Laparoscopic Repair of Large Type III Hiatal Hernia:
Objective Followup Reveals High Recurrence Rate
Majid Hashemi, FRCS, Jeffrey H Peters, MD, FACS, Tom R DeMeester, MD, FACS, James E Huprich, MD,
Marcus Quek, MD, Jeffrey A Hagen, MD, Peter F Crookes, MD, FACS, Jorg Theisen, MD,
Steven R DeMeester, MD, Lelan F Sillin, MD, FACS, Cedric G Bremner, MD, FACS
Background: Recent studies based on symptomatic
outcomes analyses have shown that laparoscopic repair
of large type III hiatal hernias is safe, successful, and
equivalent to open repair. These outcomes analyses
were based on a relatively short followup period and
lack objective confirmation that the hernia has not
recurred. The aim of this study was to compare the
outcomes of laparoscopic and open repair of large type
III hiatal hernia using both symptomatic evaluation
and barium study to assess the integrity of the repair.
Study Design: Fifty-four patients underwent repair of
a large type III hiatal hernia between 1985 and 1998.
The surgical approach was laparotomy in 13, thoracot-
omy in 14, and laparoscopy in 27. An antireflux pro-
cedure was included in all patients. Symptomatic out-
comes were assessed using a structured questionnaire
at a median of 24 months and was complete in 51 of 54
patients (94%). A single radiologist, without knowl-
edge of the operative procedure, assessed the integrity
of the repair using video esophagram. Videos were per-
formed at a median of 27 months (35 months open and
17 laparoscopic) and were completed in 41 of 54 pa-
tients (75%).
Results: Symptomatic outcomes were similar in both
groups with excellent or good outcomes in 76% of the
patients after laparoscopic repair and 88% after an
open repair. Reherniation was present in 12 patients
and was asymptomatic in 7. A recurrent hernia was
present in 12 of the 41 patients (29%) who returned
for a followup video esophagram. Forty-two percent
No competing interests declared.
Presented at the American College of Surgeons 85th Annual Clinical
Congress, San Francisco, CA, October 1999.
Received October 22, 1999; Revised December 28, 1999; Accepted January
4, 2000.
From the Departments of Surgery (Hashemi, Peters, T DeMeester, Quek,
Hagen, Crookes, Theisen, S DeMeester, Sillin, Bremner) and Radiology,
(Huprich), University of Southern California, Los Angeles, CA.
Correspondence address: Jeffrey H Peters, MD, FACS, University of South-
ern California, 1510 San Pablo St, Los Angeles, CA 90033.
© 2000 by the American College of Surgeons
Published by Elsevier Science Inc.
(9 of 21) of the laparoscopic group had a recurrent
hernia compared with 15% (3 of 20) of the open group
(p<0.001 log-rank value on recurrence-free
followup).
Conclusions: Laparoscopic repair of type III hiatal her-
nias is associated with a disturbingly high (42%) prev-
alence of recurrent hernia. More than half such recur-
rences have few, if any, symptoms. (J Am Coll Surg
2000;190:553–561. © 2000 by the American Col-
lege of Surgeons)
Herniation of a portion of the stomach through the
esophageal hiatus into the posterior mediastinum is
a common affliction of modern humans. Hiatal
hernias are classified into three types according to
their anatomic characteristics. Type I hernias in-
clude the common sliding hiatal hernia, in which
there is a dome-shaped upward migration of the
gastroesophageal junction into the posterior medi-
astinum. Type II, or true paraesophageal hernias,
have no sliding component and are characterized by
a normally positioned gastroesophageal junction
with herniation of the gastric fundus into the me-
diastinum alongside the esophagus. They are un-
common. Type III hernias, or mixed hernias, have
both a significant sliding component and a para-
esophageal component. Most so-called “paraesoph-
ageal” hernias are the mixed type (Fig. 1). Type III
hernias tend to be large, and repair has been advo-
cated regardless of symptoms because of the poten-
tial for life-threatening complications such as incar-
ceration and strangulation.1,2
Recent reports have suggested that type III her-
nias can be adequately treated by the laparoscopic
approach, and that the repair is successful in most
patients.3-7 Few of these reports contain objective
documentation of the success of hernia repair and
most have based their conclusions on short-term
ISSN 1072-7515/00/$21.00
553 PII S1072-7515(00)00260-X
 554 Hashemi et al Laparoscopic Repair of Type III Hiatal Hernia
Figure 1. Type III mixed hiatal hernia, arrow showing the gastro-
esophageal junction and dashed line the hiatus.
symptomatic followup.8 The purpose of this study
was to compare the outcomes of laparoscopic and
open repair of large type III hiatal hernias using
both symptomatic evaluation and radiographic bar-
ium studies to assess the integrity of the repair.
METHODS
Fifty-four patients with type III hiatal hernias un-
derwent primary repair between January 1985 and
May 1998. Two patients died from unrelated causes
during the followup period, one of whom had re-
sponded to a symptom questionnaire. One patient
was lost to followup. The surgical approach was
Table 1. Preoperative Findings
Characteristic
Median age (y) (range)
Median hernia size (cm) (range)
Erosive esophagitis (%)
Barrett’s esophagus (%)
Defective sphincter (%)
Positive 24-h pH (%)
J Am Coll Surg
transthoracic in 14, transabdominal in 13, and
laparoscopic in 27. The median age of the laparo-
scopic group was 68 years (range 42 to 83 years) and
of the open group, 66 years (range 39 to 80 years).
There were 18 men and 36 women. Preoperative
evaluation included video barium esophogram in
50 of 54 patients (93%), upper endoscopy in 52
(96%), esophageal manometry in 51 of 54 patients
(95%), and 24-hour ambulatory pH monitoring in
44 of 54 patients (81%). Hernia size was measured
radiographically. The sliding component was taken
as the distance from the radiographic hiatus to the
gastroesophageal junction and the paraesophageal
component as the distance from the hiatus to the
top of the fundus (shown in Fig. 1). Preoperative
lower esophageal sphincter characteristics, mano-
metric esophageal length, esophageal body motility,
24-hour pH scores, and radiographic hernia sizes
were similar in both groups (Table 1).
Surgical technique
All but one of the operations were carried out elec-
tively. The only emergency operation was by lapa-
rotomy in a 75-year-old woman who had intermit-
tent obstructive symptoms, but in whom the hernia
was easily reduced to reveal a healthy and well-
perfused stomach. Fifty-two patients had a Nissen
fundoplication, one had a Collis gastroplasty and
partial fundoplication, and one had a Toupet hemi-
fundoplication, both performed by laparotomy.
Surgical principles, irrespective of approach, in-
cluded reduction of the hernia, complete excision of
the sac, primary closure of the crura, and an antire-
flux procedure. In the laparoscopic group, the sac
was completely removed from the thoracic cavity
but not excised from its attachments at the gastro-
esophageal junction. The crura were approximated
posterior to the esophagus with a median of 7
(range 3 to 9) crural sutures. Pledgets were occa-
Laparoscopic Open p
nⴝ27 n ⴝ27 Value
68 (42–83) 66 (31–80) 0.82
7 (3–12) 8 (3–13) 0.80
11/26 (42%) 5/26 (19%) 0.13
2/26 (8%) 0/26 (0%) 0.47
16/26 (62%) 11/25 (44%) 0.33
11/22 (50%) 12/22 (55%) 0.88
 Vol. 190, No. 5, May 2000 Hashemi et al Laparoscopic Repair of Type III Hiatal Hernia 555

sionally used to aid crural suture placement. No

objective measure of crural tightness was used. No
mesh or prosthesis was used in any patient. Gas-
tropexy was not part of the procedure.

Symptomatic followup
Symptomatic followup was completed in 94% of
the patients (51 of 54 patients). All patients were
seen in person or contacted by telephone and re- Figure 2. Distribution of time for followup video esophagrams.
sponded to a structured questionnaire. Question- LAP, laparoscopy.
naires were completed in 25 of 27 patients in the
open group at a median of 34 months (range 4 to
166 months) and in 26 of 27 patients in the lapa- rical meniscus-shaped soft tissue shadowing at the
roscopic group at a median of 17 months (range 5 gastroesophageal junction that does not fill with
to 49 months). Symptomatic outcomes were contrast or air and that maintains a subdiaphrag-
graded as excellent if the patient was asymptomatic, matic position during inspiration. Fundoplication
good if there was overall symptomatic improvement disruption was identified by loss of symmetrical soft
but minor gastrointestinal symptoms were present tissue shadowing around the proximal stomach
that did not require medication for relief, fair if with or without loss of central narrowing at the level
there was overall symptomatic improvement but of the gastroesophageal junction.
gastrointestinal symptoms were present that re-
Statistics
quired medication for relief, and poor if the symp-
toms were the same as or worse than before Categorical variables were compared using the chi-
operation. square test. The Student’s t-test was used for con-
tinuous variables. Freedom from recurrence was cal-
Radiographic followup culated using the method of Kaplan and Meier.9 For
Forty-one of 54 patients (75%) agreed to followup Kaplan-Meier analysis, recurrence was considered
radiographic examination. Video esophagrams were to have occurred on the day of the study. Compar-
performed according to a protocol of five swallows ison of recurrence was performed using the log-rank
of liquid barium followed by a solid barium bolus method.
meal; anteroposterior and oblique views were ob-
tained in the upright and supine positions. All vid- RESULTS
eos were reviewed on two separate occasions 3
Symptomatic outcomes
months apart by an experienced esophageal radiol-
ogist who was unaware of the type of operative pro- Symptomatic outcomes were assessed at a median
cedure performed. Video esophagrams were com- of 24 months after surgery (34 months for open, 17
pleted in 20 of 27 patients (74%) in the open group months for laparoscopic) and are shown in Table 2.
at a median of 35 months (range 4 to 166 months) Overall, 42 patients (82%) had excellent or good
and in 21 of 27 patients (77%) in the laparoscopic
group at a median of 17 months (range 5 to 45 Table 2. Symptomatic Outcomes
months). Figure 2 shows the times of followup
video esophagrams. Anatomic recurrence was de- Laparoscopic (%) Open (%) Overall (%)
Outcomes n ⴝ26 nⴝ25 nⴝ51
fined as any evidence of herniation of the stomach
above the level of the diaphragm. Disruption of the Excellent 46 64 55
Good 31 24 27
fundoplication alone with no reherniation was Excellent/good 77 88 82
noted but not classified as a recurrence. Radiologic Fair 19 8 14
confirmation of an intact fundoplication and crural Poor 4 4 4
repair was based on the presence of the following Fair/poor 23 12 18
Satisfied 96 96 96
features: a central luminal narrowing with symmet-
 556 Hashemi et al Laparoscopic Repair of Type III Hiatal Hernia
Figure 3. Cumulative recurrence-free followup: open versus
laparoscopic.
outcomes and 9 had fair or poor outcomes. Eighty-
eight percent (22 of 25) of the open group had good
to excellent outcomes compared with 77% (20 of
26) in the laparoscopic group. When asked, 49 pa-
tients (96%) were satisfied with the outcomes of
their procedure. Symptoms correlated poorly with
the presence or absence of a recurrent hernia. Ten of
the 42 patients with relief of their symptoms had
radiologic evidence of either reherniation (7 of 42
patients) or wrap disruption (3 of 42), and 5 of the
9 patients with persistent symptoms had no evi-
dence of recurrence on imaging. One patient in the
laparoscopic group and one in the open group re-
quired postoperative dilatation.
Recurrent hernia
A recurrent hernia was present in 12 of the 41 pa-
tients (29%) who returned for a followup video
esophagram. Reherniation and disruption of the
fundoplication were present in nine and rehernia-
tion with an intact fundoplication in three. Disrup-
tion alone was seen in three patients (7%).
Forty-two percent (9 of 21) of the laparoscopic
group had a recurrent hernia compared with 15%
(3 of 20) of the open group (p⬍0.001 log-rank
value on freedom from recurrence at followup).
Kaplan-Meier estimates of recurrence-free propor-
tions are shown in Figure 3. The surgical approach
was transabdominal in two and transthoracic in one
of the three patients found to have recurrence after
open procedure, and these failures were detected on
videos carried out at 30, 35, and 64 months after
operation. The longest video followup for the open
J Am Coll Surg
group was 166 months. Failure after laparoscopic
repair was detected on videos performed at 5, 11,
12, 14, 21, 27, 41, and 45 months after operation.
The longest video followup for the laparoscopic
group was 45 months. No difference in the preva-
lence of recurrent hernia was found when the first
10 operations (both open and laparoscopic) were
compared with the last 10.
Univariate analysis of potential risk factors as-
sociated with recurrent hernia failed to find any
significant factor independent of the operative ap-
proach. Specifically, neither age, gender, presenting
symptoms, manometric esophageal length, 24-
hour pH score, nor hernia size significantly affected
the outcomes. The prevalence of erosive esophagitis
was higher in the laparoscopic group, but this dif-
ference was not statistically significant and no in-
creased risk of reherniation was imparted by the
presence of esophagitis. Two patients in the laparo-
scopic group had long segment Barrett’s esophagus;
one of these patients developed a recurrent hernia.
Operative morbidity
One patient died of respiratory failure after open
transabdominal repair. She was a 67-year-old
woman with preexisting asthma and end-stage pul-
monary disease, whose operation was uneventful.
Major complications occurred in seven patients and
minor complications in six (Table 3). There was no
difference in morbidity and mortality between the
open and laparoscopic groups. Two patients were
converted from laparoscopic to open, one because
of esophageal shortening and one after developing
pericardial tamponade. Median hospital stay, days
until removal of nasogastric tube, and mean opera-
tive times for the three procedural groups are given
in Table 4. There was a significantly shorter time of
nasogastric intubation (p⬍0.001) and a shorter
hospital stay in the laparoscopic group (p⬍0.001)
versus the open group.
DISCUSSION
This study has identified a disturbingly high prev-
alence of recurrent hernia on radiographic investi-
gation of patients after laparoscopic repair of large
type III hiatal hernias. Nearly half of the patients
returning for video esophagram after laparoscopic
repair during a followup period of 5 to 45 months
(median 17 months) were found to have a recurrent
 Vol. 190, No. 5, May 2000 Hashemi et al Laparoscopic Repair of Type III Hiatal Hernia 557

Table 3. Complications
Complication Laparoscopic (nⴝ27) Open (nⴝ 27)
Minor Urinary tract infection (n⫽2) Cellulitis of wound
Left basal atelectasis Skin wound dehiscence
Atrial fibrillation
Major Exacerbation of COPD and pneumonia Hemothorax from bleeding intercostals requiring
repeat thoracotomy
Hemothorax* Pleural effusion
Retroperitoneal bleed* Cardiorespiratory failure
Cardiac tamponade (conversion from Pneumonia
laparoscopic to open)
*In the same patient.

hernia. Most of these reherniations were asymp-
tomatic and would not have been detected other-
wise. This compares with a 15% prevalence of re-
current hernia over a followup period of 4 to 166
months (median 34 months) in patients having
open repairs. The only factor associated signifi-
cantly with reherniation was the type of surgical
approach.
Because patients were not followed routinely by
esophagram, we know only that hernia recurrence
occurred on or before the date of the study. This
could result in bias favoring the open group by vir-
tue of its longer average followup. In an effort to
detect if this bias affected our conclusions, we com-
pared freedom from recurrence between the two
groups, assuming the open group failed very early
(any observed reherniation had recurred by the end
of the first postoperative year) and the laparoscopic
group as late as possible (ie, on the date of the video
examination). This analysis would minimize the
effects of varying lengths of followup on the
Kaplan-Meier estimates of freedom from recur-
rence. Despite this hypothetical analysis, a signifi-
cant difference in the prevalence of recurrence-free
patients favoring the open group remained.
Current literature states that laparoscopic repair
of paraesophageal hiatal hernias is highly successful
Table 4. Perioperative Course
Variable Laparoscopic Open
Operating time (min) 184 176
Time to oral diet (d) 1 3
Hospital stay (d) 3 9
Mortality 0 1
Major complications* 3 4
*Number of patients with complications.
NS, not significant.
(Table 5). Most authors report symptomatic im-
provement in 80% to 90% of patients and less than
10% to 15% prevalence of recurrent hernia. The
problem of recurrent hernia after laparoscopic re-
pair of any hiatal hernia is becoming increasingly
appreciated. Recurrent hernia is now the most com-
mon cause of anatomic failure after laparoscopic
Nissen fundoplication done for gastroesophageal
reflux disease (GERD). Cornwell and colleagues15
compared the reasons for failure in patients after
laparoscopic and open Nissen fundoplication for
GERD. The reasons for failure were significantly
different. Failures after laparoscopic fundoplication
were largely from recurrent hernia (11 of 16 pa-
tients); failures after open fundoplication were from
wrap disruption or slippage or too long or tight a
fundoplication (9 of 10 patients). Soper and Dun-
negan16 likewise found recurrent hernia to be the
most common cause of anatomic fundoplication
failure. Why this is so is unclear. It may be because
of the selection of a laparoscopic approach in pa-
tients with a shortened esophagus; lack of, or break-
down of the crural closure; less extensive esophageal
mobilization; or a reduced tendency for adhesion
formation after laparoscopic compared with open
surgery. Horgan and colleagues17 have concluded
that the most important technical factors prevent-
ing recurrence in the setting of reflux disease were
effective crural closure, transhiatal esophageal mo-
p Value bilization, attention to the geometry of the fundo-
NS plication, and anchoring the wrap to the esophagus
⬍0.001 and surrounding tissues.
⬍0.001 The problem of recurrent hernia after repair of
NS
NS large type III hiatal hernias has received less atten-
tion. Outcomes after repair of these hernias are usu-
ally based on symptomatic assessment alone. Al-
 558 Hashemi et al Laparoscopic Repair of Type III Hiatal Hernia J Am Coll Surg

Table 5. Review of Reported Outcomes of Laparoscopic Repair of Paraesophageal Hernias

Mean Excellent/good Hernia
Fundoplication followup Morbidity symptomatic recurrence
Lead author Year n (%) (mo) (%) outcomes (%) (%)
Oddsdottir 4
1994 10 90 9 20 80 0
Pitcher3 1995 12 60 8 25 83 0
Willekes10 1996 30 76 — 27 — 0
Casabella11 1996 15 100 — 20 100 —
Trus12 1996 76 95 — 28 — 5
Perdikis8 1997 65 100 18 14 92 3
Gantert13 1997 55 73 10 16 — 3
Wu14 1999 38 100 12 16 79 23*
*On videos conducted at 3 to 5 months.

though recurrence rates of 6% to 13% have been
reported, they have largely been based on the need
for reoperation or investigations that are performed
on a selective basis.18 A problem is that most studies
include pure type II hernias, which may have a sig-
nificantly different propensity for reherniation
given the unaltered location of the gastroesophageal
junction. Wu14 recently reported 38 patients who
underwent repair for type II and type III hernias, 35
of whom had postoperative endoscopic or radio-
graphic examinations. At a followup of 3 months or
more postoperatively, some degree of anatomic re-
currence was noted in 23% of patients who under-
went laparoscopic repair of their hernia. He also
found no relationship of operative variables to the
prevalence of hernia recurrence.
The principles of laparoscopic repair of a large
intrathoracic hernia are analogous to those for an
open procedure, namely reduction of the hernia,
excision of the peritoneal sac, crural repair, and fun-
doplication. But there are several factors that make
laparoscopic repair of these large hernias complex.
First, volvulus of the stomach is often associated
with these hernias and makes identification of the
anatomy, in particular the location of the esopha-
gus, difficult. Second, type III hernias tend to be
large as shown by the measurements reported in this
study, and laparoscopic dissection of a large hernia
sac frequently results in sufficient bleeding to ob-
scure the field of view and impair recognition of the
anatomy. Third, the hiatal opening in a patient with
a large hernia is wide, with the right and left mus-
cular crura often separated by 4cm or more. This
can make closure problematic because of the ten-
sion required to bring the crura together. Fourth,
the right crus may be devoid of stout tissue and
sutures may pull through it easily. Last, redundant
tissue present at the gastroesophageal junction after
dissection of the sac retards creation of the
fundoplication.
Our study shows that although symptoms may
give some indication of a recurrence, most recur-
rences are not symptomatic. The ideal management
of an asymptomatic anatomic recurrence has yet to
be established. There may be risks of incarceration
and strangulation after reherniation, particularly if
the crura have been approximated and the hiatus
narrowed. During the era when fundoplications
were deliberately placed in the chest and the hiatus
was not narrowed, ulceration and perforation of an
intrathoracic stomach developed with potentially
lethal sequelae.19,20 Richardson21 has illustrated well
the problems of an intrathoracic fundoplication: of
12 patients with a fundoplication deliberately
placed in the chest because of an acquired short
esophagus, five developed serious complications
ranging from ulceration within the fundoplication
and severe gastrointestinal bleeding to fatal gastro-
bronchial fistula.
The possibility that esophageal shortening con-
tributed to the development of recurrent hernia de-
serves consideration. The presence of esophageal
shortening is almost certainly a factor responsible
for recurrent hernia formation in patients with large
type I hiatal hernias. In the setting of GERD, most
authors have identified a 3% to 5% or higher prev-
alence of the need for Collis gastroplasty to achieve
a tension-free intraabdominal repair.22,23 Although
it is possible that esophageal shortening is a contrib-
uting factor in patients with large type III hernias,
only one of the patients in the present study re-
quired esophageal lengthening. Altorki and associ-
 Vol. 190, No. 5, May 2000 Hashemi et al
ates24 reported a similar experience. Forty-seven of
52 patients with large type III hiatal hernias were
successfully repaired by the transthoracic approach
without the need for Collis gastroplasty. These ob-
servations suggest that severe esophageal shortening
is uncommon. It is possible, perhaps even likely,
that the complete esophageal mobilization that can
be achieved through a thoracic approach lessens the
need for a Collis gastroplasty.23 Properly designed
prospective studies are required to answer these
questions.
The use of prosthetic mesh as an adjunct to
repair has been advocated for both open25 and lapa-
roscopic26 repairs of large hiatal hernias. Whether its
use is beneficial or not remains controversial; we
prefer to avoid prosthetic material if possible. In
contrast to groin hernias, the esophageal hiatus is a
very dynamic area, with constant movement of the
diaphragm, esophagus, stomach, and pericardium.
Erosion of prosthetic material placed in this area
into the gastrointestinal tract will occur; the only
question is how often. Short-term followup of most
studies is insufficient to provide insight into this
problem. In one of the largest and longest term
studies reported, Carlson and coworkers25 followed
43 patients with mesh repair of large hiatal hernias
for a mean of 52 months. Mesh erosion into the
fundus of the stomach occurred in one patient
(2.3%). In practice, the issue boils down to the risk
of hernia recurrence given repair without the use of
prosthetic mesh and that of mesh erosion or recur-
rent hernia in patients with prosthetic repairs; nei-
ther are well known. The 15% prevalence of recur-
rent hernia in patients after open repair is, we
believe, not high enough to warrant placement of
prosthetic mesh. On the other hand the disturb-
ingly high recurrence in the laparoscopic group may
warrant its consideration in those undergoing lapa-
roscopic repair.
The objective evaluation carried out in this
study reveals that there is a high prevalence (42%)
of recurrent hernia after laparoscopic repair of large
type III hernias. Given the limitations of a retro-
spective analysis our data suggest that recurrent her-
nia after open repair is less common. This should be
borne in mind when selecting the operative ap-
proach in a patient with such a hernia. At present,
given otherwise healthy patients, we have returned
to an open procedure, preferring a transthoracic ap-
Laparoscopic Repair of Type III Hiatal Hernia 559
proach. This provides unimpaired esophageal mo-
bilization from the hiatus to the aortic arch and
gives ideal exposure for a firm closure of the esoph-
ageal hiatus.
References
1. Maziak DE, Todd TR, Pearson FG. Massive hiatus hernia: eval-
uation and surgical management. J Thorac Cardiovasc Surg
1998;115:53–60.
2. Skinner DB, Belsey RH. Surgical management of esophageal
reflux and hiatus hernia. Long-term results with 1,030 patients.
J Thorac Cardiovasc Surg 1967;53:33–54.
3. Pitcher DE. Successful laparoscopic repair of paraesophageal
hernia. Arch Surg 1995;130:590–596.
4. Oddsdottir M, Franco AL, Laycock WS, et al. Laparoscopic
repair of paraesophageal hernia. New access, old technique. Surg
Endosc 1995;9:164–168.
5. Huntington TR. Short-term outcome of laparoscopic para-
esophageal hernia repair. A case series of 58 consecutive patients.
Surgical Endoscopy 1997;11:894–898.
6. Koger KE, Stone JM. Laparoscopic reduction of acute gastric
volvulus. Am Surg 1993;59:325–328.
7. Edye MB, Cannin-Endres J, Gattorno F, Salky BA. Durability of
laparoscopic repair of paraesophageal hernia. Ann Surg 1998;
228;4:528–535.
8. Perdikis G, Hinder RA, Filipi CJ, et al. Laparoscopic paraesoph-
ageal hernia repair. Arch Surg 1997;132:586–589.
9. Kaplan EL, Meier P. Nonparametric estimation from incom-
plete observations. Journal Am Stat Assoc 1958;53:457–462.
10. Willekes CL, Edoga JK, Frezza EE. Laparoscopic repair of para-
esophageal hernia. Ann Surg 1997;225:31–38.
11. Casabella F, Sinanan M, Horgan S, Pellegrini CA. Systematic use
of gastric fundoplication in laparoscopic repair of paraesopha-
geal hernias. Am J Surg 1996;171:485–489.
12. Trus TL, Bax T, Richardson WS. Complications of laparoscopic
paraesophageal hernia repair. J Gastrointest Surg 1997;1:221–
228.
13. Gantert WA, Patti MG, Arcerito M, et al. Laparoscopic repair of
paraesophageal hiatal hernias. J Am Coll Surg 1998;186:428–
432.
14. Wu JS, Dunnegan DL, Soper NJ. Clinical and radiological as-
sessment of laparoscopic paraesophageal hernia repair. Surg En-
dosc 1999;13:497–502.
15. Cornwell M, Trus T, Waring JP, Hunter JG. Patterns of failure
and results of re-do fundoplication. Surg Endosc 1997;11:171.
16. Soper NJ, Dunnegan D, Anatomic fundoplication failure after
laparoscopic antireflux surgery. Ann Surg 1999;229:669–776.
17. Horgan S, Pohl D, Bogetti D, et al. Failed antireflux surgery:
what have we learned from reoperations? Arch Surg 1999;134:
809–815.
18. Williamson WA, Ellis FHJ, Streitz JMJ, Shahian DM. Para-
esophageal hiatal hernia: is an antireflux procedure necessary?
Ann Thorac Surg 1993;56:447–451.
19. Maher JW, Hocking PH, Woodward ER. The fate of the in-
trathoracic fundoplication: a review of 112 cases. In: DeMeester
TR, Skinner DB, eds. Esophageal disorders: patholophysiology
and therapy. New York: Raven; 1985.
20. Pennell TC. Supradiaphragmatic correction of esophageal re-
flux. Ann Surg 1981;173:775–781.
21. Richardson JD, Larson GM, Polk HJ. Intrathoracic fundoplica-
tion for shortened esophagus. Treacherous solution to a chal-
lenging problem. Am J Surg 1982;143:29–35.
22. Swanstrom LL, Marcus DR, Galloway GQ. Laparoscopic Collis
gastroplasty is the treatment of choice for the shortened esoph-
agus. Am J Surg 1996;171:477–481.
23. Gastal OL, Hagen JA, Peters JH, et al. Short esophagus: analysis
of predictors and clinical implications. Arch Surg 1999;134:
633–636.
 560 Hashemi et al Laparoscopic Repair of Type III Hiatal Hernia
24. Altorki NK, Yankelevitz D, Skinner DB. Massive hiatal hernias:
the anatomic basis of repair. J Thorac Cardiovasc Surg 1998;
115:828–835.
25. Carlson MA, Condon RE, Ludwig KA, Schulte WJ. Manage-
ment of intrathoracic stomach with polypropalene mesh pros-
thesis reinforced transabdominal hiatus hernia repair. J Am Coll
Surg 1998;187:227–230.
26. Frantzides CT, Richards CG, Carlson MA. Laparoscopic repair
of large hiatal hernia with polytetrafluorethylene. Surg Endosc
1999;13:906–908.
Invited Commentary
Frederick L Greene, MD, FACS
Charlotte, NC
I congratulate Dr Hashemi and his colleagues from
the University of Southern California for their pres-
entation and important contributions both today
and in the past regarding the management of esoph-
ageal disease. It was also a pleasure for me to review
the manuscript. These authors have recognized the
importance of followup in elucidating anatomic
concerns regarding the repair of combined type III
hiatus hernias, which are performed using a laparo-
scopic approach. This retrospective study has com-
pared patients undergoing laparotomy, thoracot-
omy, and laparoscopy for the management of this
illness. As others before them, these authors have
correctly assumed that symptomatic well-being
bears little or no relationship to anatomic disrup-
tion after hiatus hernia repair. The gold standard in
this particular report has been the video esopha-
gram, which was used to report anatomic recur-
rence. As you have heard, contrast studies were
completed in 75% of the patients operated on.
More specifically, 20 of 27 patients undergoing
open repair had contrast studies, and 21 of 27
(77%) of patients in the laparoscopic group were
studied radiologically. Based on these studies, a total
recurrence rate of 29% was identified. The key
point in this report is that 42% of the laparoscopic
group developed a recurrent hernia as compared
with only 15% of the group undergoing open pro-
cedures.
The authors admonish all of us to perhaps re-
think our approach to patients with type III hiatus
hernias as we consider a surgical approach. My con-
cern with this report is certainly not in the data
presented, but is based on the patients not studied.
J Am Coll Surg
In the total group, 13 patients (24%) had no fol-
lowup esophagram. This is in contrast to the fairly
complete assessment of symptoms where 94% of
the group was queried. Although appropriate
Kaplan-Meier curves may be constructed, the true
statistical evaluation may be in the confidence in-
tervals generated. The strength of the authors’ con-
clusions may, in fact, be significantly changed if all
patients were studied by contrast postoperatively.
Confidence intervals for the data presented range
from 20.4% to 61.4% recurrence rates. If one con-
structs a best-case scenario for the laparoscopically
approached patients and assumes that the six pa-
tients not undergoing contrast studies had no recur-
rence, then the confidence intervals range from
14.8% recurrence to 49.4%.
Although it may be difficult to adequately dis-
cuss a learning curve for the 27 patients undergoing
laparoscopic repair, I ask the authors to indicate
whether the patients having recurrence were early or
later in their operative experience. In addition,
when discussing patients with type III herniation,
the detrimental effect of the short esophagus has
been frequently mentioned as a potential cause for
recurrence. I would like the authors to comment on
this hypothesis and to indicate their approach if a
shortened esophagus is noted on preoperative eval-
uation. Finally, from this discussion, it is evident
that postoperative evaluation of patients with this
process should include a routine contrast study.
Would the authors comment on the timing of such
a study, which would give the most significant re-
sults in relation to recurrence.
It has been a pleasure to have the opportunity to
discuss this work. The authors present a challenge
to all of us to carefully assess our outcomes and to
report our successes and failures.
Reply
Jeffrey H Peters, MD, FACS
Los Angeles, CA
Thank you, Dr Greene, for your kind comments
and insightful questions.
The study is indeed retrospective and suffers
from all of the problems and biases associated with
a retrospective study. With that being said, however,