Acta Neur0chirurgica 59, 157--166 (1981)

ACTA NEUROCHIRURGICA
9 by Springer-Verlag 1981

Division of Neurosurgery * and Department of Pediatrics ****, University of California at San Diego, California, and Division of Neurosurgery ** and Department of Pediatrics ***, University of Kentucky Medical Center, Lexington, Kentucky

The Management of Cerebrospinal Fluid Shunt Infections A-Cflnical Experience

By

H. E. J a m e s * ,

J. W. W a l s h * * , H. D. W i l s o n * * * ,
and J . D. C o n n o r * * * *

Summary
Fifty patients with infected cerebrospinal fluid shunts were treated by one of three forms of treatment: a) Twenty-two patients had shunt removal, systemic antibiotic treatment, and either external ventricular drainage or intermittent ventricular taps for decompression and antibiotic administration. b) Seventeen patients had removal and immediate replacement of the shunt with intrashunt and systemic antibiotics. c) Eleven patients received intrashunt and systemic antibiotics without shunt removal. In the first group, antibiotics were given for a period of one week; in the second and third groups, intravenous antibiotics were administered for a minimum period of three weeks, and intraventricular antibiotics twice daily for two weeks. In all patients ventricular CSF was obtained and cultured 48 hours after cessation of antibiotic therapy, and cultures were repeated within four months after completion of therapy. Twenty-one of 22 patients in the first group, as well as 11 of 13 of the second group, were successfully treated. In the third group only four of the 11 patients responded to treatment.

Keywords: CSF shunt; infections; therapy.

Introduction
T h e r e are three c u r r e n t l y - e m p l o y e d m e t h o d s of t r e a t m e n t for the m a n a g e m e n t of c e r e b r o s p i n a l f l u i d (CSF) s h u n t infections. A l l e m p l o y a c o m b i n a t i o n of i n t r a v e n o u s a n d i n t r a v e n t r i c u l a r antibiotics because of the u n p r e d i c t a b l e c o n c e n t r a t i o n s of some antibiotics in v e n t r i c u l a r f l u i d a2, aa. I n one m e t h o d the s h u n t assembly is n o t 11 Acta Neurochirurgica,Vol. 59, Fasc.3--4

158

H . E . James et aL:

removed, and the antibiotic is administered through the shunt two weeks along with systemic antibiotic treatment for weeks 14. ~8, a:,~ A Second method consists of removal of the shunt and immediate insertion of a new shunt, followed by tv of intrashunt and three weeks of intravenous antibiotic tl 18. 19 Some authors have reported that the shunt assembi ~ be removed, an external ventricular device or a reservoir in~ and systemic and intraventricula~ antibiotics administered i period of time. When the infection has been eliminated, a new s. may then be inserted ~, ~ 26.27 In a previous report the authors presented their experience in . prospective randomized study of 30 children with infected CSI shunts and the previously-described three methods 1%This prospectiv, study could not be continued as it was felt by the authors of the study that the high incidence of failures in one of the treatmen regimes made the continuation of its application unjustified. A, cordingly, patients who subsequently presented with infected shul~ were no longer submitted to the randomization process. Here x report our experience in the management of 50 patients with shu. infections.
M a t e r i a l s and M e t h o d s
The patients were treated in the period from September 1975 to September 1980. Infection was documented by identification of bacterial organisms on culture of CSF obtained from the shunt. Thirty-nine patients also had simultaneous cultures of blood and urine as part of their preliminary evaluation. Thirty of the 50 patients formed part of the previously-reported study, and the metho of selection of the therapeutic modality in each patient is described elsewhere 1, In the remaining 20 patients the parents or guardians were told the objectives advantages, disadvantages, risks, and complications of all three modes os t h e r a and informed consent for treatment was obtained in all cases. After appropriate cultures were obtained, intravenous antibiotic theram was started pending identification and antibiotic susceptibility of the orga~:,.sro When these results were available, the most appropriate antibiotic was instit and definitive therapy was begun. In 22 patients the shunt was removed, external ventricular drainage (EVD) was instituted, with twice daily antibk doses administered through the EVD (seven patients). In another 15 pati no EVD was placed, but intermittent once daily ventricular punctures thr the anterior fontanelle were performed for decompression and antibiotic admini: tion. In each case a total of seven days of intraventricular and intra~ antibiotic therapy was administered. Following one or two days of intrav antibiotic therapy 17 patients underwent shunt removal, with insertion of a shunt at the same procedure. The remaining 11 patients did not have . shunts removed. In both of these groups antibiotics were given twice & through the shunt reservoir for a minimum period of two weeks, and by t. intravenous route for three weeks. Dosages for intraventricular and intravenous antibiotics are listed in Table 1. The concentration of antibiotics in ventricular

Therapy for Shunt Infections

159

;,d was measured at frequent intervals, and the dosage of medication was ,~sted to achieve optimal therapeutic concentrations. The antibiotic deterrninaii~were made by the method of Simon et al. 29, or of Braude et al. 2. ":~vy-eight hours following completion of therapy all patients had ventricular .-r~rations for cultures, as well as blood and urine cultures. A second CSF .,~s obtained from all patients within four months after the termination oy. An infection was said to be cured when both these CSF cultures ,~ative. The follow-up time ranged from one month to 58 months, : :~"meanof 15.3 __ 12 months. ~'~" [~Intraventricular , Cephalothin " Methicillin Nafcillin d~mpicillin ~entamycin .:~efazolin 9 '?enicillin
5

Table 1. Antibiotics and Dosage Intravenous Cephalothin Methicillin Oxacillin Ampicillin Carbenicillin Chloramphenicol Gentamycin Kanamycin Nafcillin Tobramycin Amikacin Rifampin Penicillin 150 mg/kg/day 200 mg/kg/day 200 mg/kg/day 100-200 mg/kg/day 400-600 mg/kg/day 100 mg/kg/day 5-7.5 mg/kg/day 15 mg/kg/day 150 mg/kg/day 5 mg/kg/day 15 mg/kg/day 15 mg/kg/day 250,000 units/kg/day

1-3.5 mg/kg dose 1-2 mg/kg dose 1-4 mg/kg dose 1-2 mg/kg dose 0.5-2.0 mg/kg dose 1-2 mg/kg dose 1,000 units dose

~>

Results

:' There were 50 patients in the study, with ages ranging f r o m ~" 2 m o n t h s to 9 years; there were 29 males and 21 females. Eight p a uents h a d spina bifida, which had been repaired at birth. Table 2 ~ s u m m a r i z e s the aetiology of the hydrocephalus. There were 39 2 v ntriculo-peritoneal shunts; eight patients had more t h a n one intra:!{:anial catheter. The t y p e a n d location of the shunt system is i p r e s e n t e d in Table 3. F o r t y - f i v e patients h a d single organism infections, and as in other ::~mdies s, s la. x7, 24, 25 the m a j o r i t y were due to staphylococcus ~q~able 4). Five patients h a d more t h a n one infecting organism : ~'able 5). ::> : T w e n t y - o n e of the 22 patients w h o h a d their shunts r e m o v e d and ~'~::ad either E V D or intermittent ventricular punctures were cured. s i n the g r o u p w i t h shunt r e m o v a l and immediate replacement there ~i were 15 cures a n d t w o failures. Finally, in the g r o u p in which the shunt was not r e m o v e d there were 4 successful results f r o m the
11"

160

H. E. James et al.:

Table 2. Aetiology of Hydrocephalus

Communicating
a) Secondary to intracranial haemorrhage, post-meningitis, or trauma b) Idiopathic 13 5 14

AqueductaI stenosis Arachnoid or ependymal cyst Dandy-Walker malformation Chiari malformation (spina bifida without aqueductal stenosis) Obstructive due to neoplasm Obstructive, postinfectious
Total

3 4
6

4 1
5O

Table 3. Types of Shunt Systems Shunt location : Ventriculo-peritoneal 39 Ventriculo-jugular 2 Two intracranial catheters--peritoneal shunt 8 Lurnboperitoneal shunt 1 Total 50

Table 4. Bacteriology of Infected Shunts No. of patients Single organism infection: Multiple organism infection: 45 5

Organisms identified in single bacterial type infections

Staphylococcus epidermidis: Staphylococcus aureus : Micrococcus : Pseudomonas aeruginosa: Escherichia coli: Haemophilus influenzae: Corynebacterium species: Streptococcus : Propionebacterium species : 31 5 1
1

1 2 I 2
1

Therapy for Shunt Infections

161

11 patients treated (Table 6). Three deaths occurred, two of these in the group without shunt removal; one patient developed a more virulent and antibiotic-resistent superinfection (E. coli and Pseudomonas), and another patient finally underwent a very prolonged EVD therapy course, dying when the EVD accidently became dis-

Table 5. Multiple Organism Shunt Infection Patient number Organism #7 #10 #11 #17 #23 Streptococcus p., Pseudomonas aeruginosa, Staphylococcus aureus, unidentified gram negative rod Staphylococcus epidermidis, Corynebacterium species Staphylococcus aureus, Corynebacterium species E. coli, Klebsiella p., Staphylococcus aureus E. coli, Proteus vulgaris

Table 6. Therapy and Results in CSF Infected Shunts Number of patients With removal of shunt * With removal and immediate replacement Without removal of shunt 22 17 11 Successful trials 21 15 4 Unsuccessful trials 1 2 7

Mortality 1 0 2

* Includes one patient who received intravenous antibiotics only (lumboperitoneal shunt).

connected. The third patient who died belonged to the first treatment group; he consistently grew S. epidermidis during his intermittent ventricular punctures during therapy after the shunt was removed. He was a premature child with intracranial haemorrhage and an intracerebral infected haematoma that failed to become sterile with our treatment. The shunt assembly in eight patients contained more than one intracranial tube connected to a single peritoneal catheter. Two of these patients were in the EVD group, three were in the group with immediate shunt replacement, and three were in the group without shunt removal.

162

H.E. James et ai.:

In the group that had their shunts removed without immediate replacement, shunt obstruction was noted in one patient on admission. Shunt obstruction also occurred in two patients in the group with immediate shunt replacement.
Discussion

Infections continue to be one of the most frequent and serious complications of CSF shunts. The incidence has been reported to be as high as 27 26, and 31 per cent 20 in two large series. It was initially described in ventriculo-vascular shunts 4 .*5,but subsequently many reports have discussed and extended the disease to ventriculoperitoneal and other shunts 5, 14, 15, 18, 2a 26, 2s The approach to the management of infected CSF shunts has varied. Treatment with high dose intravenous antibiotics aloneSS, intravenous antibiotics with intrashunt antibiotics without shunt removal 14, is, as, and removal and replacement of the shunt in the same operation followed by intrashunt and intravenous antibiotics 16, 19 have been advocated. Others have stated that a successful outcome can only be seen when the infected assembly is removed a-5, is, 15, s6, s7 The primary concept in present-day shunt therapy is that the treatment of ventriculities generally requires that the antibiotic be administered into a ventricle, since the ventricular penetration of many systematically administered antibiotics in infections of this nature is erratic. Intravenous therapy alone for shunt infection will seldom eradicate the disease sv. Thus the antibiotic is administered by ventricular puncture in the newborn or infant 1, 11. 21 30, or by EVD 15. s3. 24, aa, or by implanted CSF reservoirs 11, 27, or directly into the shunt system is 1.9 as, 83 in older children. Reasons for not removing the shunt are concern for shunt dependency and, in older children with functionally closed sutures, sudden deterioration due to rapid intracranial hypertension. This may occur in those situations in which the hydrocephalus is felt to have "arrested", and prompt deterioration has occurred after shunt dysfunction. Also, the concern for the preservation of the venous channel in the CSF-vascular shunts has led some authors to replace the shunt immediately upon removing the infected one in the same location, thus not sacrificing permanently the pathway for the shunt 26, 19 Increased intracranial pressure should not occur when the shunt is removed, if EVD is instituted. This also has the advantage of offering a route for the administration of the intraventricular anti-

Therapy for Shunt Infections

163

biotics. The hazards of disconnecting the system by the uncooperative child and sudden loss of CSF, as well as superinfection, are potential complications. Fluid, electrolyte, and protein replacement may be needed for a child on long-term EVD due to CSF loss. These risks are minimized in the hands of trained nursing personnel 6, 9, 15 28, 81. An alternative is the placement of reservoirs that would have skin covering yet are accessible for CSF decompression and antibiotic administration. However, at times this is not sufficient for decompression in shuntdependent children. There may also be difficulty in eradicating the infection because of the presence of the reservoir system. There is also the risk of nosocomial infections 11 Arguments in favour of shunt removal cite the principle of colonization within a foreign body (shunt) and evidence that infection will not clear unless and until the foreign body is removed. There is little doubt that treatment when the shunt is removed is often more prompt in eradicating the infection, as demonstrated in a previous randomized prospective study 10. In itself this is important because it has been noted by others that the lack of an early and effective resolution of a shunt infection is associated with a higher mortality rate 26, ~s, a finding of significance in older and debilitated patients 6, 2~. In the present study two of the eleven patients who did not have removal of their infecteds shunt died, and they constituted two of the three deaths in this series. One child died from uncontrollable systemic infection despite subsequent shunt removal and institution of EVD; the other child died after shunt removal and a prolonged EVD course, with complications resulting from disconnection of the system. There is significant variability in the antibiotic concentration in the intraventricular CSF, not only from patient to patient but also in the same patient during therapy ~. Because of this we recommend sampling antibiotic concentration in the ventricular fluid during therapy, thus adjusting the dose above the organism's minimal inhibitory concentration (MIC). Unfortunately, even with adequate concentrations of antibiotics in the CSF, there is no guarantee of resolution of the infection. Thus, in our previous report 10 we noted very high antibiotic concentrations (8.8 to 128 pg/ml), well above the bacterial MIC, in four of five staphylococcus shunt infections treated without shunt removal, without resolution of the infection. Conversely, one patient treated with the same technique had low antibiotic CSF concentrations (0.375-1.0 pg/ml), and in this particular case the staphylococcus infection was cured 1% These discrepancies may be related to the virulence of the organism, the ability of

164

H.E. James et el.:

the antibiotic to reach in adequate levels the shunt system in its entirety as well as the CSF pathways; many of these parameters are not easily assessed in each instance. Another argument for measuring intraventricular CSF antibiotic concentrations during therapy is the desire not to reach too high a level that may then produce a toxic effect on the central nervous system 7. This becomes particularly important when high levels of antibiotics are employed. We found that in most circumstances in infancy and childhood a 1-2 mg/kg dose into the ventricle through the shunt, EVD, or direct ventricular puncture, gave adequate CSF levels (Table 1). More recently, McLaurin 14 and Wald and McLaurin a~ have recommended higher doses, especially when therapy is being performed without shunt removal.
Acknowledgments

We extend our appreciation to Kathie L. Shepherd for her very patient secretarial work.
References

1. Blevins, J., Er~cksson, C. D., Ruiz-Palacios, G., Intraventricular and systemic gentamicin therapy for ventriculitis in children. Proc. of the Interscience Conf. on Antimicrobial Agents and Chemotherapy; Chicago, Illinois, 1976. 2. Braude, A. I., Bannister, J., Wright, N., Use of the gradient plate for routine clinical determinations of bacterial sensitivities to antibiotics. Antibiotics Annals (1954-1955), 1133--1140. 3. Bruce, A. M., Lorber, J., Shedden, W. I. H., Zachary, R. B., Persistent bacteraemia following ventriculo-caval shunt. Operations for hydrocephalus in infants. Dev. Med. Child. Neurol. 5 (1963), 461--470. 4. Cohen, S. J., Callaghan, R. P., A syndrome due to the bacterial colonization of Spitz-Holter valves. Brit. Med. J. 2 (1961), 677--680. 5. Drapkin, A. M., Michel, J., Sachs, T., Beller, A. J., Postoperative ventriculitis in infants. Acta neurochir. (Wien) 32 (1975), 89--100. 6. Everett, E. D., Eickhoff, T. C., Simon, R. S., Cerebrospinai fluid shunt infections with anaerobic diphtheroids (Propione-bacterium species). J. Neurosurg. 44 (1976), 580--585. 7. Fossieck, F., Jr., Parker, R. H., Neurotoxicity during intravenous infusion of penicillin. A review. J. Clin. Pharmacol. I4 (1974), 504--512. 8. Holt, R. J., Bacteriological studies on colonized ventriculo-atrial shunts. Dev. Med. Child. Neurol. 22 (Suppl.) (1970), 83--87. 9. James, H. E., Langfitt, T. W., Kumar, V. S., Gohstein, S., Treatment of intracranial hypertension. Acta neurochir. (Wien) 36 (1977), 189--200. 10. James, H. E., Welsh, J. W., Wilson, H. D., Connor, J. D., Bean, J. R., Tibbs, P. A., A prospective randomized study of therapy in cerebrospinal fluid shunt infection. Neurosurg. 7 (1980), 459--463. 11. Lee, E. L., Robinson, M. j'., Thong, M. L., Puthucheary, S. 1)., Ong, T. H., Ng, K. K., Intraventricular chemotherapy in neonatal meningitis. J. Peds. 91 (1977), 991--995.

Therapy for Shunt Infections

165

12. Lorber, J., Kalhan, S. C., Magrefte, B., Treatment of ventriculitis with gentamicin and cloxacillin in infants with spina bifida. Arch. Dis. Child. 45 (1970), 178--185. 13. Lukhardt, T., Bacterial infections in ventriculo-auricular shunt systems. Dev. Med. Child. Neurol. 22 (Suppl.) (1967), 105--109. 14. McLaurin, R. L., Infected cerebrospinaI fluid shunts. Surg. Neurol. 1 (1973), 191--195. 15. Mori, K., Raimondi, A. J., An analysis of external ventricular drainage as a treatment for infected shunts. Child's Brain 1 (1975), 243--250. 16. Nicholas, J. L., Kamal, I. M., Eckstein, H. B., Immediate shunt replacement in the treatment of bacterial colonization of Holter valves. Dev. Med. Child. Neurol. 12 (Suppl.) (1970), 110--113. 17. Nulsen, F. E., Becket, D. P., Control of hydrocephalus by valve-regulated shunt. Infections and their prevention. Clin. Neurosurg. 14 (1966), 256--273. 18. O'Brien, M., Parent, T., Davis, B., Management of ventricular shunt infections. Child's Brain 5 (1979), 304--309. 19. Perrin, J. E., McLaurin, F. L., Infected ventriculo-atrial shunts: A method of treatment. J. Neurosurg. 27 (1967), 21--26. 20. Robinson, J. S., Raimondi, A. J., Complications of ventriculo-peritoneal shunting procedures for congenital and secondary hydrocephalus in childhood. Proc. Am. Ass. Neurol. Surg., St. Louis, Missouri, 1974. 21. Salmon, J. H., Ventriculitis complicating meningitis. Am. J. Dis. Child. 40 (1972), 124--135. 22. Salmon, J. H., Adult hydrocephalus: Evaluation of shunt therapy in 80 patients. J. Neurosurg. 37 (1972), 423--428. 23. Sayers, M. P., Shunt complications. Clin. Neurosurg. 23 (1976), 393--400. 24. Scarff, T. B., Nelson, P. B., Reigel, D. H., External drainage for ventricular infection following cerebrospinal fluid shunts. Child's Brain 4 (1978), 129--136. 25. Schimke, R. T., Black, P. H., Mark, V. H., Swartz, M. N., Indolent S t a p h y l o coccus albus or aureus bacterernia after ventriculoatriostomy: Role of foreign body in its initiation and preparation. N. Engl. J. Med. 264 (1961), 264--270. 26. Schoenbaum, S. D., Gardner, P., Shillito, J., Infections of cerebrospinaI fluid shunts: Epidemiology, clinical manifestations and therapy. J. Infect. Dis. 131 (1975), 543--552. 27. Shurtleff, D. B., Christie, D., Foltz, E. L., Ventriculoauriculostomy associated infections: A 12-year study. J. Neurosurg. 35 (1971), 686--694. 28. Shurtleff, D. B., Foltz, E. L., Weeks, R. D., Loeser, J., Therapy for Staphylococcus epidermidis: Infections associated with cerebrospinal fluid shunts. Peds. 53 (1974), 55--62. 29. Simon, H. J., Yin, E. J., Microbioassay of antimicrobial agents. Appl. Microbiol. 19 (1973), 573--579. 30. Simpson, P. B., Warren, G. C., Smith, R. R., Intraventricular cephalothin in childhood ventriculitis. Surg. Neurol. 4 (1975), 279--282. 31. Smith, R. W., Alksne, J. F., Infections complicating the use of external ventriculostomy. J. Neurosurg. 44 (1976), 567--570. 32. Wald, S. L., McLaurin, R. L., Cerebrospinal fluid antibiotic levels during treatment of shunt infections. J. Neurosurg. 52 (1980), 41--46.

166

H.E. James et al.: Therapy for Shunt Infections

33. Wilson, H. D., Bean, J. R., James, H. E., Pendley, M. M., CerebrospinaI fluid antibiotic concentrations in ventricular shunt infections. Child's Brain 4 (1978), 74--82. Authors' addresses: H. E. James, M.D., and J. D. Connor, M.D., University of California Medical Center, 225 Dickinson Street, San Diego, CA 92103, U.S.A., J. W. Walsh, M.D., Ph.D., and H. D. Wilson, M.D., University of Kentucky Medical Center, Lexington, KY 40503, U.S.A.