (1994) Laparoscopy and colon cancer. Is the port site at risk? A preliminary report. Arch Surg 9: 897899 11. Saba TM, Antikatzides TG (1976) Decreased resistance to intravenous tumor cell challenge during periods of reticuloendothelial depression following surgery. Br J Cancer 34: 381386 12. Treat RM, Bessler M, Whelan RL (1995) Mechanisms to reduce incidence of tumor implantation during minimal access procedures for colon cancer. Semin Laparosc Surg 2(3): 176178 13. Walsh DC, Wattchow DA, Wilson TG (1993) Subcutaneous metastases after laparoscopic resection of malignancy. Aust N Z J Surg 7: 563565

Surg Endosc (1997) 11: 967968

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Cetrimide overdose
The intraperitoneal use of cetrimide 1% [6] has recently been advocated as an adjunct to the treatment of hydatid disease of the liver [4]. We would like to report the case of a 2-year-old boy who died following such a treatment regime. A 2-year-old Moroccan boy was admitted with general malaise and fever. Clinical examination revealed a mass in the liver. A diagnosis of chronic granulomatous disease was made but concomitant hydatid disease of the liver could not be excluded. We decided to puncture the mass in the liver under laparoscopic control, after immersing the liver in cetrimide 1%, a scolicidal agent, as has been advocated by Khoury et al. [4]. General anesthesia with endotracheal intubation was induced with intravenous thiopentone, sufentanyl, and vecuronium, and it was maintained with halothane 0.61% in oxygen-enriched air and with additional doses of sufentanyl and vecuronium. Monitoring included ECG, pulse oximetry, nasal temperature, intra-arterial and central venous pressure, and end-tidal CO2. A 5-mm trocar was inserted through the infraumbilical fold using an open technique. Pneumoperitoneum was induced with CO2 at an initial pressure of 5 mmHg and later at 8 mmHg (initial flow rate 2 l/min, later 5 l/min). No surface abnormalities of the liver were seen. The patient was positioned head down with a three-quarter right lateral tilt. The epigastrium was then punctured with a 14-gauge Abbocath and cetrimide 1% was instilled until the liver was completely submerged, which required 1 l of the solution. No fluid was obtained on ultrasound-guided laparoscopic puncture of the mass in the liver. After about 1 h, the instilled cetrimide solution was aspirated and 800 ml was easily retrieved. Before removing the laparoscope and closing the trocar wound, the peritoneal cavity was rinsed three times with 1 l of normal saline. The patient remained stable during the procedure, which lasted 134 h. No unexpected physiological changes were noted despite extensive monitoring. However, after withdrawal of halothane, the patient failed to regain consciousness, and it was noted that the pupils were dilated and did not respond to light. The patient was apneic with a general flaccid paralysis. Although supplementary doses of sufentanyl and vecuronium had not been given for more than 1 h, neostigmine, atropine, and naloxone were administered. However, this did not result in any improvement in the clinical condition of the child. Arterial blood analysis revealed a mild respiratory and a severe metabolic acidosis (pH 7.05, pCO2 53 mmHg, base excess 15 mmol/l, pO2 157 mmHg, with an FiO2 of 40%). Despite the adjustment of the ventilator and the administration of sodium bicarbonate, the metabolic acidosis continued to increase over the ensuing hours. As methaemoglobinemia formation is a known complication of the use of cetrimide, a blood sample was taken to measure this and 15 mg methylene blue was given intravenously [1, 4]. The methemoglobin level, however, later proved to be within normal limits (less than 0.6%). The laboratory detected mild hemolysis, but neither macroscopic hematureia nor hemoglobinuria was observed. Both may be associated with absorption of cetrimide [2, 3]. As the patient remained unconscious with a flaccid paralysis and fixed dilated pupils, he was transferred to the intensive care unit, where, 90 min after the end of surgery, he became cardiovascularly unstable and subsequently died 1 h later, despite vigorous resuscitation measures. Postmortem examination failed to reveal a cause for the patients death. A number of complications have been described in association with internal use of cetrimide, including coma [3], fixed dilated pupils [2, 3], flaccid paralysis [2, 3], metabolic acidosis [2, 5], cardiovascular collapse [2, 5], cardiac arrest [2], methemoglobinemia [1, 4], hemolysis [2], hemoglobinuria, and hematuria [2]. None of these reported complications, however, has been lethal or has resulted in long-term morbidity. We were not able to confirm cetrimide absorption by plasma bromide analysis, as has been suggested [5]. As bromide ions (Br) are distributed over 25% of the body weight (i.e., 4 l in our patient), absorption of 100 ml cetrimide 1% (i.e., 60 mg/kg in our patient, at least twice the reported lethal dose [3]) will elevate plasma Br to a maximum value of 0.75 mmol/l. This is below the detection limit of the commonly used gold-chloride-analysis technique for bromide. As a result of our experience, we consider that cetrimide is a dangerous and potentially lethal agent for internal application. It should certainly not be used intraperitoneally as has been recently advocated [4]. Even treatment of the hydatid cyst cavity with cetramide may result in life-threatening complications [3]. We conclude that the use of cetrimide should be restricted to external application only. References
Correspondence to: N. M. A. Bax 1. Baraka A, Yamut F, Walid N (1980) Cetrimide induced methaemoglobinaemia after surgical excision of hydatid cyst. Lancet II: 8889

968 2. Gode GR, Jayalakshmi TS, Kalla GN (1975) Accidental intravenous injection of cetrimide. Anaesthesia 30: 508510 3. Klouche K, Charlotte N, Kaaki M, Be raud JJ (1994) Coma and hemolysis after cetrimide washout of epidural hydatid cyst. Intensive Care Med 20: 613 4. Khoury G, Jabbour-Khoury A, Bikhazi K (1996) Results of laparoscopic treatment of hydatid cysts of the liver. Surg Endosc 10: 5759 5. Momblano P, Pradere B, Jarrige N, Concina D, Bloom E (1984) Metabolic acidosis induced by cetrimonium bromide. Lancet II: 1045 6. Reynolds JEF (1993) Martindale The extra pharmacopoeia. 30th ed. The Pharmaceutical Press, London, p. 787
1

Department of Pediatric Surgery University Childrens Hospital Wilhelmina P.O. Box 18009 3501 A Utrecht The Netherlands Department of Anesthesiology University Childrens Hospital Wilhelmina P.O. 18009 3501 CA Utrecht The Netherlands Department of Pharmacy University Childrens Hospital Wilhelmina P.O. 18009 3501 CA Utrecht The Netherlands Department of Intensive Care University Childrens Hospital Wilhelmina P.O. 18009 3501 CA Utrecht The Netherlands

N. M. A. Bax1 D. C. van der Zee1 N. M. Turner2 C. M. A. Rademaker3 R. J. B. J. Gemke4

Surg Endosc (1997) 11: 923927

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Acute appendicitisa clear-cut case in men, a guessing game in young women

A prospective study on the role of laparoscopy
P. J. Borgstein, R. V. Gordijn, Q. A. J. Eijsbouts, M. A. Cuesta
Department of Surgery, Academic Hospital, Vrije Universiteit, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands Received: 18 October 1996/Accepted: 2 April 1997

Abstract Background: The aggressive surgical approach to patients suspected of having acute appendicitis for fear of perforation, and the inaccuracy of available diagnostic methods lead to an unacceptably high negative appendicectomy rate, especially in young women, in whom gynecological disorders frequently mimic appendicitis. Our objectives were to determine the value of diagnostic laparoscopy in women of child-bearing age to reduce the number of negative laparotomies and establish the correct diagnosis to allow prompt and appropriate treatment. Methods: 161 consecutive adult female patients under 50 years of age with a clinical diagnosis of acute appendicitis underwent diagnostic laparoscopy prior to the planned appendicectomy. If an inflamed appendix was found, appendicectomy was usually done through a muscle-splitting McBurney incision. Other diagnoses were treated accordingly. A normal appendix was not removed. Results were compared to a group of 42 similar patients in whom the laparoscopy was omitted for various reasons, to 23 postmenopausal women, and to all 137 male adults, directly operated by the McBurney approach. Results: After laparoscopy, 55% of the patients required appendicectomy for appendicitis while in 23% a gynecological diagnosis was made in spite of previous examination by a gynecologist. Fourteen percent had a negative laparoscopy. There were no false-negative results. The negative appendicectomy rate after laparoscopy was 5% due to two false positives and eight laparoscopy failures. In the group of fertile females who escaped laparoscopy the negative appendicectomy rate was 38%. The respective rates for postmenopausal women and men were 4% and 8%. Conclusions: All women of child-bearing age suspected of

having acute appendicitis should undergo diagnostic laparoscopy prior to the planned appendicectomy, regardless of the certainty of the preoperative diagnosis. This is currently the only way to reduce the negative appendicectomy rate and establish a correct diagnosis allowing prompt and appropriate treatment. In male patients and postmenopausal women one may proceed directly to emergency appendicectomy. Key words: Acute appendicitis Diagnostic laparoscopy

Presented at the 4th International Congress of the European Association for Endoscopic Surgery, Trondheim, Norway, June 1996 Correspondence to: M. A. Cuesta

Acute appendicitis is the most common surgical emergency in the western world and demands accurate early diagnosis to prevent progression to perforation, with its associated high morbidity and mortality, and which can result in late complications such as adhesions and infertility in young women [16, 23]. The widely accepted surgical approach, appendicectomy through a muscle-splitting McBurney incision, leads to a high rate of normal appendices being removed under the suspicion of an acute appendicitis. The reported negative appendicectomy rate for men varies from 7% to 15%, whereas that for women of child-bearing age lies between 22% and 47% [4, 6]. It has long been obvious that more accurate methods of diagnosing acute appendicitis, besides clinical suspicion, are required to prevent potentially harmful, unnecessary surgery especially in young women. Several methods have been described to reduce the diagnostic error rate of acute appendicitis [13], but only three have been shown to be of value in clinical practicenamely, ultrasonography, computer-assisted diagnosis, and diagnostic laparoscopy. Ultrasonography has been used since 1986 in diagnosing acute appendicitis and has been shown, in large series, to have a sensitivity of 75% to 89% and a specificity of 86% to 100% [20]. Advantages are the noninvasive technique and the possibility of diagnosing alternative abdominal con-

924

ditions such as mesenteric adenitis, terminal ileitis, and some gynecological disorders. The main disadvantage is that ultrasonography is strongly operator-dependent. An experienced radiologist must be available 24 h a day to achieve a clinically useful sensitivity [1] (option not available in our hospital during the study). Structured data collected from history-taking and physical examination can be entered into previously programmed computers to produce a diagnostic probability for acute appendicitis or nonspecific abdominal pain [8, 10]. Although early studies reported a decline in the negative appendicectomy rate, initial enthusiasm has been dampened by various limitations and programming difficulties. Laparoscopy has long been used by gynecologists to diagnose lower abdominal pain, but prior to 1990 it did not achieve widespread attention as a diagnostic aid in acute appendicitis because the failure rates were unacceptably high (up to 28%) due to difficulties in visualizing the appendix [3, 22]. The lack of specially designed equipment and general expertise unfortunately made this diagnostic laparoscopy more a theoretical tool used by a few, without any practical adoption in most surgical departments. Since the era of laparoscopic surgery, however, the significant increase in experience with laparoscopic techniques has led to several reports on the value of laparoscopy as a method of reducing the negative laparotomy rate in women of a fertile age. The risks involved are extremely low, while important advantages are the ability to directly view the appendix and diagnose other possible causes of the acute abdomen [7]. Based on our own experience in diagnostic and therapeutic laparoscopic surgery, a prospective study was started in April 1991 with the following objectives: 1. To determine the value of diagnostic laparoscopy in all adult female patients planned to undergo laparotomy for suspected acute appendicitis 2. To compare results with appendicectomy in male patients treated directly through a muscle-splitting McBurney incision 3. To introduce the technique of laparoscopy to all residents and surgeons in our General Surgical department and yet maintain the classical McBurney appendicectomy so crucial for surgical training

Fig. 1. Scheme of the protocol. raphy to exclude obvious gynecological diseases. Patients were first seen by residents with subsequent supervision by different fellows of Gynecology (nine) and Surgery (seven). If the diagnosis of acute appendicitis was upheld, a diagnostic laparoscopy was done. Unfortunately not all patients actually underwent diagnostic laparoscopy for several reasons. Therefore, group I contains subgroup Ia (42 patients) in whom a diagnostic laparoscopy could not be performed and patients underwent directly an appendicectomy and subgroup Ib of 161 patients who underwent the diagnostic laparoscopy prior to the appendicectomy. All male patients with a clinical suspicion of acute appendicitis underwent appendicectomy, without previous diagnostic laparoscopy, through a muscle-splitting McBurney incision (Fig. 1). Diagnostic laparoscopy was performed under general anesthesia. Prophylactic antibiotics (metronidazole 500 mg) were administered intravenously and the urinary bladder was catheterized. A two-port approach was generally used (Fig. 2), infraumbilical and suprapubic, allowing adequate inspection of the lower abdomen. Occasionally an extra 5-mm trocar was required in the right abdomen to dissect a retrocecal appendix or completely visualize the upper abdomen. If an acutely inflamed appendix was found, or if it could not be excluded, appendicectomy was performed through a muscle-splitting McBurney incision or laparoscopically depending on the surgeons experience. Obvious gynecological disorders were treated by the gynecologist appropriately. A visually normal appendix was not removed.

Patients and methods

From April 1991 to September 1995, a prospective study was done on all consecutive adult patients admitted with the clinical signs of acute appendicitis and planned to undergo laparotomy. The total of 363 patients were divided into three groups: Group I Women 16 to 50 years of age (203 patients) Group II Women older than 50 years (23 patients) Group III Men 16 years and older (137 patients) The decision to subdivide the female patients into two groups was based on the relatively high frequency of gynecological disorders found in women of child-bearing age. Fifty years was the arbitrary limit chosen between preand postmenopause. Diagnostic workup included history taking, complete physical examination, and routine blood and urine analysis. Protocol dictated that once the surgeon decided that there was an indication for urgent appendicectomy, all female patients were also seen by a gynecologist, who performed specialized physical examination, cervical culture, and transvaginal ultrasonog-

Results Group I Group I consisted of 203 female patients from 16 to 50 years (median age 24) all suspected of suffering from acute appendicitis.

925 Table 1. Group Ib; diagnostic laparoscopy in women 1650 years Laparoscopic diagnosis Acute appendicitis Normal appendix Gynecological disorder Diverse diagnoses Negative laparoscopy Failed laparoscopy Total Number 88 37 5 23 8 161 % (55%) (23%) (3%) (14%) (5%)

Table 2. Group Ib; gynaecological disorders found at laparoscopy in women 1650 years (37/161 patients) Pelvic hemoperitoneum Pelvic inflammatory disease Ovarian/cyst torsion Endometriosis Unknown pregnancy/ectopic 15 13 3 4 2

Fig. 2. Placement of the trocars in diagnostic laparoscopy.

Subgroup Ia In 42 women diagnostic laparoscopy was omitted for various reasons: 31 during the first study year due to the dutysurgeons inexperience, eight instrument failures, and three advanced pregnancies (over 26 weeks). In all 42 patients a muscle-splitting incision was directly performed. Only 26 (62%) were found to have acute appendicitis, in 16 (38%) a normal appendix was removed, as confirmed by pathological examination. Subgroup Ib Some 161 women underwent diagnostic laparoscopy prior to the planned appendicectomy, including eight who were between 8 and 26 weeks pregnant. (Table 1) In 88 (55%) patients acute appendicitis was diagnosed by laparoscopy. Seventy-five subsequently underwent appendicectomy through a muscle-splitting incision; in 11 the appendix was removed laparoscopically. One case required a median laparotomy because of a perforated appendicular mass, and in another it was decided to treat the appendicular mass conservatively. There were two false-positive results due to misinterpretation of the visual aspect of the appendix. Histology showed no signs of acute inflammation. In 65 (40%) women the appendix was determined to be normal at laparoscopy and it was not removed (Table 1) except in two cases with appendicular faecoliths. In 37 of these patients (23% of subgroup Ib) a gynecological disorder was diagnosed (Table 2); four were operated laparoscopically, four required a Pfannenstiel incision and 29 could be treated conservatively. Diverse diagnoses were found in five patients: a small

liver laceration from a forgotten accident treated by laparoscopy, partial omental necrosis resected through minilaparotomy, bowel adhesions taken down laparoscopically, and two other patients had faecoliths in the appendix, this requiring appendicectomy. Twenty-three patients had a negative laparoscopy. During their hospital stay, four of these were diagnosed as having gastroenteritis and four had urinary tract infections or stones. Five patients later suffered repeated attacks of nonspecific abdominal pain. Only ten patients (6% of subgroup Ib) have remained undiagnosed at completion of the follow-up. There were no false-negative laparoscopies. None of the 63 patients in whom a visually normal appendix was left in situ were readmitted for acute appendicitis during the follow-up. In eight patients (5%) laparoscopy failed because of failure to insufflate the abdomen (three cases) or inability to completely visualize the appendix due to adhesions or a retrocecal position (five cases). All eight consequently underwent appendicectomy through a muscle-splitting McBurney incision. In only four was the appendix found to be acutely inflamed. Only one patient suffered a direct complication due to the laparoscopy. Bleeding from a trocar site required abdominal wall exploration, which was performed through the muscle-splitting McBurney incision. Group II Group II consisted of 23 postmenopausal women 50 to 86 years old (median age 54). During the learning period, 13 patients escaped laparoscopy and directly underwent appendicectomy. Acute appendicitis was diagnosed in 22 patients (96%). The remaining patient had a perforated peptic ulcer requiring an additional vertical incision. Group III Group III comprised 137 male adults, median age 23 years. All were operated on through the muscle-splitting McBurney approach for suspected acute appendicitis. An inflamed

926 Table 3. Acute appendicitis; comparison between all three groups (363 patients) Patient group Group Ia Group Ib Group II Group III Women 1650 years without laparoscopy Women 1650 years with laparoscopy Women >50 years Men >16 years (No.) (42) (161) (23) (137) Acute appendicitis 62% 56% 96% 92% (No.) (26) (90) (22) (126) Negative appendicectomy 38% 5% 4% 8% (No.) (16) (8) (1) (11)

appendix was found in 126 patients (92%). In 11 patients (11%) a normal appendix was removed. Operation time The mean operating time for diagnostic laparoscopy was 25 min while that for open appendicectomy was 45 min. Hospital stay The mean length of hospital stay was 1.7 days when diagnostic laparoscopy was the only procedure. If appendicectomy was performed the stay was similar whether the appendix was inflamed or notrespectively, 3.6 and 3.4 days. Discussion The main purpose of this prospective study was to determine the value of laparoscopy in establishing the diagnosis in young female patients suspected of having an acute appendicitis, and to reduce the negative laparotomy rate. Acute appendicitis is probably the only surgical disease where a diagnostic accuracy of only 70% to 75% is still accepted. Many methods have been described in an attempt to reduce the negative appendicectomy rateclose observation, specific laboratory tests, ultrasonography, computeraided diagnosis, scoring systems, barium enema studies, computed tomography, radioisotope imaging, and even peritoneal cytology [11, 18]. However, in spite of several promising initial reports, none of these has achieved widespread clinical application. The greatest concern is clearly in women of a fertile age in whom gynecological disorders frequently masquerade as acute appendicitis. Negative appendicectomy rates as high as 40% are no longer acceptable. Since the early 1980s, a group of authors has repeatedly pointed out the specific problem of the wide spectrum of diagnosis of acute lower abdominal pain in child-bearing age females in contrast to the simplicity of diagnosis in male patients. They suggested the use of diagnotic laparoscopy in young females with acute lower abdominal pain [18, 26]. Patients with an atypical presentation for appendicitis underwent laparoscopy, while straightforward cases directly proceeded to appendicectomy. In the latter group 25% still had a normal appendix removed, while in the former only one-third required appendicectomy. The authors also stressed the capacity of laparoscopy to diagnose other abdominal conditions, found

in 50% of their patients. Diagnosing those problems facilitated prompt and appropriate care in a majority of these patients. Despite this, a small group of 17% of patients remained undiagnosed. Since the widespread application of surgical laparoscopy in 1990, there have appeared a few prospective randomized studies on diagnostic laparoscopy in young female patients. Jadallah et al. [14, 17] have compared the diagnostic value of laparoscopy with the direct McBurney approach in female patients with clinical diagnosis of acute appendicitis. While both groups had the same percentage of 12 acute appendicitis (around 60%), after laparoscopy 38% of the patients did not require appendicectomy, and in all these patients a gynecological diagnosis was made. Our present study shows that in 161 women of childbearing age, suspected of having acute appendicitis, only 55% (88 patients) required appendicectomy after laparoscopy. In 23% (37 patients) a gynecological disorder was diagnosed in spite of prior gynecological examination. Only 6% (10 patients) ultimately remained undiagnosed. When compared to the subgroup of 42 women in whom laparoscopy was omitted, the negative appendicectomy rate was consequently reduced from 38% to 5% following diagnostic laparoscopy. (In only eight cases the removed appendix was not acutely inflamed) (Table 3). In sharp contrast, 96% of 23 women over 50 years and 92% of 137 men did indeed have acute appendicitis at operation, with respective negative appendicectomy rates of 4% and 8%. While there were two false-positive laparoscopic diagnoses of appendicitis, there have been no false-negative results after leaving a visually normal appendix in situ, even in the absence of other pathology. In this context, Wang et al. [24] recently published an interesting report on microscopical markers of inflammation found in apparently normal appendices. Traditionally, a normal appendix found at operation is nonetheless removed for two reasons: fear of missing an appendicitis or progression to it, and because a McBurney incision scar traditionally means the patient has undergone appendicectomy. These concepts may now be changed, and some authors make a plea for conservation of a normal appendix to reduce postoperative morbidity and because the appendix may be required in the future for other purposes [25]. Finally, the heated debate on whether laparoscopic appendicectomy should become the new gold standard still continues. In a recent overview of published reports by Heinzelmann et al. [12], there were no clear benefits of laparoscopic appendicectomy for operation time, hospital stay, or complication rate. Only postoperative comfort and

927

cosmetics were subjectively better. Undoubtedly, there will be advantages in certain categories of patients, but future studies are needed to define the exact role for laparoscopic appendicectomy. In conclusion, we are convinced that with the present experience and modern techniques at our disposal, all women of child-bearing age suspected of having acute appendicitis should undergo diagnostic laparoscopy in order to reduce the rate of negative appendicectomy, but also to establish the correct diagnosis allowing prompt and appropriate treatment. Laparoscopy, however, must not be used as an alternative to good clinical judgment. References
1. Editorial (1987) A sound approach to the diagnosis of acute appendicitis. Lancet 1(8526): 198200 2. Addiss DG, Shaffer N, Fowler BS, Tauxe RV (1990) The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 132(5): 910925 3. Anderson JL, Bridgewater FH (1981) Laparoscopy in the diagnosis of acute lower abdominal pain. Aust N Z J Surg 51(5): 462464 4. Andersson RE, Hugander A, Thulin AJ (1992) Diagnostic accuracy and perforation rate in appendicitis: association with age and sex of the patient and with appendicectomy rate. Eur J Surg 158(1): 3741 5. Berry J Jr, Malt RA (1984) Appendicitis near its centenary. Ann Surg 200(5): 567575 6. Chang FC, Hogle HH, Welling DR (1973) The fate of the negative appendix. Am J Surg 126(6): 752754 7. Cuesta MA, Borgstein PJ, Meijer S (1993) Laparoscopy in the diagnosis and treatment of acute abdominal conditions. Clinical review. Eur J Surg 159(9): 455456 8. De Dombal FT, Horrocks JC (1978) Use of receiver operating characteristic (ROC) curves to evaluate computer confidence threshold and clinical performance in the diagnosis appendicitis. Methods Inf Med 17(3): 157161 9. Diehl JT, Eisenstat MS, Gillinov S, Rao D (1981) The role of peritoneoscopy in the diagnosis of acute abdominal conditions. Cleve Clin Q 48(3): 325330 10. Edwards FH, Davies RS (1984) Use of a Bayesian algorithm in the computer-assisted diagnosis of appendicitis. Surg Gynecol Obstet 158(3): 219222

11. Eriksson S (1996) Acute appendicitisways to improve diagnostic accuracy (clinical review). Eur J Surg 162: 435442 12. Heinzelmann M, Simmen HP, Cummins AS, Largiader F (1995) Is laparoscopic appendectomy the new gold standard? Arch Surg 130(7): 782785 13. Hoffmann J, Rasmussen OO (1989) Aids in the diagnosis of acute appendicitis [see comments]. Br J Surg 76(8): 774779 14. Jadallah FA, Abdul-Ghani AA, Tibblin S (1994) Diagnostic laparoscopy reduces unnecessary appendicectomy in fertile women. Eur J Surg 160: 4145 15. Lewis FR, Holcroft JW, Boey J, Dunphy JE (1975) Appendicitis: a critical review of diagnosis and treatment in 1000 cases. Arch Surg 110: 677 16. Mueller BA, Daling JR, Moore DE, Weiss NS, Spadoni LR, Stadel BV, Soules MR (1986) Appendectomy and the risk of tubal infertility. N Engl J Med 315(24): 15061508 17. Olsen JB, Myren CJ, Haahr PE (1993) Randomized study of the value of laparoscopy before appendicectomy [see comments]. Br J Surg 80(7): 922923 18. Paterson-Brown S, Eckersley JR, Sim AJ, Dudley HA (1986) Laparoscopy as an adjunct to decision making in the acute abdomen. Br J Surg 73(12): 10221024 19. Pieper R, Kager L, Nasman P (1982) Acute appendicitis: a clinical study of 1018 cases of emergency appendectomy. Acta Chir Scand 148(1): 5162 20. Puylaert JB, Rutgers PH, Lalisang RI, de Vries BC, van der Werf SD, Dorr JP, Blok RA (1987) A prospective study of ultrasonography in the diagnosis of appendicitis. N Engl J Med 317(11): 666669 21. Silberman VA (1981) Appendectomy in a large metropolitan hospital. Retrospective analysis of 1,013 cases. Am J Surg 142(5): 615618 22. Spirtos NM, Eisenkop SM, Spirtos TW, Poliakin RI, Hibbard LT (1987) Laparoscopya diagnostic aid in cases of suspected appendicitis. Its use in women of reproductive age. Am J Obstet Gynecol 156(1): 904 23. Tanphiphat C, Chittmittrapap S, Prasopsunti K (1987) Adhesive small bowel obstruction. A review of 321 cases in a Thai hospital. Am J Surg 154(3): 283287 24. Wang Y, Reen DJ, Puri P (1996) Is a histologically normal appendix following emergency appendicectomy always normal? Lancet 347(9008): 10761079 25. Wheeler RA, Malone PS (1991) Use of the appendix in reconstructive surgery: a case against incidental appendicectomy. Br J Surg 78(11): 12831285 26. Whitworth CM, Whitworth PW, Sanfillipo J, Polk HC Jr (1988) Value of diagnostic laparoscopy in young women with possible appendicitis. Surg Gynecol Obstet 167(3): 187190

Surg Endosc (1997) 11: 911914

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Hemodynamics during laparoscopic extra- and intraperitoneal insufflation

An experimental study
J. J. G. Bannenberg,1 B. M. P. Rademaker,2 F. M. J. A. Froeling,3 D. W. Meijer1
1 2 3

Department of Surgery, Academic Medical Center, University of Amsterdam, PO Box 22700, 1100 DE Amsterdam, The Netherlands Department of Anesthesiology, Academic Medical Center, University of Amsterdam, PO Box 22700, 1100 DE Amsterdam, The Netherlands Red Cross Hospital, The Hague, The Netherlands

Abstract Background: Total extraperitoneal laparoscopic surgery is an alternative to the laparoscopic transperitoneal route; however, its effects on hemodynamics have not been adequately studied. This experimental study compared the effects of intraperitoneal insufflation and extraperitoneal insufflation on hemodynamics and oxygen transport. Methods: Sixteen pigs were randomly assigned for intraperitoneal insufflation or extraperitoneal insufflation with 15 mmHg carbon dioxide. Hemodynamic and oxygen transport parameters were taken during an hour of insufflation and analyzed for statistical differences. Results: During extraperitoneal CO2 pneumoperitoneum central venous filling pressures (central venous pressure, pulmonary capillary wedge pressure and mean pulmonary arterial pressure) and end-tidal CO2 increased slower but to a similar magnitude in comparison to intraperitoneal insufflation. Cardiac output and indices of oxygen consumption and oxygen delivery were equally affected by both types of insufflation. Arterial CO2 pressure increased significantly more during intraperitoneal insufflation. Conclusion: The data from this study suggest that extraperitoneal insufflation might result in less cardiovascular impairment than intraperitoneal insufflation. Key words: Laparoscopic surgery Pneumoperitoneum Hemodynamics Extraperitoneal Intraperitoneal

inguinal hernia repair, and bladder neck colposuspension have been introduced into the surgical spectrum for clinical use [15, 10, 12]. The directness of the extraperitoneal approach and contraindications for transperitoneal endoscopic surgerysuch as obesity, inadequate bowel preparation, and intraperitoneal adhesionsare factors that might make the extraperitoneal approach more attractive. This type of laparoscopic surgery may, however, be technically more demanding than its transperitoneal counterpart because of the limited view, the restricted working space, and the absence of familiar landmarks. This might result in prolonged procedures affecting the circulatory basis of the patient. Pneumoperitoneum for laparoscopic surgery has been associated with hemodynamic changes. Clinical studies during intraperitoneal insufflation with carbon dioxide showed that arterial blood pressure increases and cardiac output decreases [8, 13]. Few data exist on the hemodynamic effects of laparoscopy in an extraperitoneally created cavity. If an extraperitoneal laparoscopic procedure results in more circulatory depression, its potential usefulness in humans would be limited. In this experimental study we compared the effects of extraperitoneal and intraperitoneal laparoscopic insufflation on the hemodynamic parameters in a porcine model.

Materials and methods

Sixteen pigs (2735 kg) were used in this study. Anesthesia was induced with an i.m. injection of azaperon (12 mg/kg) and atropine (1 mg). After endotracheal intubation the lungs of the animals were mechanically ventilated at a rate of 12 breaths/min with a tidal volume of 10 ml/kg, with a mixture of oxygen in air (FIO2 0.4). Anesthesia was maintained with 0.71.0% halothane (inspired concentration) and the infusion regimen consisted of lactated Ringers solution at a rate of 4 ml/kg/h during the experiment. The right internal jugular vein was exposed and a flow-directed pulmonary arterial catheter (Baxter, Americ Edwards Laboratories, Irvine, CA, U.S.A.) was inserted and floated into the pulmonary artery. A Wallace 16-gauge catheter was placed in the right brachial artery for arterial pressure measurements and arterial blood-gas sampling.

Laparoscopic procedures on extraperitoneally located anatomical structuressuch as nephrectomy, adrenalectomy, lumbar sympathectomy, para-aortic lymph node sampling,
Correspondence to: J. J. G. Bannenberg, Department of Surgery, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands

912

Fig. 1. Hemodynamic variables during extraperitoneal () and intraperitoneal () carbon dioxide insufflation. Each data point represents the mean SD (n 8).

On a random basis eight pigs were assigned to extraperitoneal insufflation and eight to intraperitoneal insufflation with carbon dioxide. Through a modified open Hasson technique, as described by Horattas [7], a disposable 10-mm trocar was placed, extraperitoneally or intraperitoneally, and connected to a pressure-controlled carbon dioxide insufflator (Electronic Laparoflator 263400-20, Storz-Endoskop, Switzerland) set at 15 mmHg pressure. We attempted to create an extraperitoneal cavity of approximately 1 l of CO2 gas. From previous animal experiments we noted that this volume is sufficient to create a cavity large enough to be able to perform endoscopic retroperitoneal nephrectomy, para-aortic lymph-node sampling, or lumbar sympathectomy [13]. On average, the amount of gas necessary to create a 15-mmHg intraabdominal pressure in these animals was 5 l CO2. Hemodynamic measurements were recorded using disposable transducers (Gould, U.S.A.). Measurements were made for cardiac output (mean of four determinations by thermodilution, using room-temperature normal saline injectate (5 ml), Edwards Laboratories, Santa Ana, CA, U.S.A.), heart rate, mean arterial blood pressure, central venous pressure, mean pulmonary arterial pressure, pulmonary capillary wedge pressure, and endtidal carbon dioxide. Arterial and central venous blood samples were taken for gas analyses of arterial oxygen, arterial carbon dioxide pressure, and mixed venous oxygen saturation (ABL II, Radiometer, Copenhagen, Denmark). After preoperative workup the pigs were positioned in the supine position and allowed a stabilization period of 20 min before control measurements were taken. After starting the insufflation, measurements were made as follows; 1 min after insufflation with carbon dioxide, 5 min, 10 min, 15 min, 30 min, 45 min, and 60 min after insufflation. After the last sample was taken the pneumoperitoneum was desufflated and a control sample was taken after 10 min of desufflation. Blood samples were drawn during each measurement. Results are expressed as mean SD. Data was analyzed with two-way analysis of variance for repeated measures. When indicated, differences between means were analysed using paired t-tests with Bonferroni correction for multiple comparisons. Blood-gas measurements were analyzed with the Mann-Whitney U test; p values of <0.05 were considered statistically significant.

pulmonary arterial pressure, and pulmonary capillary wedge pressure increased significantly during the first minutes of intraperitoneal insufflation, reaching a plateau after approximately 10 min. These pressures showed a more gradual increase during extraperitoneal insufflation, reaching a plateau after 45 min. Differences in this respect were significant between both methods. The gas exchange parameters are shown in Fig. 2. There is a rapid increase of end-tidal CO2 during the first few minutes of intraperitoneal insufflation, reaching a plateau after 10 min. End-tidal CO2 during extraperitoneal insufflation increases at a slower pace and the magnitude of the increase is significantly lower than with intraperitoneal insufflation. During both insufflation methods arterial CO2 pressure increases, but the increase during intraperitoneal insufflation is significantly larger in magnitude than with extraperitoneal insufflation. pH decreased similarly in both groups. The arterial oxygen pressure and central venous oxygen saturation did not change significantly during either of the insufflation procedures. Discussion Extraperitoneal laparoscopic surgery is rapidly becoming an established route for surgical procedures. The effects of extraperitoneal insufflated carbon dioxide on gas exchange have been studied previously; however, the effects on hemodynamics are not yet clear. Our data concerning hemodynamic changes suggests that extraperitoneal insufflation with carbon dioxide is associated with hemodynamics similar to those observed during intraperitoneal insufflation. Extraperitoneal insufflation is associated with less rapid increases in central venous filling pressures compared to intraperitoneal insufflation. The mechanisms that are responsible for the hemodynamic changes during laparoscopic surgery appear to be multifactorial. Both pressure and pharmacological effects of the insufflated gas may affect hemodynamics. Pressure gradients will affect venous return. Increases in venous return secondary to increased intraperitoneal pressure have been reported to augment cardiac output. In general, however,

Results Hemodynamic changes with both insufflation methods are shown in Fig. 1. Heart rate did not change significantly during either of the insufflation methods. Although mean arterial blood pressure and cardiac output did increase significantly during intraperitoneal insufflation, but not during extraperitoneal insufflation, the differences between both methods were not significant. Filling pressures such as central venous pressure, mean

913

Fig. 2. Gas exchange variables during extraperitoneal () and intraperitoneal () carbon dioxide insufflation. Each data point represents the mean SD (n 8).

venous return is expected to decrease during intraperitoneal insufflation, as indicated by decreases in cardiac performance. This apparent contradiction may be explained by a time-dependent phenomenon. Initially, blood will be squeezed out of the abdominal cavity to the heart, causing a transient increase in cardiac output. Secondly, sustained increases in intraabdominal pressure will ultimately impede venous return and depress cardiac output. In addition to the pressure effects, pharmacological effects of the insufflation gas will affect hemodynamics during laparoscopy. The absorption of CO2 from the insufflation cavity, resulting in mild hypercarbia, is associated with sympathetic stimulation, which may increase heart rate, blood pressure, systemic vascular resistance, and cardiac output. On the other hand, severe acidemia and hypercapnia may depress heart performance, secondary to decreases in myocardial inotropy. We did observe significant increases in cardiac output and blood pressure during intraperitoneal insufflation, but not during extraperitoneal insufflation. Heart rate did not increase with either method. A clinical study [15] recently confirmed this data. However, in this nonrandomized study, measurements were only taken just before and at the end of the insufflation period. Furthermore, cardiac output and central venous filling pressures were not measured. Our results concerning cardiac output are in contrast to most human studies, which report moderate decreases in cardiac output during intraperitoneal insufflation. Studies performed in pigs may give different hemodynamic results compared to studies in humans. There are no data on human studies measuring cardiac output and central filling pressures during extraperitoneal laparoscopy. We could not demonstrate a significant difference between the extra- and the intraperitoneal approach with respect to cardiac output, blood pressure, and heart rate. Indeed, central venous oxygen saturation, which reflects global oxygen delivery to the tissues, was not affected by either approach. Therefore, it seems reasonable to assume that cardiac performance is affected similarly by both the extra- and the intraperitoneal approach. Central venous filling pressures increased during both intra- and extraperitoneal insufflation. However, during intraperitoneal insufflation, central filling pressures increased

faster than during extraperitoneal insufflation. Maximum pressures were reached on average 30 min earlier during the intraperitoneal approach. A possible explanation for this phenomenon might be that pressures are transmitted more quickly into the thorax during intraperitoneal insufflation. The soft tissues that cover the extraperitoneal cavity may serve as a buffer and may therefore delay the transmission of the pressure into the thorax. In theory, this might be advantageous for patients with limited cardiac reserve. However, whether the differences between both methods with respect to the central venous filling pressures are of any clinical consequence remains to be determined. The duration of an extraperitoneal approach is on average longer than 45 min. Therefore, it is to be expected that during an extraperitoneal procedure central venous filling pressures are elevated throughout a significant part of the procedure. Gas-exchange parameters showed an earlier and stronger tendency toward the development of respiratory acidosis in the intraperitoneal insufflation group. These findings are in agreement with others that showed, in an experimental study in dogs, that if the insufflated gas is limited to the retroperitoneal space, the absorption of CO2 appears to be reduced compared to intraperitoneal insufflation [14]. These findings contrast with a study that showed more marked CO2 diffusion into the body during extraperitoneal than during intraperitoneal CO2 insufflation [11]. However, in that study three different operations were evaluated, which might have influenced the results. Two mechanisms contribute to the development of hypercarbia during laparoscopy: absorption from the intra- or extraperitoneal cavity and increased physiologic dead space ventilation. Lister showed in an experimental study on pigs that CO2 absorption from the abdominal cavity depends on the level of intraabdominal pressure [9]. He hypothesized that by increasing the intraabdominal pressure from 0 to 10 mmHg a progressively larger area of the peritoneum is exposed to CO2, resulting in an increase in the diffusion area. Increasing intraperitoneal insufflation pressures above 10 mmHg will not result in any further increase in the diffusion area, causing a plateau in CO2 absorption. Nevertheless, a further increase in arterial CO2 pressure may occur as a result of increases in physiologic dead space ventilation. The greater absorptive capacity of the peritoneal membrane may also be reflected in a

914

greater systemic absorption of CO2 [6]. Our results indicate that the extraperitoneal insufflation is associated with lower arterial CO2 pressure values compared to intraperitoneal insufflation. A possible explanation might be that during extraperitoneal insufflation a relatively small cavity was created, resulting in a small diffusion area for CO2. A possible criticism of the study protocol is that the two different experiments were not compared in the same animal. Performing both intra- and extraperitoneal measurements in the same pig would only be feasible in the half of the animals where the intraperitoneal experiment was carried out first. The opposite order of experiments is, however, impossible. The creation of an extraperitoneal area causes the peritoneum to detach from the fascia of the abdominal muscles. This would have biased the intraperitoneal measurements. For this reason the randomized model was selected in which to perform the comparative measurements. In conclusion, our results indicate that extraperitoneal insufflation with CO2 for laparoscopy is associated with hemodynamic changes similar to those observed during intraperitoneal insufflation. Intraperitoneal insufflation is associated with more rapid changes in central venous filling pressures. Gas-exchange variables indicate a stronger tendency toward the development of respiratory acidosis during conventional intraperitoneal insufflation. The data from this study suggests that extraperitoneal insufflation might result in less cardiovascular impairment than intraperitoneal insufflation. References
1. Bannenberg JJ, Hourlay P, Meijer DW, Vangertruyden G (1995) Retroperitoneal endoscopic lumbar sympathectomy: laboratory and clinical experience. Endosc Surg Allied Technol 3: 1620

2. Bannenberg JJ, Meijer DW, Klopper PJ (1994) The prone position. Using gravity for a clear view. Surg Endosc 8: 11151116 3. Bannenberg JJ, Meijer DW, Klopper PJ (1995) Extraperitoneal laparoscopic paraaortic lymph node sampling in prone position: development of a technique. Laparoendosc Surg 5: 4146 4. Brunt L, Molmenti E, Kerbl K, Soper N, Stone A, Clayman R (1993) Retroperitoneal endoscopic adrenalectomy: an experimental study. Surg Laparosc Endosc 3: 300306 5. Clayman R (1993) Retroperitoneoscopy. In: Clayman R, McDougall E (eds) Laparoscopic urology. Quality Medical, St Louis, pp 383394 6. Collins J (1981) Inert gas exchange of subcutaneous and intraperitoneal gaspockets in piglets. Respir Physiol 46: 391404 7. Horattas M, Rosser R (1993) A new and simple approach to open laparoscopy. Surg Gynecol Obst 176: 287289 8. Joris J, Noirot D, Legrand M, Jacquet N, Lamy M (1993) Hemodynamic changes during laparoscopic cholecystectomy. Anesth Analg 76: 10671071 9. Lister D, Rudston-Brown B, Wriner B, McEwin J, Chan M, Walley K (1994) Carbon dioxide absorption is not linearly related to intraperitoneal carbon dioxide insufflation in pigs. Anesthesiology 80: 129136 10. Mandressi A, Buizza C, Antonelli D, Belloni M, Chisensa S, Zaroli A, Bernasconi S (1993) Retro-extraperitoneal laparoscopic approach to excise retroperitoneal organs: kidney and adrenal gland. Minimally Invasive Ther 2: 213220 11. Mullet C, Viale J, Sagnard P, Miellet C, Ruynat L, Counioux H, Motin J, Boulez J, Dargent D, Annat G (1993) Pulmonary CO2 elimination during surgical procedures using intra- or extraperitoneal CO2 insufflation. Anesth Analg 76: 622626 12. Raboy A, Hakim LS, Ferzli G, Antario JM, Albert PS (1993) Extraperitoneal endoscopic vesicourethral suspension. Laparoendosc Surg 3: 505508 13. Westerband A, Water JM VD, Amzallag M, Lebowitz PW, Chardavoyne R, Abou-Taleb A, Wang X, Wise L (1992) Cardiovascular changes during laparoscopic cholecystectomy. Surg Gynecol Obst 175: 535538 14. Wolf JJ, Carrier S, Stoller ML (1995) Intraperitoneal versus extraperitoneal insufflation of carbon dioxide as for laparoscopy. J Endourol 9: 6366 15. Wright DM, Serpell MG, Baxter JN, ODwyer PJ (1995) Effects of extraperitoneal carbon dioxide insufflation on intraoperative blood gas and hemodynamic changes. Surg Endosc 9: 11691172

News and notices

Surg Endosc (1997) 11: 971973

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

New Address for the European Association for Endoscopic Surgery (E.A.E.S.)
Effective January 1, 1997, the new correspondence, telephone, and fax numbers of the E.A.E.S. office are: E.A.E.S. Office, c/o Mrs. Ria Palmen Luchthavenweg 81 Unit 1.42 5657 EA Eindhoven The Netherlands or: P.O. Box 335 5500 AH Veldhoven The Netherlands Tel: +31 40 2525288 Fax: +31 40 2523102

Fellowships in Minimally Invasive Surgery The University of Pittsburgh Medical Center Pittsburgh, PA, USA
One year fellowships in advanced minimally invasive surgery in both general and thoracic surgery are being offered at the University of Pittsburgh Medical Center beginning on July 1, 1997. Requirements include completion of residency training programs in the desired area. The fellowships include a competitive salary and travel allowance. Interested candidates should send a letter of inquiry with curriculum vitae to: Philip R. Schauer, MD (General Surgery) or James Luketich, MD (Thoracic Surgery) The University of Pittsburgh Medical Center 3471 Fifth Avenue Suite 300 Pittsburgh, PA 15213-3221

Volunteer Surgeons Needed Northwestern Nicaragua Laparoscopic Surgery Teaching Program, Leon, Nicaragua
Volunteer surgeons are needed to tutor laparoscopic cholecystectomy for this non-profit collaboration between the Nicaraguan Ministry of Health, the National Autonomous University of Nicaragua, and Medical Training Worldwide. The program consists of tutoring general surgeons who have already undergone a basic laparoscopic cholecystectomy course. Medical Training Worldwide will provide donated equipment and supplies when needed. For further information, please contact: Medical Training Worldwide Ramon Berguer, MD, Chairman Tel: 707-423-5192 Fax: 707-423-7578 e-mail: berguer.r@martinez.va.gov

Fellowships in Laparoscopic Surgery Staten Island University Hospital Staten Island, NY USA
A one year fellowship, to start July 1, 1997, in advanced laparoscopic surgery is being offered at Staten Island University Hospital. The selected fellow will be exposed to many advanced general laparoscopic surgeries including: hiatal hernia repair, splenectomy, adrenalectomy, bowel resection, and others. Participation in research projects will be encouraged. For further information, please contact: Barbara Coleman Coordinator, Surgical residency program Tel: 718-226-9508

Fellowship in Minimally Invasive Surgery George Washington Medical Center Washington, DC USA
A one-year fellowship is being offered at the George Washington University Medical Center. Interested candidates will be exposed to a broad range of endosurgical Education and Research Center. Active participation in clinical and basic science research projects is also encouraged. For further information, please contact: Carole Smith 202-994-8425 or, send curriculum vitae to: Dr. Jonathan M. Sackier Director of Endosurgical Education and Research George Washington University Medical Center Department of Surgery 2150 Pennsylvania Avenue, N.W. 6B-417 Washington, DC 20037, USA

Essentials of Laparoscopic Surgery Surgical Skills Unit University of Dundee Scotland, UK

Under the direction of Professor A. Cuschieri the Surgical Skills Unit is offering a three-day practical course designed for surgeons who wish to undertake the procedures such as laparoscopic cholecystectomy. This intensely practical program develops the necessary operating skills, emphasizes safe practice, and highlights the common pitfalls and difficulties encountered when starting out. Each workshop has a maximum of 18 participants who will learn both camera and instrument-manipulation skills in a purpose-built skills laboratory. During the course there is a live demonstration of a laparoscopic cholecystectomy. The unit has a large library of operative videos edited by Professor Cuschieri, and the latest books on endoscopic surgery are on display in our Resource area. Course fee including lunch and course materials is $860.

972 For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042 recreational pursuits which can be arranged by the facility to suit your personal agenda. For further details please contact: Carole Smith: Department of Surgery 2150 Pennsylvania Avenue NW 6B Washington, DC 20037, USA Tel: (202) 994-8425

Advanced Endoscopic Skills Surgical Skills Unit University of Dundee Scotland, UK

Each month Professor Cuschieri Surgical Skills Unit offers a 412 day course in Advanced Endoscopic Skills. The course is intensely practical with hands on experience on a range of simulated models. The program is designed for experienced endoscopic surgeons and covers advanced dissection techniques, extracorporeal knotting techniques, needle control, suturing, internal tying technique, stapling, and anastomotic technique. Individual workstations and a maximum course number of 10 participants allows for personal tuition. The unit offers an extensive collection of surgical videos and the latest books and publications on endoscopic surgery. In addition, participating surgeons will have the opportunity to see live advanced laparoscopic and/or thoracoscopic procedures conducted by Professor Cuschieri and his team. The course is endorsed by SAGES. Course fee including lunch and course materials is $1850. For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

Courses at the Royal Adelaide Centre for Endoscopic Surgery

Basic and Advanced Laparoscopic Skills Courses are conducted by the Royal Adelaide Centre for Endoscopic Surgery on a regular basis. The courses are limited to six places to maximize skill development and tuition. Basic courses are conducted over two days for trainees and surgeons seeking an introduction to laparoscopic cholecystectomy. Animal viscera in simulators is used to develop practical skills. Advanced courses are conducted over four days for surgeons already experienced in laparoscopic cholecystectomy who wish to undertake more advanced procedures. A wide range of procedures are included, although practical sessions can be tailored to one or two procedures at the participants request. Practical skills are developed using training simulators and anaesthetised pigs. Course fees: $A300 ($US225) for the basic course and $A1,600 ($US1,200) for the advanced course. For further details and brochure, please contact: Dr. D. I. Watson or Professor G. G. Jamieson The Royal Adelaide Centre for Endoscopic Surgery Department of Surgery Royal Adelaide Hospital Adelaide SA 5000 Australia Tel: +61 8 224 5516 Fax: +61 8 232 3471

The Practical Aspects of Laparoscopic Fundoplication Surgical Skills Unit University of Dundee Scotland, UK
A three-day course, led by Professor Cuschieri, designed for experienced laparoscopists wishing to include fundoplication in their practice. The course covers the technical details of total and partial fundoplication using small group format and personal tuition on detailed simulated models. There will be an opportunity to observe one of these procedures live during the course. Maximum course number is six. Course fee including lunch is $1850. For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

Advanced Laparoscopic Suturing and Surgical Skills Courses MOET Institute San Francisco, CA, USA
Courses are offered year-round by individual arrangement. The MOET Institute is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide continuing medical education for physicians and designates these CME activities for 2040 credit hours in Category 1 of the Physicians Recognition Award of the American Medical Association. These programs are also endorsed by the Society of Gastrointestinal Endoscopic Surgeons (SAGES). For further information, please contact: Wanda Toy, Program Administrator Microsurgery & Operative Endoscopy Training (MOET) Institute 153 States Street San Francisco, CA 94114, USA Tel: (415) 626-3400 Fax: (415) 626-3444

Courses at George Washington University Endosurgical Educational and Research Center

George Washington University Endosurgical Educational and Research Center is proud to offer a wide range of surgical endoscopy courses. These courses include advanced laparoscopic skills such as Nissen fundoplication, colon resection, common bile duct exploration, suturing, as well as subspecialty courses. Individual surgeons needs can be met with private tuition. The Washington D.C. area is a marvelous destination to visit for

Courses at WISE Washington Institute for Surgical Endoscopy Washington, DC, USA
The Washington Institute of Surgical Endoscopy is pleased to offer the following courses:

973 Laparoscopic antireflux and hiatal hernia surgery (July 1415, 1997); Laparoscopic management of the common bile duct and difficult cholecystectomy (May 1516, August 1112, November 1011, 1997); Laparoscopic colon and rectal surgery (June 2021, September 1516, December 45, 1997). Also, courses for operating room nurses and technicians will be run on a monthly basis and personal instruction and preceptorship is available. For further information, please call: Carole Smith Washington Institute of Surgical Endoscopy 2150 Pennsylvania Avenue, N.W. Washington, DC 20037 Tel: 202-994-9425

9th International Meeting Society for Minimally Invasive Therapy July 1416, 1997 Kyoto, Japan
Scientific program to include: Plenary, Parallel, Poster, and Video sessions. Host Chairman: Professor Osamu Yoshida, Department of Urology, Kyoto University, 54 Shogoin Kawahara-sho, Sakyo, Kyoto 606, Japan. Phone: +81 75 751-3328, Fax: +81 75 751-3740. This meeting coincides with the Gion Festival in Kyoto, one of the greatest festivals in Japan. For further information, please contact: Secretariat of SMIT 9th Annual International Meeting c/o Academic Conference Planning 383 Murakami-cho Fushimika, Kyoto 612 Japan Tel: +81 75 611-2008 Fax: +81 75 603-3816

Call for Abstracts Society of American Gastrointestinal Endoscopic Surgeons (SAGES) 1998 Annual Meeting April 14, 1998 Seattle, WA, USA
Abstract deadlines: Oral and Poster abstracts: September 12, 1997 Video Submissions: September 18, 1997 For further information, or to obtain an abstract form, please contact: SAGES Program Committee Society of American Gastrointestinal Endoscopic Surgeons Suite #3000 2716 Ocean Park Boulevard Los Angeles, CA 90405 Tel: (310) 314-2404 Fax: (310) 314-2585 e-mail: SAGESMail@AOL.com

Colorectal Disease in 1998 February 1921, 1998 Fort Lauderdale, FL, USA Symposium Director: Steven D. Wexner, MD
Cleveland Clinic Florida presents its ninth annual postgraduate course. Provides an intensive, in-depth, analytical review of all aspects of colorectal disease, including laparoscopy; colorectal carcinoma screening and genetics, inflammatory bowel disease; and pouch surgery. There will be a review of both basic and advanced principles of diagnosis and management of disease. Video techniques will be shown as well. The faculty is internationally represented and includes leading experts in the field. Simultaneous Spanish and Italian translation is available. For more information, please contact: Cleveland Clinic Florida Department of Education 2950 West Cypress Creek Road Fort Lauderdale, FL 33309-1743 Tel: 800-359-6101, ext. 6066 Fax: 954-978-5539

European Course on Laparoscopic Surgery (English language) November 1821, 1997 Brussels, Belgium
Course director: G.B. Cadiere For further information, please contact: Administrative Secretariat Conference Services s.a. Avenue de lObservatoire, 3 bte 17 B-1180 Bruxelles Tel: (32 2) 375 16 48 Fax: (32 2) 375 32 99

Courses Offered at the University of Minnesota Minneapolis, Minnesota, USA September 17, 1997: Fourth Annual Conference, Molecular Biology of Colorectal Cancer September 17, 1997: Sixth Annual Conference, Endorectal Ultrasonography September 1820, 1997: Sixtieth Annual Conference, Principles of Colon and Rectal Surgery
For further information, please contact: Continuing Medical Education University of Minnesota 615 Washington Avenue SE, Suite 107 Minneapolis, MN 55414 Tel: 800-776-8636 Fax: (612) 626-7766

6th World Congress of Endoscopy Surgery Roma 98 6th International Congress of European Association for Endoscopic Surgery June 36, 1998 Rome, Italy
The program will include: the latest, original high quality research; symposia; plenary lectures; abstract presentations (video, oral, and posters); EAES and SAGES postgraduate courses, OMED postgraduate course on therapeutic endoscopy; working team reports; educational center and learning corner; meeting of the International Society of Nurses and Associates; original and non original scientific reports; and a world expo of new technology in surgery. For further information, please contact: Congress Secretariat: Studio EGA Viale Tiziano, 19 00196 Rome, Italy Tel: +39 6 322-1806 Fax: +39 6 324-0143

Technique
Surg Endosc (1997) 11: 961962

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

A new fine probe for electric cautery during endoscopic surgery

Y. Yanagita,1,2 T. Asao,1 R. Katoh,1,2 M. Takahashi,1,2 Y. Nagamachi1
1 2

First Department of Surgery, Gunma University, 3-39-15 Showa-machi Maebashi 371, Japan Department of Surgery, Gunma Cancer Center, Ota 373, Japan

Abstract. The present paper introduces a new fine probe for electric cautery (1.65 mm in diameter, 22 cm long) that can be connected to a conventional cylindrical handcontrolled cautery holder, which is monopolar and widely used in general surgery. When cautery was required, a 14gauge intravenous catheter was inserted at an appropriate site under the guidance of a videoscope. After removing the steel inner needle, the extra tube was used as the fine surgical port for the cautery probe. The position of insertion could be altered according to the operating field. Cautery was performed by conventional methods. There was no bleeding or air leakage at the site of puncture during or after surgery. The puncture wound was closed without any sutures. Based on these results, the new fine probe for cautery can reduce the number of surgical ports required for instruments during video-assisted surgery, thus improving the ease and safety of endoscopic surgery. Key words: Endoscopic surgery Cautery probe Cautery holder

During endoscopic surgery, all surgical manipulation must be performed through a limited number of fixed surgical ports, which contributes to the difficulty associated with laparo- or thoracoscopic surgery [1, 2]. Therefore, we developed a new cautery probe that is sufficiently fine to allow it to enter through the extra tube of a 14-gauge intravenous catheter. The use of this new probe was evaluated during laparo/thoracoscopic surgery. Materials and methods
The cautery probes were 1.65 mm in diameter and 22 cm long. With the exception of the tip, the whole shaft was insulated with a contractive tube purchased from Ohm Co. (Tokyo, Japan). Three different tip shapes were

Fig. 1. Surgical techniques using fine cautery probe. developed (Spatula, L-hook, and Needle-shaped). The probes were connected to a conventional cylindrical hand-controlled cautery holder from Valleylab Inc. (Colorado, USA). When cautery was required, a 14-gauge intravenous catheter was inserted at an appropriate site under videoscopic guidance. After removing the steel inner needle, the extra tube was used as the surgical port for the fine cautery probe (Fig. 1).

Results and discussion The probe was used during eight endoscopic surgical procedures (two cholecystectomies, two colectomies, and four

Correspondence to: T. Asao

962

thoracoscopic operations). No complications associated with the probe were observed in any of the patients. The cautery could be manipulated as during conventional surgery. The position of insertion could be altered according to the operating field. Insertion had to be repeated several times in order to allow suitable manipulation during thoracoscopic surgery due to the narrow space between the ribs. No bleeding or air leakage was observed at the puncture site during or after surgery. The puncture wound was closed without any sutures and scar formation was not induced. Based on these results, the new fine probe for cautery can

reduce the number of surgical ports required for instruments during video-assisted surgery, thus improving the ease and safety of endoscopic surgery. References
1. Rohlf S (1995) Electrosurgical safety considerations for minimally invasive surgery. Minim Invasive Surg Nurs 9: 2629 2. Waxman K, Birkett DH, Sackier JM, Este-McDonald J, Duquette J (1994) Clinical and laboratory evaluation of an electrosurgical laparoscopic trocar. Surg Endosc 8: 10761079

Surg Endosc (1997) 11: 957960

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Sonographic blood flow measurements in malignant breast tumors

A potential new prognostic factor
C. Sohn, F. Beldermann, G. Bastert
Department for Prenatal and Gynecological Ultrasound Diagnosis and Treatment, Clinic of OB/GYN, University of Heidelberg, Heidelberg, Germany Received: 16 August 1996/Accepted: 27 January 1997

Abstract Background: The aim of this study was to find a possible relationship between the biological behavior of carcinomas of the breast and sonographically detectable blood flow after first studies showed a correlation between blood flow and prognostic factors. Method: 259 patients with ductal invasive breast cancer were examined using MEM (i.e., the Maximum Entropy Method), a new sonographic blood flow measurement technique capable of discerning considerably slower blood flow velocities than Doppler sonography. Due to the lack of objective methods for quantifying the blood flow, the findings were divided into three classes dependent upon the visual color information obtained. The blood flow was correlated with the size of the tumor, lymph node and receptor status, ploidy and S-phase fraction. Results: Most of the patients with small tumors, without lymph node metastases, with positive receptors, with a diploid genome, and with a low S-phase fraction belonged to the group with the lowest blood flow. Conclusion: The close relationship between the established prognostic factors and the sonographic blood flow measurements using MEM might be indicative of a new preoperative prognostic factor; this must, however, be confirmed by larger studies. Key words: Ultrasound blood flow MEM technique Prognostic factors

individual colored points) [15, 16]. This appeared to correlate with the established prognostic factors. This study was conducted in order to confirm this initial impression; 259 patients with ductal invasive breast cancer were examined sonographically on the preoperative day and the sonographic data was compared with the postoperative findings. The utilization of a new ultrasound technique, capable of distinctly more sensitive blood flow diagnosis than conventional Doppler sonography, appears to be decisive for this study. Materials and methods Sonographic technique
Ultrasound system: CSA manufactured by Acoustic Imaging, Phoenix, USA Color technique: MEM (Maximum Entropy Method) The maximum entropy method is a nonlinear method of spectral analysis initially developed by J.P. Burg for the US Naval Forces for the location of boats. A modification of the algorithms used in the color Doppler results in an improved spectral resolution and noise perception. MEM forestalls the signals from becoming imperceivable due to the noise by optimizing the separation of the flow information and the disturbing noise. MEM is thus even capable of detecting signals which are weaker than the background noise. The detection of extremely slow blood flow velocities, which are imperceivable to the Doppler technique, is thus possible. Slowest discernible blood flow velocities (phantom measurements): MEM: 0.2 mm/s Color Doppler: 1020 mm/s The MEM technique is thus capable of discerning the slowest of blood flow velocities, as present in the minutest of blood vessels. The following problem does, however, arise: To date the color information concerning the blood flow has been quantified using an integrated pulsed Doppler, which searched for the individual colored points. The thus obtained Doppler spectra were analyzed and the usual Doppler parameters could be calculated. The MEM technique is, however, capable of detecting considerably slower blood flow velocities than the pulsed Doppler can register; hence, this type of quantification is no longer possible. One thus merely has the possibility of directly registering the color information and evaluating it. This led to the following division of the tumor blood flow velocities into three classes:

During sonographic determination of the blood flow within tumors of the breast we noticed that there are malignant tumors in which the carcinomas display a very rapid blood flow (in the form of numerous colored areas) and carcinomas with much less blood flow (in the form of merely
Correspondence to: C. Sohn, Universita ts-Frauenklinik, Voss Str. 9, D-69115 Heidelberg, Germany

958

Fig. 1. Blood flow within a carcinoma of the breast corresponding to class I.

Fig. 2. Blood flow within a carcinoma of the breast corresponding to class II.

Class I: Merely individual colored points within the tumor or on its edge are discernible. Class II: One colored area is discernible within the tumor Class III: Numerous colored areas are discernible within the tumor. This is a subjective classification. Thus, we tested its reproducibility in advance: Two physicians were presented with five examples, as color depictions, from each class. These 15 examples were all correctly classified by both colleagues; hence, we could assume a good reproducibility [16].

Patients
Some 259 patients with ductal invasive breast cancer were included in the study. Only patients with ductal invasive carcinomas were included in order to obtain a homogenous group as far as histology is concerned. All patients were examined on the preoperative day and their inclusion in the study was confirmed postoperatively by the histologic findings of an invasive ductal carcinoma. The size of the tumor, the lymph node and the receptor status, the ploidy, and the S-phase fraction were registered together with the histology of the tumor. The examinations were all carried out by the same person. For statistical analysis the Kruskal-Wallis Test and Fishers Exact Test were used.

Fig. 3. Blood flow within a carcinoma of the breast corresponding to class III.

Results 1. The classification of the 259 patients into the three blood flow classes: Class I: 106 (40.9%) Class II: 60 (23.1%) Class III: 93 (36%) 2. The correlation of the T1 tumors with the blood flow classes: There were 49 T1 tumors in class I (46.2% of the tumors in class I), 36 in class II (60% of the tumors in class II), and 30 in class III (32.2% of the tumors in class III) (Figs. 13). 3. The correlation of the lymph node status with the blood flow classes: More than 14 axillary lymph nodes were removed from all patients and histologically examined; 125 patients had no lymph node metastases; 134 patients had metastases in a lymph node; 72 (58%) of the 125 patients without lymph node metastases belonged to blood flow class I, 21 (17%) to class II, and 32 (25%) to class III; 34 (25%) of the 134 patients with metastases in a lymph node were classified as class I, 39 (29%) as class II, and 61 (46%) as class III (p 0.022).

4. The correlation of the receptor status with the blood flow classes: Information concerning the receptor status was only available for 203 patients examined by the same laboratory. These patients were classified as having both a positive estrogen and progesterone receptor status, only one positive receptor status, or as both receptors having a negative status. Both receptors were found to be positive in 73 patients: 33 (47%) of these belonged to blood flow class I, 14 (19%) to class II, and 25 (34%) to class III. In 44 patients only one receptor (estrogen or progesterone receptor) was positive: 15 (34%) of these patients belonged to class I, 10 (23%) to class II, and 19 (43%) to class III. Both receptors were negative in 63 patients: 20 (32%) of them belonged to class I, six (9%) to class II, and 37 (59%) to class III (p 0.034). 5. The correlation of the ploidy with the blood flow classes: The ploidy was only known in 200 patients examined by the same laboratory; 79 patients had a diploid genome: 38 (48%) of these belonged to class I, 19 (24%) to class II, and 22 (28%) to class III; 121 patients had an aneuploid genome in the tumor: 37 (31%) of them were in class I, 33 (27%) in class II, and 51 (42%) in class III (p 0.035). 6. The correlation of the S-phase fraction with the blood

959 Table 1. Lymph node statusa LN negative: n 125 LN positive: n 134

a

Table 2. Receptor statusa 1: 2: 3: 1: 2: 3: 72 21 32 34 39 61 Pat Pat Pat Pat Pat Pat 58% 17% 25% 25% 29% 46% ER and PR positive: n 73 ER or PR positive: n 44 ER and PR negative: n 63
a

CL CL CL CL CL CL

The correlation of the lymph node status with the blood flow classes: n 259.

CL CL CL CL CL CL CL CL CL

1: 2: 3: 1: 2: 3: 1: 2: 3:

34Pat 14Pat 25Pat 15Pat 10Pat 19Pat 20Pat 6Pat 37Pat

(47%) (19%) (34%) (34%) (23%) (43%) (32%) (9%) (59%)

flow classes (n 243): 243 patients were divided into two groups with an S-phase fraction greater or smaller than 5. The S-phase was less than 5 in 137 cases; 86 (63%) of them belonged to class I, 24 (17%) to class II, and 27 (20%) to class III. The S-phase fraction was greater than 5 in the remaining 106 patients: 21 (20%) of them belonged to blood flow class I, 22 (21%) to class II, and 63 (59%) to class III (p 0.022).

The correlation of the receptor status with the blood flow classes: n 180.

Table 3. Ploidya Aneuploid genome: n 121 Diploid genome: n 79

a

CL CL CL CL CL CL

1: 2: 3: 1: 2: 3:

37 33 51 38 19 22

Pat Pat Pat Pat Pat Pat

31% 27% 42% 48% 24% 28%

The correlation of the ploidy with the blood flow classes: n 200.

Discussion The results presented show that the intensity of the blood flow within a carcinoma of the breast (determined using a new and extremely sensitive color method) clearly correlates with the established prognostic factors. This can be impressively demonstrated by the correlation with the lymph node status: Almost 60% of all patients with a poor blood supply within the carcinoma of the breast were free of axillary metastases of the lymph nodes, whereas only 25% of the patients with a good blood supply within the tumor were free of metastases in a lymph node. The correlation of a low blood flow with the ploidy, S-phase fraction, and receptor statusand thus a favorable prognosisis also evident: a low S-phase fraction was mostly evident in the blood flow class I, whereas tumors with an aneuploid genome mostly had a lot of color pixels or areas. The detection of estrogen and progesterone receptors also correlated with less blood flow within the tumor, a lack of receptor correlating with a good blood supply [7, 15, 16]. These results are not so clear as shown in our first studies on this field, but show a correlation between color pixels and prognostic factors in malignant breast tumors [15, 16]. The prerequisite for this diagnosis appears to be the utilization of a new sonographic color techniquethe MEM color technique. Thanks to its modified color algorithms it is capable of detecting distinctly slower blood flow velocities than conventional color Doppler sonography can. It is thus possible to demonstrate blood flow within notably smaller vessels than was previously the case. The theoretic reason for the examination of such slow blood flow velocities is based on the fact that the neovascularization of malignant tumors leads to the development of the minutest of blood vessels without a tunica. These minute vessels have a very low vascular resistance and thus an extremely slow blood flow velocity. The MEM technique appears to be capable of detecting such blood flow velocities [16]. The subjective classification into the three described blood flow classes is undoubtedly problematic. Although we could demonstrate that this type of classification is reTable 4. S-Phasea S-phase < 5%: n 137 S-phase > 5%: n 106
a

CL CL CL CL CL CL

1: 2: 3: 1: 2: 3:

86 24 27 21 22 63

Pat Pat Pat Pat Pat Pat

63% 17% 20% 20% 21% 59%

The correlation of the S-phase with the blood flow classes: n 243.

producible, it is imperative that further objectivization be attained. To this end we are currently testing an automatic mathematic classification system, according to the classes described above, which places the color pixels in relationship to the black and white surface [4]. The study presented here illustrates a new field of utilization of color-coded sonographic blood flow measurements. To date this type of diagnosis has been used within the framework of diagnosis of a tumors nature [13, 5, 6, 8,11, 13, 14, 16], which does, however, strike us as being less promising than the utilization described here. B scan sonography is capable of determining the nature of a tumor with a relatively high scan accuracy (85% to 90%); blood flow diagnosis can further increase this accuracy. An absolute accuracy in diagnosis can, however, not be attained; this is true for all scanning methods [1, 2, 6, 12].

Conclusion It has to be critically evaluated whether or not the number of operations of benign findings in the breast can be reduced using this new method. But as we mentioned above, absolute accuracy in diagnosis can and will not be attained in the near future. Thus, it makes sense not only to use this suitable highly sensitive ultrasound technique for diagnosis but also to evaluate the prognosis of malignant neoplasms. Our results present preliminary and still somewhat un-

960

certain indications concerning a new prognostic method for carcinomas of the breast [16]. Long-term studies on large groups of patients must confirm these preliminary results. This examination opens an entirely new field in the prognostic evaluation of malignant breast cancerwhen preoperative assessment of the prognosis thus appears to be possible for the first time. Should our first results be confirmed the following question arises: Could this new preoperative prognosis assessment perhaps even influence the surgical procedure? References
1. Bamber JC, Sambrook M, Minasian H, Hill CR (1983) Doppler study of blood flow in breast cancer. In: Jellins J, Kobayashi T (eds) Ultrasonic examination of the breast. John Wiley, pp 371378 2. Blohmer J (1991) Mammasonographie. In: Sohn Ch, Holzgreve (eds). Ultraschall in Gyna kologie und Geburtshilfe. Chapman & Hall, Buch, Druck 3. Burns PN, Virjee JM, Gowland M, et al (1983) The origin of doppler shift signals from breast tumors. In: Jellins J, Kobayashi T (eds) Ultrasonics examinations of the Breast. John Wiley, pp 379384 4. Delorme S, Anton H-W, Knopp MV, Betsch B, Trost U, Junkermann I, Fournier DV, Van Kaick G (1991) Vaskularisation des Mammakarzinoms: Quantitative und morphologische Beurteilung mittels farbcodierter Dopplersonographie. Abstract 423 Ultraschall Klin Prax 6: S.219 5. Heilenktter U, Jagella P (1993) Farbdopplersonographie exstirpationsbedu rftiger MammatumorenDarstellung einer Untersuchungsmethode. Geburtshilfe Frauenheilkd 53: 247252 6. Jellins J (1988) Combining imaging and vascularity assessment of breast lasions. Ultrasound Med Biol 14: 121130

7. Kaufmann M, Minckwitz GV, Finn HP, Schmid H, Goerttler K, Bastert G ( ) Combination of grading and new biological factors (S-phase fraction and epidermal growth factor receptor) can predict relapse and survival in patients with node-negative primary breast cancer. Onkologie 17: 166172 8. Madjar H, Sauerbrei W, Mnch S, Prmpeler H, Schillinger H (1990) Methodenanalyse zur Doppleruntersuchung der weiblichen Brust. Ultraschall Med 4: 196201 9. Madjar H, Jellins J, Schillinger H, Hillemanns HG (1986) Differenzierung von Mammakarzinomen durch CW-Doppler-Ultraschall. Ultraschall Med 7: 183184 10. Madjar H, Pro mpeler H, Wilhelm CH (1991) Doppler zur Diagnostik und Therapie von Brusterkrankungen. Abstract 425 Ultraschall Klin Prax 6: S.220 11. Minasian H, Bamber JC (1982) A preliminary assessment of an ultrasonic doppler method for the study of blood flow in human breast cancer. Ultrasound Med Biol 8: 357364 12. Sohn CH, Stolz W (1991) Dopplersonographische Durchblutungsmessung von Brusttumoren. Abstract 424 Ultraschall Klin Prax 6: S.219 13. Sohn CH, Grischke EM, Wallwiener D, Kaufmann M, Fournier DV, Bastert G (1992) Die sonographische Durchblutungsdiagnostik gutund bo sartiger Brusttumoren. Gerbutshilfe Frauenheilkd 52: 397403 14. Sohn CH, Stolz W, Grischke EM, Wallwiener D, Bastert G, Fournier DV (1992) Die dopplersonographische Untersuchung von Mammatumoren mit Hilfe der Farbdopplersonographie, der Duplexsonographie und des CW-Dopplers. Zentralbl Gyna kol 114: 249253 15. Sohn CH, Grischke EM, Stolz W, Bastert G (1993) Untersuchungen zum Zusammenhang zwischen dem Grad der Durchblutung und dem biologischen Verhalten von Mammatumoren. Ultraschall Klin Prax 8: 1114 16. Sohn CH, Thiel C, Baudendistel A, Kaufmann M, Bastert G (1996) Is the sonographic degree of blood supply a new prognostic factor? Breast 5: 110112

Surg Endosc (1997) 11: 907910

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Hormone-cytokine response
Pneumoperitoneum vs abdominal wall-lifting in laparoscopic cholecystectomy
T. Yoshida,1 E. Kobayashi,1,2 Y. Suminaga,1 H. Yamauchi,1 T. Kai,1 N. Toyama,1 H. Kiyozaki,1 A. Fujimura,2 M. Miyata1
1 2

Department of Surgery, Omiya Medical Center, Jichi Medical School, 3311-1 Kawati-gun, Tochigi 329-04, Japan Department of Clinical Pharmacology, Jichi Medical School, 3311-1 Kawati-gun, Tochigi 329-04, Japan

Received: 10 November 1996/Accepted: 19 February 1997

Abstract Background: Changes in blood hormone and cytokine were investigated in patients who underwent laparoscopic cholecystectomy via insufflation (CO2 group) vs those who had abdominal wall-lifting (Air group). Methods: Seventeen female patients with cholecystolithiasis were randomly divided into two groups. Peripheral blood samples were obtained during perioperative period, and plasma hormone levels (ACTH, cortisol) and serum cytokine levels (TNF, IL-1, IL-6, IL-10) were measured. Results: The number of circulating lymphocytes significantly decreased at 1 h after surgery in both groups, but the decrease in the CO2 group was significantly smaller than that in the Air group. There was no significant difference in hormone elevation between groups. Serum concentrations of IL-6 and IL-10 in the Air group were significantly higher than in the CO2 group. Conclusions: CO2 insufflation may reduce cytokine production in laparoscopic cholecystectomy. Key words: Laparoscopic cholocystectomy Pneumoperitoneum Wall-lifting Hormone Cytokine

cytokine production in patients who underwent laparoscopic cholecystectomy using the pneumoperitoneum method and abdominal wall-lifting method. The former procedure was performed by insufflation with carbon dioxide to protect the peritoneum from exposure to air, while in the latter, air was freely circulated. The results indicate that the pneumoperitoneum method produced less cytokines, while no significant difference in physiological conditions was observed between the two groups. The mechanisms of minimization of surgical damage by laparoscopic surgery are discussed. Materials and methods Patients
Seventeen female patients (35 to 65 years old; 49.0[2.25], mean[SEM]) with cholecystolithiasis without choledocholithiasis were selected. They were diagnosed by ultrasonographic examination and drip infusion cholangiography. Endoscopic retrograde cholangiography was also done in the cases where common bile duct was not clear from former examinations. Patients with liver dysfunction, acute inflammation, metabolic diseases, and cardiopulmonary diseases were excluded from this study. The subjects were randomly selected to undergo laparoscopic cholecystectomy using the carbon dioxide insufflation method (CO2 group) or the abdominal walllifting method [17] (Air group). The patients gave their informed consent after the protocol was approved by the appropriated Review Board. Selection took place on the day before the surgery, using numbered and sealed envelopes.

In recent years, laparoscopic cholecystectomy (LC) has become accepted world-wide because it is believed to be less traumatic than open cholecystectomy. A number of reports have shown that LC reduces postoperative pain and length of hospital stay and minimally impairs respiratory function, muscle performance, and immune function [7, 16, 19, 21]. There is a possibility that protection of the peritoneal cavity from exposure to air might be important to limit surgical damage due to the reduced production of cytokines after surgery [13, 22]. In this study, we compared hormonal and
Correspondence to: E. Kobayashi

Operation
All patients were anesthetized according to the same protocol; after fasting for 8 h, general anesthesia was achieved using thiopental sodium, vecuronium bromide, and sevoflurane. After the operation, pentazocine was given for pain relief if necessary. No blood products were required. Fluid replacement during and 24 h after surgery was similar in each group. In both groups, laparoscopic cholecystectomy was performed using 10-mm and 5-mm ports, as performed routinely [4]. Intraoperative cholangiography was performed on all patients. In the CO2 group, the intraabdominal pressure was kept under 8 mmHg during the operation as usually performed. In the Air group, a port for the laparoscope was initially inserted into the abdominal cavity through open minimal laparotomy; two Kirschner wires were placed through the skin, one in the upper portion of

908 Table 1. Patient characteristics and clinical parametersa Mean [SEM] CO2 group (n 9) Age (years) Height (cm) Weight (kg) Operation time (min) Anesthesia time (min) Pentazocine request (times) Time to full diet (days) Hospital stay (days)
a b

Air group (n 8) 50.5 [2.7] 155.3 [1.6] 61.6 [3.4] 121.3 [8.8] 202.5 [5.3] 1.4 [0.5] 1.8 [0.2] 6.0 [0.1]

p b 0.700 0.498 0.149 0.562 0.923 0.689 0.804 0.920

47.7 [3.7] 158.0 [1.0] 54.7 [2.6] 114.4 [9.3] 198.9 [10.7] 1.1 [0.4] 1.7 [0.1] 6.0 [0.4]

There was no significant difference in all characters and parameters Mann-Whitney U test

the umbilicus and the other in the midclavicular line at the right costal margin. The ends of those wires were pulled by means of a lifting handle and a lifting bar to provide a good view of the surgical side [17]. One surgeon (Y. S.), who is proficient in the technique for both methods, performed all procedures.

Blood samples and analyses

Peripheral blood samples were drawn at 5 points: at 6 A.M. on the day of the operation (Preop.), during the operation (Intraop.), 1 h after the operation (Postop.), and at 6 A.M. on the 1st and 3rd postoperative days (1POD, 3POD). Each sample was divided into three tubes. One of the blood samples in the EDTA tube was used for counting peripheral white blood cells (WBC) and lymphocytes using an auto-analysis system. The plasma obtained from the blood was stored at 20C until the measurement of adrenocorticotropic hormone (ACTH) and cortisol. Serum samples were used for measuring C-reactive protein (CRP), interleukin-1 (IL-1), interleukin-6 (IL-6), interleukin-10 (IL-10), and tumor necrosis factor (TNF). Cortisol and ACTH levels were determined using immunoradiometric assay (IRMA) [1, 2]. CRP was measured by turbidimetric immunoassay [5]. Concentrations of serum IL-6, IL-10, and TNF were determined using an enzyme-linked immunosorbent assay [8]. IL-1 was also determined using IRMA.

Fig. 1. Change of peripheral lymphocyte number in CO2 vs Air groups over time. The decrease of lymphocyte number in the Air group was significantly larger at 1 h and 1 day after operation than in the CO2 group. CO2 group; Air group; *p < 0.05, **p < 0.01 (Mann-Whitney U test). : Lymphocyte number of preoperative sample was equaled to 100%. Fig. 2. The changes in serum CRP after laparoscopic cholecystectomy. The serum CRP level in the air group was significantly higher at 1 and 3POD than those in the CO2 group. : CO2 group, : Air group; *p < 0.05(Mann-Whitney U test). : The minimal detectable value of CRP was 0.3 mg/dl.

Statistical analysis
Data are expressed as mean (standard error of the mean [SEM]) for each study group. Statistical analysis was done by the Mann-Whitneys U test, and the significant difference was set at less than 5%.

Results The clinical parameters of the patients are summarized in Table 1. There was no significant difference in those values between the CO2 and Air groups. The changes in the circulating lymphocyte count after the operation are shown as percentages: [number of peripheral lymphocytes in postoperative course/number of lymphocytes in preoperative condition] 100 (%) and are summarized in Fig. 1. The lymphocyte counts declined transiently in both groups. However, the rate of decrease in the Air group was significantly large at Postop. and 1POD (CO2 vs Air group: 82.1% [3.6] vs 69.2 [4.03] p 0.027, and 100.1 [3.9] vs 82.9 [2.72] p 0.008 at Postop. and 1POD, respectively; mean [SEM]). The changes in the level of serum CRP are shown in Fig. 2. The serum CRP level in the Air group was significantly

higher than in the CO2 group on 1POD and 3POD (0.94 mg/dl [0.26] vs 3.92 [1.12] p 0.027, and 1.7 [0.43] vs 6.8 [1.44] p 0.012, at 1POD and 3POD, respectively). The changes in the level of plasma hormones after surgery are shown in Fig. 3. Both ACTH and cortisol increased during and immediately after surgery, then decreased to the preoperative level on 1POD. There was no significant difference in the levels of these two hormones between the CO2 and the Air groups. No definite changes in serum IL-1 and TNF concentration were detected in either group. Serum IL-6 and IL-10 were detected in the peripheral blood and the changes in serum concentrations of these cytokines are shown in Fig. 4. The concentration of IL-6 in the Air group significantly increased postoperatively, in comparison with the CO2 group (6.64 pg/ml [1.38] vs 22.1 [7.38] p 0.0039 and 10.4 [2.75] vs 71.4 [39.33] p 0.0021, at Postop. and

909

Fig. 3. Hormonal changes after laparoscopic cholecystectomy. Changes of serum ACTH and cortisol showed a similar pattern in both groups. : CO2 group, : Air group. : The minimal detectable value of ACTH was 5 pg/ml. Fig. 4. The changes of serum IL-6 and IL-10 after laparoscopic cholecystectomy. : CO2 group, : Air group; *p < 0.05, **p < 0.01 (Mann-Whitney U test). The concentrations of IL-6 and IL-10 in the Air group showed higher level groups showed higher level at 1 day after surgery than those in the CO2 group. : The minimal detectable value of IL-6 and IL-10 were 0.4 pg/ml and 0.5 pg/ml, respectively.

1POD, respectively). Serum IL-10 concentration in the Air group was also significantly higher than in the CO2 group on 1POD (2.23 pg/ml [0.86] vs 6.27 [1.44] p 0.031). Discussion LC has rapidly become the standard operation for cholecystolithiasis around the world. Recent reports have demonstrated that LC is less traumatic than open cholecystectomy from the standpoint of physiological, immunological, and inflammatory response [911, 16, 19]. It has been suggested that the reduction of surgical damage might be due to the smaller incision required [9, 10]. There is also a possibility that protecting the peritoneal cavity from exposure to air might minimize the inflammatory response [13]. We compared two methods for performing LC: the pneumoperitoneum method, which utilizes insufflation with carbon dioxide, and the abdominal wall-lifting procedure, in which air circulates from the ports. The former has been considered to

provide a good view of the gallbladder and the latter has the merit of causing smaller changes in cardiopulmonary function during the operation [15]. No significant difference was noted postoperatively between the two groups in physiological factors such as postoperative pain and length of hospital stay. However, the changes in the number of circulating lymphocytes, CRP, IL-6, and IL-10 were significantly larger in the Air group than in the CO2 group. Surgical damage impairs both quality and quantity of peripheral lymphocytes, and this impairment causes immunosuppression after surgery [14, 18]. It has been reported that surgical treatment triggers the release of inflammatory cytokines from peritoneal macrophages and the synthesis of acute-phase proteins, and that these cytokines might be indicative of the degree of surgical damage [6]. Recent experimental studies have shown that air contamination of the peritoneal cavity regulates early inflammatory responses and induces immunosuppression and tumor growth [3, 20]. In another paper, we compared the changes in circulating lymphocytes, hormones, and cytokines that occur in human

910

subjects following major and minor surgery and showed that the rate of decrease in the number of peripheral lymphocytes and the increase of IL-6 and IL-10 were well correlated with the degree of surgical damage [12]. In this paper, we demonstrated that the abdominal wall-lifting method causes a more productive response of those cytokines and greater changes in circulating lymphocytes than the pneumoperitoneum method. The present study is the first clinical observation to compare the hormone and cytokine response to air exposure vs CO2 insufflation of the peritoneal cavity.
Acknowledgments. We thank Prof. I. Sakurabayashi (Department of Clinical Laboratory, Jichi Medical School, Omiya Medical Center) for technical assistance and advice. This work was partially supported by a Grant of Public Trust Fund for the Promotion of Surgery.

References
1. Addison GM, Hales CN (1971) Two-site assay of human growth hormone. Horm Metab Res 3: 5960 2. Addison GM, Hales CN, Woodhead JS, et al. (1971) Immunoradiometric assay of parathyroid hormone. J Endocrinol 49: 521530 3. Allendorf JDF, Bessler M, Kayton ML, et al. (1995) Increased tumor establishment and growth after laparotomy vs laparoscopy in a murine model. Arch Surg 130: 649653 4. Bailey RW, Zucker KA, Flowers JL, et al. (1991) Laparoscopic cholecystectomy. Ann Surg 214: 531540 5. Bergstrom K, Lefvert AK (1980) An automated turbidimetric immunoassay for plasma proteins. Scand J Clin Lab Invest 40: 637640 6. Cruickshank AM, Fraser WD, Burns HJG, et al. (1990) Response of serum interleukin-6 in patients undergoing elective surgery of varying severity. Clin Sci 79: 161165 7. Delaunay L, Bonnet F, Cherqui D, et al. (1995) Laparoscopic cholecystectomy minimally impairs postoperative cardiorespiratory and muscle performance. Br J Surg 82: 373376 8. Engvall E, Perlmann P (1972) Enzyme-linked immunosorbent assay, ELISA.III. Quantitation of specific antibodies by enzyme-labeled antiimmunoglobulin in antigen-coated tubes. J Immunol 109: 129135

9. Glaser F, Sannwald GA, Buhr HJ, et al. (1995) General stress response to conventional and laparoscopic cholecystectomy. Ann Surg 221: 372380 10. Griffith JP, Everitt NJ, Lancaster F, et al. (1995) Influence of laparoscopic and conventional cholecystectomy upon cell-mediated immunity. Br J Surg 82: 677680 11. Kloosterman T, von Blomberg ME, Borgstein P, et al. (1994) Unimpaired immune functions after laparoscopic cholecystectomy. Surgery 115: 424428 12. Kobayashi E, Yamauchi H (1997) Interleukin-6 and a delay of neutrophil apoptosis after major surgery. Arch Surg 132 (in press) 13. Kobayashi E, Yoshida T, Yamauchi H, et al. (1995) Immune function in the patients undergoing open vs laparoscopic cholecystectomy. Arch Surg 130: 676 14. Lennard TWJ, Shenton BK, Borzotta A, et al. (1985) The influence of surgical operations on components of the human immune system. Br J Surg 72: 771776 15. McDermott JP, Regan MC, Page R, et al. (1995) Cardiorespiratory effects of laparoscopy with and without gas insufflation. Arch Surg 130: 984988 16. McMahon AJ, Russell IT, Baxter JN, et al. (1994) Laparoscopic versus minilaparotomy cholecystectomy: a randomised trial. Lancet 343: 134138 17. Nagai H, Kondo Y, Yasuda T, et al. (1993) An abdominal wall-lift method of laparoscopic cholecystectomy without peritoneal insufflation. Surg Laparosc Endosc 3: 175179 18. Park KS, Brody JI, Wallace HA, et al. (1971) Immunosuppressive effect of surgery. Lancet :5355 19. Redmond HP, Watson WG, Houghton T, et al. (1994) Immune function in patients undergoing open vs laparoscopic cholecystectomy. Arch Surg 129: 12401246 20. Watson RWG, Redmond HP, McCarthy J, et al. (1995) Exposure of the peritoneal cavity to air regulates early inflammatory responses to surgery in a murine model. Br J Surg 82: 10601065 21. Yoshida T, Kobayashi E, Yamauchi H, et al. (1995) Laparoscopic cholecystectomy minimally impairs postoperative cardiorespiratory and muscle performance. Br J Surg 82: 996997 22. Yoshida T, Kobayashi E, Miyata M, et al. (1996) Increased tumor establishment and growth after laparotomy vs laparoscopy in a murine model. Arch Surg 131: 219

Surg Endosc (1997) 11: 942943

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Hand-assisted laparoscopic splenectomy for hydatid cyst

K. E. W. Ballaux, J. M. Himpens, G. Leman, M. R. P. Van den Bossche
Department of Abdominal and Thoracic Surgery, Sint-Blasius Ziekenhuis, Kroonveldlaan 50, 9200 Dendermonde, Belgium Received: 27 September 1996/Accepted: 19 November 1996

Abstract. Splenic hydatidosis is a rare condition. We performed a hand-assisted laparoscopic splenectomy for a large hydatid cyst localized in the center of the spleen. We discuss the advantages of the helping hand. Key words: Hydatidosis Laparoscopy Spleen Splenectomy

Discussion Although the procedure is difficult, advantages of laparoscopic splenectomy are obvious: less pain, fast recovery of bowel function, short hospital stay, and improved cosmetic outcome [4, 5]. The Dexterity Glove and Pneumosleeve (Dexterity Inc.) is a recent adjunct in laparoscopic surgery that allows hand manipulation with easier exposure of the structures and better control of minor or major bleedings [3]. In laparoscopic splenectomy, seizing and manipulating the spleen is a major problem, especially in the case of splenomegaly [6, 7]. The Dexterity Glove and Pneumosleeve (Dexterity Inc.) proved to be a valuable tool in dealing with this problem [3]. The 8-cm minilaparotomy used for the helping hand was necessary to remove the spleen intact since morcellation could have spread the echinococcus scolices into the abdominal cavity. Introduction of the enlarged spleen into a bag may prove extremely difficult and can be avoided by this technique [2].

Splenic hydatidosis accounts for 0.58% of total hydatidosis, with liver and pulmonary foci accounting for 6070% and 3035%, respectively [1, 8, 9].

Case report
A 26-year-old Turkish man was admitted to our Emergency Department with diffuse urticarial wheals the day after taking a copious, spicy meal. Upon admission the patient was nauseous and had a temperature of 38.7C. Biochemistry was normal except for eosinophilia of 9.9% (760/mm3). Xray of the abdomen demonstrated two shell-shaped calcifications in the left subcostal region. Ultrasonography of the abdomen revealed a 5-cm septated cystic mass in the spleen. A CAT scan showed an enlarged spleen containing a multilocular and partially calcified unruptured cyst. Volume of the spleen was estimated at 1,000 cc. Echinococcal radioallergosorbent test (RAST) was positive. A hand-assisted laparoscopic splenectomy was performed. Four trocar portals were made (Fig. 1). An 8-cm incision was made in the left lower quadrant for application of an airtight system allowing introduction of ones hand and forearm (Dexterity Glove and Pneumosleeve; Dexterity Inc., Carpinteria, CA, USA.). This system allowed us to manipulate the spleen and reflect the colon and small intestine in order to expose the splenic hilum. The spleen was mobilized gradually: Gastrocolic and colosplenic ligaments were divided and the hilum was dissected. The splenic artery was ligated first and the whole spleen turned blue. The splenic vein was clipped and divided. Finally the splenodiaphragmatic attachments were released. The spleen was readily removed intact through the minilaparotomy. The procedure lasted 200 min. Blood loss was estimated at 100 cc. The postoperative course was uneventful; the patient left our department on the 4th postoperative day.

Correspondence to: J. M. Himpens

Fig. 1. Sites of trocar portals and minilaparotomy.

943

this case because the cyst was large compared to the splenic volume and was localized in the center of the spleen [8] (Fig. 2). References
1. Abi F, El Fares F, Khaiz D, Bouzidi A (1989) Les localisations inhabituelles du kyste hydatique. A propos de 40 cas. J Chir 126: 307312 2. Dexter SPL, Martin IG, Alao D, Norfolk DR, McMahon MJ (1996) Laparoscopic splenectomy. The suspended pedicle technique. Surg Endosc 10: 393396 3. Gossot D, Meijer D, Bannenberg J, De Wit L, Jakimovicz J (1995) La sple nectomie laparoscopique revisite e. Ann Chir 49: 487489 4. Grossbard ML (1996) Is laparoscopic splenectomy appropriate for the management of hematologic diseases? Editorial Surg Endosc 10: 387 388 5. Phillips EH, Carroll BJ, Fallas MJ (1994) Laparoscopic splenectomy. Surg Endosc 8: 931933 6. Terrosu G, Donini A, Silvestri F, Petri R, Anania G, Barillari G, Baccarani U, Risaliti A, Bresadola F (1996) Laparoscopic splenectomy in the management of hematological diseases. Surgical technique and outcome of 17 patients. Surg Endosc 10: 441444 7. Trias M, Targarona EM, Balague C (1996) Laparoscopic splenectomy: an evolving technique. Surg Endosc 10: 389392 8. Uriarte C, Pomares N, Martin M, Conde A, Alonso N, Bueno M (1991) Splenic hydatidosis. Am J Trop Med Hyg 44: 420423 9. Vara-Thorbeck R, Rosell J, Ruiz-Morales M (1991) Milzechinokokkose. Ihre konservative chirurgische Behandelung. Zentralbl Chir 116: 14111416

Fig. 2. Longitudinal section of the spleen with large hydatid cyst in its center.

Recently some authors have advocated conservative spleen surgery for echinococcus cysts, i.e., cyst enucleation and marsupialization [9]. This possibility was excluded in

New technology
Surg Endosc (1997) 11: 946956

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Technology and principles of tomographic image-guided interventions and surgery

A. Melzer, A. Schmidt, K. Kipfmu ller, D. Gro nemeyer, R. Seibel
MRI, Institute for Diagnostic and Interventional Radiology, Medical Computer Science University Witten/Herdecke, Schulstrasse 10, 45468 Mu lheim/Ruhr, Germany Received: 17 October 1996/Accepted: 21 October 1996

Abstract. Image guidance of instruments can be performed by fluoroscopy, ultrasound imaging, computed tomography (CT), magnetic resonance imaging (MRI), and by stereotactic navigation. During the last year, there has been significant progress in MRI instrument guidance in the field of interventional radiology. Our first 168 clinical cases of MRI-guided interventional procedures, e.g., aspiration biopsy of neoplasms and tomographic microtherapy with local interstitial chemo-ablation, confirm the feasibility of MRI guidance. An expansion of MRI guidance during surgical endoscopic procedures is currently under evaluation and the initial results of this development are presented. Tomographic-guided surgery and the implementation of MRI or CT scanning within the environment of an operating room (OR) entail specific technological requirements and OR design considerations. Fast sequences, interventional protocols, in-room monitor, as well as MR-compatible probes, cannulae, catheters, instruments, endoscopes, and auxiliary equipment that are necessary for this combined surgical image-guided approach are described. Key words: Interventional MRI guidance Tomographic microtherapy Endoscopic surgery Image-guided surgery

In open surgery the experienced operator is able to interpret the patients pathological anatomy below the surface view within the operative field. By virtue of the tactile and force feedback, he is able to identify and dissect the correct tissue planes and avoid damage to vital structures. However, the surgical open access is invasive and a magnified wellilluminated view of the operative field is difficult. Endoscopic surgery provides a detailed, close-up, magnified view of the operative field. However, the remote
Correspondence to: A. Melzer

nature of the access leads to manipulative restrictions [3]. Precise remote dissection has to be performed by the surgeon in the presence of decoupling of the visual from the motor axes, reduced tactile and force feedback, and kinematic restriction imposed by the reduced degrees of freedom of the instruments. The evaluation of local anatomy and pathology is entirely dependent on visual inspection of the image on the video monitor screen. The image is generated only from the light reflected from the surfaces of organs and tissue. The surfaces of the operative field in conjunction with underlying supporting tissues and organs present the volume of interest. Tomographic visualization of the complete operative tissue volume can be achieved if telescopic inspection is combined with an additional intraoperative tomographic imaging system. Two main issues then arise. The first concerns the type of imaging technology which best suits the operative task. The second relates to the need for and benefit to be derived from a stereotactic navigation system that would virtually generate the instrument within peroperatively gained tomographic data. CT and MRI carry the greatest potential for the provision of high-quality on-line and real-time intraoperative tomographic visualization of the operative tissue volume. This new approach can be defined as tomographic therapy, or surgical tomography, or tomographic surgery. In the following account, current imaging systems, techniques, and technologies for surgical tomography are described with special emphasis on MRI-guided interventions. Visualization of the operative field Modern video-endoscopes and laparoscopes provide an excellent image of the surfaces of the operative field in real time. However, precise and targeted intraoperative control of the surgical process and navigation of the tools also require on-line visualization of the tools, anatomical planes, pathology, and physical-chemical changes without any adverse effects. Four of the current diagnostic imaging

947

Fig. 1. Fluoroscopic images are acquired on-line and in real time but involve a significant amount of radiation exposure. Display of a nonspatially defined summation image of the objects within the x-ray beam is obtained.

techniquesfluoroscopy, ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI) meet these requirements to a varying extent and, furthermore, can be deployed during surgery. With the exception of fluoroscopy, these imaging techniques have, until recently, been used primarily for diagnostic imaging. Technical innovations, particularly in the field of MRI, have enabled these imaging modalities to be used to image-guide radiological and surgical interventions. In endoscopic surgery the provision of a colored surface view by the telescope combined with near-real-time tomographic information on the tissue composing the operative volume can lead to a significant increase in the precision of the therapeutic effect, together with an enhancement of the safety margin, particularly in anatomically crowded and dangerous areas. The appropriate intraoperative application of fluoroscopy, US, CT, and MRI to the respective surgical procedure requires analysis and evaluation of each imaging technique as well as further dedicated development.

Fig. 2. Ultrasound imaging requires acoustic coupling to the scanned tissue and is altered by sound shadows. Only the tip of an instrument can be displayed.

Ultrasound Ultrasound imaging is based on the attenuation and scattering of the reflected sound waves generated by a transducer array which emits and receives the reflected waves. A sectional image is gathered by the transducer array and displayed in real time. The image is not tomographic because the radiation is unidirectional and the scanner is not moved around the patient. Thus the US image lacks spatial resolution and does not permit a complete view behind structures that produce a sound shadow such as gallstones or airfilled compartments (Fig. 2). Penetration of bone is minimal and differentiation of soft-tissue planes is limited. Resolution and penetration depth are strongly dependent on the wave length. The Doppler shift is used to detect the velocity of a moving column of blood, and systems are availabel which display in color (e.g., blue for venous and red for arterial blood) flow to and away from the probe (color Doppler). Ultrasound contrast media (micro air bubbles) facilitate tissue discrimination significantly. The advances in 3-D US rendering during the past 5 years are promising but require further refinement before routine clinical application. Contact ultrasound of organs with high-resolution linear-array probes (7.510 Mhz) during laparoscopic surgery is now used routinely in many centers during biliary surgery, for the staging of tumors, during laparoscopic cryotherapy, and for the localization of islet cell tumors. External ultrasound-guided insertion of the primary trocar cannula has been used in the retroperitoneal approach to endoscopic adrenalectomy, and this technique of ultrasoundguided port insertion carries considerable benefits. As the current systems are not stationary, interactive instrument guidance requires manual control or coupling of the instrument, e.g., biopsy needle or cryoprobe, to the US scanning transducer head. Evaluation of the percutaneous access tra-

Fluoroscopy The fluoroscopic image represents a summation of the x-ray absorption by all the structures that fall within the radiation beam. Display of soft tissue and vascular structures and hollow visceral organs is inadequate and requires the use of contrast media. Intraoperative fluoroscopy is used routinely in trauma surgery for the reduction of fractures. Intraoperative cholangiography identifies anatomical variations and pathology of the biliary system (Fig. 1). Intravascular interventions are routinely performed with flouroscopic guidance. The major advantages of fluoroscopy are real-time and on-line visualization and optimal spatial resolution. Its drawbacks include significant x-ray exposure, administration of allergenic contrast media, difficulties with 3-D rendering, and reproducibility of the summated images because spatial orientation is limited, although the use of biplanar fluoroscopy overcomes this limitation to some extent.

948

Fig. 3. CT scanning is provided in axial and angulated paraxial direction. A The gantry opening of a CT scanner limits intraoperative imaging and use of instruments. However, the excellent spatial resolution provides accurate display of both the instruments. However, the excellent spatial resolution provides accurate display of both the instrument and operative volume, B, C, CT-guided endoscopic thoracic sympathectomy.

jectory with US guidance is difficult and requires considerable expertise, and its efficiacy is thus user dependent. Despite these current limitations, which are likely to be overcome as the technology advances, US scanning provides a noninvasive image guidance system of proven value for percutaneous biopsy, puncture, and drainage procedures. More recently, US guidance has been reported during neurosurgery and for spinal fusion [24].

mm. The major drawbacks relate to X-ray exposure and to only monoplanar imaging in the axial and paraxial directions (Fig. 3). Electron beam tomography (EBT, Siemens/ Imatron) has the advantage over conventional CT scanning in terms of the scanning time, which is reduced to 30 msec by this machine (Fig. 4). In addition, the gantry opening is larger. However, the current image reconstruction time by EBT is approximately 70 s. Thus the latest conventional CT scanners, e.g., the Somatom S4 real time (Siemens), are faster in terms of the actual image representation.

Computed tomography CT provides reproducible high-resolution images with good differentiation of soft tissues. As the x-ray beam tube is rotated around the patient, the gathered data on absorption in relation to the position of the beam allow the calculation of pixels that are displayed in gray shades. In conjunction with the slice thickness, the primary two-dimensional nature of the tomographic images permits spatial orientation and 3-D rendering. During helical CT scanning the patient is moved while the x-ray beam is rotated. The acquired data spiral greatly facilitates 3-D reconstruction. Although the patient has to be moved in and out of the scanner and the opening of the CT gantry limits surgical interventions, CT is a contact-free means of reproducible and precise interventional imaging with a resolution < 1.0 Magnetic resonance Imaging In contrast to the other imaging systems, MR images are reconstructed from signals emitted from relaxation of hydrogen proton spins. The signal intensity depends on the hydrogen proton content and type of the chemical binding in the respective tissue. Thus the image is not only morphological but also contains information on the physical and chemical characteristics. Aside from superior soft-tissue contrast, various effects, i.e., temperature changes, tissue necrosis, and blood flow, can be exploited. Instruments, however, are difficult to display with MRI. Due to the volumetric nature of the MR images, multiplanar and 3-D images can be obtained. The application of fast and

949

Fig. 4. Electron beam tomography (EBT) dispenses with the need for rotating x-ray tubes and provides ultrafast scanning of different planes simultaneously within 30100 msec, whereby motion artifacts and radiation exposure are reduced significantly.

ultrafast sequences enables breathholding and fluoroscopy-like imaging. The MR image acquisition parameters can be externally controlled via computer interfaces, enabling an interactive slice orientation to the instrument as this is moved. Scanning time with MRI is relatively unlimited as IDA regulations limit the time of exposure to MRI to 7.6 Tesla/hour/day ionizing radiation is not involved, and no side effects have been reported provided precautions are taken to remove all metal objects and exclude certain categories of patients. The new open magnet designs improve access to the patient (Fig. 5AF).

quently, the 3-D data are used to display the anatomical internal structures by superimposing them onto the patients skin (Fig. 6B). In this fashion, the precise outline of the tumor and important or vital anatomical structures which must be preserved during surgery are displayed throughout the operationproviding a safety road map to the surgeon. Stereotactic navigation One of the principles of image-guided surgery is based on stereotactic methods which have been used traditionally by neurosurgeons. The trajectory of the insertion path is evaluated initially by conventional radiological skull x-rays or CT or MRI scans. The stereotactic procedure is then performed by means of a frame mounted on the patients head that allows advancement of a probe along the line of the trajectory to the predetermined distance. The instrument is advanced and positioned without direct control. Only predetermined distances and angles verify the correct position. The concept of stereotactic instrument guidance has been advanced by developments that involve 3-D rendering of a whole set of CT or MR images and optoelectronic tracking of the instrument. Small infrared LEDs (light-emitting diodes) are attached to the instruments handle and these are tracked by linear cameras. These feed this spatial x,y,z information to the controlling computer, which thus receives information on the instruments position and orientation. The system then displays a virtual instrument (corresponding to the real instrument used by the surgeon) on the respective image of the patient, and thus maneuvering of the instrument can be controlled with great precision and without the need of a stereotactic frameframeless stereotaxis (Fig. 7). This technology is currently being applied in ENT surgery, orthopedics, and neurosurgery [14, 20]. However, changes within the real operative volume during surgery are not detected. Thus either update imaging or an additional on-line imaging system is required. Techniques of tomographic-guided interventions CT image guidance Intraoperative CT scanning has no significant influence on the operative environment and the nature of the instruments

Three-dimensional imaging The imaging systems previously described provide crosssectional uniplanar images. Difficult diagnosis and preoperative surgical planning are facilitated by 3-D reconstruction. For this purpose, the tomographic data are subjected to further computed processing. This includes identification and selection of the tissue and organs to be reconstructed. The classification of tissues relates to the respective gray shade. The preselected gray shades are displayed in color codes. This segmentation is usually performed manually in each of the tomographic slices. Automated segmentation requires high-performance workstation computers and errors due to artifacts are likely with the current generation of software. The final step in the reconstruction is the 3-D rendering of the respective color-coded segments to the volumetric model that can be displayed in various angles or cuts (Fig. 6A). For diagnostic applications, the time required for the 3-D reconstruction of the cross-sectional images is not important as it is for the conduct of the surgical operations. The intraoperative utilization of 3-D reconstruction of the pathological anatomy requires fast processing of the image data in near real time. In the future 3-D images can be used intraoperatively by means of the registration process whereby the images are projected on to the patients corresponding anatomic region. A video image of the patient or a laser surface scan is aligned with the computer-derived surface of the rendered 3-D reconstruction. Prominent anatomical landmarks serve for the video optic matching of the two images. Subse-

950

Fig. 5. New open MRI scanners facilitate access to the patient during scanning and provide multiplanar images with superior soft-tissue contrast, display of physical and biochemical changes, and angiography without contrast agents. Current designs of open and interventional MRI units: A ACCESS Toshiba, San Francisco, CA, USA. B Magnetom Open, Siemens, Erlangen, Germany. C Signa SP, GE Medical Systems, Milwaukee, Wis, USA. D Signa Profile, GE Medical Systems, Milwaukee, Wis, USA. E Horizon, Picker, Cleveland OH, USA. F Airis, Hitatchi, Tokio, Japan.

used during the procedure. Selection of patients is also not restricted. Compatibility with CT guidance of materials for instruments and devices is governed by their radio-opacity. Large metal objects within the scanning field should be avoided. Titanium, aluminum, and synthetic and ceramic materials can be used. Smaller instruments and catheters must be sufficiently radio-opaque to be visible during the

procedure and can be doped with gold or platinum for this purpose. The major drawback of CT interventional guidance is the movement of the patient in and out of the gantry opening, as this restricts surgical instrument access to the operative area. Thus CT scanning has to be performed intermittently during these interventions. This limitation is largely

951

Fig. 6. A3-D reconstruction of CT images of intraspinal micro-endoscopy. B Projection of an MR image on the skin of the patient within an open magnet (F. Jolesz, Brigham and Womens Hospital, Boston, MA, USA).

sympathectomy, as well as local tumor biopsy and interstitial therapy are currently performed under combined CT guidancefluoroscopic control [8, 21, 22].

MR image guidance The first MR image-guided procedures were performed in 1986 with conventional closed (circular magnets) MR units [9, 13, 19]. The closed configuration of conventional MR units allows high field strength but restricts access to the patient. Phillips then developed a combination of a fluoroscopy unit and a high field scanner with 1.5 teslas (T) providing fast and ultrafast imaging. The technical realization of an open magnet MRI unit with large gap and adequate field strength (0.20.5 T) is governed by physical limitations and requires either cost-intensive superconducting coils or heavyweight resistive or permanent magnets weighing 1050 tons.
Fig. 7. Opto-electronic Instrument Navigation System SPOCS (Aesculap, Tuttlingen Germany).

Open MRI units (Fig. 5AF) In 1986 the first open imaging unit with a temple-like configuration (Access, 0.067 T) was developed by Toshiba (San Francisco CA, USA). The low magnet field reduces artifacts and facilitates application but has limited imaging capabilities compared to higher field units. Since then, new fast imaging with keyhole sequences and techniques using

overcome by combination with a flouroscopic C-arm (hybrid CT-fluoroscopy) which provides a real-time instrument guidance and adds coronal or sagittal viewing. Percutaneous laser decompression of spinal disk herniation, endoscopic treatment of intraspinal scars after disc surgery, endoscopic

952

MR-guided procedures have been established [7]. The new model of the Access will be equipped with superconducting magnets and a field strength of 0.35 T. The new Magnetom OPEN (Siemens, Erlangen, Germany), comprises an 0.2T resistive magnet with an opening gap of 440 mm and 260 circumferential horizontal access. Fast, good-quality imaging is provided through active field stabilization and water-cooled gradient coils. Interventional equipment is available that comprises a LCD in-room monitor console, special knee chair with heel pedals, foot pedal, cold light source, operative field illumination, and special fast sequence protocols. The autodisplay function permits instant view of the images while the measurement protocol is still running, thereby providing near-real-time visualization of the instrument during its advancement. Picker has released an open MR unit (Horizon) with a design similar to that of the Siemens Magnetom OPEN. The magnetic field strength is slightly higher (0.23 T) and the opening gap is 5 cm wider. The two-axis patient bed can be removed. Interventional equipment is under development. A dedicated MR-therapy system has been realized by GE Medical Systems (Milwaukee, Wisconsin USA) in collaboration with the group of F. Jolesz, Boston. The vertical open magnet configuration comprises two opposed superconducting 0.5-T magnet rings that enable vertical access through the operating gap of 560 mm. The patients bed can be moved through the rings in the longitudinal direction only. An additional MR observation monitor and an optoelectronic navigation system are incorporated in the upper section between the magnet rings [23].

Fig. 8. Intratumoral ethanol ablation of a sacral tumor.

Application of vertical or horizontal open MR configuration The selection of the appropriate opening direction of an MRI unit depends on the nature of the intervention to be performed. Two types of procedures can be performed currently: perioperative MRI-assisted and intraoperative MRI-guided interventions. MR-assisted procedures. The horizontal open magnets are ideal for perioperative MRI procedures that require frequent movement of the patient; i.e., surgery is performed outside of the magnet and scanning is conducted only for intermittent assessment of the operative tissue volume. This approach is ideal for procedures in which an additional on-line means of visualization is applied, e.g., video-endoscopy. Configurations that allow free access from three sides in ca. 260 and two axes of patients bed movement ease the change from MR scanning to the operating table. Horitzontal open Siemens scanners have been installed in the neurosurgical OR at the university hospitals of Heidelberg and Erlangen in Germany. MRI-guided procedures. Percutaneous interventions benefit from performance under MR control, preferably by means of interactive MR fluoroscopy (Fig. 8). We have

carried out 163 MRI-guided procedures such as aspiration biopsy, peridural and periarticular corticoid injection, facet joint treatment, chemical sympathicotomy, and intratumoral ethanol infiltration in 131 patients. This experience has confirmed the feasibility and benefits of MRI guidance. Procedures where drugs or energy are delivered to a specific site should be guided either by MR fluoroscopy or intermittent scanning with temperature-sensitive sequences (turbogradient echoes). If horizontal access to the operative volume is possible, the horizontal open-design MRI unit is appropriate. However, the vertical open MRI design gives more degrees of freedom and an ergonomically better access to the operative site. However, the transfer of the patient from the operating table to the MRI unit is more difficult because the patient has be to moved in and out through the magnet rings. This interferes with the instruments, supply lines, and anesthetic tubing. Intraoperative MRI guidance is also important during the placement of surgical instruments or endoscopes within the operative field, e.g., neurosurgical access to the ventricular system or cystic malformations. In Heidelberg the first MR-guided transcranial approach to an intracerebral cystic malformation has been performed successfully. At the Brigham and Womens Hospital in Boston, 50 CNS biopsies, sinus surgery, and open brain surgery have been performed under MR guidance using the vertical open unit [11]. Prerequisites and techniques of MRI guidance There are certain specific technological requirements for MRI-guided techniques and these procedures differ considerably from other interventional techniques. Development of MR-compatible instruments, devices, and ancillary equipment as well as optimization and adaptation of software and hardware of the current MR systems are the major objectives of the current developmental program. We have been working on the open 0.067-T Access (Toshiba) since 1987 [6] and in 1995 we started a collaboration program with Siemens (Erlangen, Germany) to evaluate the feasibility of interventions in the Magnetom OPEN system [18].

953

Patient selection and patient monitoring. In contrast to CTor US-guided interventions patient selection for MRIguided procedures is governed by physical considerations. The following conditions constitute absolute contraindications for MRI guidance:

Pacemakers Neurostimulation devices Known or suspected ferromagnetic aneurysm clips Large ferromagnetic implants

Patients with known history of ischemic heart disease or with symptoms of angina and ST changes should not be subjected to MRI-guided procedures because ST and T waves are distorted by the magnetic field, and for this reason, detection of a myocardial ischemic episode during the intervention may be obscured [10]. MRI-compatible anesthetic equipment, optoelectronic pulse/ECG sensors, pulse oximetry, and blood gas analysis are available (Magnetic Resonace Equipment Corporation, Bay Shore, NY, USA; Daum, Schwerin, Germany; and others). However, a complete and safe intraoperative monitoring system for patients during MRI-guided surgery requires additional development. Interventional MR sequences. Adaptation of fast sequences and the protocols to the nature of the intended procedure is necessary. Time of acquisition, TR, is the time between each pulse sequencean image with 256 lines requires 256 pulse sequences. Reduction of the lines to a rectangular field of view reduces the sequence time. Echo time is the time between end of the pulse sequence and the start of signal acquisition and leads to the different weighing of spin echo sequences. The motion artifacts are reduced by short echo times. Standard (SE) spin echo pulse sequences have long acquisition times but provide excellent soft-tissue contrast. T2-weighted spin echo images have long echo times. Gradient echo (GE) sequences are faster but more subject to artifacts. The weighting of fast gradient echoes is influenced by the flip angle. Thus 90 enhances the weighting of T1 and angles <30 result in T2* contrast. GE sequences are more sensitive to inhomogeneities of the magnetic field and thus the artifacts of instruments are larger than when SE sequences are used. The current fast sequences are 2-D FLASH (fast low-angle shot) and 2-D FISP (free induction steady precession). 3-D sequences acquire a volume (slab) that is segmented into slices. 3-D sequences are more sensitive to motion artifacts. The crucial component of fast and ultrafast imaging is the quality, strength, and speed of the gradient coils. The ultrafast echo planar imaging (EPI) has acquisition times in the millisecond range because all 256 lines of the image are acquired within one TR time. At present, EPI is applicable only to high field units. New approaches to faster imaging are keyhole and waveletencoded sequences. Both are based on the principle of avoiding complete acquisition of the data set required for the whole image. Instead acquisition is restricted to the volume of interest where, for example, the interventional instrument is moved [1, 25]. Although heat generation can be directly displayed by means of Turbo flash 2-D echoes, their image quality is, in

Fig. 9. MRI-compatible cannula with variable curvature (SmartGuide, Daum Schwerin).

low-field systems, still suboptimal. The signal changes detected by FISP 2-D or flash 2-D echoes are useful to evaluate to extent to localized tissue destruction. Image-guided tumor ablation requires precise control with standard T1 and T2 sequences. MRI-compatible materials, cannulae, and catheters. Instruments and probes that are inserted within the image plane reduce the signal intensity. Magnetic field inhomogeneties around the instrument by virtue of their paramagnetic effects decrease the relaxation time of the protons encircling the instrument such that their signal becomes undetectable within the echo time. Control of the artifact size can be achieved by electric current application [5]. The size of the artifact depends on the material, the applied sequences, and the angle of the probe within the magnetic field. Current MRI-compatible instruments are manufactured from antimagnetic stainless steel, titanium, and nickel-titanium alloys. MRI-compatible needles and microtrocar/cannulae with controllable artifact size and a variable-curvature guiding cannula (SmartGuide, Daum Medical, Schwerin) manufactured from these materials have been developed and tested. SmartGuide consists of a superelastic internal precurved cannula that is constrained inside a straight outer sleeve. The same concept has been applied to the variable curvature spatual according to Cuschieri. When the inner superelastic component is advanced beyond the end of the restraining outer sleeve, round the corner tissue passage is achieved (Fig. 9). A further development based on the SmartGuide principle has been applied to catheters for image-guided interventions. Intravascular interventions pose special problems for the imaging of static elements because the MR angiography sequences display the blood flow selectively. Conventional synthetic catheter materials are not visible in MR images. Various dotting techniques aimed at generation of the appropriate artifact which would allow MRI detection of these catheters are under evaluation.

Video optic equipment, trocars, and instruments for MRguided endoscopic surgery. The nature of MRI specifically imposes special requirements for instruments, devices, and electric equipment that can be used within the proximity of

954

Fig. 10. A Experimental setup of MR-guided laparoscopy. B MRI representation of an MRI-compatible trocar, laparoscope (Olympus Winter&Ibe, Hamburg, Germany), and a superealistic laparoscopic forceps (ND&C, Fremont, CA, USA).

the magnetic field. In collaboration with Olympus Winter&Ibe we have developed MRI-compatible rod-lens endoscopes. New light cables with a length of 5 m can be connected to a powerful cold light source placed outside the MR suite. A novel MR cold light source (Siemens) dispenses with the need for a ventilator, but its maximal current power is only 100 W. Prototypes of MRI-compatible endocameras have been released for evaluation by Olympus. The adaptation of the new in-room MRI monitor as an endoscopy monitor is possible and should provide adequate image quality because of high resolution and multisync capability. There is a need for MRI-compatible trocar/cannulae for MR-guided laparoendoscopic surgery. The MRI compatibility of trocar/cannulae has been evaluated in vitro (Fig. 10A,B). The ferromagnetic characteristic of forged or hardened stainless steel excludes the majority of trocars. In our tests we found the Step system (Innerdyne, Sunnyvale Ca, USA) [17] to be the most MRI-compatible access port. The initial puncture with a Veress needle and insufflation can be performed under MR fluoroscopic control. MR visualization of balloon dissection procedures [2] can be achieved by filling the balloon with a 23 mmol/l gadolinium/saline contrast agent. Major redesign and development work is needed for endoscopic instruments such as probes, hooks, forceps, and scissors to achieve MR compatibility. The substitution for the steel components of the endoscopic instruments such as the hinge complex with superelastic shape-memory Ni-Ti is an obvious solution [15, 16]. Interactive image control. The visualization of instruments during MRI-guided procedures is grossly inferior to that obtained by CT. The instrument has to be aligned completely with the orientation of the image plane. During MRIguided work, any movement by the patient during the insertion of the instrument (e.g., needle) leads to loss of the

needle artifact within the MRI image. This can be overcome by changing the slice orientation and position when a representation of the changed needle path can be achieved. This process is, however, time consuming and awkward. The principle of optoelectronic frameless instrument navigation can be implemented in the control system of an MRI unit whereby the tracking of an instrument within the magnetic field can be directly transferred to the slice orientation. (Fig. 11) Images can then be generated according to the position in alignment with the instrument or perpendicular to the distal tip of the instrument. The image plane orientation can be changed and redefined without moving the instrument itself. This ensures constant precise control of a surgical instrument during advancement. Care must be taken to avoid any bending of the instrument as the distorts the actual position of the instrument tip [23]. Tip tracking. The volumetric nature of the MRI signal detection allows orientation of specific signal frequencies in space. With the incorporation of a miniature radiofrequency coil at the tip of an instrument, local proton spin signals can be acquired. The signal source representing the tip of the instrument can then be localized within the coordinate system of the magnetic field and thus within the human body [4, 12]. This technique has been evaluated experimentally and is referred to as tip tracking. As the instrument is moved, the position of the signal moves accordingly and can be tracked with great precision. The current approach to tip tracking by Doumulin, Leung, and Debatin involves the need for two workstations to compute the three-dimensional coordinates of the coil and to display the position as a moving symbol in two planes of an acquired MR image. Thus direct on-line control is not possible. We have started a tip-tracking project that allows direct display of the instrument signal in a complete MR image, thereby enabling in-

955

Fig. 11. Schematic representation of opto-electronic instrument navigation in combination with tip tracking. This approach provides interactive MRI imaging in an axial or perpendicular direction relative to the instrument movement.

teractive slice orientation in the axial or perpendicular planes with the instrument. Conclusion The goal of safe conduct of minimal-access therapeutic procedures requires continuous reference image data sets and correspondence between these and the instruments position in relation to the anatomical structures within the operative tissue volume. Accurate control of each step of the procedure must be provided and unexpected changes in the anatomical geometry within the operative volume should be detected on-line. A step toward these ideals is the utilization of stereotactic navigation systems that display a virtual instrument within presurgically gained image data. The relatively stable proportions of the operative volume in sinus and intracranial surgery permits appropriate intraoperative localization of the instruments position. However, there are still problems to be solved during neurosurgery as removal of cerebrospinal fluid, hemorrhage, and cerebral edema induce shifts in the spatial configuration of the operative volume, and the actual process of tissue removal is not displayed. During thoracic and abdominal procedures there is a continuous and significantly greater shift of structures due to breathing, heart beat, intestinal peristaltic activity, etc. With the current systems, external scanning performed preoperatively cannot determine the displacement of the underlying thoracic or abdominal structures. The gross anatomical displacements that occur during surgery could only

be deduced from computer simulation. This situation is inadequate since despite being informative, it does not guarantee safety during the actual intervention. There is a need for intraoperative imaging capable of fast upgrading of the original data set of the operative volume during the surgical intervention. The optimal solution to this problem entails the optoelectronic instrument navigation with tip tracking as the reference for acquisiton of accurate intraoperative image planes (Fig. 11). Accurate online imaging of the operating instrument is crucial to the progress of image-guided surgery. The ideal imaging system should permit:

Unrestricted conduct of the interventional or surgical procedure Reproducibility On-line and real-time visualization Appropriate soft-tissue and bone discrimination Precise display of the instrument Volumetric data and spatial references Facility for multiplanar slicing and interactive selection of image orientation Flow and spatiothermal sensitivity

Ultrasound, fluoroscopy, and CT are established in clinical routine and have specific indications and advantages. US is simple to apply during procedures where the acoustic coupling to the tissue is possible. The high resolution of CT (< mm) and its excellent bone delineation make it the ideal image-guiding system in high-risk areas and when bony structures are part of the operative volume, i.e., intraspinal

956

and skull-base microsurgery. The instrument display within the image is significantly better with CT than MRI, and the position of the functional tip is visualized with much greater precision. For intravascular procedures, fluoroscopy is still unrivaled, although recent progress in MR angiography (MRA) has clearly demonstrated the great potential for MRA in the evaluation and treatment of vascular disease [16]. In principle MRI is capable of meeting most, if not all, the above-mentioned requirements, but the instrument display need significant improvement. What is needed is full integration of the imaging systems with the technology used during surgical operations and with the environment of the operating room. The selection of the imaging tool depends on the procedure to be performed and requires extensive clinical evaluation. Tomographic guidance and endoscopic surgery involve complementary technologies. The endoscope permits a colored, high-resolution, on-line, and real-time image of the operative field, i.e., accurate visualization of the surface anatomy of the internal organs and tissues. However, it does not provide any information on the parenchyma of organs, supporting tissues, and intraparenchymal pathology, blood vessels, and other vital structures (operative tissue volume). Visualization of the trajectory of the instruments pathway through the tissues cannot be provided by endoscopic imaging. Intraoperative tomographic imaging reveals all the essential components of the parenchymal structures within an operative volume. As soon as the operative field is accessed and exposed, the patient can be moved out of the scanner, and the operation can be conducted without any spatial restriction. During the course of the operation, should further tomographic visualization of the operative tissue volume become necessary, the patient is scanned again to obtain refreshed tomographic images. The optical navigation and tip-tracking systems can be used to give a baseline orientation within the operative volume in between the scanning periods. The majority of imaging systems are currently in hands of radiologists who do not usually perform surgical procedures. On the other hand, surgeons who perform the vast majority of operations and interventions are missing out since they undertake complex endoscopic operations without the benefit of tomographic imaging. The future lies in close interdisciplinary collaboration between these two specialties with the establishment of tomographic imageguided endosurgery. Even in the absence of such multidisciplinary units, as we approach the next millennium, surgeons have to become proficient with tomographic image guidance. The safety and precision of their surgical interventions will be enhanced; otherwise, they will lose out to the radiological surgeons or surgical radiologists of the next century. References
1. Bornsted A, Busch M, Wendt M, Seibel R, Gro nemeyer D (1996) Turbo accelerated asymmetric phase read exchange keyhole (Phreak) imaging as a method for fast and high contrast MR-guidedinterventions. MAGMA 4, Supplement ESMRMB 96, Nr 260, p 142 2. Coptcoat MJ (1995) Overview of extraperitoneal laparoscopy. End Surg 3: 13

3. Cuschieri A (1992) General principles of laparoscopic surgery. In: Cuschieri A, Buess G, Pe rissat J (eds) Operative manual of endoscopic surgery. Springer, Heidelberg pp 172175 4. Davis CP, McKinnon GC, Debatin JF, von Schulthess GK (1996) Ultra-high-speed MR Imaging. Eur Radiol 6: 297311 5. Glowinski A, Adam G, Bu cker A, Neuerburg J, van Vaals J, Gu nther RW (1996) Catheter visualization for interventional MR by actively controlled locally induced field inhomogeneities. Procedings, 51, ISMRM, New York 6. Groenemeyer DHW, Seibel RMM, Busch M, Rothschild PA, Kaufmann L (1989) Interventionelle Kernspintomographie. In: Groenemeyer DHW, Seibel RMM (eds) Interventionelle Computertomographie. Ueberreuter, Wien-Berlin, pp 296313 7. Gro nemeyer D, Seibel R, Kaufman L (1991) Low-field design eases MRI-guided biopsies. Diagn Imaging 3: 161165 8. Groenemeyer DHW, Seibel RMM, Melzer A, Schmidt A (1995) Image guided access techniques. End Surg 3: 6975 9. Hajek PC, Gylys-Morin VM, Stava J, van Sonnenberg E (1987) Localization grid for MR guided biopsy. Radiology 165: 825826 10. Hughes CW, Bell C (1993) Anesthesia equipment in remote hospital location. In: Rogers MC, Covino BJ, Tinker JH, Longecker DE (eds) Principles and practice of anesthesiology. Mosby, St. Louis, MO, pp 38 11. Jolesz F (1996) MRI guided therapy. MAGMA 4, Supplement ESMRMB 96, Nr 280, p 150 12. Leung DA, Debatin JF, Wildermuth S, Heske N, Dumoulin CL, Darrow RD, Hauser M, Davis CP, von Schulthess GK (1995) Real-time biplanar needle tracking for interventional MR imaging procedures. Radiology 197: 485488 13. Lufkin R, Teresi L, Hanafee W (1987) New needle for MR-guided aspiration cytology of the head and neck. AJR Am J Roentgenol 149: 380382 14. Melzer A (1996) Instruments for laparo-endoscopic surgery: conventional and intelligent. In: Tuoli D., Gossot J., Hunter J (eds) Endosurgery 15. Melzer A, Stoeckel D (1995) Performance improvement of surgical instrumentation through the use of Ni-Ti materials. Proceedings SMST-94, Monterey, CA, USA, pp 401410 16. Melzer A, Buess G, Cuschieri A (1994) Instruments for endoscopic surgery. In: Cuschieri A, Buess G, Perissat J (eds) Operative manual for endoscopic surgery, vol 2. Springer, Berlin 17. Melzer A, Kipfmu ller K, Groenemeyer D, Seibel R, Bue G (1995) Ports, trocars/cannulae and other access techniques. Semin Laparosc 2: 170201 18. Melzer A, Schmidt A, Kipfmu ller K, Deli M, Sto ckel D, Seibel R, Gro nemeyer D (1996) Prerequisites for magnetic resonance imageguided interventions in endoscopic surgery. Min Inv Ther Allied Techn 4: 260266 19. Mueller PR, Starck DD, Simeone JF (1986) MR guided aspiration biopsy: needle design and clinical trials. Radiology 161: 605609 20. Nolte LP, Zamorano LJ, Jiang Z, Wang Q, Langlotz F, Berlemann U (1995) Image-guided insertion of transpedicular screws. A laboratory set-up. Spine 20: 497500 21. Seibel R (1996) Image guided minimally invasive therapy. Surg End 22. Seibel RMM, Groenemeyer DHW (1994) Technique for CT guided micro-endoscopic dissection of the spine. End Surg 3(4): 226230 23. Silverman SG, Collick BD, Figueira MR, Khorasani R, Adams DF, Newmann RW, Topulos GP, Jolesz FA (1995) Interactive MR-guided biopsy in an open configuration MR imaging system. Radiology 197: 175181 24. Wenda K, Degreif J, Hu wel N, Kessel G (1995) Die Reposition der frakturierten Wirbelko rperhinterkante unter sonographischer Kontrolle. Langebecks Arch Chir 95: 477 25. Wendt M, Busch M, Lenz M, Seibel R, Gro nemeyer D (1996) Dynamic tracking algorithm for interventional MRI using wavelet encoding in 3D gradient-echo-sequences. MAGMA 4, Supplement ESMRMB 96, Nr 261, p 143

Case reports
Surg Endosc (1997) 11: 936938

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Covered metal stent for tumor obstruction of efferent loop recurrence after gastrectomy
H. Y. Cheung, S. C. S. Chung
Department of Surgery, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong Received: 13 June 1996/Accepted: 31 July 1996

Abstract. Reports the case of a 60-year-old woman who underwent R2 total gastrectomy, and subsequent palliation of painful symptom recurrence via a membrane-covered metal stent. Key words: Covered metal stents Efferent loop recurrence Gastrectomy

mal manner, and the position of the stent can be adjusted by pulling back the introducer shaft as necessary. The proximal end of the stent was left in the esophagus above the esophagojejunal anastomosis (Fig. 3). The deployment process was monitored by x-ray and an endoscope was inserted alongside the stent. At the end of the procedure injection of contrast down the endoscope showed free flow (Fig. 4). The patient was able to take solid food until she succumbed from her disease 6 months later.

Discussion Dysphagia resulting from jejunal loop obstruction by recurrent tumor can be a challenging clinical problem. We describe successful palliation in a patient with distressful dysphagia by implantation of a membrane-covered expandable metal stent. Case report
A 60-year-old woman underwent R2 total gastrectomy with Roux-en-Y esophagojejunostomy for carcinoma of the proximal stomach. Pathological examination showed T2N0 disease and the resection margins were free of tumor. She made an uneventful recovery but re-presented 2 years and 4 months later with progressive dysphagia. Ultrasound examination showed multiple liver metastasis. Endoscopy showed a healthy anastomosis but the efferent loop was invaded by tumor 5 cm from the esophagojejunal anastomosis. The endoscope cannot be passed through the obstruction. Injection of contrast material down the channel of the endoscope showed a short, narrowed segment with acute angulation (Fig. 1). The obstruction was negotiated with a Zebra guide wire (Microvasive, Boston Scientific Corporation, USA) and the narrowed segment was dilated with a 15-mmdiameter balloon catheter. The proximal and distal limits of the obstructed segment were then marked by injection of 0.5 ml of Lipiodol (Lipiodol Ultra-Fluid, Guerbet, France). A covered nitinol stent (Covered Ultraflex, Microvasive Boston Scientific Corporation, USA) with a 12-cm-long membrane-covered segment was then implanted in the following manner. A Savary guide wire (Savary Gillard, Wilson-Cook Medical, Inc., USA) is passed into the jejunum and the stent is inserted over the guide wire (Fig. 2). With the distal extent of the stent in a straight segment of the jejunum, the stent was deployed by removing the crocheted thread that held the stent onto the introducer. Deployment started distally and progressed in a proxi-

Correspondence to: S. C. S. Chung

After gastrectomy for carcinoma of stomach, local recurrence causing dysphagia occurs in about 20% of patients. Reoperation is seldom justified because of the limited life expectancy. Anastomotic recurrence may be amenable to endoscopic laser therapy if the obstruction if predominantly intraluminal and exophytic [1, 4, 7, 8]. Tumor invasion of the efferent loop is even more difficult to manage as the main bulk of the recurrent tumor is extraluminal or from extrinsic compression. Dilatation alone will only give very short-lived relief. Because the obstructed segment is often angulated, implantation of conventional plastic stents is hazardous and is unlikely to be successful because the straight stents cannot conform to the curvature of the jejunal loop. The availability of self-expanding metal stents has made esophageal stenting much safer because the stricture does not need to be dilated to a large size and the process of stent insertion is less traumatic. Perforation, the most feared complication of stenting, can largely be avoided. Knyrim [6] reported that no perforation occurred in metal stent placement in his randomized study. He concluded that selfexpanding metal stents are safer and more cost-effective than conventional plastic prostheses. However, tumor ingrowth through the instices of the uncovered expandable stent can cause recurrent dysphagia 26 months later [3, 5]. The second-generation expandable metal stents have membrane covers to overcome this problem [9, 10]. At the moment there are several types of commercial expandable stents available. They differ in the method of deployment, the radial force they generate, the mechanism of anchorage, and their ability to conform to a curvature [2].

937

Fig. 1. Short, narrowed segment with acute angulation (arrow) at the efferent loop. Fig. 2. The undeployed stent is in place with the proximal and distal limits of stricture marked by Lipiodol.

Fig. 3. The proximal end of the stent is placed in the esophagus above the anastomosis. Fig. 4. Free flow of contrast into the jejunal loop.

These factors must be considered in choosing the right stent for a particular patient. In our case, because of the acute angulation in the obstructed segment, the ability to withstand a sharp bend without buckling is the most important factor. The particular stent we used is deployed by pulling on a thread crocheted onto the stent to bind it to the introducer. The process of deployment starts distally, so it is possible to adjust the position of the stent proximally by pulling on the introducer during the deployment process.

This is important when precise deployment of the stent is necessary. Nitinol is an alloy of nickel and titanium that has thermal as well as shape memory; in addition, its superelastic behavior enables the stent to fit the contour of the stricture without buckling. The longer stent lengths (150 mm) as compared to Wallstent (110 mm) contributed an advantage in managing tortuous efferent loop recurrence and prevent stent migration. The main drawback of expandable metal stents is their

938

high cost. Whether their use is justified in patients with disseminated malignancy and limited life span remains a difficult financial and ethical problem. References

6.

7. 1. Buset M, des Marez B, Baize M, Bourgeois N, de Boelpaepe C, de Toeuf J, Cremer M (1987) Palliative endoscopic management of obstructive esophagogastric cancer: laser or prosthesis? Gastrointest Endosc 33(5): 357361 2. Chung SCS, Shin FG, Chan ACW, Lau JYW, Sung JY, Li AKC (1996) Physical properties of expandable esophageal stents. Gastrointest Endosc 43 (4): 332 3. Cwikiel W, Stridbeck H, Tranberg KG, von Holstein CS, Hambraeus G, Lillo-Gil R, Willen R (1993) Malignant esophageal strictures: treatment with a self-expanding nitinol stent. Radiology 187: 661665 4. Ell C, Riemann JF, Lux G, Demling L (1986) Palliative laser treatment of malignant stenoses in the upper gastrointestinal tract. Endoscopy 18 (Suppl 1): 2126 5. Ell C, Hochberger J, May A, Fleig WE, Hahn EG (1994) Coated and 8.

9.

10.

uncoated self-expanding metal stents for malignant stenosis in the upper GI tract: preliminary clinical experience with Wallstents. Am J Gastroenterol 89(9): 14965001 Knyrim K, Wagner HJ, Bethge N, Kermling M, Vakil N (1993) A controlled trial of expansible metal stent for palliation of esophageal obstruction due to inoperable cancer. N Engl J Med 329(18): 1302 1307 Krasner N, Barr N, Skidmore C, Morris AI (1987) Palliative laser treatment of malignant dysphagia. Gut 28(7): 792798 Mellow MH, Pinkas H (1985) Endoscopic laser therapy for malignancies affecting the esophagus and the gastroesophageal junction. Analysis of technical and functional efficacy. Arch Intern Med 145 (8): 14431446 Song HY, Do YS, Han YM, Sung KB, Choi EK, Sohn KH, Kim HR, Kim SH, Min YI (1994) Covered, expandable esophageal metallic stent tubes: experiences in 119 patients. Radiology 193: 689695 Wu WC, Katon RM, Saxon RR, Barton RE, Uchida BT, Keller FS, Rosch J (1994) Silicone-covered self-expanding metallic stent for the palliation of malignant esophageal obstruction and esophagorespiratory fistulas: experience in 32 patients and a review of the literature. Gastrointest Endosc 40: 2233

Surg Endosc (1997) 11: 966

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The author replies

I appreciate the comments made by Dr. Deitel and can only agree to all of his statements. After sharing the experience of Dr. W. Clark we have also started to perform the Rouxen-Y gastric bypass with the laparoscopic approach. With the laparoscopic instrumentation available today this procedure seems to be quite feasible and may be even easier to do than the loop gastrojejunostomy. The main concern today is, as stated by Dr. Deitel, to avoid any disastrous consequence that could be related to the laparoscopic technique. The long-term outcome must of course be at least as good as the results seen in open surgery. The adjustable band has rapidly become very popular since this procedure is easy to perform with the laparoscopic technique. There is still very limited long-term experience with this new procedure. I agree with Dr. Deitel that we should be very cautious when applying new techniques and not repeat mistakes already made by ourselves and others in open surgery.

H. Lo nroth
Department of Surgery Sahlgrens University Hospital S-413 45 Gothenburg Sweden

Surg Endosc (1997) 11: 915918

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The optimum pneumoperitoneum pressure for laparoscopic surgery in the rat model
A detailed cardiorespiratory study
R. Berguer, T. Cornelius, M. Dalton
Department of Surgery, University of California Davis, Sacramento, CA, USA Received: 18 December 1996/Accepted: 13 March 1997

Abstract Background: There is increasing interest in the rat model of laparoscopic surgery. This study evaluates the cardiorespiratory effects of increasing CO2 pneumoperitoneum (PP) in the rat. Methods: Nine Sprague-Dawley rats were subjected to CO2 PP at pressures of 2, 5, 10, and 15 mmHg or control (no PP) under anesthesia. Catheters were placed in the femoral artery and the jugular vein to measure heart rate (HR), blood pressure (MAP), and arterial pH, PCO2, PO2, and HCO3. A thermistor probe in the aortic arch measured cardiac output (CO) and blood temperature (BT). Results: CO2 PP had no effect on CO, MAP, or BT at any pressure. CO2 PP greater than 5 mmHg caused significant bradycardia and CO2 PP greater than 10 mmHg caused significant respiratory acidosis. Conclusions: CO2 PP pressures above 10 mmHg in rat should be avoided when performing laparoscopic surgery in the rat model. Key words: Rats Pneumoperitoneum Cardiovascular Laparoscopy

toneum decreases tidal volume and assisted hyperventilation is needed to compensate for this effect under anesthesia. Survival laparoscopic procedures in the rat also require a pneumoperitoneum but, due to the difficulty of endotracheal intubation, the anesthetized rat must breathe spontaneously. Under these conditions, laparoscopic surgery in rodents with a standard pneumoperitoneum pressure of 15 mmHg presents a clear risk of respiratory depression from the combined effects of anesthesia and the pneumoperitoneum which could result in circulatory shock and respiratory acidosis. Therefore, the purpose of this study is to determine the maximum pressure for CO2 pneumoperitoneum pressure in the anesthetized and spontaneously breathing rat that allows normal cardiac and respiratory function.

Materials and methods

Male Sprague-Dawley rats weighing between 330 and 450 g were studied under general anesthesia consisting of 12% inhaled isoflurane in a 50% mixture of O2 and N2O delivered at a gas flow rate of 2 l/min via a semiclosed anesthetic circuit and close-fitting masks. We selected this anesthetic regimen because we have previously demonstrated that this inhaled anesthetic mixture results in excellent analgesia and muscle relaxation during laparoscopic and open surgery in rats while causing significantly less respiratory depression than standard pentobarbital anesthesia [7]. The time from the initial administration of anesthetic until surgical anesthesia was achieved was recorded as the induction time. Rats were placed on an electric warming blanket to maintain their rectal temperature in the range of 36.537.5C. A polyethylene femoral arterial catheter (outside diameter 0.965 mm) was placed for blood pressure recording and arterial blood gas sampling. The internal jugular vein was similarly cannulated for central venous pressure (CVP) measurements and saline infusion. No intravenous infusion was given in the present study since rats that undergo experimental survival laparoscopic procedures receive no maintenance intravenous fluid. A thermistor probe was placed via the right carotid artery into the aortic arch for measurements of cardiac output using a Cardiomax 2, model 95 cardiac output monitor (Columbus Instruments, Columbus, Ohio). The cardiac output was measured using the thermodilution method by injecting 0.1 ml of room temperature saline into the internal jugular catheter. Each cardiac output measurement was calculated

Since the first reports of laparoscopy in the rat [5, 8] there has been increasing interest in using a rodent model for research [14, 11, 20] and, to a lesser extent, for technical training [16] in laparoscopic surgery. If valid experimental comparisons of the tissue effects of laparoscopic and open procedures are to be made in a rodent model, the animals cardiorespiratory physiology under anesthesia must be comparable during the two procedures. In humans and large animals it is well known that a 15 mmHg CO2 pneumoperiCorrespondence to: R. Berguer, 150 Muir Road (112), Martinez, CA 94553, USA

916 as the average of three successive readings at each time point. Following placement of the three catheters, baseline measurements were obtained for heart rate (HR, beats/min), mean arterial pressure (MAP, mmHg), cardiac output (CO, ml/min), arterial pH, PCO2, PO2, base excess (BE), HCO3, and blood temperature (BT). Arterial blood gas measurements were made by withdrawing 0.1 ml of blood from the femoral arterial line through a capillary tube into a 1-ml syringe. The excess blood was then flushed into the circulation via the arterial catheter with an additional with 0.2 ml of saline solution. The pneumoperitoneum was created by making a 2-mm midabdominal skin incision, placing a standard Veress needle (Ethicon Endosurgery, Cincinnati, OH) into the peritoneal cavity, and insufflating CO2 through a standard laparoscopic insufflator (Karl Storz, Germany) set at the minimum flow rate. Five rats were studied at pneumoperitoneum pressures of 0, 2, 5, and 10 mmHg and four additional rats underwent anesthesia without pneumoperitoneum for control purposes. Cardiorespiratory measurements were obtained in all animals following catheterization (baseline), 10 min after reaching each new level of pneumoperitoneum, and again 10 min after desufflation. At the end of the experiments the rats were sacrificed with intravenous KCl. Data for each physiologic variable at each time point are expressed as the mean value SEM for each experimental group. Baseline data were compared to values obtained at each time point and the pneumoperitoneum group was compared to control animals using a twoway repeated measures ANOVA and Tukeys Honest Significant Difference Post Hoc test. Statistical significance was defined as p < 0.05.

Results There were no significant differences in animals body weight, time to induction, and total experimental time between the pneumoperitoneum and control groups. Blood temperature demonstrated a global decrease from baseline to desufflation (means of 37.3 and 36.9, respectively, p 0.043) that was not significantly different between groups. Heart rate decreased significantly with increasing pneumoperitoneum pressure and was significantly lower than in control rats at a pneumoperitoneum pressure of 10 mmHg (Fig. 1A). MAP did not change significantly over time and was similar between groups (Fig. 1B). Cardiac output was significantly decreased in the pneumoperitoneum group compared to the control group only at a pressure of 5 mmHg and then demonstrated a significant elevation after desufflation compared to baseline values (Fig. 1C). CVP measurements in the pneumoperitoneum group ranged from 3.2 to 4.0 cmH2O and were not significantly different over time, reflecting a total average of 3.3 0.24 ml of intravenous saline given during CO measurements in the pneumoperitoneum group. Rats demonstrated CO2 retention with increasing pneumoperitoneum pressure. The arterial pCO2 at 10 mmHg pneumoperitoneum was significantly increased in the pneumoperitoneum group compared to baseline values and was dramatically higher than in control animals (Fig. 2A). Arterial pCO2 levels did return to baseline values after desufflation. Blood pH and BE data mirrored the changes in arterial pCO2 in both groups (Fig. 2B,C). Arterial pO2 levels did not change significantly over time and O2 saturation in all rats remained near 100% throughout the experiment (data not shown).
Fig. 1. Cardiovascular variables in anesthetized rats subjected to CO2 pneumoperitoneum pressures of 0, 2, 5, and 10 mmHg (n 5) or no pneumoperitoneum (Control, n 4). The last pressure of 0 mmHg corresponds to values after desufflation. Individual data points represent the mean SEM (error bars) for each group. *p < 0.05 compared to control animals, p < 0.05 compared to baseline values at each time point. (a) Heart rate (HR). (b) Mean arterial pressure (MAP). (c) Cardiac output (CO). (, control, , pneumo).

Discussion The purpose of this study was to compare the cardiorespiratory effects of CO2 pneumoperitoneum pressures ranging

917

from 2 to 10 mmHg in spontaneously breathing anesthetized rats. The results of this study may assist investigators who use rodent models of laparoscopic surgery in selecting a pneumoperitoneum pressure that allows comparably normal cardiac and respiratory function. In humans and large animals, studies have reported both increases [12] and decreases [13, 15] in cardiac output during CO2 pneumoperitoneum, depending on the combined effects of intraabdominal pressure on preload and afterload [18]. In the present study we measured a slight decrease in CO only at a pneumoperitoneum pressure of 5 mmHg, followed by a significant increase in CO following desufflation. A decrease in CO at 5 mmHg but not at 10 mmHg of pneumoperitoneum is probably due to experimental variability, and thus we interpret our data as indicating no immediate effect of increasing pneumoperitoneum on CO. These results probably reflect an adequate preload during the insufflation as measured by CVP. The significant increase in CO following desufflation suggests that animals subjected to pneumoperitoneum nevertheless incurred a metabolic debt, perhaps due to the respiratory acidosis. In this regard, it is notable that the anesthetic regimen itself did not cause any significant respiratory depression in control animals despite experiment durations of 188.8 14.5 min. Our results should be interpreted with some caution because inhalational anesthesia with 12% isoflurane in a 50% N2O/O2 mixture causes significantly less respiratory depression than other anesthetics [7, 9]. The respiratory effects of CO2 pneumoperitoneum seen in larger animals and humans are progressive increases in arterial PCO2 and decreases in arterial pH during intraperitoneal CO2 insufflation at 15 mmHg [6, 10, 13, 14, 17, 18]. This hypercarbia usually requires controlled hyperventilation to maintain a normal arterial PCO2 under anesthesia. The present study in spontaneously breathing rats demonstrated a similar significant respiratory acidosis during CO2 pneumoperitoneum. The marked respiratory acidosis at a pneumoperitoneum pressure of 10 mmHg is probably the result of the inability of spontaneously breathing rodents to compensate for the combined hydrostatic effects of the pneumoperitoneum and the increased absorbed CO2 load. Although our observations could have been partly due to the effects of accumulated CO2 over the course of the experiment, we believe this is unlikely since the pH and pCO2 rapidly returned to baseline values after desufflation. In previous studies of survival of laparoscopic surgery in rats we have found that a 2 mmHg CO2 pneumoperitoneum in the rat causes no significant cardiorespiratory alterations [7] and permits the investigator to perform complex intraabdominal procedures [2, 3]. The present study confirms the safety of a 2 mmHg pneumoperitoneum while demonstrating that pneumoperitoneum pressures of 10 mmHg cause significant bradycardia and respiratory acidosis. These findings are relevant to other laparoscopic research studies employing rats or mice where pneumoperitoneum pressure of 46 mmHg [1] and 68 mmHg [19] are reported. In one study the authors introduced 6 ml of CO2 into the peritoneal cavity of CD-1 mice without reporting the actual intraperitoneal pressure [21]. In summary, we conclude that CO2 pneumoperitoneum pressures of 10 mmHg or greater should not be used during

Fig. 2. Respiratory variables in anesthetized rats subjected to CO2 pneumoperitoneum pressures of 0, 2, 5, and 10 mmHg (n 5) or no pneumoperitoneum (Control, n 4). The last pressure of 0 mmHg corresponds to values after desufflation. Individual data points represent the mean SEM (error bars) for each group. *p < 0.05 compared to control animals, p < 0.05 compared to baseline values at each time point. (a) Arterial pCO2. (b) Arterial pH. (c) Arterial base excess (BE). (, control, , pneumo).

918

laparoscopic surgery in the spontaneously breathing rat due to bradycardia and severe respiratory acidosis. References
1. Allendorf JD, Bessler M, Kayton ML, Whelan RL, Treat MR, Nowygrod R (1995) Tumor growth after laparotomy or laparoscopy. A preliminary study. Surg Endosc 9: 4952 2. Berguer R, Gutt C, Stiegmann GV (1993) Laparoscopic surgery in the rat. Description of a new technique. Surg Endosc 7: 345347 3. Berguer R, Gutt CN (1994) Laparoscopic colon surgery in a rat model: a preliminary report. Surg Endosc 8: 11951197 4. Bouvy ND, Marquet RL, Hamming JF, Jeekel J, Bonjer HG (1996) Laparoscopic surgery in the rat: beneficial effect on body weight and tumor take. Surg Endosc 10: 490494 5. Crane LH, Martin L (1991) Postcopulatory myometrial activity in the rat as seen by video-laparoscopy. Reprod Fertil Dev 3: 685698 6. Cunningham AJ, Turner J, Rosenbaum S, Rafferty T (1993) Transoesophageal echocardiographic assessment of haemodynamic function during laparoscopic cholecystectomy. Br J Anaesth 70: 621625 7. Dalton M, Hildreth J, Matsuoka T, Berguer R (1996) Determination of cardiorespiratory function and the optimum anesthetic regimen during laparoscopic surgery in the rat model. Surg Endosc 10: 297300 8. Filmar S, Gomel V, McComb PF (1987) Operative laparoscopy versus open abdominal surgery: a comparative study on postoperative adhesion formation in the rat model. Fertil Steril 48: 486489 9. Flecknell PA (1987) Laboratory animal anesthesia. Academic Press, San Diego 10. Graham AJ, Jirsch DW, Barrington KJ, Hayashi AH (1994) Effects of intraabdominal CO2 insufflation in the piglet. J Pediatr Surg 29: 1276 1280 11. Gutt CN, Berguer R, Stiegmann GV (1993) Laparoscopic surgery in the rat: description of a new technique. Zentralbl Chir 118: 631634

12. Hashimoto S, Hashikura Y, Munakata Y, Kawasaki S, Makuuchi M, Hayashi K, Yanagisawa K, Numata M (1993) Changes in the cardiovascular and respiratory systems during laparoscopic cholecystectomy. J Laparoendosc Surg 3: 535539 13. Ho HS, Gunther RA, Wolfe BM (1992) Intraperitoneal carbon dioxide insufflation and cardiopulmonary functions. Laparoscopic cholecystectomy in pigs. Arch Surg 127: 928932; discussion 932923 14. Ishizaki Y, Bandai Y, Shimomura K, Abe H, Ohtomo Y, Idezuki Y (1993) Changes in splanchnic blood flow and cardiovascular effects following peritoneal insufflation of carbon dioxide. Surg Endosc 7: 420423 15. Joris JL, Noirot DP, Legrand MJ, Jacquet NJ, Lamy ML (1993) Hemodynamic changes during laparoscopic cholecystectomy. Anesth Analg 76: 10671071 16. Kayton M, Morales A, Chen M, Treat M, Nowygrod R (1994) Laparoscopic surgery in the rat: a model for teaching laparoscopic suturing techniques. Surg Endosc 8: 547 17. Odeberg S, Ljungqvist O, Svenberg T, Gannedahl P, Backdahl M, von Rosen A, Sollevi A (1994) Haemodynamic effects of pneumoperitoneum and the influence of posture during anaesthesia for laparoscopic surgery. Acta Anaesthesiol Scand 38: 276283 18. Safran DB, Orlando RR (1994) Physiologic effects of pneumoperitoneum. Am J Surg 167: 281286 19. Trokel M, Allendorf J, Treat M, Whelan R, Nowygrod R, Bessler M (1994) Inflammatory response is better preserved after laparoscopy vs laparotomy. Surg Endosc 8: S30 20. Trokel MJ, Bessler M, Treat MR, Whelan RL, Nowygrod R (1994) Preservation of immune response after laparoscopy. Surg Endosc 8: 13851387; discussion 13871388 21. Watson R, Redmond H, McCarthy J, Burke P, Bouchier-Hayes D (1995) Exposure of the peritoneal cavity to air regulates the early inflammatory responses to surgery in a murine model. Br J Surg 82: 10601065

Surg Endosc (1997) 11: 939941

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

A tactile sensor for laparoscopic cholecystectomy

S. Matsumoto, R. Ooshima, K. Kobayashi, N. Kawabe, T. Shiraishi, Y. Mizuno, H. Suzuki, S. Umemoto
Department of Surgery, Second Teaching Hospital, Fujita Health University, 3-6-10 Otohbashi, Nakagawa-ku, Nagoya, Aichi 454, Japan

Abstract. During laparoscopic surgery, surgeons observe the three-dimensional abdominal cavity on a twodimensional TV monitor, which is a limitation. Another limitation is that surgeons are unable to estimate the softness of organs or tissues during laparoscopic surgery as they are only allowed to use instruments which touch objects and direct palpation is not permitted during the procedure. The tactile sensor which we used displays the object softness immediately as a digital score, which can then be superimposed on a TV monitor as a graph. With the tactile sensor, we were able to ascertain the presence of a gallstone in the gallbladder or cholecystic duct during laparoscopic cholecystectomy and also able to discriminate between a stone and an air bubble during intraoperative cholangiography. We were convinced that the tactile sensor would be useful in laparoscopic surgery, which does not permit surgeons to palpate objects with human fingers. Key words: Tactile sensor Laparoscopic cholecystectomy Cholelithiasis Diagnosis Ultrasound

We recently had opportunities to use an ultrasonic tactile sensor probe, which has been developed by Professor Omata and Olympus Co. [1, 2], as an instrument to detect a gallstone in the gallbladder as well as an auxiliary means to confirm the absence of a stone in the cholecystic duct. We felt that this instrument was also supportive for intraoperative cholangiography, given our experiences of 20 laparoscopic cholecystectomy cases. Case report
A 60-year-old male patient underwent a laparoscopic cholecystectomy. The tactile sensor was used to scan the surface of the gallbladder from the fundus to the cholecystic duct to see whether it could detect a gallstone from outside. Three scans were made, and the bottom line shows the typical curve, which shifts upward in the later scan. This is evidence of hard material in the gallbladder under the line which the probe scanned (Fig. 1).

Materials and methods Principles of ultrasonic tactile sensor

The sensor tip contains a piezoelectric transducer (PZT) which consists of ceramics such as lead zirconate itanate and a vibration pickup. When the tip comes into contact with body tissue, the resonance condition changes according to the hardness of the tissue, and the resonance frequency shifts in a negative direction. It has been proven scientifically that the shift increases according to the softness of the tissue. The shift in resonance frequency can be displayed on a computer monitor as a graph in which the horizontal axis is time. The quantity of value shift at each point could be measured by scanning a probe.

Laparoscopic surgery started with laparoscopic cholecystectomy, and assessments have since been made of the possibility of its application to various ailments typically treated by surgery. Both laparoscopy and thoracoscopy allow easy observation and identification of objects as long as they are located on surface areas. However, there is the disadvantage that they cannot identify lesions which do not produce changes on the surface. It is obvious that intraluminal lesions cannot be observed from outside. If an intraluminal object is large and sufficiently hard, it may be possible to use a dissector or other instrument to touch and identify it from the outside. It is a defect of this method of surgery that, unlike with human hands, one cannot determine with laparoscopic graspers or other instruments whether objects are soft or hard. We were frustrated when we could not be sure of the location of the lesion when performing laparoscopic or thoracoscopic local resection for early gastric, colonic cancer, or pulmonary metastases.
Correspondence to: S. Matsumoto

System structure
We have constructed a laparoscopic probe which is 295.5 mm long whose sensor tip is 8 mm in diameter and whose shaft diameter is 10 mm (Figs. 2, 3.). The tactile sensor tip is installed in the front of the probe and faces forward. The circuit generating the vibration is contained in the probe handpiece. Power supply to the probe is contained in and calculation of the resonance frequency is made in the measuring section, and the resonance frequency which is detected according to the hardness of the organic tissue is digitally processed in the frequency counter and sent to a personal computer through a universal counter (GP-1B). In the personal computer, records of time changes in digitalized resonance frequency information and

940

Fig. 3. Ultrasonic tactile sensor probe.

Fig. 1. Graph superimposed on the screen during scanning of gallbladder. The bottom line indicates the gallbladder softness during scanning from the fundus to the neck. The curve shows an upward shift at the right side, and this indicates a gallstone.

Fig. 4. A gallstone was embedded in the silicone gum. The black round material shows a gallstone.

Fig. 2. System structure. The ultrasonic tactile sensor probe is connected to a personal computer through a universal counter. The result is superimposed on the monitor through a scan converter.

data are converted into a graph. The graph of resonance frequency data is superimposed on a laparoscopic monitor image through a scanner converter.

Preliminary test and results

In order to check whether the probe was capable of detecting a gallstone, a gallstone was buried in the preliminary test in silicone gum as illustrated in Fig. 4 and the tactile sensor probe was moved as shown in Fig. 5. The black material in the photograph shows the gallstone. Figure 5 shows that the curved line shifts upward when the probe is passed over the gallstone, and we can thus tell that there is a hard object beneath the sensor.

Discussion In this study, we confirmed that with the tactile sensor we could detect hard material in the gallbladder provided that we could place the probe on the surface of the gallbladder from the fundus to the cystic duct and could scan correctly
Fig. 5. Model of the ultrasonic tactile sensor probe scan. When the surface of the silicone gum is scanned after a gallstone has been embedded in it, the curved line, which shifts toward a minus direction according to the softness of silicone gum, shifts upward only when the probe scans over the embedded gallstone.

941

over the gallstone. Through experience we learned that a stone could be missed if the sensor was not placed over a stone when there were only a few small stones in a large gallbladder in which stones could move about during scanning. But as the lumen of the cystic duct is narrow and the 8-mm probe is wide enough to cover the width of the whole cystic duct lumen, it seems unlikely that we would miss a stone impacted in the cystic duct. The tactile sensor thus was useful when making a decision as to at which point the cystic duct should be cut when there are many stones passing from the gallbladder to the cystic duct. We also should be able to examine stones in the common bile duct and in the hepatoduodenal ligament with the tactile sensor, which can scan the common bile duct in that area. We often have problems with air bubbles because we cannot distinguish them from stones during intraoperative cholangiography. It is difficult to get good images of dissected free cystic duct with the ultrasound diagnostic device. In such cases, it will be unnecessary to repeat cholan-

giography if the tactile sensor has not shown a positive stone reaction. Clinically, this is very useful. Based on our experiences with a case of cholelithiasis, we think the tactile sensor would be useful for finding a lesion which was not exposed on the organ surfacefor example, intraluminal lesions or metastatic lesions in the lung or the liver. The tactile sensor is particularly advantageous when locating pulmonary lesions because ultrasound diagnostic devices are not helpful when detecting lesions.

References
1. Omata S, Terunuma Y (1992) New tactile sensor like the human hand and its application. Sensors Actuaters A 35: 915 2. Omata S, Yoshida S, Constantinou C, Kayata K, Yamaguchi O, Shiraiwa Y (1994) New medical sensor for detecting compliance of living tissue and its applications. Technical Digest of the 12th Sensor Symposium, pp 245248

Surg Endosc (1997) 11: 928932

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Predictive factors for synchronous common bile duct stones in patients with cholelithiasis
A. Alponat,1 C. K. Kum,1 A. Rajnakova,1 B. C. Koh,2 P. M. Y. Goh1
1 2

Department of Surgery, National University Hospital, 5 Lower Kent Ridge Road, Singapore Biostatistic Consultancy Unit, Clinical Research Centre, Singapore

Received: 29 November 1996/Accepted: 9 April 1997

Abstract Background: To determine the predictive factors of synchronous common bile duct (CBD) stones, data from 878 consecutive patients who underwent cholecystectomy in a university clinic from June 1991 to June 1996 were retrospectively analyzed. Methods: Based on clinical, biochemical, and ultrasonographic criteria, 194 patients were selected for ERCP, 180 preoperative and 14 postoperative. Results: Cannulation of CBD was successful in 192 (99%) patients. Stones were identified in 62 (32%) patients and sphincterotomy was performed in 56 (90%). Duct clearance was achieved in 43 (77%) cases. There was a high predictive value for the presence of CBD stones in patients with cholangitis, present jaundice, and dilated CBD with evidence of stones on ultrasound (75%, 72%, and 67% respectively). A dilated CBD without stone on ultrasound and elevated liver enzymes had less than 40% positive predictive value. History of previous jaundice, pancreatitis, previously raised liver enzymes, and present pancreatitis was predictive in less than 20% of the cases. Univariate analyses revealed that clinical findings of cholangitis and obstructive jaundice, elevated liver enzymes (previous and present), and ultrasonographic findings of stones in a dilated CBD were significant positive predictors. Subanalysis of each elevated liver enzyme revealed that alanine transaminase, aspartate transaminase, alkaline phosphatase, and gamma glutamyl transpeptidase were significant predictors. Both elevated conjugated and total bilirubins were also significant predictors for CBD stones. Conclusion: Multivariate logistic regression analysis on these significant predictors showed that cholangitis (odds ratio [OR]: 10.5), dilated CBD with evidence of stones on ultrasound (OR: 7.4), elevated aspartate transaminase (OR: 2.9), and conjugated bilirubin (OR: 5.3) were jointly significant. The likelihood of having stones in the duct without

any of these predictors was 7%, but 99% when all the predictors were positive. Key words: Predictive factors Synchronous common bile duct stones Cholelithiasis

Correspondence to: P. M. Y. Goh

The introduction of endoscopic retrograde cholangiopancreaticography (ERCP) and sphincterotomy in the early 1970s changed the practice of biliary surgery [4, 13]. Endoscopic sphincterotomy (ES) was first introduced for removal of recurrent or retained stones after cholecystectomy [5]. However, its rapid success has expanded the indications. Endoscopic sphincterotomy with stone extraction became a wellestablished procedure in the management of choledocholithiasis in high-risk elderly patients in whom a noninflamed gallbladder might be left in situ [18]. Patients with cholangitis not responding to antibiotics and patients with severe gallstone pancreatitis not responding to conservative treatment have also become the candidates for ERCP and sphincterotomy during the acute illness [19]. In the last decade, laparoscopic cholecystectomy (LC) has gained widespread acceptance and become the new gold standard for the treatment of symptomatic gallbladder disease. Laparoscopic management of associated choledocholithiasis via a transcystic or transcholedochal route has also been advocated [1]. However, it remains a demanding approach that cannot be applied by the majority of surgeons. Technical difficulties and necessity of advanced skills to manage synchronous CBD stones laparoscopically have fueled the use of preoperative ERCP and sphincterotomy prior to LCan approach which fits in well with the minimally invasive concept [14]. ES and duct clearance are successful in 9095% of patients with an acceptable morbidity and mortality rate [15]. Most of the complications are self-limited pancreatitis. However, CBD stones are detected in only 2550% of the

929

patients who undergo ERCP. Therefore, knowledge of reliable factors predictive of synchronous CBD stones is essential to avoid unnecessary ERCP and reduce avoidable morbidity and costs. The aim of this study was to determine such factors and their reliability in predicting CBD stones. Patients and methods
Between July 1991 and July 1996, 878 consecutive patients underwent cholecystectomy (783 laparoscopic and 95 open), of which 194 were selected for either preoperative or postoperative ERCP. Data collected prior to ERCP included age, sex, history of right hypochondrial pain, indication for procedureprevious or present elevated serum liver enzymes, clinical findings of cholangitis, jaundice, pancreatitis, dilated CBD over 6 mm with or without stones on ultrasound, serum level of each liver enzyme (aspartate transaminase [AST], alanine transaminase [ALT], alkaline phosphatase [ALP], gamma glutamyl transpeptidase [GGT], lactate dehydrogenase [LDH])bilirubins, and ultrasonographic findings. Age was categorized as young (<65) and old (65 and above). Biochemical analysis of blood was evaluated as abnormal when liver enzymes and bilirubin levels were elevated greater than two times normal. Elevated liver enzyme tests were considered present when any three of AST, ALT, ALP, GGT, and LDH were elevated. ERCP was performed with an Olympus JFIT side-viewing duodenoscope under fluoroscopic control. If sphincterotomy was indicated, the bile duct was cannulated with a papillotome (Wilson-Cook Med., Inc., NC, USA); a combination of coagulation and cutting diathermy was used. Stones were extracted with the Dormia basket or balloon catheter. Cholecystectomy was done between 24 to 72 h after ERCP.

Table 1. Indications for endoscopic retrograde cholangiopancreaticography and positive predictive values Number of patients (%), n 194 31 (16) 7 (4) 8 (4) 139 (72) 31 (16) 52 (27) 8 (4) 53 (27) 36 (19) Positive predictive value (%) 16 14 13 37 19 67 75 36 72

Indication Previous LFTa elevation Previous pancreatitis Previous jaundice Present LFTa elevation Present pancreatitis Present jaundice Cholangitisb Dilated CBDc without stone on ultrasound Dilated CBDc with stone on ultrasound
a b

Liver function test Jaundice with fever, right hypochondrial pain c Common bile duct

Statistical analysis
All statistical analyses were performed with the SPSS for Windows version 6.0. In the preliminary analysis, univariate statistical methods were used to determine which factors were significantly related to the presence of CBD stones. The chi-square test with Yates correction or Fishers exact test was used for categorical variables; p < 0.05 was considered significant. All significant factors from the univariate analysis were subsequently included in the multivariate logistic regression procedure. The FORWARD automatic variable selection procedure was used in determining which of these variables were predictors of presence of CBD stones. Presence of CBD stones on ERCP was the dependent variable. Four of 194 patients with missing data were excluded from the analysis. The missing data was ultrasonographic evaluation of CBD in two and ERCP findings due to failure in cannulation in two.

Results There were 73 males (38%) and 121 females (62%). The average age was 52 (range 1386) years. Cannulation of CBD was successful in 192 cases (99%). Cannulation failure occurred in two patients (1%) in whom the ampulla was located in a large diverticulum; they were excluded from the study. Some 180 patients (93%) had preoperative ERCP and 14 (7%) had postoperative ERCP. The indication for postoperative ERCP was obstructive jaundice in six patients retained stones were found in all of them. Elevated liver enzymes and serum amylase levels were the indications in another six patientsERCP findings were normal in all. In the remaining two patients, incomplete clearance of CBD in open exploration required postoperative ERCP. Stone extraction was successful in these cases.

Stones were found in 62 cases out of 192 (32%). Six patients with multiple large stones (>15 mm) in the CBD were referred to surgery for open CBD exploration without endoscopic sphincterotomy attempt during our early experience with interventional ERCP. Endoscopic sphincterotomy was performed in 56 patients (90%). Both wire-guided cannulation and pre-cut sphincterotomy techniques were employed to improve success rate. Stone retrieval was successful in 43 (77%) cases. Failure of stone extraction occurred in eight patients (14%) because of the large size of stones (mean: 18.3 mm, ranging from 15 to 23 mm). These eight patients underwent open exploration of CBD. Two patients who had single small stones (<5 mm) had only sphincterotomy and subsequently underwent LC. Three patients who had septic cholangitis underwent sphincterotomy, one of which also required stenting. The mean duration for diagnostic ERCP was 25 (SD: 16 min) and 42 (SD: 21.5 min) when sphincterotomy and stone extraction were performed. Indications for ERCP and positive predictive values (PPV) for these predictors are shown in Table 1. Complications included pancreatitis in seven cases (4%), all of which resolved spontaneously. Pancreatitis occurred following sphincterotomy in two and diagnostic ERCP in five. One patient (0.5%) had a retained stone basket which required surgery and duodenotomy. Two patients (1%) had minor bleeding that stopped spontaneously without the need for blood transfusion. Univariate analysis of each set of pre-op ERCP data was carried out to assess its significance in predicting the presence of CBD stones (Table 2). Cholangitis, jaundice, past and present elevated liver enzymes, and ultrasonographic finding of CBD stones were predictors on univariate analysis even after using the Bonferroni-Holm method to adjust for multiple comparisons. Analysis of individual enzymes revealed that elevated AST, ALT, ALP, GGT, and raised bilirubins (conjugated and total) were significant predictors. Multiple logistic regression analysis on these 11 variables showed that only cholangitis, dilated CBD over 6 mm with stone(s) on ultrasound, elevated AST, and conjugated bilirubin were jointly significant (Table 3). The probability of having stones in CBD is thus: log odds ratio (OR) 2.62 + (2.26 cholangitis) + (2 dilated CBD with stone

930 Table 2. Analysis of preoperative parameters for statistical significance by using chi-square or Fishers exact test* CBDa stone Predictors Age (years) >65 <65 Sex (F percent) Past history Right upper quadrant pain Right upper quadrant pain radiating back Indication Cholangitis Resolved pancreatitis Previous jaundice Previously elevated liver enzymes Present elevation of liver enzymes Present pancreatitis Present jaundice CBD >6 mm without stone on ultrasound CBD >6 mm with stone on ultrasound Laboratory invest Aspartate transaminase Alanine transaminase Alkaline phosphatase Gamma glutamyl transpeptidase Lactate dehydrogenase Conjugated bilirubin Total bilirubin Ultrasonography findings GB stone 1 >1 GB st. size <1 cm >1 cm
a b

Absent (n 130)b 41 (32) 89 (68) 83 (63) 130 (100) 60 (46) 2 (2) 6 (5) 7 (5) 26 (20) 88 (67) 25 (19) 17 (13) 34 (26) 10 (8) 78 (60) 80 (62) 70 (54) 102 (78) 68 (54) 28 (22) 22 (17) 11 (9) 111 (87) 102 (84) 19 (16)

Present (n 62) 29 (47) 33 (53) 38 (61) 61 (98) 23 (37) 6 (10) 1 (2) 1 (2) 5 (8) 52 (84) 6 (10) 35 (56) 19 (31) 26 (42) 55 (89) 54 (87) 52 (84) 54 (87) 42 (68) 42 (71) 38 (61) 6 (11) 52 (90) 42 (81) 10 (19)

Significance 2 p

3.574 0.002 1.058

0.06 0.93 0.54* 0.19 0.02* 0.43* 0.44* 0.04* 0.02* 0.14* 0.01 0.56 0.01 0.01* 0.03* 0.01* 0.05* 0.06 0.01 0.01 0.78* 0.66*

37.826 0.331 31.189

3.481 41.678 36.425

CBD, common bile duct Values in parantheses are percentages

Table 3. Predicted odds ratio of having stones in the common bile duct based on the multivariate logistic regression model Predictor Cholangitis Dilated common bile duct with stone on ultrasound Elevated aspartate transaminase Elevated conjugated bilirubin Odds ratio (OR) 10.5 7.4 2.9 5.3 95% confidence interval 1.55 2.85 1.25 2.35 71.79 18.99 6.88 11.83 p value 0.02 0.01 0.01 0.01

on ultrasound) + (1.08 elevated AST) + (1.66 elevated conjugated bilirubin), in which, cholangitis 1 if present or 0 if not; dilated CBD over 6 mm with stone on ultrasound 1 if yes or 0 if not; elevated AST 1 if yes or 0 if not; and elevated conjugated bilirubin 1 if present and 0 if not.

Discussion CBD stones are present in 1015% of patients undergoing cholecystectomy and the incidence rises to 30% in those over 80 years of age [6]. In the prelaparoscopic cholecystectomy era, the traditional management for CBD stones was to perform a choledochotomy and stone extraction. However, open CBD exploration is associated with significant morbidity and retained stone rate [5, 18]. Preoperative

ERCP and sphincterotomy were introduced to avoid the complications associated with open CBD exploration [11]. To date, the endoscopic approach has never been convincingly proven to be superior to open exploration [19, 22]. After widespread acceptance of LC, the management of CBD stones has regained renewed interest. Several strategies have emerged to manage synchronous CBD stones. The first strategy is to ignore them. But it has been shown that 55% of untreated CBD stones became symptomatic, and half had complications [12]. The second option is routine preoperative cholangiography and sphincterotomy if stones are present in a dilated duct. However, this approach is not cost-effective and routine ERCPs are not without danger. ERCP has a morbidity rate of 36.4% and a mortality rate of 0.050.1%, increasing to a morbidity rate of 510% and a mortality rate of 13.1% when sphincterotomy is performed [6]. In addition, it is

931 Table 4. Predicted probability of having stones in CBD based on the pattern of predictors present in any individual patient Model 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
a

Cholangitis No Yes No No No Yes Yes Yes No No No Yes Yes Yes No Yes

Dilated CBD with stone on ultrasounda No No Yes No No Yes No No Yes Yes No Yes Yes No Yes Yes

Elevated aspartate transaminase No No No Yes No No Yes No Yes No Yes Yes No Yes Yes Yes

Elevated conjugated bilirubin No No No No Yes No No Yes No Yes Yes No Yes Yes Yes Yes

Predicted probability (%) 7 43 35 18 28 85 69 80 61 74 53 94 97 92 89 99

CBD, common bile duct

expensive [7]. Therefore, routine use of preoperative ERCP cannot be recommended in all patients undergoing cholecystectomy. The third strategy is to perform intraoperative cholangiogram and retrieve CBD stones laparoscopically. Intraoperative cholangiography is successful in 88% of patients with a positive predictive value of 6392% [16]. This procedure also delineates the biliary anatomy and may prevent bile duct injury. Furthermore, complete treatment of biliary lithiasis in one operation may avoid the joint risk of two procedures (ERCP and LC) [17]. Although routine use of intraoperative cholangiography adds about 20 min to the LC, fluoroscopic cholangiogram can be completed in an average time of 6.9 min [2]. However, it has a false-positive rate ranging from 2.1 to 67%, which leads to unnecessary CBD exploration or conversion to open surgery [16]. The fourth strategy is preoperative ERCP in selected cases. Preoperative ERCP reduces the need for intraoperative cholangiography, avoids the need for most postlaparoscopic studies, and provides important information for the bile duct anatomy prior to LC. However, the significant difference in developing recurrent biliary symptoms following sphincterotomy (21%) compared to the biliary surgical group (6%) and high cost of ERCP should be kept in mind [8, 24]. Experience with selective preoperative ERCP and sphincterotomy prior to LC has been accumulating [3, 810, 20, 26]. Clinical, biochemical, and ultrasonographic evidence of stones has been utilized as criteria in patient selection for ERCP. Some studies reported that only the severity of patients initial illness (acute cholangitis, persistent obstructive jaundice, and acute gallstone pancreatitis) had higher than 85% positive predictive value [21, 23, 25]. Elevated liver enzymes correlated well with duct calculi with a positive predictive value of 3060% in other studies [9, 11]. Ultrasonographic finding of a dilated duct was predictive in 2070% of the cases, whereas evidence of stones in a dilated duct on ultrasound was almost uniformly accurate [3, 20, 26]. Multivariate analysis in this study has identified cholangitis, dilated CBD over 6 mm with stone(s) on ultrasound,

elevated AST, and conjugated bilirubin as the main predictive parameters. A combination of predictors increases the odds of having stones in the CBD. By applying the derived formula, likelihood of having stones in the duct can be estimated; thereby, the number of the unnecessary ERCPs can be reduced. According to the formula, probability of having stones in the duct without any of these predictors is 7% whereas it increases to 99% if all four predictors are present (Table 4). This study has identified important predictors of synchronous CBD stones in patients who require cholecystectomy. Selective use of ERCP in these patients will ensure a high yield rate and thus improve cost-effectiveness. The small number of patients who are false negatives, i.e., have associated CBD stones but also do not have any abnormal parameters, usually have small stones that are likely to pass spontaneously. If not, postoperative ERCP can deal with most of them. References
1. Arregui ME, Davis CJ, Arkush AM, Nagan RF (1992) Laparoscopic cholecystectomy combined with endoscopic sphincterotomy and stone extraction or laparoscopic choledochoscopy and electrohydraulic lithotripsy for management of cholelithiasis with choledocholithiasis. Surg Endosc 6: 1015 2. Carroll BJ, Phillips EH, Rosenthal R, Gleischman J, Bray JF (1996) One hundred consecutive laparoscopic cholangiography. Results and conclusions. Surg Endosc 13(3): 319323 3. Cisek PL, Greaney GC (1994) The role of endoscopic retrograde cholangiopancreaticography with laparoscopic cholecystectomy in the management of choledocholithiasis. Am Surg 60: 772776 4. Classen M, Demling L (1974) Endoskopische sphinkterotomie der papilla Vateri und steinextraktion aus dem ductus choledochus. Disch Med Worchensahr 99: 496497 5. Danilewitz MD (1989) Early postoperative endoscopic sphincterotomy for retained common bile duct stones. Gastrointest Endosc 35: 298299 6. Fletcher DR (1994) Changes in the practice of biliary surgery and ERCP during the introduction of laparoscopic cholecystectomy to Australia: their possible significance. Aust N Z J Surg 64: 7580 7. Frazee RC, Roberts J, Symmonds R, Hendricks JC, Snyder S, Smith R, Custer MD, Stoltenberg P, Avots A (1993) Combined laparoscopic

932 and endoscopic management of cholelithiasis and choledocholithiasis. Am J Surg 166: 702705 Graham SM, Flowers JL, Scott TR, Biley RW, Scovill WA, Zucker KA, Imbembo AL (1993) Laparoscopic cholecystectomy and common bile duct stones. Ann Surg 216: 6167 Hainsworth PJ, Rhodes M, Gompertz RHK, Armstrong CP, Lennard TWJ (1994) Imaging of common bile duct in patients undergoing laparoscopic cholecystectomy. Gut 35: 991995 Hawasli A, Llyoyd L, Pozios V, Veneri R (1993) The role of endoscopic retrograde cholangio-pancreaticogram in laparoscopic cholecystectomy. Am Surg 59: 285289 Heinerman MP, Boeckl O, Pimpl W (1988) Selective ERCP and preoperative stone removal bile duct surgery. Ann Surg 209: 267272 Johnson AG, Hosking SW (1987) Appraisal of the management of bile duct stones. Br J Surg 74: 555560 Kawai K, Akasaka Y, Murakami M (1974) Endoscopic sphincterotomy of the ampulla of Vater. Gastrointest Endosc 20: 148151 Kum CK, Goh PMY (1996) Preoperative ERCP in the management of common bile duct stones before laparoscopic cholecystectomy. Eur J Surg 162: 205210 Leese T, Neoptolemos JP, Carr-Locke DL (1985) Successes, failures, early complications and their management following endoscopic sphincterotomy: results in 394 consecutive patients from a single centre. Br J Surg 72: 215219 Madhavan KK, Macintyre IMC, Wilson RG, Saunders JH, Nixon SJ, Hamer-Hodges DW (1985) Role of intraoperative cholangiography in laparoscopic cholecystectomy. Br J Surg 82: 249252 Millat D, Deleuze A, Atger J, Briandet H, Fingerhut A, Marrel E, de-Seugin C, Soulier P (1996) Treatment of common bile duct lithiasis under laparoscopy. A prospective multicenter study in 189 patients. Gastroenterol Clin Biol 20 (4): 339345 Neoptolemos JP, Carr-Locke DL, Fossard DP (1987) Prospective randomised study of preoperative endoscopic sphincterotomy versus surgery alone for common bile duct stones. Br Med J 294: 470474 Neoptolemos JP, London N, Slater ND, Carr-Locke DL, Fossard DP, Moosa AR (1986) A prospective study of ERCP and endoscopic sphincterotomy in the diagnosis and treatment of gallstone acute pancreatitis. Arch Surg 121: 697702 Rieger R, Wayand W (1995) Yield of prospective, noninvasive evaluation of the common bile duct combined with selective ERCP/ sphincterotomy in 1390 consecutive laparoscopic cholecystectomy patients. Gastrointest Endosc 42: 612 Rijna H, Borgstein PJ, Meuwissen SGM, De Brauw LM, Wildenborg NP, Cuesta MA (1995) Selective preoperative endoscopic retrograde cholangiopancreaticography in laparoscopic biliary surgery. Br J Surg 82: 11301133 Stiegmann GV, Goff JS, Mansour A, Pearlman N, Reveille RM, Norton L (1992) Precholecystectomy endoscopic cholangiography and stone removal is not superior to cholecystectomy, cholangiography, and common duct exploration. Am J Surg 163: 227230 Surick B, Washington M, Ghazi A (1993) Endoscopic retrograde cholangiography in conjunction with laparoscopic cholecystectomy. Surg Endosc 7: 388392 Targarona EM, Ayuso RM, Bordes JM, Ros E, Pros I, Martinez J, Teres J, Trias M (1996) Randomized trial of endoscopic sphincterotomy with gallbladder left in situ versus open surgery for common bile duct in high risk patients. Lancet 347 (9006): 926929 Vanneman W, Kingsbury R, Duberman E, Lee M (1992) When is ERCP indicated before laparoscopic cholecystectomy. Gastrointest Endosc 38: 265 Vitale GC, Larson GM, Wieman TJ, Cheadle WG, Miller FB (1993) The use of ERCP in the management of common bile duct stones in patients undergoing laparoscopic cholecystectomy. Surg Endosc 7: 911

8. 9. 10. 11. 12. 13. 14. 15.

19.

20.

21.

22.

23. 24.

16. 17.

25. 26.

18.

Surg Endosc (1997) 11: 933935

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Impact of laparoscopic cholecystectomy on indications for surgical treatment of gallstones

E. W. Steinle,1 R. L. VanderMolen,1 A. Silbergleit,1,2 M. M. Cohen2,3
1 2

Department of Surgery, St. Joseph Mercy Hospital, Pontiac, MI, USA Wayne State University, Detroit, MI, USA 3 Rose Medical Center, Denver, CO, USA Received: 25 September 1996/Accepted: 14 March 1997

Abstract Background: The objective of this study was to compare the histology of gallbladders removed prior to the introduction of laparoscopic cholecystectomy with that found after the introduction of the laparoscopic technique to determine if there has been a change in the indications for surgical treatment of gallbladder disease. Methods: A retrospective review of all patients undergoing cholecystectomy during 1989, 1992, and 1993 was completed at two large community teaching hospitals in two different geographic regions of the United States. Patients who underwent cholecystectomy as the primary procedure were studied. A total of 1,815 cases met the criteria for analysis. Histological diagnoses were categorized as acute cholecystitis with or without cholelithiasis, or chronic cholecystitis with cholelithiasis. Results: The number of cholecystectomies performed increased by 58% from 1989 to 1993 (p < 0.05). The number of cholecystectomies for acute cholecystitis did not change. Conclusions: With the advent of laparoscopic cholecystectomy, the number of cholecystectomies significantly increased and the proportion of cholecystectomies performed for chronic disease also increased. There has been a significant change in the surgical management of gallbladder disease with increased willingness to recommend elective cholecystectomy. Further study is needed to determine if there is real benefit from earlier elective cholecystectomy. Key words: Cholecystectomy Cholecystitis Laparoscopy Surgery rate surgical indications

Laparoscopic cholecystectomy accounts for 8083% of the cholecystectomies performed nation wide [1], and since its introduction in 1987 [5] its use has been accompanied by a major increase in the cholecystectomy rate [8, 11, 15]. The procedure has been shown to have a low morbidity and mortality rate [1, 2, 11, 15, 16]. Laparoscopic cholecystectomy has also reduced the total cost per case of cholecystectomy by decreasing the hospital stay [9]. The National Institutes of Health (NIH) consensus conference recommended laparoscopic cholecystectomy as the treatment of choice for symptomatic gallstones [10]. The NIH also concluded that the indications for laparoscopic cholecystectomy, in general, are similar to those for open cholecystectomy. Indeed, the availability of laparoscopic cholecystectomy should not expand the indications for gallbladder removal [10]. Despite this warning, published reports have shown large increases (2960%) [8, 11, 15] in the total number of cholecystectomies since the introduction of the laparoscopic technique compared to the cholecystectomy rate in the 1980s [3]. These data have led some to speculate that the indications for cholecystectomy have changed, leading not only to an increase in the total number of procedures but also in the total cost of treatment of gallbladder disease [4, 8]. This study was designed to determine if the introduction of laparoscopic cholecystectomy has led to a change in the indications for cholecystectomy. Patients and methods
During the years 1989, 1992, and 1993, a total of 1,815 patients underwent cholecystectomy as the primary procedure at our two hospitals. Both hospitals are large community teaching hospitals, one (St. J) is in Pontiac, Michigan, the other (RMC) is in Denver, Colorado. Patients having incidental cholecystectomy, or cholecystectomy for chronic acalculous cholecystitis or trauma were excluded. Cases of carcinoma of the gallbladder were also excluded. The average age of the patients was 51 and 52 years old, respectively, for 1989 and 1992. The percentage of patients who were female was 70% in 1989 and 71% in 1992. Permanent sections from each case were reviewed by attending pathologists. Standard criteria for acute and chronic cholecystitis were used.

Correspondence to: M. M. Cohen, Department of Surgery, Grace Hospital, 6071 W. Outer Drive, Detroit, MI 48235, USA

934

Fig. 1. The number of cholecystectomies performed as the primary procedure during the years 1985 to 1993.

Fig. 2. The number of open and laparoscopic (Lap) cholecystectomies performed each year since the introduction of laparoscopic cholecystectomy at each hospital during 1990.

Gallbladders were designated as acute cholecystitis with or without cholelithiasis or chronic cholecystitis with cholelithiasis. Results were analyzed using the chi-squared test.

Results Figure 1 shows the total number of cholecystectomies performed at the two institutions during the years 1985 through 1993. The number of patients was relatively stable over the 5 years prior to the introduction of the laparoscopic technique; 1989 was the last year in which no laparoscopic cholecystectomies were performed. By 1991, all general surgeons with privileges at the two hospitals had obtained the required credentials to perform laparoscopic cholecystectomy. The average number of cholecystectomies performed from 1985 to 1989 was 429 per year. In 1993 there were 679, a 58% increase (p < 0.05). Throughout the period 1985 to 1993, the hospitals remained the same size, the populations for which they provided service were stable, and the general surgical staffs were not significantly expanded. This stability of the population being served was confirmed by demonstrating that the total number of appendectomies (considered a good indicator of the need for surgical care) did not change from 1989 to 1992. In 1989 a total of 362 appendectomies were performed at the two hospitals whereas in 1991 a total of 350 were performed. It is also interesting to note the remarkable penetration of the laparoscopic technique at both hospitals within about 2 years (Fig. 2). No laparoscopic procedures were performed in 1989, yet by 1993, 86% of the cholecystectomies were performed laparoscopically. The ratio of acute to chronic disease of the gallbladders during these years also changed from 0.17 in 1989 to 0.1 in 1992 and 0.12 in 1993. As can be seen in Fig. 3, while the total number of patients with acute disease remained constant, the entire increase in cholecystectomies was due to a large and significant increase in the number of cases with chronic disease. Chronic cholecystitis accounted for only 338 cases in 1989 compared to 546 cases in 1993. This represents an increase of 62%. In 1992, 22% of the open cholecystectomies were performed for acute cholecystitis and by 1993 this had increased to 33%.

Fig. 3. The number of gallbladders showing acute or chronic disease during the years 1989, 1992, and 1993.

Discussion In the early 1990s laparoscopy produced a remarkable change in the technique of cholecystectomy. This has been accompanied by improved patient recovery time, decrease in cost per procedure, and an acceptable morbidity rate. Recently, several authors have noted that along with these changes, there has also been a substantial increase in the total number of patients undergoing cholecystectomy [8, 11, 15]. We have seen a similar increase in the cholecystectomy rate at both of our institutions. However, the absolute number of cases of acute cholecystitis has remained unchanged. The observed increase was entirely among patients with chronic cholecystitis and cholelithiasis. There is a large population of patients with cholelithiasis who are either asymptomatic or who never opt for surgical treatment [13]. It seems clear that the recent increase in the cholecystectomy rate comes primarily from the group of patients with mild or asymptomatic disease. There were probably initially some patients with moderate to severe symptoms who were more willing to undergo a laparoscopic than an open operation. But the higher cholecystectomy rate has been main-

935

tained even after that pool of patients has been treated. Escarce et al. [6] have produced data from Medicare patients in Pennsylvania over the period 1986 to 1993 which strongly suggests that a 22% increase in the rate of cholecystectomy in that state was accompanied by a lowering of the threshold for performing surgery. Lowering the threshold for cholecystectomy is not necessarily inappropriate if the patients benefit overall. But in mildly symptomatic or asymptomatic patients, for benefit to outweigh risk the operation must be extremely safe, as serious complications occurring in the course of the natural history are rare in such individuals [14]. Apparently physicians, both primary care and specialist, are now more likely to recommend laparoscopic cholecystectomy than previously. This is probably due to the rapid recovery and the avoidance of a large abdominal incision. This trend has already eliminated the anticipated cost savings of laparoscopic cholecystectomy for the payors [4]. By contrast, Ho et al. [7] in a study of cholecystectomy from 1988 to 1994 reported that their cholecystectomy rate had stabilized at a level only 11% higher than in the years preceding the introduction of laparoscopic cholecystectomy. This finding could be explained by the fact that theirs is a tertiary referral center and may not reflect what is happening in the country as a whole. As stated in the NIH consensus conference, the challenge to the clinician is ascertaining which symptoms are and which are not due to gallstones [10]. A risk of operating on patients with stones whose symptoms are not caused by the stones is that the symptoms will frequently recur. As has recently been observed [12], if we broaden the indications for cholecystectomy to include patients with minimal or even no symptoms, the postcholecystectomy syndromes will become of much greater clinical relevance. More importantly, the small but real risk of bile duct injury and its sequelae must be weighed very carefully against the potential benefit for a patient with minimal or equivocal symptoms. It is perhaps too easy to recommend a minimally invasive operation which can be done as an ambulatory procedure. A cost/benefit analysis is needed to determine if the US population has been well served by the

introduction of laparoscopic cholecystectomy and by the significant increase in the cholecystectomy rate that this has produced. References
1. Baird D, Wilson J, Mason E, Duncan TD, Evans JS, Luke JP, Ruben DM, Lucas GW (1992) An early review of 800 laparoscopic cholecystectomies at a university-affiliated community teaching hospital. Am Surg 58: 206210 2. Cohen MM (1992) Initial experience with laparoscopic cholecystectomy in a teaching hospital. Can J Surg 35: 5963 3. Diehl A (1987) Trends in cholecystectomy rates in the United States. Lancet 2: 683 4. Diehl A (1993) Laparoscopic cholecystectomy: too much of a good thing? J Am Med Assoc 270: 14691470 5. Dubois F, Icard P, Berthelot G, Levard H (1990) Coelioscopic cholecystectomy. Ann Surg 211: 6062 6. Escarce JJ, Chen W, Schwartz JS (1995) Falling cholecystectomy thresholds since the introduction of laparoscopic cholecystectomy. J Am Med Assoc 273: 15811585 7. Ho HS, Mathiesen KA, Wolfe BM (1996) The impact of laparoscopic cholecystectomy on the treatment of symptomatic cholelithiasis. Surg Endosc 10: 746750 8. Legorreta A, Silber J, Constantino G, Kobylinski R, Zata S (1993) Increased cholecystectomy rate after the introduction of laparoscopic cholecystectomy. J Am Med Assoc 270: 14291432 9. McIntyre RC, Zoeter MA, Weil KC, Cohen MM (1992) A comparison of outcome and cost of open vs laparoscopic cholecystectomy. J Laparoendosc Surg 2: 143148 10. NIH consensus development panel on gallstones and laparoscopic cholecystectomy (1993) Gallstones and laparoscopic cholecystectomy. J Am Med Assoc 269: 10181024 11. Orlando R, Russell J, Lynch J, Mattie A (1993) Laparoscopic cholecystectomy: a statewide experience. Arch Surg 128: 494499 12. Phillips SF (1996) Diarrhea after cholecystectomy: if so, why? Gastroenterology 111: 816818 13. Ransohoff D, Gracie W (1990) Management of patients with symptomatic gallstones: a quantitative analysis. Am J Med 88: 154160 14. Ransohoff DF, Gracie WA (1993) Treatment of gallstones. Ann Intern Med 119: 606619 15. Steiner CA, Bass EB, Talamini MA, Pitt HA, Steinberg EP (1994) Surgical rates and operative mortality for open and laparoscopic cholecystectomy in Maryland. N Engl J Med 330: 403408 16. The Southern Surgeons Club (1991) A prospective analysis of 1518 laparoscopic cholecystectomies. N Engl J Med 324: 10731078

Surg Endosc (1997) 11: 919922

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Comparative retrospective study of surgical treatment of spontaneous pneumothorax

Thoracotomy vs thoracoscopy
R. Jime nez-Mercha n, F. Garc a-D az, C. Arenas-Linares, J. C. Giro n-Arjona, M. Congregado-Loscertales, J. Loscertales
Department of General and Thoracic Surgery, Hospital Universitario de Sevilla, Avda. Dr. Fedriani s/n, 41071 Seville, Spain Received: 23 September 1996/Accepted: 28 February 1997

Abstract Background: Since May 1992 we have used video-assisted thoracic surgery (VATS) for the treatment of the spontaneous pneumothorax. In this study we review the results obtained with this technique and we compare them with those obtained by conventional surgery (19761992). Methods: 110 patients (98 males and 12 females, age 1583 years) were operated by VATS, and 627 patients (567 males and 60 females, age 1489 years) by conventional surgery. The surgical technique and the complications are compared. No one died. Results: Although the number of complications was greater in the first group than in the second one (29.5% vs 15.1%), this is attributed to the lack of experience among surgeons in the first cases; which complications nearly disappeared in the last 60 patients. Less pain, better recovery, and shorter hospital stay resulted. Conclusions: For us VATS is the ideal technique with which to treat spontaneous pneumothorax. Key words: Spontaneous pneumothorax VATS Open surgery

only a medium-risk intervention with low morbidity and generally no mortality, many patients complain of considerable pain in the postoperative period that tends to take a long time to disappear [6]. With both options, the loss of working days is important. Since the introduction of videothoracoscopy and videoassisted thoracic surgery (VATS), the problems inherent to this type of surgery have been notably reduced. Postthoracotomy pain is minimized by using only three working ports, and this results in a greater ability of the patient to perform respiratory physiotherapy. Benefits are fewer pleuropulmonary problems and shorter postoperative hospital stay. All this contributes to a more rapid recovery and an earlier return to work [14, 18]. However, long-term results of the VATS approach are not well known. From the technical point of view, blebs and bullae are easily identifiable with this approach; sometimes vision is better than in open surgery. But blebs and bullae do not generate problems when removed with the help of the new endostapler devices [3], either, or in the realization of the pleural abrasion [5] or pleurectomy [4]. The aim of this study is to revise the results obtained in our series of patients operated by VATS and to compare them to a previous series of patients operated by thoracotomy.

Spontaneous pneumothorax is one of the afflictions most commonly treated by thoracic surgeons. The aim of treatment is to completely reexpand the lung and avoid recurrence. Until a shorttime ago, there were two options: placement of endopleural drainage or surgical treatment by thoracotomy. With the former, a failure and recurrence rate that surpasses the 30% has been reported [2, 9, 15]. The surgical treatment tends to thoroughly solve this problem; recurrence does not surpass 1% [11, 17]. Nevertheless, despite being
Correspondence to: J. Loscertales

Materials and methods

Until May 1992 in our department 627 patients (567 men and 60 women) had been operated on for spontaneous pneumothorax using conventional open surgery. The average age of these patients was 31 years (range, 14 to 89 years). In 324 cases the pneumothorax was located in the right side, in 262 in the left side; 11 were bilateral synchronous and 30 bilateral metachronous. Though we have used several approaches (Table 1), the most commonly used one has been Noirclercs lateral thoracotomy [13]. This thoracotomy allows entry into the thorax with only section of the intercostal muscle, since the opening of the serratus is made without sectioning it, only opening its fibers in their own course.

920 Table 1. Type of thoracotomy in open surgery Axillary thoracotomy Lateral thoracotomy Posterolateral thoracotomy Middle sternotomy 14 603 2 8 Table 2. Spontaneous pneumothorax; findings at operation by thoracotomy Section of adhesions Exeresis of Bullae-blebs Electrocoagulation of blebs Exeresis of giant bullae Exeresis of apical scars Others 174 (27.7%) 580 (92.5%) 31 (4.9%) 27 (4.3%) 105 (16.7%) 47 (7.5%)

Since May 1992, VATS has been used in the treatment of spontaneous pneumothorax in our department: 110 patients (98 men and 12 women) with an average age of 29.5 years (range, 15 to 83 years) have been operated. On 55 occasions the pneumothorax was located on the right side (50%) and in 49 on the left hemithorax (44.6%); in four cases it was bilateral metachronous (3.6%) and in two bilateral synchronous (1.8%). There are no significant differences regarding the age of the patients and the location of the pneumothorax between groups. We use the following technique for videothoracoscopy: patient on lateral decubitus opposite to the affected side, under general anesthesia, with double lumen tracheal tube. During the operation the lung collapse is maintained through the occlusion of one of the lumens of the double lumen tube. On certain occasions the anesthetist is asked to insufflate the lung in order to observe small blebs/bullae in a better way, and at the end to check if there are any air leaks. Three ports are opened with 12-mm trocars: the first one, in seventheighth intercostal space, middle axillary line, for the camera (0 straight vision), using two high-resolution monitors; thereafter two other ports are opened at the level of the thirdfourth space, anterior axillary line, and fifthsixth space under the tip of the scapula, through which the working instruments are introduced. Once the adherences of the lung to the wall, if any, are freed, the bullae or blebs are located, and they are extirpated with the endostapler device Endo-Gia. (Between one and eight loaders, an average of three, have been used.) The change of position of the camera in the anterior and/or posterior port must be always accomplished, since bullae/blebs can be found in the posterior segment, and these bullae cannot be properly visualised with the camera in its initial position. After that, a plastic sponge with its size adapted to the ports is introduced to accomplish the abrasion of the parietal pleura, and thereafter we instillate 150200 ml of 0.5% iodine alcohol solution to bath the whole visceral pleura. The possible air leaks are proven with this liquid, after asking the anesthetist to insufflate the lung, if there is any possible doubt about the line of suture. The action of the sponge in the pleurodesis is reinforced due to the irritation produced by the iodine. This is an original procedure of ours [8]. Upon ending the intervention, through the inferior port we introduce a 28 F drainage tube connected to the aspiration system with a water seal, and another small-caliber drainage-type Pleurocath, through the anterosuperior port, checking its location and the total lung reexpansion with the video camera. Normally, the fine drainage is withdrawn in 2448 h, while the tube in the inferior location is withdrawn in 4872 h if there is no air leak or important pleural effusion. The patient leaves the hospital the 1st day after the second tube removal, sending the patient to the Rehabilitation Department in order to continue with the assisted respiratory physiotherapy over 2030 days. The technique we have used in conventional surgery is similar to the one used here with VATS: removal of the lesion with staplers, also followed by parietal pleural abrasion with plastic sponge and bath of the visceral pleural with iodine solution [6]. After leaving the hospital, the patients are followed up in the policlinic. The first visit is at 1 month; visits then occur every 3 months during the 1st year and every 6 months until 5 years. The patients are annually followed up thereafter.

Table 3. Spontaneous pneumothorax; findings at operation by VATS Section of adhesions Exeresis of bullae-blebs Electrocoagulation of blebs Exeresis of giant bullae Exeresis of apical scars 37 (33.6%) 101 (91.8%) 9 (8.2%) 4 (3.6%) 15 (13.6%)

Results The pathologic findings both in open surgery and in the patients operated by videothoracoscopy are presented in Tables 2 and 3, respectively. There are no significant differences. The complications that appeared in the postoperative period of the patients operated by videothoracoscopy were more frequent in our first 30 cases (learning curve), since the technique was initiated in our department for the treatment of this disease. One must also indicate the direct re-

lationship between the complications and the lack of active collaboration on the part of the patient in the physiotherapy; in the last 60 patients we have had practically no complications. The global number of complications has been 26 (23.6%), distributed as follows: eight air leaks of more than 5 days of duration (8.2%); eight transitory apical air chambers (8.2%), which disappeared with physiotherapy in 7 days maximum, and among these ones only two needed a new pleural drainage Pleurocath; on four occasions the chamber was not eliminated either with physiotherapy or with the placement of the new drainage; six patients presented a few quantity pleural effusion, which has evolved toward its resolution after a drainage by thoracocentesis. In two cases (1.8%) the pneumothorax had a recurrence in 1530 days after the intervention. The four patients with the unresolved air chamber and the two ones with recurrences were operated through a lateral thoracotomy. In the former ones, bullae were not in evidence, having enough with a new pleural abrasion and a forced expansion of the lung to solve the problem. In the two patients with recurrence, one could be observed to have a bulla not removed (this was our second case) and the other one to have the suture staplers open, perhaps due to the improper use of a white loader (vascular) with 3-mm staples, instead of a blue one (parenchymatosous) with 3.5-mm staples. The exeresis of the bulla in the first patient and the resuture with the blue loader in the second one solved the problem. In the follow-up of the patients, between 1 month and 4 years, no more complications, or sequelae, or recurrences appeared. In the patients operated by thoracotomy the following complications appeared: 46 air leaks of more than 5 days of duration (7.6%); 32 apical air chambers (5.1%), and 10 of them required new placement of endopleural drainage; 12 pleural effusions (1.9%), which were solved with drainage; five wound surgical infections (0.8%). No recurrences have been observed in these patients. The duration of the operation in the patients operated by videothoracoscopy has gone from 100 min in the first cases to 3040 min nowadays, which is similar to open surgery. Postoperative hospital stay has decreased from 8.5 days mean in the open surgery to 4.5 days mean in those operated by thoracoscopy. The analgesia used in the patients that underwent to thoracotomy was 2 g intravenous magnesic metamizol four

921

times a day during the first 3 days, followed by one 575-mg pill every 6 h after that. This was carried out during the whole hospital stay; analgesia was maintained during an extended postoperative period (about 2 months) in 50% of the patients. However, the patients operated by VATS required maximum analgesic doses (magnesic metamizol, or, more recently, 30 mg intravenous ketorolac trometamol, four times a day) only the 1st day, and later one 10-mg pill every 6 h during the following 34 days, generally until the withdrawal of the posteroinferior drainage, continuing with analgesia thereafter only in 10% of the patients after leaving the hospital. The reintegration of the patients to their working lives is important and it is closely related to the postoperative pain and to the respiratory rehabilitation. The patients operated by videothoracoscopy are immediately sent to the Rehabilitation Department after they leave the hospital, and this keeps them from 20 to 30 days in respiratory physiotherapy programs. After that, most of the patients return to activities that they were involved in before the operation. On the contrary, more than the 60% of the patients operated by open surgery delay the return to their working activities 2 or more months, which implies an important loss of working days. Though less important than the parameters previously indicated, we should not forget the cosmetic results. While we can operate by videothoracoscopy with only three incisions of about 15 mm, in open surgery we need a thoracotomy of at least 1015 cm, and this implies worse cosmetic results.

Discussion Since May 1992 we have performed surgical treatment of spontaneous pneumothorax using videothoracoscopy [8]. Until then, we were strong supporters of surgical treatment by limited lateral thoracotomy since we consider that the resection of the bullae and the pleural abrasion is the only effective treatment against this affliction. We do not agree with treatment based on instillation of irritant agents in the pleural space, such as tetracycline or talc, because results are poor, although on certain occasions we are obliged to use them in patients of advanced age with concomitant important diseases that contraindicate the surgical treatment [7]. On the other hand, conventional thoracotomy causes considerable postoperative pain, and this implies less collaboration by the patient during the postoperative period, which increases the postoperative complications and hospital stay. With the introduction of videothoracoscopy, these problems are enormously minimized, reducing pain and stay in the hospital [1, 12, 16]. When we compare the results in our experience obtained by VATS (n 110) with those for thoracotomy (n 627), we see that although it is difficult to specify the intensity of the pain, the patients in the first group complain much less and carry out the respiratory physiotherapy with more intensity and less trouble; the analgesic doses, equal during the 1st day, are reduced thereafter in the VATS group. Equally, in the follow-up, they present less pain and they do not present dysesthesia, which customarily occurs after the thoracotomy.

With regard to pleurodesis, we do not create any scarificationneither with a laser (because we do not have any) nor with electrocoagulation because considerable pain is reported in such cases [10]. Neither do we accomplish apical pleurectomy since it does not provide more benefits than the abrasion with sponge and the iodine instillation, and it sometimes produces serious bleeding problems. Upon analyzing the complications presented between both groups, we observe that, of the 627 patients operated by thoracotomy, we had a complication rate of 15.1%, smaller than in the thoracoscopy group (23.6%). Nevertheless, one must take into account that VATS is a new technique that we began to use precisely in this type of pathology, the complications having almost practically disappeared in the last 60 patients. It is quite possible that the intensity with which we accomplished the pleurodesis at the beginning was not enough to achieve a right pleural symphysis. On the other hand, we also began to place a Pleurocath in the inferior port, but we abandoned it because we realized it was not enough to evacuate the effusion generated after the abrasion. With regard to the duration of the intervention, even though at the beginning it lasted 100 min mean, currently it is 3040 min mean, similar to the time used with thoracotomy. No one died in either group and the hospital stay mean has decreased from 8.5 days (range, 522) for those operated by thoracotomy to 4.5 days (range, 214) for those treated by VATS. In conclusion, we consider that VATS must currently be the elective method of choice for treatment of spontaneous pneumothorax, even in the first episode. Minimal thoracic incisions reduce the postoperative pain; this results in better recovery of the patient, decrease in the postoperative hospital stay, earlier return to working life, and consequent socioeconomic savings. Finally cosmetic results are also good.

References
1. Bertrand PC, Regnard JF, Spaggiari L, Levi JF, Magdeleinat P, Guibert L, Levasseur P (1996) Immediate and long-term results after surgical treatment of primary spontaneous pneumothorax by VATS. Ann Thorac Surg 61: 16411645 2. Elfeldt RJ, Schroeder D, Meinicke O (1991) Spontanpneumothorax uberlegungen zur aetiologie und therapie. Chirurg 62: 540546 3. Hazelrigg SR, Landreneau RJ, Mack JM, Acuff T, Seifert PE, Aver JE, et al. (1993) Thoracoscopic stapled resection for spontaneous pneumothorax. J Thorac Cardiovasc Surg 105: 389393 4. Inderbitzi R, Leiser A, Furrer MU (1994) Three years experience in video-assisted thoracic surgery for spontaneous pneumothorax. J Thorac Cardiovasc Surg 107: 14101415 5. Liu HP, Lin PJ, Hsieh MJ, Chang JP, Chang CH (1995) Thoracoscopic surgery as a routine procedure for spontaneous pneumothorax. Results from 82 patients. Chest 107: 559562 6. Loscertales J, Ayarra FJ, Garc a D az FJ, Arenas Linares C, Rico Alvarez A (1988) Neumoto rax esponta neo. Ed Elba S.A. 7. Loscertales J, Garc a D az F, Jime nez Mercha n R, Ayarra FJ, Arenas Linares C, Giro n Arjona JC (1994) Tratamiento del neumoto rax esponta neo en pacientes mayores de 70 an os. Arch Bronconeumol 30: 344347 8. Loscertales J, Jime nez Mercha n R, Ayarra Jarne FJ, Garc a D az FJ, Arenas Linares C, Giro n Arjona JC (1995) Nuestra experiencia en el

922 tratamiento del neumoto rax esponta neo por videotoracoscopia. Cir Espan ola 57: 526529 Mercier C, Page A, Verdant A, Cossette R, Dontigny L, Pelletier C (1976) Outpatient management of intercostal tube drainage in spontaneous pneumothorax. Ann Thorac Surg 22: 163165 Mitchell RL (1990) The lateral limited thoracotomy incision standard for pulmonary operation. J Thorac Cardiovasc Surg 590596 Murray K, Matheny R, Howanitz P, Myerowitz D (1993) A limited axillary thoracotomy as primary treatment for recurrent spontaneous pneumothorax. Chest 103: 137142 Nathanson LK, Shimi SM, Wood RAB, Cuschieri A (1991) Videothoracoscopic ligation of bulla and pleurectomy for spontaneous pneumothorax. Ann Thorac Surg 62: 316319 Noierclerc M, Dor V, Chauvin G (1973) Extensive lateral thoracotomy without muscle section. Ann Chir Thorac Cardiovasc 12: 181186 14. Sampietro R, Biraghi T, de Angelis G, dUrbano C, Fuertes F (1993) Videotoracocirug a: un nuevo me todo para el tratamiento de la patolog a pulmonar. Cir Espan ola 53: 287291 15. Van Test WF, Roukema JA, Verpalen MC, Lobach HJ, Palmen FM (1991) De spontane pneumothorax; operen of niet? Ned Tijdschr Geneeskd 135: 11741178 16. Waller DA, Forty J, Morritt GN (1994) Video-assisted thoracoscopic surgery versus thoracotomy for spontaneous pneumothorax. Ann Thorac Surg 58: 372377 17. Weeden D, Smith G (1983) Surgical experience in the management of spontaneous pneumothorax, 197282. Thorax 38: 737743 18. Yim APC, Ho JK, Chung SS, Ng DYC (1994) Video-assisted thoracoscopic surgery for primary spontaneous pneumothorax. Aust NZ J Surg 64: 667670

9. 10. 11. 12. 13.

Surg Endosc (1997) 11: 944945

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

A new approach in the management of incarcerated hernia

Emergency laparoscopic hernia repair
K. Liao, J. Ramirez, S. Carryl, G. W. Shaftan
Department of Surgical Services, The Brookdale University Hospital and Medical Center, Linden Boulevard and Brookdale Plaza, Brooklyn, NY 11212, USA Received: 18 November 1996/Accepted: 26 December 1996

Abstract. The authors describe the case of a 74-year-old male presenting with an incarcerated epigastric hernia. An algorithm for successful management of such a case is proposed. Key words: Emergency laparoscopy Incarcerated hernia

Although laparoscopy has been widely used for elective hernia repair, little is known about its role in the emergency management of incarcerated hernias. A case of incarcerated epigastric hernia which was successfully managed by laparoscopy is reported here. Case report
A 74-year-old black male with a past medical hisotry of hypertension, diabetes, heart disease, and two episodes of stroke with resultant right hemiplegia was brought to the Emergency Department of the Brookdale University Hospital and Medical Center after his primary care physician found a painful mass in the abdominal wall about 14 h previously. Prior to this visit, the patient had abdominal discomfort and recurrent vomiting during the past month. His wife recalled that the patient had a golf-ball-size mass at the midabdomen about 2 weeks ago but it disappeared spontaneously 5 h after she massaged the mass. The patient had three episodes of vomiting after he arrived at the hospital; his last bowel movement was the night before coming to the Emergency Department. Physical examination revealed a moderately obese man lying supine who was not sufficiently oriented to answer any questions. The vital signs were: blood pressure 136/80 mmHg, pulse 89 per min, respiration 20, and temperature 36.3C. The abdomen was slightly distended with poor abdominal muscle tone, especially on the right side. A 5 6 6 cm supraumbilical mass was palpated that was firm, tender, and irreducible. The surface of the mass was smooth and it was normothermic. There was moderate guarding around the mass and in the right lower quadrant but there was no obvious rebound. Bowel sounds were audible; rectal examination was unremarkable and stool guaiac was negative. Hemoglobin was 12.8 g/dl, hematocrit 39.8%, white blood cell 7,200/mm3, and urinalysis was normal. Abdominal X-ray

showed a few slightly dilated small-bowel loops with no air-fluid levels and air was noted in the colon. A diagnosis of incarcerated epigastric hernia was made. Emergency operation was planned; however, during the transport of patient to the operating room, the hernia reduced spontaneously. A 2.5-cm fascial defect was found in the midline, 2 cm above the umbilicus. There still was some tenderness and guarding in the supraumbilical area and in the right lower quadrant. Since ischemic bowel could not be ruled out based on the clinical findings, laparoscopy was performed. After nasogastric and bladder intubation, a 1-cm skin incision was made along the right anterior axillary line at the level of midabdomen through which a 10-mm trocar was placed into the abdominal cavity by an open approach. The abdomen was insufflated with carbon dioxide to maintain 15 mmHg pneumoperitoneum and then the camera was inserted. Another two 5-mm trocars were placed under direct vision at the right upper and right lower quadrants. A 2.5 cm linear defect was seen in the linea alba just above the umbilicus and a 8 cm segment of small bowel was found to be erythematous and edematous with patchy ecchymoses on the serosa which involved the entire circumference. No free fluid was found inside the abdomen. With the help of the graspers, a piece of 4 8 cm2 Gore-Tex DualMesh (W. L. Gore & Associates, Inc., Flagstaff, AZ) was inserted inside the abdomen and was laid flat over the defect. Two 2-0 nylon straight needle sutures were placed percutaneously at about 1 cm above and below the defect to anchor the mesh at 6 and 12 oclock: The needle passed through the abdominal wall and the mesh and with the help of the grasper inside; the needle then was pushed back through the mesh and the abdominal wall, exiting close to the entrance. The knots were tied just underneath the skin. A Tacker (Origin Medsystems, Inc., Menlo Park, CA) was inserted via a 5-mm trocar and multiple titanium tacks were fired to fix the mesh around the defect. Good covering of the mesh over the fascial defect was achieved after 12 tacks were placed. The previous hernia defect area appeared tight when the pressure was applied from outside. The compromised smallbowel segment was reevaluated before the camera was removed, and the color of the bowel has improved dramatically compared to its appearance at the beginning of the operation. Good peristalsis also was noted in that segment of bowel. After removal of all the trocars, the fascia and the skin openings were closed. Postoperatively the patient recovered without complications and he was discharged from the surgical service on the 5th postoperative day.

Discussion Traditionally, the decision to perform closed taxis or emergency operative repair for an incarcerated hernia is based on the presence or absence of the clinical signs of impending or

Correspondence to: K. Liao

945

Fig. 1. Clinical approach to the incarcerated hernia. H + P, evaluated by history and physical examination; +, success; , failure.

actual strangulation of the hernial contents. These include pain in the region of the hernial swelling, tenderness to palpation, and discoloration of the tissues over the swelling [4]. The clinical ability to objectively identify strangulation of bowel is far from ideal, especially in the aged patients with multiple medical problems and compromised mental status. Too-aggressive emergency surgical intervention for incarcerated hernias, especially when laparotomy is required, carries a much higher mortality and morbidity than do those operated electively [3, 5]. On the other hand, toovigorous efforts to reduce an incarcerated hernia may result in reductio en masse or reduction of strangulated, nonviable bowel [4]. Most surgeons have found it very difficult to choose between emergency operation and observation if confronted with the clinical situation of an unreliable, disoriented patient. Laparoscopy should be considered in incarcerated hernias under certain circumstances. Although incarcerated epigastric hernia is rare, similar principles of management also should apply with other abdominal hernias. With laparoscopic herniorrhaphy one can thoroughly evaluate the viability of reduced hernial contents without exposing them to the external environment, and, at the same

time, repair the hernia; there is much less trauma for the patient compared with open repair, especially when celiotomy is required to evaluate bowel viability [1, 2]. In the era of laparoscopic surgery and managed care, the application of laparoscopy to the management of incarcerated hernias certainly offers an extra choice for the surgeon. An algorithm for management is proposed (Fig. 1). References
1. Barkun JS, Wexler MJ, Hinchey EJ, Thibeault D, Meakins JL (1995) Laparoscopic versus open inguinal herniorrhaphy. Surgery 118: 703 710 2. Feldman MG, Russell JC, Lynch JT, Mattie A (1994) Comparison of mortality rates for open and closed cholecystectomy in the elderly: Connecticut statewide survey. J Laparoendosc Surg 4(3): 165172 3. Hjaltason E (1981) Incarcerated hernia. Acta Chir Scand 147: 263267 4. Nyhus MN, Bombeck CT, Klein MS (1991) Hernias. In: Sabiston DC (ed) Textbook of surgery. 14th ed. WB Saunders, Philadelphia, pp 11341148 5. Stephens BJ, Rice WT, Koucky CJ, Gruenberg JC (1992) Optimal timing of elective indirect inguinal hernia repair in healthy children: clinical considerations for improved outcomes. World J Surg 16: 952 957

Review article
Surg Endosc (1997) 11: 883893

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Laparoscopic anatomy of the region of the esophageal hiatus

G. G. R. Kuster,1 F. A. Innocenti2
1 2

Division of General Surgery, MS213, Scripps Clinic and Research Foundation, 10666 North Torrey Pines Road, La Jolla, CA 92037, USA Department of Surgery, University of Concepcion School of Medicine, Concepcion, Chile

Received: 24 October 1996/Accepted: 26 December 1996

Abstract Background: The wide patient acceptance of hiatal and gastroesophageal surgery performed by laparoscopy has suddenly generated a large volume of procedures frequently done by surgeons with limited experience in this area. This has resulted in an excessive number of complications. Knowledge of the normal and pathologic laparoscopic anatomy is essential for safe dissection around the esophageal hiatus. Methods: This description is based on the experience gained during 850 open and 150 laparoscopic surgeries in and around the hiatus and on the review of the literature. Results: Laparoscopic approach, dissection, and accessibility of the hiatus and surrounding organs are different than those experienced through cadaveric dissection and open surgeries. Conclusions: Clear understanding of the normal and pathologic anatomy and its variations facilitates laparoscopic dissection of the hiatus and neighboring structures and should help the surgeon avoid complications. Key words: Laparoscopic anatomy Esophageal hiatus Diaphragm Laparoscopic fundoplication Gastroesophageal junction

Complications such as excessive bleeding [12, 16, 20, 21]; esophageal and gastric perforations [9, 12, 20, 21, 41, 49, 54]; pneumothorax [12, 16, 20, 21, 24, 54]; pneumodiastinum [10]; paraesophageal herniation [10, 21, 54]; dysphagia [12, 16, 21, 38]; and delayed gastric emptying and diarrhea possibly due to vagal injury [12, 37, 48] have occurred more often after laparoscopic procedures in the hiatal area than after open surgery. This is due, in part, to the surgeons lack of familiarity with the laparoscopic anatomy of the region.
Presented at the 4th International Congress of the European Association for Endoscopic Surgery, Trondheim, Norway, June 1996 Correspondence to: G. G. R. Kuster

Many procedures have been performed by surgeons with great expertise in advanced laparoscopic procedures but limited experience with open surgery of the hiatal area [13]. A clear understanding of the normal and pathologic laparoscopic anatomy of the hiatus and surrounding organs should minimize the risks of injuries and poor functional results, making the laparoscopic approach comparable to or better than the open procedure [34, 37, 39, 46, 50, 51, 54]. Experience gained with laparoscopic cholecystectomy has demonstrated that knowledge of the anatomy and its variations, as well as the distortions resulting from disease, is essential to anticipate and prevent technical complications [1, 11, 29, 42]. The region of the hiatus involves several organs, which in some pathologic conditions may vary significantly in their relationship, shape and size, tissue consistency, and fragility. This study is based on the anatomic findings of 850 open and 150 laparoscopic surgical dissections performed in the area of the hiatus and on a review of the literature. The detailed anatomic descriptions of the surgical dissections recorded by the senior author were reviewed for location, size, consistency, variations, pathologic distortions, and other elements of interest encountered in the structures around the hiatus. These operations included hiatal hernia repairs; antireflux procedures; distal esophagectomies; esophageal myotomies; local removal of tumors, diverticuli, or cysts; total and proximal gastrectomies; splenectomies; vagotomies; and miscellaneous procedures involving the diaphragm and perihiatal areas. The anatomic description will be presented in the order that the surgeon encounters the organs and planes of the tissue when dissecting the area for various surgical procedures. Special emphasis is placed on the frontal view of the anatomic structures as seen and felt with the laparoscopic instruments. Patient position and trocar insertion for hiatal approach The patient is placed in a semilithotomy position with or without a lumbar cushion [18] and support under the gluteal

884

area to prevent the patient from sliding when placed in steep Fowlers position. The laparoscope can be introduced through any area of the upper abdominal wall. However, the fact that the diaphragmatic hiatus is quite posterior and cephalad within the abdominal boundaries should be considered. The location of a mid-distance between the xiphoid and the umbilicus is used for most operations. A 0 angle scope placed at the level of the umbilicus or below will offer a tangential view of the hiatus and will not provide a good retroesophageal visualization. This can be offset by the use of a 30 or 45 angle scope. The exact position of the scope should be tailored to the size of the patient and the actual procedure. The size of the lateral segment of the left lobe of the liver will also influence the position of the scope. This is particularly important for the location of the port which will accommodate the liver retractor. It is a good practice to introduce the first trocar in a site which is not critical, like the left subcostal margin, and to insert the rest of the trocars under direct vision after a complete abdominal exploration, tailoring the site and angle of insertion according to the function of each port. For procedures that will involve only the front part of the hiatus, umbilical access is adequate [30]. A general view of the upper abdomen will disclose the front portion of the dome of the diaphragm, the lateral segment of the left lobe of the liver, and part of the stomach. In large paraesophageal hiatal hernias, half or more of the stomach is in the chest cavity and the general examination of the upper abdomen may not reveal any portion of the stomach without lifting the liver and/or reducing the stomach by exerting traction on the gastrocolic omentum [27, 28]. In the majority of adult patients, the spleen is not visualized during this general examination of the upper abdomen. Part of the transverse colon might also be seen, depending on the thickness of the greater omentum, which may preclude the frontal view of the entire transverse colon, particularly in obese people. The falciform ligament is also well visualized, and an examination of the right side of the upper abdomen will disclose the diaphragm, the rest of the liver and, most of the time, the fundus of the gallbladder. The rest of the ports are placed in the subcostal areas. Depending on the body build of the patient, most access cannulas will be located well below the coronal plane of the hiatus, except for those entering at the right or left side of the xiphoid, which may reach the hiatus directly posteriorly in individuals with a wide habitus.

Fig. 1. Exposure of the hiatal area by retracting the lateral segment of the left lobe of the liver (a); (b) triangular ligament; (c) stomach and esophagus being retracted to the left; (d) gastrohepatic omentum with its thin area (window) between a branch of the left inferior phrenic artery (above) and the hepatic division of the vagus nerve and an aberrant hepatic artery (below).

Left lobe of the liver For exposure of the hiatus, the lateral segment of the left lobe of the liver needs to be retracted upward with a nontraumatic instrument (Fig. 1) or folded to the right after dividing the left triangular ligament. The triangular ligament does not insert exactly along the posterior border of the left hepatic lobe but instead on its superior surface, about 1 cm from the posterior border [35]. The insertion on the liver reaches the posterior border as it approaches the inferior vena cava. This ligament is thin and contains minor vessels, particularly at the left end, but it widens at the right to accommodate the left hepatic veins and the inferior vena cava. The left triangular ligament measures an average of

1.5 cm wide in its left half and narrows to the right. The length of the left triangular ligament varies widely, following the transverse length of the left lobe of the liver. There are cases of long ligaments extending well beyond the left end of the liver to the lateral abdominal wall, as well as cases of short or rudimentary ligaments. If the triangular ligament is transected, attention should be given to avoid injury to the upper end of the spleen (particularly in the presence of a very long left hepatic lobe reaching over and beyond the upper pole of the spleen), as well as the liver, the fundus of the stomach and esophagus, the left suprahepatic veins, and branches of the left and right diaphragmatic arteries and veins. Variations in the size and shape of the left lobe of the liver are frequent. The liver as seen by frontal radiocolloid scintigraphy shows a smaller craniocaudal dimension in supine position than standing position [22]. This is particularly true in the left lobe of normal individuals (27% decrease is seen). Cirrhosis allows less pliability, and therefore the decrease in size is only 18%. The left lobe of the liver may present anomalous lobulations, accessory lobes with a pedicle, ectopic hepatic tissue without pedicle, and hyperplasia, hypoplasia, or absence of the lateral segment. Accessory hepatic lobes presenting as large and pedunculated masses are rare [25]. Jaques et al. [23] described ten cases of hypoplasia or aplasia of one or both segments of the left lobe among 517 abdominal computed tomographic scans. The left lobe is frequently elongated and embraces the spleen [4]. The left lobe may be congentially atrophic or totally absent. Congenital atrophy has also been described in association with hypertrophy of the caudate lobe, with gastric volvulus [3], and with hepatodiaphragmatic colon interposition (Chilaiditis syndrome) [23]. Acquired atrophy of the left lobe of the liver can be focal or diffuse. Volume diminution is a frequent finding in cirrhosis. The consistency, texture, and color of the liver vary in

885

normal as well as pathologic conditions. The increased rigidity noted in cirrhosis in relation to the degree of fibrosis may interfere with the mobility and bendability of the organ. The presence of soft and fragile liver should be of major concern to the laparoscopist because of the possibility of surgical trauma resulting in intra- and postoperative hemorrhage, bile leaks, or tissue necrosis. Some elderly patients and patients with fatty infiltration of the liver, acute hepatitis, hypercortisonism, or lymphoma may present with an extremely fragile liver. The surgeon and assistants should be extremely gentle with liver retractors and the blind introduction of instruments that could injure the liver. The Gleason capsule may detach easily from the liver parenchyma when traction is applied to adhesions or hepatic ligaments.

Gastrohepatic omentum The lesser omentum, also known as the gastrohepatic omentum, ligament, or membrane, extends from the porta hepatis to the lesser curvature of the stomach, the antero-right wall of the esophagus, and the diaphragm. The gastrohepatic omentum is best visualized by exerting traction on the stomach toward the left side and caudally (Fig. 1). It contains branches of the left gastric artery and vein, the hepatic branch of the anterior vagus nerve, the posterior and anterior Latarjet nerves, and some lymphatic vessels and nodes. It also frequently carries branches of the right gastric artery and vein, an aberrant hepatic artery, and branches of the left inferior phrenic artery and vein. The caudal portion of the lesser omentum corresponds to the hepatoduodenal ligament, extending from the liver to the first 2.5 cm of the duodenum. The free edge of this ligament forms the anterior boundaries of the hiatus of Winslow and contains the hepatic artery, the portal vein, the extrahepatic biliary tract, lymphatic vessels, and nodes. The cephalad end of the lesser omentum extends to the diaphragm, where its anterior layer reflects upward behind the posterior border of the left lobe of the liver to join the triangular ligament [35]. The thickness of the gastrohepatic omentum changes from the diaphragm to the hiatus of Winslow. The cephalad portion (0.5 to 2 cm long by 2 to 3 cm wide) is fibrous, opaque, and tense. It joins the phrenoesophageal ligament and frequently contains branches of the left inferior phrenic and left gastric vessels. Caudally to this thick segment is an area 3 to 6 cm long by 3 to 4 cm wide which forms a transparent window, with only the anterior and posterior peritoneal layers and minimal fat surrounding a loose network of very thin vessels. This area, called the Kuster window by the trainees, permits the laparoscopic view of the caudate lobe of the liver and the right crus of the diaphragm. The window portion of the lesser omentum is a very important landmark in laparoscopic dissection of the area of the hiatus since it is a constant finding and can be easily torn with or without minimal bleeding, gaining immediate access to the right side and the posterior portions of the hiatus and its contents (Fig. 2). In large hiatal hernias, particularly paraesophageal, the lesser omental window may not be readily apparent until reducing the stomach to its intraabdominal position. Caudal to the window, the lesser omentum becomes
Fig. 2. The thin window portion of the gastrohepatic omentum has been opened to show the (a) caudate lobe of the liver and (b) right crus of the diaphragm. The dissection of the hiatus is initiated by blunt dissection of the free border of the right crus with a Peanut sponge mounted on a rod (c).

thicker with fat and contains the hepatic vagal branch crossing from the lesser curvature to the liver and the aberrant hepatic artery, which is present in 11 to 23% of the cases [17, 20].

Caudate lobe of the liver The caudate lobe can be seen through the transparent gastrohepatic omentum and can be reached by opening the lesser omentum at the window (Fig. 2). This lobe is situated between the inferior vena cava to the left and the right crus of the diaphragm to the right. Occasionally, the caudate lobe protrudes to the right between the porta hepatis and the inferior vena cava, constituting the so-called caudate or papillary process ]19, 52]. Particular attention should be given to avoid injury to the caudate lobe during dissection of the right crus of the diaphragm and during repair of hiatal hernias at the time of placement of sutures between the left and right crus of the diaphragm behind the esophagus.

Esophageal hiatus The esophageal hiatus is an inverted teardrop-shaped opening located about 1 cm to the left of midline, at the level of the tenth thoracic vertebra. Its anterior and lateral borders are formed by the muscular and fibrous edge of the diaphragmatic crura, and the posterior angle is formed by the medial arcuate ligament. The crura start in the anterolateral surface of the first lumbar vertebrae, the intervertebral joints, and the anterior longitudinal ligament. The right crus inserts on the anterior surface of L1 to 3 or 1 to 4 and the left crus on L1 to 2 or 1 to 3, as well as on the intervertebral disks and the arcuate ligament. The crura extend cephalad and anteriorly to form the hiatus, inserting anteriorly into the transverse ligament of the central tendon of the dia-

886

Fig. 3. View of the hiatus after complete dissection: (a) anterior and posterior vagus nerves retracted together with the esophagus; (b) right and left crura; (c) upper pole of the spleen.

phragm. There are many variations in the formation of the hiatus [32]. In more than 90% of cases, the right crus forms the right and left margins of the hiatus, or the left crus contributes to the right margin. The right crus is immediately behind the gastrohepatic omentum and is located caudally in relation to the left crus (Fig. 3). The hiatus closely surrounds the esophagus. There may be space between the esophagus and the hiatus posteriorly, filled with areolar or fatty tissue, but normally no more than 2 cm. There should not be any sac herniation of peritoneum between the esophagus and the hiatus in any place, unless there is a hiatal hernia. The orientation of the hiatus in relation to the esophagus is neither perpendicular nor symmetrical. The elliptical opening of the hiatus is slanted to the right in its posterior end in such a way that the esophagus passes in front of the left crus. In addition, the hiatal opening is oriented in a sagittal view from anterior-cephalad to posterior-caudal, with the esophagus passing into the abdomen first in its anterior wall and then its lateral walls obliquely and finally its posterior wall at a subdiaphragmatic position 2 to 2.5 cm further caudally. In large paraesophageal hernias the mere surgical displacement of the esophagus from its posterior location to the anterior border of the hiatus brings 2 to 3 cm of esophagus to an intraabdominal position. The laparoscopic dissection of the hiatus should be done bluntly, since there are no important vessels between the hiatal border and the esophagus (Fig. 2). This is greatly facilitated by the use of a gauze sponge mounted at the tip of a rod (Peanut sponge, Ideas for Medicine, Inc., Clearwater, FL 34622) impregnated with a 1:10,000 epinephrine solution [26]. The dissection is started by opening the lesser omental window and cutting the peritoneum along the medial border of the right crus. The right crus is then pushed to the right and the hiatal contents are pushed to the left, separating the two until the arcuate ligament is reached, so the dissection of the left crus may be started from below. This dissection is performed retroperitoneally. To reenter

the peritoneum in the left side of the hiatus, a rent should be made in the peritoneum between the left crus and the right wall of the esophagus. This can be done from behind the esophagus or from the left side by retracting the cardia and esophagus caudally and to the right. If the dissection is advanced cephalad to the edge of the left crus, the mediastinum is entered, and further dissection will perforate the left pleura. The anterior border of the hiatus is dissected by detaching and lifting the phrenoesophageal ligament from the anterior wall of the esophagus or by cutting through the phrenoesophageal ligament transversely at the edge of the diaphragm until the transverse muscle bundles of the hiatus are identified in front of the longitudinal fibers of the esophagus and the anterior vagus nerve. Dissection of the hiatus is extended into the mediastinum, where one encounters the pericardium in the front, the right and left mediastinal pleuras laterally, the aorta posteriorly, and the esophagus with its vessels, lymph nodes, and vagus nerves in the center. In open surgery, the dissection of the hiatus is usually started anteriorly. In laparoscopic surgery, the dissection should be started posteriorly; otherwise drops of blood might drain down from the front, obscuring the tissues: red muscle fibers, yellow fat, grayish fascia and ligaments, and white vagus nerves. Both right and left crus can be, and have been, mistaken for the esophagus and can be dissected, sometimes entering the pleural cavity. This error is avoided by paying attention to the direction of the muscle fibers and by moving the cardial end of the stomach right and left; the esophagus follows the stomach while the crura are fixed posteriorly. The right and left crus can be transected at midlevel if the hiatus needs to be enlarged (i.e., to accommodate the stomach or colon to replace a resected esophagus). In this case the pleura should be retracted to avoid entering the chest cavity and branches of the left inferior phrenic artery may need ligature. The crura on the sides and behind the esophagus have little or no tendinous fibers. The hiatal border is stronger anteriorly because of the endothoracic and endoabdominal fascia and the central tendon of the diaphragm. The crura become tendinous as they reach their vertebral origin and their fusion with the arcuate ligament. Most of the length of the right and left crus is formed by a 11.5-cm-thick bundle of very soft parallel muscle fibers which tear easily if grasped or sutured and tied firmly.

Phrenoesophageal ligament or membrane The phrenoesophageal ligament is formed by the cephalad extension of the endoabdominal fascia and the caudal extension of the mediastinic fascia fusing at the hiatus. The fibers which come from the endoabdominal fascia go through the hiatus and insert in the adventitia and the connective tissue located between the muscle fibers of the esophagus 1 to 2 cm above the hiatus, completely encircling the esophagus. A second sheet of this same fascia descends to insert in the adventitia of the anterior half of the perimeter of the abdominal esophagus, 1 to 2 cm caudal to the hiatal border (Fig. 4). The fibers that come from the endothoracic

887

Fig. 4. The phrenoesophageal ligament (a) covers the anterior muscular border of the hiatus and the proximal 1 to 2 cm of the abdominal esophagus. The distal anterior and lateral walls of the esophagus are covered by peritoneum, making the esophagus a retroperitoneal organ (b).

Fig. 5. The muscle layers of the distal esophagus and gastroesophageal junction: (a) external longitudinal muscle bundles of the esophagus and stomach; (b) internal transverse layer of the esophagus; (c) crisscrossing of muscle fibers starting at the cardia.

fascia are thinner and less constant than the abdominal fascia, and they join the endoabdominal fibers. During fetal life, these two fascias are very strong, with fibrous and muscular components [5]. They regress during the maturation of the fetus although muscle fibers may persist. With the active swallowing and respiratory movement after birth, this union is modified and becomes more lax; connective and fatty tissue fill the space between the two layers. The membrane is less evident in the adult except in its anterior portion, where a very distinct, strong, white fascia is clearly seen. Patients with large hiatal hernias may show a very elongated, thin, and practically nonexistent phrenoesophageal ligament [17]. Vessels between esophageal and inferior phrenic arteries and veins pass over the ligament and need control before detaching the phrenoesophageal ligament from the esophagus.

Esophagus The length of the abdominal segment of the esophagus which can be seen by laparoscopy varies from 0.5 to 2.5 cm. It is located between T11 and L11, but this position varies with height and body build of the individual. Before dissection, the location of the esophagus is not clearly apparent by laparoscopy. The abdominal esophagus is in contact with the left lobe of the liver anteriorly, with the caudate lobe and part of the right crus on the right side, with part of the left crus and fundus of the stomach on the left, and with the left crus of the diaphragm, the arcuate ligament, and the aorta posteriorly. The upper pole of the spleen may reach the left wall of the esophagus in cases of splenomegaly. The abdominal esophagus is essentially a retroperitoneal organ, since it is covered with peritoneum anteriorly and partially covered at its lateral walls, whereas the posterior

wall is not covered by peritoneum (Fig. 4) and is separated from the retroesophageal elements described above by areolar or fatty tissue. In sliding hiatal hernias, the entire abdominal esophagus may be found in the mediastinum, or it may float between the abdomen and the supradiaphragmatic area according to the prevailing pressures between the chest and abdominal cavities. In pure paraesophageal hernias the gastroesophageal junction remains in a subdiaphragmatic position, with herniation of the stomach and/or other tissues alongside the esophagus, between the esophagus and the hiatus. Retroesophageal herniations of retroperitoneal fat and posterior gastric wall do not create a peritoneal sac in the mediastinum, whereas paraesophageal hiatal hernias located anteriorly will form a sac by elongation of the phrenoesophageal ligament and peritoneum [27, 28]. The distal mediastinal esophagus can be mobilized down to the abdomen for 3 to 15 cm (Fig. 3). This may require detachment from periesophageal areolar mediastinal tissue, right and left pleura, aorta, and pericardium, and may require division of arteries coming directly from the aorta. The vagus nerves limit the amount of mobilization. Maximal mobilization is obtained by division of the anterior and posterior vagus nerve in the lower mediastinum.

Gastroesophageal junction The gastroesophageal junction is normally located to the left of the midline at the level of T10, at the level of the anteroposterior projection of the seventh costal cartilage. The longitudinal muscle layers of the esophagus run straight through the gastroesophageal junction into the stomach where they spread like a fan (Fig. 5). The internal transverse muscular sheet of the esophagus is formed by much shorter bundles of incomplete circles around the submucosa. These two layers are very distinct and can be easily

888

dissected apart from each other and from the submucosa of the esophagus. The fibers of the inner muscle layer change their direction at the gastroesophageal junction to create a pattern of long and short bundles with different orientation. The long bundles form the so-called gastric sling embracing the anterior and posterior gastric walls to the left of the cardial end of the stomach. They create the cardiac notch or angle of His between the esophagus and the fundus of the stomach. The short bundles form the so-called semicircular clasps and maintain the transverse orientation of the semicircular inner esophageal layer, embracing the lesser curvature. Both sling and clasp fibers are more abundant at the gastroesophageal junction, forming a thicker layer than in the esophagus [47]. The crisscrossing of longitudinal, oblique, and transverse fibers is seen at the entrance to the stomach when a longitudinal esophageal myotomy is performed. In addition, the separation of muscle layers and the dissection away from the gastric mucosa is notoriously more difficult than in the esophagus. These features greatly facilitate the laparoscopic myotomy which is frequently performed for achalasia of the esophagus. However, the fibrosis and inflammation of the esophageal wall observed in peptic strictures preclude an easy distinction between esophagus and stomach, since the longitudinal and transverse layers may be fused and adherent to the mucosa. The same problem is seen in posttraumatic strictures of the lower esophagus and after perforations. Endoscopic biopsies of the esophageal mucosa may result in the same problem of fixation to the muscular layers. Preoperative mucosal biopsies over a leiomyoma or lipoma may lead to penetration of the esophageal lumen during laparoscopic enucleation of the tumors. A circular or annular muscular formation constituting a sphincter does not exist in the human gastroesophageal junction [47]. The manometric high-pressure zone noted at the level of the lower esophageal sphincter appears to be related to the sling and clasp fibers at the gastroesophageal junction. Myotomy of these fibers abolishes the highpressure zone and the gastroesophageal competence [33, 34].

this report. The change of position of the stomach in hiatal hernias varies from sliding of the cardia into the lower mediastinum to complete rolling of the stomach into a large hernial sac in paraesophageal hiatal hernias. In some cases the entire stomach is herniated, assuming an upsidedown position with significant elongation of the phrenoesophageal, gastrophrenic, and gastropancreatic ligaments. The posterior wall of the upper stomach may extend behind the hernial sac and remain in the mediastinum after apparent reduction of the stomach down to the abdomen and total evacuation of the sac [28]. Complete mobilization of the fundus of the stomach requires transection of the gastrosplenic omentum containing the short gastric vessels, the gastrophrenic ligament, and the portion of gastropancreatic ligament extending from the left of the left gastric artery toward the spleen. The anterior wall of the fundus is the most distensible and mobile area of the stomach, allowing its passage to the level of the caudate lobe, in front or behind the esophagus, without the need to divide any gastric ligament in the great majority of individuals, thus allowing the Rossetti modification of Nissen fundoplication [40]. Gastric wall thickness The wall of the gastroesophageal junction is thicker at the side corresponding to the major curvature of the stomach than toward the lesser curvature (average 4.5 mm vs 3.8 mm). It is also slightly thicker in its posterior than its anterior walls (average 3.5 mm vs 3.3 mm). The esophagus becomes thinner as it leaves the gastroesophageal junction in a cephalad direction [6, 7, 47]. During surgery, iatrogenic perforation of the distal esophagus and gastroesophageal junction may happen during blind and traumatic dissection. Special consideration should be given to the fact that the anterior gastric wall is the thinnest area of the stomach, particularly the proximal centimeter below the cardia. Traction on the upper part of the anterior wall of the stomach should be avoided when the stomach is mobilized. A generous grasp of the stomach and part of the lesser curvature with nontraumatic forceps is preferable. Arcuate ligament The arcuate ligament is an area of fusion of the right and left crus of the diaphragm behind the esophagus (Fig. 6). When this fusion contains a tendinous portion of the crura, the ligament is a fibrous cord located 13 cm behind the esophagus and in front of the aorta, at the level of the first lumbar vertebra. The arcuate ligament is located just cephalad to the origin of the celiac axis. When this fusion is formed only by muscle fibers, the arcuate ligament may be difficult to identify [31]. The arcuate ligament may be absent in other cases, leading to a common hiatal opening for the esophagus and aorta (Fig. 7). This is seen frequently in large paraesophageal hernias [28]. Aortic hiatus The aorta, the thoracic duct, and sometimes the azygos vein cross the diaphragm through the aortic hiatus at the level of

Stomach The stomachs anterior surface is at least partly covered by the liver and transverse colon. In most cases, the major curvature extends for 5 to 25 cm below the xiphoid. The stomach is very mobile, depending on the position of the patient and the amount and weight of the gastric contents. The gastroesophageal junction is fairly fixed in normal subjects by the retroperitoneal fixation of the esophagus and the gastropancreatic ligament. Some mobility limitations are observed in the upper part of the greater curvature due to the gastrophrenic and gastrosplenic ligaments. The proximal part of the lesser curvature and upper posterior wall are fixed due to the gastropancreatic ligament and left gastric vessels [15]. The posterior wall of the stomach is adjacent to the diaphragm, spleen, left kidney, left adrenal gland, pancreas, and transverse mesocolon. The pathologic changes of the stomach due to cancer and peptic ulcer disease are vast and beyond the scope of

889

Fig. 6. Arcuate ligament (a); (b) aorta. Fig. 7. Large hiatal opening common to the esophagus (a) and aorta (b). Fig. 8. The elements behind the esophageal hiatus: (a) aorta and aortic hiatus, (b) cisterna chyli, (c) thoracic duct, (d) esophagus, (e) inferior vena cava. Fig. 9. Arteries to the hiatus and esophagus: (a) left inferior phrenic artery originating in the aorta; (b) esophageal branch of the left gastric artery. The left inferior phrenic artery may originate in the celiac trunk. Fig. 10. Veins of the hiatal region: (a) left inferior phrenic vein; (b) left adrenal vein; (c) inferior vena cava. Fig. 11. Vagus nerves.

890

the 12th thoracic vertebra (Fig. 8). The aortic hiatus is located posterior to the esophageal hiatus. The median arcuate ligament separates these two openings, and the crura forms its lateral borders. The aorta has been mistakenly dissected for the esophagus in large hiatal hernias with absent arcuate ligament. Arteries of the hiatal region The proximal portion of the stomach and the abdominal esophagus are supplied by the esophageal branches of the left gastric artery. These branches join the esophageal arteries coming directly from the aorta above the diaphragm. Frequently, the distal esophagus also receives branches provided by the left inferior phrenic artery. The borders of the hiatus are supplied by a branch of this artery [17]. The left inferior phrenic artery may originate from the aorta or from the celiac trunk (Fig. 9). Veins of the hiatal region The left inferior phrenic vein is seen coursing parallel to the left triangular ligament, passing in front of the hiatus and draining into the inferior vena cava (Fig. 10). The left gastric vein can be identified 2 or 3 cm distal to the hiatus; it receives tributary branches of the esophagus and may anastomose with the left inferior phrenic vein. In portal hypertension, portosystemic collateral channels are formed in the lower esophagus and stomach. A network of tortuous and hypertensive veins are seen which can bleed easily and are difficult to control. Vagus nerves The vagal trunks cross the hiatus next to the esophagus (Fig. 11). They are distinctly identified as anterior and posterior branches in 8085% of the cases [2, 44]. The anterior vagus nerve is located more to the left than the posterior, but both nerves, seen through a sagittal anterior/posterior view, are located on the right half of the esophagus in the majority of instances. The posterior branch may be situated closer to the aorta than to the esophagus. In 1220% of the cases, the vagus nerves divide into multiple smaller branches well above the hiatus [14]. The anterior vagus nerve divides into two branches: anterior gastric and hepatic. The anterior gastric branch provides the anterior Latarjet nerve, which runs 0.5 to 1 cm to the right of the lesser curvature. The hepatic branch of the anterior vagus nerve usually separates from the anterior gastric branch at the level of the abdominal esophagus, advancing toward the liver through the lower border of the avascular window of the gastrohepatic omentum. Frequently, it is formed by multiple parallel bundles. The posterior vagus nerve also divides into two branches: the posterior gastric and the celiac. The posterior gastric branch provides the posterior Latarjet nerve, with branches to a more cephalad portion of the stomach than the anterior Latarjet nerve. The celiac branch is always a single, thick nerve extending through the gastropancreatic ligament to join the celiac plexus.

Unless some type of vagotomy is intended as part of the surgical procedure, the vagus nerves should be protected against injury during dissection of the hiatus, the esophagus, and surrounding tissues. The use of blunt dissection with a Peanut sponge (as described above) will rarely result in damage to the main vagus trunks while dissecting the borders of the crura away from the hiatal contents or when dissecting the esophagus in the lower mediastinum. The anterior vagus nerve, and particularly the hepatic and Latarjet branches, may be mistakenly divided when performing esophageal myotomy (Heller procedure) for achalasia [49]. It is preferable to perform the myotomy to the left of the anterior vagus nerve [36], since the branches extending to the fundus of the stomach can be divided without clinical consequences. The posterior vagus nerve should be retracted with the esophagus (Figs. 3 and 7). Its celiac branch sometimes limits the caudal mobilization of the posterior wall of the cardia. The hepatic branch is frequently sacrificed when opening the gastrohepatic omentum. Celiac plexus The celiac ganglia and plexus are located below the origin of the celiac artery. The exact location is variable. They are situated between T12L1 intervertebral disk and the middle of L2. The celiac plexus is formed by one to five ganglia with a diameter of 0.5 to 4.5 cm [53] and reaches the area of the insertion of the diaphragmatic crura. Damage to the celiac plexus during surgery in this area (particularly during dissection of the crura of the diaphragm in front of the aorta and during approximation of the crura in hiatal hernia repairs) does not appear to produce clinically significant effects. Injection of neurolytic solutions into this plexus has been used for pain control in carcinoma of the pancreas. The laparoscopic injection is performed by opening the window of the gastrohepatic ligament and retracting the lesser curvature of the stomach to the left. The neurolytic solution is applied to the front and to the right and left lateral periaortic tissues between the crura and the upper border of the pancreas. Lymphatics The cisterna chyli, when it is present, lies in front of the body of the first and second lumbar vertebra between the right crus of the diaphragm and the aorta (Fig. 8). From there, the thoracic duct advances in a cephalad direction. Perforation of this duct may lead to chylous ascites. On the other hand, its ligation does not result in a significant problem. Gastrosplenic ligament The membrane that joins the spleen with the greater curvature of the stomach is called the gastrosplenic ligament or omentum. There is a space between the anterior and posterior leaflets of the cephalad portion of this ligament, leaving a small segment of the posterior wall of the stomach uncovered by peritoneum. This area is situated immediately in front of the left crus of the diaphragm and may contain

891

vessels between the diaphragm, the pancreas, and the posterior wall of the stomach. To the left in this area is the left adrenal gland. The gastrosplenic omentum contains the short gastric vessels in its upper portion and the left gastroepiploic vessels as well as some branches from the splenic artery in its most caudal area. There are lymphatic vessels and nodes all along the gastrosplenic omentum. Gastrophrenic ligament Cephalad to the gastrosplenic omentum is the gastrophrenic ligament extending to the esophagus. This contains none or very few vessels toward the esophagus, allowing a blunt dissection to approach the left and posterior wall of the distal esophagus. The gastrophrenic ligament, however, contains many gastric and phrenic vessels extending for 2 to 3 cm to the left of the angle of His until joining the short gastric vessels to the spleen. This ligament joins the gastropancreatic ligament and upper border of the pancreas. Gastropancreatic ligament Seldom described under this name since it is formed by a wide area containing the left gastric vessels, the gastropancreatic ligament separates the portion of the lesser sac located behind the gastrocolic omentum from the portion behind the posterior wall of the stomach. The only communication between these two areas is an orifice caudal to the gastropancreatic ligament known as the omental foramen. The cephalad portion of the gastropancreatic ligament is lost in the retroperitoneum below the esophageal hiatus and behind the cardia. The right wall is covered by the peritoneum extending from the proximal part of the lesser curvature to the right of the aorta and head of the pancreas. The left boundaries continue with the posterior leaflet of the gastrophrenic ligament. The gastropancreatic omentum contains the left gastric vessels; the posterior gastric artery, which is a branch of the proximal segment of the splenic artery; lymphatic vessels and nodes; the celiac division of the posterior vagus nerve; and small retroperitoneal, diaphragmatic, and pancreatic vessels. The gastropancreatic and gastrophrenic ligaments may be very elongated and may extend into the mediastinum in large hiatal hernias. The presence of thick, fatty, and vascular tissue pulled into the posterior mediastinum together with the cardial end of the stomach may confuse the surgeon when trying to identify and dissect the posterior wall of the esophagus. The key to avoiding difficulties is to first dissect the crura and then retract all tissue interposed between the junction of the right and left crus and the esophagus into an intraabdominal position. Diaphragm The central tendon of the diaphragm ends posteriorly at the level of the anterior portion of the crura. The tendinous portion gives way to muscle fibers on the sides of the crura. There are several diaphragmatic vessels in the vicinity of the esophageal hiatus. The left inferior phrenic artery comes

from the celiac trunk or directly from the aorta (Fig. 9). Rarely it arises from the left gastric, left renal, or accessory left hepatic artery. It crosses between the left crus and the esophagus and distributes its branches to the left hemidiaphragm and the left adrenal gland. The left inferior phrenic artery may give branches to the esophagus and anastomose with branches of the left gastric artery and thoracic esophageal arteries. Attention should be given to these vessels when dissecting the esophagus away from the hiatus. Some of them may require clipping before transection. The left inferior phrenic vein may drain into the inferior vena cava or the left adrenal vein (Fig. 10). In cases in which part or all of the drainage is into the vena cava, this vein passes transversely in front of the hiatus, usually below the triangular ligament. The left inferior phrenic vein may also anastomose with esophageal branches of the left gastric vein and with veins of the greater curvature of the stomach through the gastrophrenic and gastropancreatic ligaments. Congenital diaphragmatic defects should be recognized during laparoscopic procedures since they can be easily repaired [30] and may lead to complications like pneumothorax if left uncorrected [8]. Transverse colon In one particular congenital syndrome, the colon may be located between the liver and the diaphragm. This is known as Chilaiditis syndrome, and may be found associated with hypoplasia or absence of the left hepatic lobe as described above. The colon is frequently found in the sac of large paraesophageal hernias and in Morgagni hernias [27, 28, 30]. Spleen Although not in direct relation with the hiatus, except in cases of splenomegaly, the spleen is an important organ to consider in any laparoscopic dissection of the hiatus or its surroundings because of the frequent risk of injury, tears of its capsule, or bleeding from the many fragile vessels contained in its ligaments. The surgeon and assistants must constantly be aware of the position of the spleen and the dangers of exerting traction in the gastrosplenic, splenocolic, and splenophrenic ligaments [45]. These ligaments vary in size and site of insertion in the spleen. The upper pole of the spleen is seen after dissecting the gastrophrenic ligament or during dissection of the left crus (Fig. 3). Trauma to the upper pole of the spleen should be avoided when passing instruments behind the esophagus from right to left. To avoid this, the instrument is advanced in front and just to the left of the left crus, the esophagus is then retracted to the right, and the instrument is further advanced under direct visualization. The upper pole of the spleen may be in direct contact with the lateral end of the left lobe of the liver and the left triangular ligament which sometimes surround it [4]. Pancreas The pancreas is not in direct contact with the esophageal hiatus. Its upper border is caudal to the celiac trunk and the

892

caudal insertions of the crura are well behind the body of the pancreas. However, the pancreas may be traumatized during dissection in the hiatal area, producing postoperative elevation of serum amylase. The pancreas does not slide up into the mediastinum in hiatal hernias. Conclusion Familiarity with the normal and pathologic laparoscopic anatomy is essential for safe surgical interventions in the area of the esophageal hiatus. Training in the animal laboratory is technically valuable but may lead to excessive confidence due to the anatomic simplicity of the animal model. Some laparoscopic complications, like pneumothorax and esophageal and gastric perforations, rarely seen during open procedures, are usually the result of the difficulties encountered in finding proper anatomic landmarks or the consequences of struggles to provide adequate exposure for dissection. This review of the laparoscopic anatomy of the area of the esophageal hiatus, based on published material and on personal experience, emphasizes the anatomic and pathologic variations that the surgeon may encounter, the laparoscopic approach for dissection, and the technical maneuvers that may help the surgeon avoid complications.
Acknowledgments. This work was supported in part by the Mr. and Mrs. A. Foroni and Family funds and by the Mr. and Mrs. M. C. Weiner funds.

References
1. Adams DB (1993) The importance of extrahepatic biliary anatomy in preventing complications at laparoscopic cholecystectomy. Surg Clin North Am 73: 861871 2. Aharinejad S, Firbas W (1989) The innervation of the human esophagus. Acta Anat 136:715720 3. Ahmed AF, Bediako AK, Rai D (1988) Agenesis of the left hepatic lobe with gastric volvulus. NY State J Med 88: 327328 4. Arenson AM, McKee JD (1986) Left upper quadrant pseudolesion secondary to normal variants in liver and spleen. J Clin Ultrasound 14: 558561 5. Ashour M, Butchart EG, Breckenridge IM (1984) Phreno-oesophageal muscle anomaly or rudimentary oesophageal sphincter? J R Coll Surg Edinb 29: 387388 6. Bolondi L, Casanova P, Santi V, Caletti G, Barbara L, Labo G (1986) The sonographic appearance of the normal gastric wall: an in vitro study. Ultrasound Med Biol 12: 991998 7. Caletti G, Bolondi L, Labo G (1984) Anatomical aspects in ultrasonic endoscopy for the stomach. Scand J Gastroenterol Suppl 94: 3442 8. Childers JM, Caplinger P (1995) Spontaneous pneumothorax during operative laparoscopy secondary to congenital diaphragmatic defects. A case report. J Reprod Med 40: 151153 9. Collet D, Cadiere GB (1995) Conversions and complications of laparoscopic treatment of gastroesophageal reflux disease. Formation for the development of laparoscopic surgery for gastroesophageal reflux disease group. Am J Surg 169: 622626 10. Coster DD, Bower WH, Wilson VT, Butler DA, Locker SC, Brebrick RT (1995) Laparoscopic Nissen fundoplicationA curative, safe, and cost-effective procedure for complicated gastroesophageal reflux disease. Surg Laparosc Endosc 5: 111117 11. Crist DW, Gadacz TR (1993) Laparoscopic anatomy of the biliary tree. Surg Clin North Am 73: 785798 12. Cuschieri A, Hunter J, Wolfe B, Swanstrom LL, Hutson W (1993) Multicenter prospective evaluation of laparoscopic antireflux surgery: preliminary report. Surg Endosc 7: 505510

13. Dallemagne B, Weerts JM, Jehaes C, Markiewicz S (1996) Causes of failure of laparoscopic antireflux operations. Surg Endosc 10: 305310 14. Dia A, Ouedraogo T, Zida M, Sow ML (1994) Anatomic pitfalls of vagotomies. Ann Chir 48: 452461 15. Erenbourg L, Reggiani P (1985) A new ligament: the gastropancreatic. Anat Clin 7: 143144 16. Fontaumard E, Espalieu P, Boulez J (1995) Laparoscopic NissenRossetti fundoplication. Surg Endosc 9: 869873 17. Gray SW, Rowe JS, Skandalakis JE (1979) Surgical anatomy of the gastroesophageal junction. Am Surg 45: 575587 18. Hay JM, Magdeleinat P, Regnard JF (1993) Positioning of the patient for oesophageal hiatus surgery. Ann Chir 47: 8083 19. Heloury Y, Leborgne J, Rogez JM, Robert R, Barbin JY, Hureau J (1988) The caudate lobe of the liver. Surg Radiol Anat 10: 8391 20. Hinder RA, Filipi CJ, Wetscher G, Neary P, DeMeester TR, Perdikis G (1994) Laparoscopic Nissen fundoplication is an effective treatment for gastroesophageal reflux disease. Ann Surg 220: 472483 21. Hunter JG, Trus TL, Branum GD, Waring JP, Wood WC (1996) A physiologic approach to laparoscopic fundoplication for gastroesophageal reflux disease. Ann Surg 223: 673687 22. Jackson ML, Zuckier L, Goldfarb CR, Ongseng F (1986) Effect of patient positioning on liver size estimation. J Nucl Med 27: 16321634 23. Jacques PF, Warshauer DM, Keefe B, Mauro MA, McCall JM (1974) Variations in liver-colon anatomic relationship: relevance to interventional radiology. J Vasc Interv Radiol 5: 637641 24. Joris JL, Chiche JD, Lamy ML (1995) Pneumothorax during laparoscopic fundoplication: diagnosis and treatment with positive endexpiratory pressure. Anesth Analg 81: 9931000 25. Kuster G, Biel F (1967) Accuracy of laparoscopic diagnosis. Am J Med 42: 388393 26. Kuster GG, Fischer B (1993) Pharmacologic hemostasis in laparoscopy: topical epinephrine facilitates cholecystectomy. Am Surg 59: 281 284 27. Kuster GGR, Gilroy S (1993) Laparoscopic repair of paraesophageal hiatal hernia. Surg Endosc 7: 362363 28. Kuster GGR, Gilroy S (1993) Laparoscopic technique for repair of paraesophageal hiatal hernia. J Laparoendosc Surg 3: 331338 29. Kuster GGR, Gilroy SBC (1995) Intraoperative trans-gallbladder cholangiography intended to delineate bile duct anatomy. J Laparoendosc Surg 5: 377384 30. Kuster GGR, Kline LE, Garzo G (1992) Diaphragmatic hernia through the foramen of Morgagni: laparoscopic repair case report. J Laparoendosc Surg 2: 93100 31. Lindner HH, Kemprud E (1971) Clinicoanatomical study of the arcuate ligament of the diaphragm. Arch Surg 103: 600605 32. Listerud MB, Harkins HN (1958) Anatomy of the esophageal hiatus. Anatomic studies on two hundred four fresh cadavers. Arch Surg 76:835 33. Mattioli S, Pilotti V, Felice V, Di Simone MP, DOvidio F, Gozetti G (1993) Intraoperative study on the relationship between the lower esophageal sphincter pressure and the muscular components of the gastroesophageal junction in achalasic patients. Ann Surg 318: 635 639 34. McKernan JB, Laws HL (1994) Laparoscopic Nissen fundoplication for the treatment of gastroesophageal reflux disease. Am Surg 60: 8793 35. Min PQ, Yang ZG, Lei QF, Gao XH, Long WS, Jiang SM, Zhou DM (1993) Peritoneal reflectons of left perihepatic region: radiologicanatomic study. Radiology 182: 553557 36. Oddsdo ttir M (1996) Laparoscopic management of achalasia. Surg Clin North Am 76: 451458 37. Patti MG, Arcerito M, Pellegrini CA, Mulvihill SJ, Tong J, Way LW (1995) Minimally invasive surgery for gastroesophageal reflux disease. Am J Surg 170: 614718 38. Peters JH, Heimbucher J, Kauer WKH, Incarbone R, Bremner CG, DeMeester TR (1995) Clinical and physiologic comparison of laparoscopic and open Nissen fundoplication. J Am Coll Surg 180: 385393 39. Richards KF, Fisher KS, Flores JH, Christensen BJ (1996) Laparoscopic Nissen fundoplication: cost, morbidity and outcome compared with open surgery. Surg Laparosc Endosc 6: 140143 40. Rossetti M, Hell K (1977) Fundoplication for treatment of gastroesophageal reflux in hiatal hernia. World J Surg 1: 439 41. Schauer PR, Meyers WC, Eubanks S, Norem RF, Franklin M, Pappas TN (1996) Mechanisms of gastric and esophageal perforations during laparoscopic Nissen fundoplication. Ann Surg 223: 4352

893 42. Scott-Conner CEH, Hall TJ (1995) Anatomic pitfalls of laparoscopic surgery in the right upper quadrant. Contemp Surg 47: 307314 43. Siewert JR, Jennewein HM, Waldeck F (1973) Experimental studies on the function of the lower esophageal sphincter following intrathoracic displacement, myotomy and circular myectomy. Bruns Beitr Klin Chir 22: 818828 44. Skandalakis LJ, Donahue PE, Skandalakis JE (1993) The vagus nerve and its vagaries. Surg Clin North Am 73:769784 45. Skandalakis PN, Colborn GL, Skandalakis LJ, Richardson DD, Mitchell WE Jr (1993) The surgical anatomy of the spleen. Surg Clin North Am 73:747768 46. Stein HJ, Feussner H, Siewert JR (1996) Failure of antireflux surgery: causes and management strategies. Am J Surg 171: 3640 47. Stein HJ, Liebermann-Meffert D, De Meester TR, Siewert JR (1995) Three-dimensional pressure image and muscular structure of the human lower esophageal sphincter. Surgery 117: 692698 48. Swanstrom L, Wayne R (1994) Spectrum of gastrointestinal symptoms after laparoscopic fundoplication. Am J Surg 167: 538541 49. Swanstrom LL, Pennings J (1995) Laparoscopic esophagomyotomy for achalasia. Surg Endosc 9: 286292 50. Swanstrom LL, Pennings JL (1995) Safe laparoscopic dissection of the gastroesophageal junction. Am J Surg 169: 507511 51. Trus TL, Laycock WS, Branum G, Waring JP, Mauren S, Hunter JG (1996) Intermediate follow-up of laparoscopic antireflux surgery. Am J Surg 171: 3235 52. Van Minh T, Galizia G, Lieto E (1992) Anatomy of the caudate lobe of the liver. New aspects and surgical applications. Ann Chir 46: 309318 53. Ward EM, Rorie DK, Nauss LA, Bahn RC (1979) The celiac ganglia in man: normal anatomic variations. Anesth Analg 58: 461465 54. Watson DI, Jamieson GG, Devitt PG, Matthew G, Britten-Jones RE, Game PA, Williams RS (1995) Changing strategies in the performance of laparoscopic Nissen fundoplication as a result of experience with 230 operations. Surg Endosc 9: 961966

Surg Endosc (1997) 11: 969

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

The author replies

We have read with care the letter to the editor of Professor Bax concerning the intraperitoneal use of cetrimide and the unfortunate death of this 2-year-old patient. We do feel, with the author, that this complication may be a cetrimiderelated toxicity; however, we would like to clarify several issues related to the clinical use of cetrimide: The diagnosis in this 2-year-old boy was not documented by a CT scan, which gives the typical appearance of a cyst with multiple daughter cysts, which is pathognomonic of hydatid disease [2]. Also, there was no mention of the indirect hemagglutination test or any other hydatid studies, which usually are positive. Hydatid cysts at this age are rare, and other possibilities, including hepatoblastoma may be considered. We have reported in our experience [5, 6] that only cysts surfacing to the peritoneal cavity should be treated laparoscopically. The cyst is punctured in the fibrous capsule rather than in the liver tissue, cetrimide or hydatid fluid in direct contact with bloodstream [3], or highly vascular areas like liver or epidural space [7] could result in anaphylactic shock. Especially in this 2-year old boy, whose estimated blood volume is 8001000 ml. The cetrimide was left in the peritoneal cavity for 1 h; 200 ml of this fluid was not retrieved and could have been absorbed. Concerning the liver puncture, it was not clear whether cetrimide was injected inadvertently into the liver. The findings suggest an obvious case of cetrimide overdose, although not documented by methemoglobinemia [1]. In our described technique the cetrimide instillment does not exceed 1,000 ml in adults; the hydatid fluid aspirated from the cyst is also replaced by equal amounts of cetrimide. In a vast experience with more than 1,000 cases in our institution [4] of using cetrimide since 1972, no cetrimide overdose has been reported and it is a safe practice to use 1% cetrimide, which is not toxic in this concentration [8]. We conclude that this is an obvious case of cetrimide overdose in a 2-year-old child, but we still consider the use of cetrimide 1% a safe practice provided it is not injected in highly vascular areas or in the bloodstream and is used in moderate amounts for a short period of time in the peritoneal cavity. Needless to say, it should be used in hydatid asepsis after documentation of hydatid disease. References
1. Baraka A, Yamut F, Walid N (1980) Cetrimide induced methaemoglobinaemia after surgical excision of hydatid cyst. Lancet 11: 8889 2. Beggs I (1985) The radiology of hydatid disease. AJR Am J Roentgenol 145: 63948 3. Gode GR, Jayalaksmi TS, Kalla GN (1975) Accidental intravenous injection of cetrimide. Anesthesia 30: 508510 4. Jidejian YD (1979) Hydatid disease. Dar El-Mashreq, Beirut, Lebanon 5. Khoury G, Geagea T, Hajj A, Jabbour-Khoury S, Baraka A, Nabbout G (1994) Laparoscopic treatment of hydatid disease of the liver. Surg Endosc 8: 11031104 6. Khoury G, Jabbour-Khoury S, Bikhazi K (1996) Results of laparoscopic treatment of hydatid cyst of the liver. Surg Endosc 10: 5759 7. Klouche GR, Charlotte N, Kaaki M, Beraud JJ (1994) Coma hemolysis after cetrimide washout of epidural hydatid cyst. Intensive Care Med 20: 613 8. Martindale (1996) The extra pharmacopocia. 22nd ed. The Pharmaceutical Press, London, p 1122

G. Khoury S. Jabbour-Khoury K. Bikhazi

Department of Surgery American University of Beirut Medical Center P.O. Box 113-6044 Beirut, Lebanon

Correspondence to: G. Khoury

Letters to the editor

Surg Endosc (1997) 11: 965

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Laparoscopic bariatric surgery

We read with interest the article Laparoscopic gastric bypass: another option in bariatric surgery by Lo nroth et al. [Surg Endosc (1996) 10: 636638]. The gastric bypass technique which they describe (loop gastrojejunostomy, with afferent-efferent jejunojejunostomy for diversion of bile and pancreatic juice) is acceptable. However, this method has been largely replaced over the past 15 years by the Roux-en-Y gastric bypass. The Roux-limb facilitates technically the movement of the jejunum to a high position. However, with the advent of, enthusiasm for, and progressive application of video-assisted laparoscopic surgery, caution must be advised with respect to bariatric surgery. The authors mention that laparoscopic surgery should of course be based on experiences already reached in the area of open surgery, but this statement demands further emphasis. In 1993, Robert J. Greenstein of Mount Sinai Medical School and Joseph F. Amaral of Brown University spearheaded a move to formulate basic guidelines regarding the application of laparoscopic techniques to obesity surgery in order to avoid disastrous consequences. This resulted in the recommendation by the American Society for Bariatric Surgery (ASBS) that Laparoscopic obesity operations should be undertaken only by surgeons who are experienced both in video-laparoscopic technique and in the complexities of open bariatric operations and the field of morbid obesity. This recommendation has been circulated by the ASBS to the chiefs of surgery in the USA and Canada. Bariatric surgery has become a specialized field in general surgery. Expertise is necessary in dealing with the massively obese patient population, in their selection and comorbidities, in the specific measurements required to avoid failure, and in the postoperative treatment of sequelae and nutrition. In the event that a laparoscopic bariatric operation has to be converted, experience in the open abdominal surgery is required. We see a parallel here to the action of SAGES, when laparoscopic cholecystectomy catapulted into the surgical realm 6 years ago. SAGES then issued a statement with the logical recommendation that laparoscopic cholecystectomy be performed only by surgeons experienced in open cholecystectomy. This recommendation helped to avoid major complications which would have devalued early videolaparoscopic surgery; it also deterred a number of medical gastroenterologists experienced in laparoscopy who were anxious to undertake laparoscopic cholecystectomies. SAGES was instrumental in setting criteria for credentialling laparoscopic surgery in hospitals and provided many comprehensive courses by experts. In like fashion, sessions on laparoscopic bariatric surgery have been an integral part of the ASBS meetings for the past 4 years. Lo nroth, Dalenba ck, Haglind, and Lundell from Gothenburg are experienced bariatric abdominal surgeons and laparoscopists [7]. However, there are learning curves with laparoscopic gastric banding [1, 5] vertical banded gastroplasty [2, 3], Roux-en-Y gastric bypass [9], and ileogastrostomy [4]. These procedures have been done by devoted individuals and groups experienced in both the open and laparoscopic approaches, thus avoiding catastrophes and setbacks for bariatric surgery. The laparoscopic procedures, where exposure and the ability to measure may be less optimal in the massively obese, must be shown to be at least as effective and safe as the open procedures. A number of groups are using the laparoscopic adjustable gastric band in a controlled study, but further evaluations are necessary [6, 8]. References
1. Broadbent R (1993) Lessons learned from laparoscopic gastric banding. Obes Surg 3: 415420 2. Catona A, Gossenberg M, Mussini G, La Manna L, De Bastiani T, Armeni E (1995) Videolaparoscopic vertical banded gastroplasty. Obes Surg 5: 323 3. Chua TY, Mendiola RM (1995) Laparoscopic vertical banded gastroplasty: the Milwaukee experience. Obes Surg 5: 7780 4. Cleator IGM, Litwin DEM, Phang PT, Brosseuk DT, Rae AJ (1994) Laparoscopic ileogastrostomy. Obes Surg 4: 4043 5. Favretti F, Cadiere GB, Segato G, Bruyns G, De Marchi F, Himpens J, Foleto M, Lise M (1995) Laparoscopic adjustable silicone gastric banding: technique and results. Obes Surg 5: 364371 6. Greenstein RJ, Rabner JG, Green S, Hodge-Penn G, Kaiser S, Halpen NA (1996) The laparoscopic adjustable gastric band (Lap-band): observations from the preliminary Mount Sinai experience of a nationwide multi-center, FDA-moderated study. Obes Surg 6: 124 (abstract) 7. Lo nroth H, Dalenba ch J, Haglind E, Lundell L (1995) Laparoscopic vertical banded gastroplasty: 1-year follow-up. Obes Surg 5: 249 (abstract) 8. OBrien P, McMurrick P (1995) Posterior gastric wall prolapse after Lap-band placement. Obes Surg 5: 247 (abstract) 9. Wittgrove AC, Clark GW (1996) Laparoscopic gastric bypass, Rouxen-Y: experience of 27 cases, with 318 months follow-up. Obes Surg 6: 5457, 329

M. Deitel
Department of Nutritional Sciences and Surgery University of Toronto 2238 Dundas Street West, Suite 201 Toronto, Ontario M6R 3A9, Canada

Surg Endosc (1997) 11: 970

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Erroneous CO2 pressure readings in a morbidly obese patient

Since the initial reports in 1992 [2] laparoscopic splenectomy has become an accepted technique for patients with idiopathic thrombocytopenic purpura (ITP) [4]. Simultaneously, increasing experience with laparoscopic techniques has broadened patient selection criteria to include obses patients [6], who had previously been excluded. Experience has also dictated alterations in patient positions and trocar sites to facilitate exposure and manipulation of the spleen, resulting in decreasing operative times [9]. Refinements in technology, such as the harmonic scalpel, have also facilitated dissection and hemostasis [7]. With each refinement, however, new potential problems can arise and must be identified. In the case reported here the combination of lateral position and morbid obesity presented the potential for unrecognized overinsufflation of CO2. A 37-year-old man with ITP inadequately responsive to medical therapy was referred for splenectomy. The patient was morbidly obese, weighing 318 lb on a 510 frame. Due to anticipated exposure problems with either open or supine laparoscopic approaches, the lateral or hang spleen laparoscopic approach was selected [1]. The patient size limited the usual table rotation position [3] and he was in almost a true lateral position. The Hassan open technique was utilized for placement of a supraumbilical trocar with the blunt finger technique of Filipi [5] used to access the peritoneal cavity due to the difficult exposure. Following trocar placement and confirmation of position by endoscopic visualization of omental tissue the laparoscope was withdrawn and CO2 insufflation was initiated. Insufflation was at 6 l per min via a Stryker high-flow insufflator with a maximum pressure setting at 15 cm of water. It was noted that although several liters had been instilled, the abdominal pressure reading remained zero. Although our initial thought was a loose tubing connection, the scope was immediately reinserted to check for position. This afforded visualization of fatty tissue flapping over the end of the trocar like a flutter valve mechanism. When the scope was advanced beyond this fatty tissue a large pneumoperitoneum was visualized and the pressure reading jumped to 18 cm of water. This was immediately released to the proper pressure reading of 15. The remainder of the splenectomy was completed uneventfully using a fourtrocar technique, the 30 laparoscope, and the harmonic scalpel. Blood loss was negligible and a 72-g spleen was morsalated and removed via an 18-mm trocar site (the weight is artificially decreased due to blood aspiration during the morsalation process). The total procedure time was 118 min. The patient did well postoperatively and was discharged on the 2nd postoperative day. While for the experienced laparoscopic surgeon operations in morbidly obese patients can be simplified relative to the open technique due to better visualization and depth of access, other problems can arise. In this case, I believe the patients morbid obesity combined with the lateral position and midline trocar site resulted in intermittent occlusion of the trocar tip to the pressure measurement while allowing continued insufflation. If this situation is not recognized, CO2 overinsufflation with dire consequences can result [8]. The patients obesity further obscured the problem by limiting visual external inspection of the abdomen. Our scope is often removed during initial insufflation to allow better flow and to facilitate setup of other equipment. In a morbidly obese patient continuous visualization during insufflation appears safer. Special precautions should be taken during insufflation of CO2 via a midline trocar with a morbidly obese patient in the lateral position. Low pressure readings should not be assumed to be accurate as a flutter valve effect of tissue over the trocar can occur. Visual endoscopic inspection during insufflation is indicated and pressure tubing should be changed to a more lateral trocar when this site is available. References
1. Delaitre B (1995) Laparoscopic splenectomy: the hanged spleen technique. Surg Endosc 9: 528529 2. Delaitre B, Maignien B (1992) Laparoscopic splenectomytechnical aspects. Surg Endosc 6: 305308 3. Dexter SPL, Martin IG, Alao D, Norfolk DR, McMahon MJ (1996) Laparoscopic splenectomy: the suspended pedicle technique. Surg Endosc 10: 393396 4. Grossbard ML (1996) Is laparoscopic splenectomy appropriate for the management of hematologic and oncologic diseases? Surg Endosc 10: 387388 5. Laparoscopic Surgery Update. May 1996: Vol. 4, No. 5: 4960 6. Robles AE, Owens CC, Bianchi C, Pitombo C, Garberoglio C, Tabuenca A (1996) Outpatient laparoscopic cholecystectomy is safe for the obsese and morbidly obese patient. SAGES Meeting; Poster Abstract; 205 7. Rothenberg SS (1996) Laparoscopic splenectomy using the harmonic scalpel. J Laparoendosc Surg 6: S61S62 8. Smith I, Benzie RJ, Gordon NLM, Kelman GR, Swapp GH (1971) Cardiovascular effects of peritoneal insufflation of carbon dioxide for laparoscopy. Br Med J 3: 410411 9. Trias M, Targarona EM, Balague C (1996) Laparoscopic splenectomy: an evolving technique. Surg Endosc 10: 389392

K. D. Stalter
449 Main Street Oneonta, NY 13820 USA

Original articles
Surg Endosc (1997) 11: 894898

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Short-term outcome of laparoscopic paraesophageal hernia repair

A case series of 58 consecutive patients
T. R. Huntington
Department of Surgery, St. Lukes Regional Medical Center Department of Surgery, Boise Veterans Affairs Medical Center, 222 N. 2nd Street, Suite 107, Boise, ID 83702, USA Received: 24 July 1996/Accepted: 20 November 1996

Abstract Background: The purpose of this study is to determine the morbidity, mortality, and short-term outcomes associated with laparoscopic paraesophageal hernia repair (LPHR). Methods: A series of 58 consecutive LPHRs performed by the author were reviewed with an average 1-year follow-up. Morbidity and mortality rates were compared with historical series of open repairs. Anatomy and technical considerations pertinent to LPHR were reviewed. Results: There were no procedure-related or perioperative deaths in this series of patients undergoing LPHR. Four major complications occurred (7%), two of which required reoperation, all in urgently repaired patients. One patient required conversion to laparotomy (1.7%). Based on symptoms, there were no reherniations. No patients had longterm dysphagia worse than preoperatively. Preoperative symptoms of chest pain, esophageal obstruction, hemorrhage, and reflux were resolved in all patients. Conclusions: LPHR is safe, effective, and compares favorably to historical series of open paraesophageal hernia repair. Key words: Paraesophageal hernia Laparoscopy Laparoscopic Hiatal hernia Gastroesophageal reflux Fundoplication

short-term outcome. This is the largest series of laparoscopic repairs reported to date.

Materials and methods

Patient population
Between August 1992 and October 1995, 218 of the authors patients from St. Lukes Regional Medical Center and the Boise Veterans Affairs Medical Center had laparoscopic repair of hiatal hernias. Twenty-six of these were operated on for symptoms or radiographic findings of a paraesophageal hernia (chest pain, dysphagia, esophageal obstruction, hemorrhage, and aspiration). Of the remaining 187 who had a chief complaint of gastroesophageal reflux, 32 of these were identified as having a significant paraesophageal hernia (PH) either preoperatively or at the time of surgery. The referral base was biased in favor of paraesophageal hernias and large hiatal hernias due to somewhat selective referral of this type of hernia. A PH was defined as a peritoneal lined defect through the esophageal hiatus typically containing at least one-quarter to one-third of the stomach within the hernia sac. The most superior portion of the fundus was located superior to the gastroesophageal junction. All patients had preoperative esophagogastroduodenoscopy (EGD) performed by various physicians. Only 38% were specifically noted as having a PH on EGD. The remainder were described as having a large or huge hiatal hernia, sometimes associated with a description of torsion, obstruction, or complete intrathoracic stomach. One patient, not included in the 58 cases reported here, had an incorrect preoperative diagnosis of PH based on symptoms and EGD; he underwent fundoplication for reflux symptoms. Of those patients repaired primarily for symptoms of PH, 82% had preoperative upper gastrointestinal series (UGI) performed at various institutions. Seventy percent were interpreted by their radiologist as having a PH. The remainder were described as having hiatal hernias that were huge, large, completely or largely intrathoracic, or in some as being partially torsed. One patient with a PH demonstrated on EGD and confirmed at surgery had an UGI interpreted as normal. Esophageal manometry was done in elective cases when technically feasible (43 cases). Four of these had dysmotility demonstrated. All 58 patients underwent laparoscopic repair, although one required conversion to laparotomy. During the 3 years covered by this study, no patients seen by the author for gastroesophageal reflux or paraesophageal hernia were denied laparoscopic repair in favor of repair by laparotomy. Ten patients in this series underwent urgent repair.

Numerous case reports of laparoscopic paraesophageal hernia repairs (LPHRs) appearing in the literature [13, 79, 11, 14] have confirmed that the procedure is technically feasible and safe. This report describes the authors series of 58 consecutive LPHRs. The overall morbidity and mortality in this series were compared with large historical series of paraesophageal hernia repair (PHR) by laparotomy [4, 5, 10, 18] or thoracotomy [13]. The large case series reported here provides data regarding technique, morbidity, mortality, and

895

Fig. 1. Transverse view of esophageal hiatus showing the diaphragmatic crura, herniated gastroesophageal junction, and the herniated lesser omental sac posterior to the gastrohepatic ligament and stomach. C, caudate lobe of liver; GH, gastrohepatic ligament; E, esophagus; F, fundus; Sp, spleen.

Fig. 2. Sagittal view through the mediastinum showing the paraesophageal hernia with lesser sac hernia posterior to fundus (F ). The initial dissection of the hernia is through the plane posterior and external to the lesser (posterior) hernia sac. St, stomach; E, esophagus.

Five of the 58 patients had had previous failed fundoplication or paraesophageal repair, four by laparotomy and one by laparoscopy. The average follow-up was 12 months (383 days) and ranged from 6 weeks to 36 months. An attempt was made to telephone or examine all survivors between January and March 1996, during which time 45 patients were successfully interviewed or examined by the authors. Data on the remaining 13 patients, including five deceased, were taken from the most currently available medical records. Patients were evaluated for symptoms of recurrent herniation, gastroesophageal reflux, dysphagia, hemorrhage, and complications related to surgery.

Technique
The surgical approach utilized a standard five-port placement consisting of one 5-mm subxyphoid port and four 11-mm ports placed in a near straight line crossing the upper abdomen approximately 10 cm below the xyphoid. The patient was supine with the surgeon standing on the right utilizing the two midline ports. The assistant, or in later cases, the robotic arm (AESOP, Computer Motion, Goleta, CA), was on the patients left. Gentle traction was applied to the hernia contents, reducing it as much as possible. The most superior aspect of the fundus, which was retroperitoneal, was frequently not reducible. The peritoneum was divided at the hiatus anterior to the esophagus and stomach except in reducible hernias, in which case the sac was divided adjacent to the cardioesophageal junction. Short gastric vessels were found stretched and frequently elongated over the left side of the hiatus and did not require division. The gastrohepatic ligament was divided over the caudate lobe of the liver exposing the lesser sac. It is important to recognize that in a large PH the lesser sac extends into the mediastinum. The mediastinal extension of the lesser sac, or lesser sac hernia, is bounded anteriorly by the gastrohepatic ligament, the herniated stomach, and the herniated gastrosplenic ligament. In extreme forms these structures form a transverse plane dividing the hiatus hernia into two compartments, the greater and lesser sac. Figure 1 illustrates a cross section of the esophageal hiatus demonstrating the two components of the hernia. The lesser sac hernia into the mediastinum was the key to easy reduction of the paraesophageal hernia. After entering the lesser sac through the gastrohepatic ligament, the most posterior layer of lesser sac peritoneum was identified and entered where it was draped over the left and right crura to the right of and posterior to the esophagus. Further dissection was undertaken posterior and external to the lesser sac (Fig. 2). This allows reduction of the lesser sac and coincidentally mobilizes the posterior esophagus and herniated stomach. The greater sac hernia was approached through the same plane, ergo, from posteriorly and to the right, again keeping external to the sac. This

plane of dissection was utilized to remove the entire sac and the sliding portion of the fundus. The entire sac and stomach were thus reduced. The esophagus was mobilized as far as necessary for adequate length, sometimes past the point at which the esophagus passes posterior to the posterior aspect of the aortic arch. Of note, all patients had adequate esophageal length, although some required extensive mobilization of the esophagus. The hernia sac was excised, or at a minimum was incised down to the cardioesophageal junction. Crural repair was performed by placing sutures between the right and left crura posterior to the esophagus and, if needed, anterior to the esophagus. In the event of excessive tension on the crura, a relaxing incision was made in the right crus. The relaxing incision defect was covered with mesh affixed by hernia staples. Following this, a Nissen or Toupet fundoplication completed the procedure in all but two cases. Short gastric vessels were usually stretched and elongated by the herniated fundus. These were divided only in the rare instance that there was tension on the fundoplication. A Toupet plication was chosen in favor of a Nissen plication in those patients with esophageal dysmotility demonstrated on manometry (peak mean body amplitude <40 mm Hg or less than 50% primary peristalsis). In later cases a Toupet was used in all patients who did not have preoperative manometry. Although no problems were encountered with esophageal damage from nasogastric tubes or dilators, increasing awareness of this possible complication strongly militated against intraoperative instrumentation of the esophagus. If there was difficulty identifying the esophagus, a Malony dilator was cautiously introduced by the anesthesiologist while the surgeon straightened the esophagus and esophagogastric-gastric junction as much as possible. Tightness of the warp was judged by unimpeded passage of a Babcock clamp between the wrap and the esophagus. If there was any question about esophageal or gastric integrity at any point in the procedure, 250 ml of methyl-blue-stained saline was introduced by NG tube in an attempt to identify any leakage.

Results

All patients Fifty-eight patients with a mean age of 67.3 years, range of 24 to 90 years, underwent LPHR over a 40-month period (Table 1). Fifty-seven had successful completion of the pro-

896 Table 1. Results of current series of 58 LPHR Elective LPHR Perioperative mortality Morbidity Length of staya Recurrence Previous failed repair Total intrathoracic stomach Preoperative GER Postoperative GER Total patients
a

Urgent LPHR 0 4 7.4 0 2 6 4 0 10

Total 0 4 2.8 0 5 12 46 0 58

% of total 0 7% 0 9% 21% 79% 0 100%

0 0 1.9 0 3 6 42 0 48

Minor complications consisted of urinary retention, a single episode of congestive heart failure requiring an additional 2 days of hospitalization, and one patient requiring esophageal dilation for dysphagia at 2 months postoperative. Mild dysphagia lasting 4 weeks or, rarely, up to 12 weeks, was common. One patient with normal esophageal motility had repair of an epiphrenic diverticulum 6 months postoperatively. This was either missed or not present on the initial EGD. No splenic injuries were encountered.

Previous hiatal hernia repair As noted above, five patients had previous failed hiatal hernia repairs (one laparoscopic and four by laparotomy) performed by other surgeons. The patient with a previous failed laparoscopic repair had an aborted attempt at repair without laparotomy. All normal tissue planes were obliterated, requiring identification of the esophagus by manual palpation. Those patients with a previous failed repair by laparotomy had extensive adhesions in the upper abdomen and mediastinum, making the cause of failure difficult to determine. One of these patients had an essentially untouched left crus, suggesting that the left crus was not properly identified at the first procedure.

Includes preoperative stay

cedure laparoscopically; one patient required conversion to a laparotomy. All but two patients had either Toupet (eight) or Nissen (48) fundoplication. One patient had reduction and excision of the sac with crural repair only. Another patient, who had end-stage Alzheimers dementia, had reduction of her torsion and gastrostomy only. She died 6 weeks postoperatively. Average total length of stay for all patients was 2.8 days. There were four major complications, two of which required laparoscopic reoperation for esophageal obstruction. One patient had a pulmonary embolism (survived) and one patient required a feeding gastrostomy for esophageal dysmotility and central dysphagia. All four major complications were in urgently repaired patients. There were no procedure-related deaths, although there were four early deaths from unrelated causes: two of Alzheimers dementia, one of ovarian carcinoma, and one of lung carcinoma at 12 months, 6 months, 2 months, and 6 months from surgery, respectively. One additional patient died of a myocardial infarct 13 months post-LPHR. At a mean of 12 months there were no known recurrences or symptoms of recurrence. Gastroesophageal reflux Forty-six patients had complained of gastroesophageal reflux symptoms preoperatively. All of these patients underwent fundoplication and reported resolution of reflux symptoms postoperatively. Total intrathoracic stomach PH A total of 12 patients had herniation of the entire stomach into the chest; six were repaired urgently. These six urgently repaired patients with intrathoracic stomach had three of the four major complications encountered in this series. Elective LPHR All 48 patients undergoing elective repair had either Nissen or Toupet fundoplication. Forty-two (87%) of these had symptoms of gastroesophageal reflux (GER) preoperatively. There were no major complications in the patients undergoing elective repair. Mean operative time was 115 min and average total length of stay (LOS) was 1.9 days.

Urgently repaired LPHR Of the 10 patients who underwent urgent LPHR, two were operated on for pain suggestive of incarceration or strangulation; one of these was strangulated but not infarcted. Eight had esophageal obstruction; four of these had aspiration pneumonia. Only one of the 10 underwent preoperative manometric studies. This patient had an uneventful postoperative course. Two patients with torsed total intrathoracic stomach PH developed postoperative esophageal obstruction with aspiration pneumonia. One of these, who did not undergo fundoplication, had infarcted the remnants of the excised hernia sac. This was corrected by dividing the remnant of the sac down to the serosa/muscularis at the cardioesophageal junction at the angle of His. The other patient had a Nissen fundoplication and developed edema in the area of the wrap and cardioesophageal junction. This was converted to a Toupet repair with prompt relief of obstruction. Both reoperations were done laparoscopically; they were discharged at 5 and 6 days following the second procedure. The additional two complications in this group of urgently repaired PHs were a pulmonary embolism at 3 weeks postoperative (survived) and one patient with dysphagia requiring a postoperative percutaneous endoscopic gastrostomy (PEG). This patient with Alzheimers and a total intrathoracic stomach hernia was unable to swallow preoperatively. Postoperatively she was able to swallow much better but still required supplemental tube feedings. Her dysphagia was from dysmotility demonstrated on esophagogram and was compounded by a failure to eat secondary to Alzheimers dementia. No evidence of mechanical obstruction was present; she died of complications of Alzheimers dementia 4 months postoperatively. Average LOS for these urgently repaired patients, including preoperative stay, was 7.4 days.

897 Table 2. Results of previous paraesophageal hernia repair by laparotomy and thoracotomy Total patients Ellis [4] Haas [5] Hill [6] Menguy [10] Ozdemir [12] Pearson [13] Treacy [17] Williamson [18] Averaged This study
a b

Emergent 10 a 12 12 1 18 7 20% 17%

a

Morbidity 13 4 a 2 a 5 a 14 14% 7%
b

Mortality 1 4 a 0 2 1 2 2 3% 0

Length of stay 9.5 11.4 a a a a a a 10.1 2.8

Recurrence 4 0 0 0 3 0 a 13 6% 0

51 21 22 30c 31 53 53 119 46 58

Not stated Does not include six splenectomies c All total intrathoracic stomach d Percentage of those stated

Discussion Elective LPHR had a very low morbidity with no major complications. Although follow-up is fairly limited (12 months), there have been no known recurrences. GER had been either cured or greatly improved in all cases. Elective hospital stay is commensurate with other laparoscopic procedures. Large series of PHR infrequently included morbidity rates or length of stay. Morbidity, when given, sometimes uses a different perspective than we presently use. Although carefully reported, Ellis [4] did not consider six splenectomies as morbidity and Menguy [10] did not consider 2 weeks of pseudo-obstruction or a gastric fistula as major complications. Comparison of the authors series to historical series must be kept in perspective. Open PHR morbidity rates averaged 14% [4, 5, 10, 13, 18] and mortality averaged 3% [4, 5, 10, 12, 13, 17, 18]. Length of stay, when noted, averaged 10.1 days [4, 5] for open repairs. LPHR in the authors series had a morbidity rate of 7%, no mortality, and an average LOS of 2.8 days (Table 2). Three of the four morbid events in the authors series were esophageal obstruction or dysphagia, all occurring in the group of urgently repaired patients. These three were probably to some extent preventable. Two occurred in torsed PH with total intrathoracic stomach undergoing Nissen fundoplication who did not have preoperative motility studies. From this experience, it is clear that if a fundoplication is used in urgent repairs without preoperative manometry, it should be a Toupet plication. The third patient, who did not have a fundoplication, was a failure to fully divide the remnants of the hernia sac. The 12-cm wide rim of sac infarcted and obstructed the cardioesophageal junction. Division of this rim down to the muscularis/serosa of the cardioesophageal junction is mandatory. Patients without preoperative esophageal manometry should be considered for Toupet fundoplication. The true definition of PH as been discussed by nearly every paper written on PH. Laparoscopy may offer some advantage over laparotomy in establishing the diagnosis of PH. The mere presence of a peritoneal defect adjacent to the esophagus does not presently qualify as a true PH, although it is hard to call this a pure sliding hernia. This type of defect is probably much more noticeable through a laparoscope for

two reasons. First, exposing the hiatus by laparotomy usually involves retracting the stomach inferiorly, reducing small paraesophageal peritoneal defects prior to visualizing them. Second, positive-pressure inflation during laparoscopy should make potential defects more prominent. In this study only patients with herniation of significant portions of intraperitoneal stomach in which the fundus was superior to the cardioesophageal junction were considered PH [14]. Preoperative UGI and EGD often described typical findings of PH but failed to actually label the diagnosis as a PH. These studies were done by numerous physicians and suggest that the definition and ramifications of PH are not universally understood. Whether or not to plicate PHs remains controversial. This series and some others [9, 13, 16, 17] have a high incidence of preoperative GER. Treacy [17], Pearson [13], and Styger [16] advised fundoplication in all PHR. Anatomically this makes sense. The diameter of the distal esophagus with associated wrap is significantly larger than the repaired hiatus and should help prevent subsequent reherniation. Also of concern is the rearrangement of anatomy resulting from PHR, which may allow reflux to develop when it was not formerly present. Other large series [4, 9, 10, 15, 16] conclude that there is a significant (10% overall) incidence of GER developing in patients who do not have antireflux procedures and presumably did not have preoperative GER (Table 3). Postoperative GER appears in several studies [4, 9, 10, 15, 16] as one of the sources of postoperative morbidity. With the exception of full wraps in acute torsions, the addition of fundoplication did not appear to add to morbidity in LPHR. Technically, fundoplication was simple and straightforward, the majority of dissection having been already completed during reduction of the hernia. The gastric fundus was uniformly floppy, likely because of elongation of the gastrosplenic ligament. Short gastric vessels were usually not divided. Fundoplication carries a low risk with a significant potential benefit in LPHR. Although only eight patients had preoperative esophageal obstruction, experience with them strongly suggests that patients with complete esophageal obstruction should be repaired promptly in spite of any underlying pneumonia. Pulmonary status and SaO2 improved dramatically within

898 Table 3. Patients with gastroesophageal reflux (GER) after paraesophageal hernia repair without antireflux procedures Total hiatal hernias Ellis [4] McKernon [9] Menguy [10] Streitz [15] Styger [16] Williamson [18] Total This study
a

Paraesophageal 51 16 30 13 28 119 257 58

Anatomic repair without anti-reflux procedure 49 6 21 3 19 100 198 2

Postoperative GERb 4 1 1 1 6 6 19 (10%) 0

222 283 a a a a 218

aware of and familiar with techniques for LPHR. Frequent conversion to laparotomy is denying these patients a procedure of lower morbidity. One-half of the procedures in this series were encountered in patients with a primary diagnosis of gastroesophageal reflux. Surgeons attempting laparoscopic fundoplication should have adequate experience and expertise to consistently complete these procedures laparoscopically. Although long-term follow-up is needed, the laparoscopic approach should be strongly considered in patients undergoing PH repair. References
1. Cloyd DW (1994) Laparoscopic repair of incarcerated paraesophageal hernias. Surg Endosc 8: 893897 2. Congreve DP (1992) Laparoscopic paraesophageal hernia repair. J Laparoendosc Surg 2: 4548 3. Cuschieri A (1993) Laparoscopic antireflux surgery and repair of hiatal hernia. World J Surg 17: 4045 4. Ellis FH, Crozier RE, Shea JA (1986) Paraesophageal hiatus hernia. Arch Surg 121: 416420 5. Haas O, Rat P, Christophe M (1990) Surgical results of intrathoracic gastric volvulus complicating hiatal hernia. Br J Surg 77: 13791381 6. Hill LD, Tobias JA (1968) Paraesophageal hernia. Arch Surg 96: 735744 7. Koger KE, Stone JM (1993) Laparoscopic reduction of acute gastric volvulus. Am Surg 59: 325328 8. Kuster GG, Gilroy S (1993) Laparoscopic technique for repair of paraesophageal hiatal hernias. J Laparoendosc Surg 3: 331338 9. McKernon JB, Champion JK (1995) Laparoscopic antireflux surgery. Am Surg 61: 530536 10. Menguy R (1988) Surgical management of large paraesophageal hernia with complete intrathoracic stomach. World J Surg 12: 415422 11. Oddsdottir M, Franco AL, Laycock WS, Waring JP, Hunter JG (1995) Laparoscopic repair of paraesophageal hernia. Surg Endosc 9: 164 168 12. Ozdemir IA, Burke WA, Ikins PM (1973) Paraesophageal hernia. A life-threatening disease. Ann Thorac Surg 16: 547554 13. Pearson FG, Cooper JD, Ilves R (1983) Massive hiatal hernia with incarceration: a report of 53 cases. Ann Thorac Surg 1: 4551 14. Pitcher DE, Curet MJ, Martin DT (1995) Successful laparoscopic repair of paraesophageal hernia. Arch Surg 130: 590596 15. Streitz JM, Ellis FH (1990) Iatrogenic paraesophageal hiatus hernia. Ann Thorac Surg 50: 446449 16. Styger S, Ackermann C, Schuppisser JP, Tondelli P (1995) Reflux disease following gastropexy for para-esophageal hiatal hernia. Schweiz Med Wochenschr 125: 12131215 17. Treacy PJ, Jamieson GG (1987) An approach to the management of para-esophageal hiatus hernia. Aust N Z J Surg 57: 813817 18. Williamson WA, Ellis FH, Streitz JM, Shahian DS (1993) Paraesophageal hiatal hernia: is an antireflux procedure necessary? Ann Thorac Surg 56: 447521

Not stated b Of those not having had an antireflux procedure who subsequently required therapy for GER

hours of repair in these patients. Delaying surgery to correct malnutrition almost inevitably resulted in the patient developing aspiration pneumonia. The use of a laparoscope to visualize the hiatus and mediastinum allows visualization and appreciation of anatomical details that are difficult to appreciate at laparotomy. Dissection of the hernia sac and subsequent reduction of the stomach is straightforward and relatively easy if the anatomy is understood. The dissection into the mediastinum should be started by incising the posterior peritoneum of the lesser sac. The lesser sac is dissected free and reduced from external to the sac. The greater sac hernia is likewise approached externally. The esophagus is easily identified and mobilized by this technique. The low morbidity and low mortality of LPHR, as noted in this study, are significant in that old and debilitated patients who had been refused open repair were included in this series. The low morbidity of LPHR should allow repair in patients formerly thought to be too risky for PHR by laparotomy or thoracotomy. Although many surgeons advocate repair of all PH, it is apparent that there are many physicians taking the view advocated by Treacy [17] of watchful waiting in elderly or debilitated patients. The low morbidity of LPHR has convinced many of our local physicians that early repair should be undertaken. Indeed, it is a change in this threshold that has contributed to the large numbers of LPHR in this series. As more and more laparoscopic antireflux procedures are done by increasing numbers of surgeons, we need to be

Surg Endosc (1997) 11: 963964

Surgical Endoscopy
Springer-Verlag New York Inc. 1997

Extracorporeal knotting simplified with a new instrument

E. Croce, S. Olmi
1st Department of General and Thoracic Surgery, Center for Minimally Invasive Surgery, Fatebenefratelli and Oftalmico Hospital, C.so Porta Nuova 23, Milan, Italy Received: 11 June 1995/Accepted: 26 January 1996

Abstract. We report a quick, reliable, inexpensive method that uses a new, reusable instrument which can be used in laparoscopic and thoracoscopic surgery to execute any kind of extracorporeal suture. Key words: Extracorporeal laparoscopic suturing Roeder slip-knot Surgical knot

In minimally invasive surgery the most common technique used for suturing (for example, in Nissen Rossetti fundoplication, of a blood vessel, a cystic duct, or the base of the appendix) is the standard extracorporeal knot technique. An instrument created by us allows speed and ease of execution and the possibility of doing different types of knots with any type of suture material that is at least 90 cm long. Is it possible not only to do some autostopping slipknots in an easier way, equivalent to a Roeder knot, but, more importantly, all the simple and double surgical knots which one traditionally used can be executed with ease? After having made the knot, the instrument works as a knot pusher.

Technique
The instrument for tying laparoscopic knots (produced by Karl Storz Gmbh) is made of steel; it is about 55 cm long, with a proximal part 3 mm in diameter and 6 cm long that ends in a flattened portion which contains an aperture. The distal part, 5 mm in diameter and 49 cm long, has a rounded end that allows its use as a small probe for palpation (Fig. 1). Let us now examine the technique of simplified external knotting. Having introduced the needle into the cavity, while the other end of the suture stays external to the trocar, it is passed through the two edges in order to stitch and is then extracted via the same trocar in which it had been introduced. The surgical knot or slip knot is made quickly and autonomously. There is no need for an assistants finger to remain on the trocar where the two ends come out, to keep them separated and to prevent the

escape of CO2such as occurs with the Roeder slip knot, which engages both the surgeon and an assistant. Thus the needle is passed through the hole and, while the right hand is holding both the instrument and the part of the thread ending with the needle (Fig. 2), the left hand is used to perform the simple or double surgical knot. Having tied the knot, it is sufficient to advance the instrument into the abdominal cavity so that the knot can be tied. In this manner, the instrument behaves very much as an extension of the surgeons finger (Figs. 3 and 4). Once the first knot has been tightened, the instrument is pulled out and the procedure is repeated until the desired number of knots have been created. As mentioned earlier, it is also possible to create a slip knot similar to a Roeder slip knot. After tying a surgical knot, the thinner part of the instrument is put just under the knot, holding the two ends of the suture material close to it. The shorter end of the instrument is circled around the end with the needle four times so that it is pegged together to the last coil and then passed in the hold (Fig. 5). The instrument is then pulled out, which allows preparation of knots similar to a Roeder knot (Fig. 6). Successively introducing the needle in the cavity, the knot can be pushed into the abdominal cavity and will position itself around the desired structure (Fig. 7). Another similar slip knot is obtained by locating both ends of the thread parallel to the instrument and winding the shorter end round the end with the needle five or six times with (Fig. 8) or without (Fig. 9) pegging together the third coil. Passing the short end in the hole, the instrument is pulled out, thus tying the knot being made, which can be pushed into the cavity (Figs. 6 and 7). It is possible, in order to obtain extra hold, to execute a surgical knot in front of the slip-knot. Moreover, by executing a slip knot it is possible to employ the new instrument in the usual fashion as an endoloop, using the preferred thread and knotting and tightening the knot everywhere due to the length of the instrument.

Results We successfully use this instrument and these suturing techniques in any situation in which execution of an extracorporeal knot is required. In about 30 s, especially when employing the usual surgical knot, even an inexpert laparoscopic surgeon is able to

Correspondence to: S. Olmi, Via Solferino 42, 20121 Milano, Italy

964

Fig. 1. Laparoscopic knot-tying instruments. Fig. 2. Needle is passed through the hole of the instrument to tie knot. Fig. 3. Once tied, the instrument acts as an extension of the surgeons finger.

execute extracorporeal suturing extremely safely and much more simply, quickly, and cheaply than with any other method.
Fig. 4. The instrument may now be advanced into the abdominal cavity.

Discussion One of the most problematic areas in the field of laparoscopic surgery is knot tying. In certain instances, this hurdle is sufficiently great that laparoscopic procedures have to be abandoned. This is particularly true where organs such as the proximal stomach require suturing in the procedures such as Nissen fundoplication; the method described here makes it possible to execute blood vessel bindings, a cholecystectomy, a Heller-Dor, a gastric perforation closure, a reinforcement of stapled-bowel anastomosis, and intestinal anastomosis. In addition, the length of the instrument allows one to tighten a knot anywhere either in pelvic rectness or at esophageal hiatus level, even in obese patients. This paper describes an instrument which greatly facilitates the execution of knot tying during laparoscopic interventions. This instrument acts in essence as an extension of the surgeons hand and, given its rigid structure, allows the surgeon full control of the process of knot tying. As a result, this technical device simplifies knot tying and may help to reduce the frustration often associated with intracorporeal suturing during laparoscopic surgery. References
1. Buess G, Cuschieri A, Manuckle K (1993) Technique and preliminary results of laparoscopic cardiomyotomy. End Surg 1: 7681

Figs. 57.

Technique for preparation of knots similar to a Roeder knot.

Figs. 89. Technique for preparation of slip knots with (Fig. 8) or without (Fig. 9) pegging together the third coil.

2. Cadiere G, Houben J, Bruyns J, Himpens J, Panzer J, Gelin M (1994) Laparoscopic Nissen fundoplication; technique and preliminary results. Br J Surg 81: 400403 3. Cuschieri A, Shimi S, Nathanson L (1992) Laparoscopic reduction, crural repair, and fundoplication of large hiatal hernia. Am J Surg 163: 425430 4. Meilahn JE (1992) The need for improving laparoscopic suturing and knot tying. J Laparoendosc Surg 2: 267 5. Nathanson LK, Nathanson PDK, Cuschieri A (1991) Safety of vessel ligation in laparoscopic surgery. Endoscopy 23: 206210 6. Nathanson LK, Easter DW, Cuschieri A (1991) Ligation of the structures of the cystic pedicle during laparoscopic cholecystectomy. Am J Surg 161: 350354 7. Roder H (1918) Die tecknik der mandelgesundungbestrebungen. Aertzl Rundshau 57: 169171 8. Semm K (1978) Tissue puncher and loop ligation. New aids for surgical-therapeutic pelviscopy endoscopy intraabdominal surgery. Endoscopy 10: 119124 9. Shimi S, Cuschieri A, Vander Velpen G (1994) Comparative study of the holding strength of slipknots using absorbable and nonabsorbable ligature materials. Surg Endosc 8: 12851291 10. Soper NJ, Hunter JG (1992) Suturing and knot tying in laparoscopy. Surg Clin North Am 72: 11391152 11. Sung-Tao K, Airan MC (1994) Therapeutic laparoscopic suturing techniques. Surg Endosc 6: 4146 12. Szabo Z (1993) Laparoscopic suturing and tissue approximation. In: Minimally invasive surgery. McGraw Hill, New York, pp 141155