J Neurosurg 77:875-880, 1992

Shunt implantation: reducing the incidence of shunt

infection

MAURICE CHOUX, M.D., LORENZO GENITORI, M.D., DOROTHY LANG, M.D., F.R.C.S.,
AND GABRIEl, LENA, M.D.
Department of Pediatric' Neurosurgery, Hbpital des Enfants, La Timone, Marseille, France

~" Shunt infection remains the foremost problem of shunt implantation after mechanical malfunctions.
Diversionary cerebrospinal fluid shunt implantation has a high complication rate, with 5% to 15% of such
shunts becoming infected. Of these infections, 70% are diagnosed within 1 month after surgery and more than
90% within 6 months. Shunt infection in the vast majority of cases is therefore a complication of shunt
surgery. The authors review their experience with shunt implantation during two time periods. From January,
1978, to December, 1982, 302 children with hydrocephalus underwent 606 operations. Among these children,
47 (15.56%) developed a proven shunt infection, with an incidence of infection per procedure of 7.75%. As a
result of this study, a new protocol for shunt procedures involving modifications in the immediate pro-, intra-,
and postoperative management of children undergoing shunt implantation was initiated. With this new
protocol, 600 children underwent a total of 1197 procedures between January, 1983, and December, 1990.
The incidence of shunt infection decreased dramatically, with two infections (0.33%) in 600 patients and a
per-procedure rate of 0.17%. The overall annual risk of a shunt infection in the pediatric neurosurgical unit is
currently 1.04%.

KEY WORDS 9 hydrocephalus 9 ventriculoperitoneal shunt shunt infection 9

operative protocol

NFECTION remains a serious complication of shunt shunt infection rate in Marseille, France, revealed that

I implantation, with a patient mortality rate rang-

ing from 30% to 40%? 3.42while those who survive
risk intellectual, cognitive, and neurological deficits. 35
Infection has been reported to occur in 10% to 15% of
shunt procedures; ~-''32'36 however, some authors have
described lower infection rates ranging from 2% to
5%.I,10,11,27,39,40 Many factors have been associated with
shunt infection, including the age of the patient, 3"~3"2~'3~
the etiology of the hydrocephalus, j3 and the type of
shunt imp!anted. 7''5'2s Other authors have emphasized
the importance of the surgeon's experience? 3 Most
studies concerning the use of prophylactic antibiotic
medications are subject to criticism, 1'2"6"9'10"16-19'2L23'
27.31.41.42and there is no definitive evidence that prophy-
lactic antibiotic medications reduce shunt infection
rates. Some other factors (possibly influenced by the
surgical team), such as timing of the operation, duration
of surgery, number of operations per patient, number
of people in the operating room, and the length of time
during which the shunt material is exposed to the
atmosphere, have not been studied until now.
In November, 1982, a retrospective analysis of the
a high proportion (7.75%) of shunts became infected;
99% of these infections were diagnosed within 6 months
of shunt implantation, supporting a basic premise that
infection of a shunt is a complication of shunt surgery.
As a result of this study, there was a review involving
all personnel concerned with the pro-, intra-, and post-
operative care of patients undergoing a shunt proce-
dure. This included the neurosurgeons, neuroanes-
thesiologists, nursing staff, and microbiologists. The
baseline incidence of shunt infection and the type of
responsible organisms were known; using this infor-
mation, together with data from the literature identify-
ing risk factors for shunt infection, we devised a proto-
col for shunt implantation. This prospective study was
initiated in January, 1983. The effect of this new pro-
tocol on the incidence of shunt infection was then
audited prospectively. This study demonstrates the ef-
ficacy of this protocol.
Clinical Material and Methods
For the purposes of this study, infection rates were
compared in two consecutive series of patients with

J. Neurosurg. / Volume 77/December, 1992 875


hydrocephalus studied prior to and following the intro-
duction of a new protocol designed to reduce the inci-
dence of shunt infections. The first group of patients
was admitted to the pediatric neurosurgical unit be-
tween January, 1978, and December, 1982, and the
second group between January, 1983, and December,
1990. In each patient, hydrocephalus was diagnosed by
ultrasonography, computerized tomography, or mag-
netic resonance imaging.
Patients studied included those undergoing primary
shunt implantation and those admitted for a shunt
revision. In each case, the age and sex of the child, the
etiology of the hydrocephalus, the operative procedure,
the indication for operation in patients having a shunt
revision, and the type of shunt system implanted was
recorded on a computer printout for subsequent anal-
ysis. In the vast majority of patients (99%), a ventricu-
loperitoneal shunt was inserted, with the most fre-
quently used valve being a membrane type. During
shunt insertion, cerebrospinal fluid (CSF) was sent for
Gram staining and subsequent culture; patients found
to have infected CSF were excluded from subsequent
analysis.
Diagnosis of Shunt Infection
A postoperative shunt infection was defined as an
infection confirmed within 6 months of operation and
was diagnosed if there was inflammation along the
length of the shunt, wound discharge, or wound dehis-
cence. Signs of meningitis, ventriculitis, or nonfocal
systemic indications of infection were also investigated
for possible infection of the implanted material. Pa-
tients with abdominal symptoms or signs, including
distention, intra-abdominal cysts, poor absorption of
CSF, or abdominal pain or tenderness, were also inves-
tigated because these could be due to a shunt infection.
Peripheral blood samples and CSF were analyzed (dif-
ferential white blood cell count and Gram staining) and
cultured for aerobes and anaerobes in suspected cases
of shunt infection. Serum C-reactive protein and anti-
staphylococcal antibody titers were studied. After dis-
charge from the hospital, each patient was reviewed on
an outpatient basis and follow-up monitoring ranged
from 6 months to 8 years.
Protocol for Shunt Implantation Initiated in 1983
An outline of the protocol for shunt implantation
used in our hospital beginning in 1983 is presented in
Table 1.
Preoperative Preparation of Patients. Children ad-
mitted for either shunt insertion or shunt revision were
carefully assessed from a general medical point of view;
if an intercurrent infection was diagnosed or if there
was a localized skin problem, shunt insertion was de-
ferred until the patient's condition had improved. In
patients where urgent treatment of hydrocephalus was
required and a general medical problem contraindi-
cated shunt placement, a period of temporary external
ventricular drainage was preferred over definitive shunt
876
M. Choux, et al.
TABLE 1
Outline of revisedprotocolfor shunt implantation
preoperativeperiod
assessment of patient
localized skin problem
general medical condition
no shaving(except in older children undergoingrevision)
skin preparation (povidoneiodine or chlorhexidine)
no antibiotic medications
shunt implantation
theater schedules/timing
shunts early in the morning, before other operations
neonates & infants before older children
no more than four shunt proceduresper day
length of operation: 20 to 40 rains
theater staff
only four people in operating room (surgeon, assistant, anesthe-
siologist, circulatingnurse)
no scrub nurse
experienced neurosurgeon
shunt material
careful selection of shunt
opening sterile packagingat last moment
no testing of valve
surgical technique
two skin incisions
meticulous hemostasis
careful siting of valve/reservoir
quality of skin closure
antibiotic medications
prophylacticintravenous antibiotic 30 mins beforeskin incision
postoperativeperiod
head position-- avoid pressure on valve
no antibiotic medications
two shampoos
approximate length of stay:
4 days (first shunt)
2 days (shunt revision)
implantation until the problem had been rectified. An-
tibiotic medications were not routinely given in the
preoperative period.
All patients admitted for shunt insertion or revision
underwent a hair shampoo the evening before and the
morning of the operation using an aqueous povidone-
iodine (Betadine) preparation. Neonates and infants
were not routinely shaved regardless of whether the
operation was for implantation of a new shunt or
revision of an existing shunt. In older children admitted
for a shunt revision, a limited area of the head was
shaved by the surgeon in the anesthesia room immedi-
ately before the patient was transferred to the operating
theater. The scalp and skin were then prepared with
aqueous povidone-iodine or chlorhexidine in neonates
while the surgeon scrubbed.
Intraoperative Precautions Against Shunt Infec-
tion. Operating schedules were prepared so that shunt
implantation was carried out early in the day prior to
other neurosurgical procedures. Neonates, infants, and
young children were operated on before older children,
regardless of whether the procedure was a primary
shunt insertion or a revision. No more than four shunt
J. Neurosurg. / Volume 77/December, 1992
Reduction of infection in shunt implantation
procedures were planned per day. It was proposed that
each shunt procedure should take no more than 20 to
40 minutes.
The number of people in the operating room was
reduced to four: the surgeon and assistant, the anesthe-
siologist, and a circulating nurse. To avoid possible
contamination of the instrumentation and materials,
there was no scrub nurse; the sterile equipment was
opened at the last minute by the surgeon or assistant.
Shunt implantation was carried out by a neurosurgeon
with considerable experience with both the selection
and use of shunt materials and the operative techniques
required for shunt insertion.
The surgeon chose shunt equipment appropriate for
the age of the patient and the etiology of the hydro-
cephalus. Presterilized and packaged material was not
opened until just prior to implantation to avoid pro-
longed exposure of the shunt material and possible air
contamination. After the packaging was opened and
before insertion, the shunt was immersed in a genta-
micin bath. Formal testing of the valve for patency or
verification of the opening pressure was avoided be-
cause of the risk of introducing infection.
A sterile adhesive drape was used to secure conven-
tional drapes and expose the minimum area possible.
Two small skin incisions were made without an inter-
mediate incision. Hemostasis was secured and subcu-
taneous hemorrhage avoided. The abdominal incision
was made first to minimize the length of time the cranial
wound was open. The preassembled shunt was placed
in position and the exposed distal catheter was covered
by a sterile drape during cannulation of the lateral
ventricle and insertion of the ventricular catheter. The
valve was carefully positioned in the subgaleal space to
avoid skin damage from the shunt material. After shunt
insertion, the cranial wound was closed in a single layer
using interrupted resorbable sutures. The peritoneal
catheter was placed in the right paracolic gutter after
opening of the peritoneum under direct vision, and the
abdomen was closed in layers in the normal manner.
Throughout the operation the wounds were irrigated
with a diluted solution of aqueous povidone iodine,
and all shunt materials were rinsed with gentamicin
prior to implantation. Each patient was given cloxacillin
or oxacillin in a dose of 100 mg/kg intravenously before
the skin incisions were made, usually at the time of
anesthesia induction.
Postoperative Period. The nursing staff ensured
that the patient's head was positioned correctly to avoid
pressure on the cranial wound and shunt valve. This
was particularly important in premature babies and
neonates. Antibiotic medications were not administered
in the postoperative period. The patient's hair was
washed the morning after surgery and again before
discharge using aqueous povidone iodine. Dressings
were changed every 2 days using strictly aseptic tech-
niques. Patients were hospitalized for a mean of 4 days
for a primary shunt implantation and a mean of 2 days
for a shunt revision.
J. Neurosurg. / Volume 77/December, 1992
Results
Patients Treated Before Revision of Protocol
From January, 1978, to December, 1982, 606 oper-
ations involving shunt implantation were performed in
302 children (130 patients presented with hydrocepha-
lus secondary to myelomeningocele). During this pe-
riod, 47 children (22 boys and 25 girls) developed a
proven shunt infection. Infection was more common
in infants; 26 patients were under 2 years of age and,
of these, 17 were under 1 year. Twenty-seven patients
(57%) developed the infection within 1 month of the
operation; infection was confirmed in 40 patients (85%)
within 4 months of surgery and in only seven patients
(15%) after the 4th month. Of those with a shunt
infection, hydrocephalus was congenital in 23, was ac-
quired in 13, was obstructive because of an intracranial
tumor in 10, and was of unknown origin in one.
In total, 59 organisms were found in the 47 children,
as multiple organisms were cultured in nine. The most
common organism was Staphylococcus epidermidis
(44%), while S. aureus accounted for 23% of the infec-
tions. The other organisms grown were Enterabacter,
Enterococcus, S. haemolyticus, Acinobacter, Pseu-
domonas aeruginosa, and Candidas albicans. Over-
all, infection occurred in 47 patients (15.56%) and the
incidence of infection per procedure was 7.75%.
Patients Treated After Revision of Protocol
From January, 1983, to December, 1990, 600 pa-
tients (345 boys and 255 gifts) had a total of 1197
operations (43 children presented with hydrocephalus
secondary to myelomeningocele). Of the 600 patients,
482 underwent primary shunt implantation. During the
study period, 226 underwent only one operation and
374 underwent a total of 971 revisions. In two patients
the shunts were revised after treatment of a shunt
infection, and in two the revisions followed unexplained
pyrexia. We did not find that shunt reinsertion after
infection was associated with an increased risk of shunt
infection.
The majority (59% or 353) of the patients were more
than 1 year old; 194 patients (32%) were less than 6
months old and 53 (9%) were aged between 6 months
and 1 year. One-third (200) of the patients developed
hydrocephalus because of an intracranial tumor; the
other causes of hydrocephalus included intraventricular
hemorrhage, infection, spina bifida aperta, or an intra-
cranial malformation. In approximately 20% of the
cases the etiology of the hydrocephalus was unknown.
Overall, infection was confirmed in two patients,
resulting in an infection rate of 0.33% per patient and
0.17 % per procedure. Two additional patients had post-
operative pyrexia and, despite repeated cultures of CSF
and peripheral blood, no causative organism was grown.
Nevertheless, each had the shunt removed, a temporary
period of external ventricular drainage, and implanta-
tion of a new shunt when afebrile. One of the two
patients with a confirmed shunt infection was a 13-
year-old girl with von Recklinghausen's disease and an
877
 M. C h o u x , el a /

optic pathway tumor; this patient developed hydro- TABLE 2

cephalus and underwent shunting. As a result of a lmidence oj shunl it~,ction in previot~sl.r reported serie,~
mechanical shunt problem, she had four shunt revisions Cases Operations
7 months later and immediately postoperatively devel-
oped a fever; on analysis, E n t e r o b a c t e r was isolated Authors & Year % With % With
No. Infections No. Infections
from the CSF. After a temporary period of external
drainage and systemic antibiotic medications, a new Shurtleff, el al., 1973 102 27 299 14
shunt was inserted when the CSF was sterile and the Schoenbaum, et al., 1975 289 22 743 13.2
George, et al., 1979 410 21.7 840 13
white blood cell count was less than 10/cu mm. The Keucher & Mealey, 1979 228 22 754 7
other patient was an infant born prematurely at 31 O'Brien, et al., 1979 245 15 778 2.7
weeks who weighed 1.4 kg, with a stage III intraven- Welch. 1979 404 3.2 624 2.9
tricular hemorrhage and secondary hydrocephalus. McCullough, el al., 1980 257 8.9 435 2.6
Ajir, elal., 1981 -- -- 171 12.1
Two weeks after shunt implantation, the infant became Alvares-Garijo & Mengual, 1982 90 17.8 -- --
pyrexial and a shunt infection due to S. a u r e u s was Duret, et al., 1983 48 4.1 56 3.5
confirmed by CSF analysis and blood culture. This Fitzgerald & Connolly, 1984 43 4.6 82 2.4
baby was also treated by shunt removal, external ven- Renter, et at., 1984 802 -- 1174 7.9
tricular drainage, and systemic antibiotic medications Waiters, et al., 1984 1477 18 5179 5
Ammirati & Raimondi, 1987 431 22 1485 6
and was then reshunted 2 weeks later after normaliza- Meirovilch, et al., 1987 -- 33 -- 22
tion of the CSF; the infant made good progress there-
after.

Discussion
Surgical Technique
We have demonstrated here that, with a meticulous
surgical technique and modifications to the pre-, intra-,
and postoperative care, it is possible to reduce signifi-
cantly the incidence of shunt infection. In our recent
experience, we have seen that introduction of a rela-
tively simple protocol for shunt implantation has re-
duced the per-procedure infection rate from 7.75% to
0.17% in patients undergoing a primary shunt place-
ment or a shunt revision. The majority of shunt infec-
tions are observed within 3 months of shunt insertion
and are a result of direct contamination at the time of
surgery. We agree with Welch 4~ that shunt infection is
a potentially preventable complication of shunt surgery.
In accordance with data in other reports collected be-
tween 1978 and 1982 (Table 2) in which the infection
rate per patient ranged from 3.2% to 2 7 % , 1"2"10"13"20"23"27"40
the infection rate was 15.56% in our patients treated
within that period. The number of patients included in
each of these studies ranged from 48 to 410; thus,
infection rates per procedure ranged from 2.6% to 14%
as compared to 7.75% in our series.
After introduction of our new protocol in 1983, the
incidence of shunt infection was dramatically reduced
from 15.56% to 0.33% per patient and from 7.75% to
0.17% per procedure. Overall, these results are encour-
aging and lend support to the hypothesis that shunt
infection is potentially preventable. In addition, our
results compare favorably with infection rates per pro-
cedure reported in contemporary series, including those
of Fitzgerald and Connolly ~ (2.4%), McCullough, et
al. 23 (2.6%), Waiters, et a l ) ~ (5%), Ammirati and
Raimondi 3 (6%), and Meirovitch, et al. 25 (22%).
In agreement with other studies, more than one-half
of the infections in the group treated between 1978 and
1982 occurred in children less than 2 years old; 3'~3'~9
25.27.30therefore, we operate on neonates and infants be-
fore older children. However, contrary to the findings of
Boynton, et al., s who reported infection in 50% of
shunted premature babies, of the 73 premature babies
in our series who were shunted between 1983 and 1990,
only one (1.4%) developed a shunt infection. George,
et al., 13 showed that, after two operations, the risk of
infection increased from 10% to 26%. Meirovitch, et
al., 25 reached the same conclusion; in contrast, Renter,
et al.,29 did not find that children with multiple proce-
dures had a higher risk of infection. We operate on
children undergoing initial shunt placement before
those having shunt revision.
Haase, et al., ~5 and Raimondi, et al,, 28 found that
the use of a one-piece shunt system diminished the rate
of infection, especially in patients aged less than 2 years;
consequently, we avoid the use of connectors whenever
possible. In addition, investigation of shunt malfunc-
tion by tapping the shunt reservoir was strictly limited
in this study to patients in whom an infection was
suspected because of the inherent risk of introducing
infection. This view is in contrast to that of Noetzel
and Baker, 26 who in a combined retrospective and
prospective study showed that no early or late compli-
cations occurred in 91 patients with a total of 209
diagnostic tap procedures. Leggate, et aL,22 reported no
complications in cases in which a separate reservoir had
been inserted for diagnostic tap procedures.
In accord with the report of George, et al., ~3 who
found that the experience of the surgeon was the single
most important factor in the reduction of shunt infec-
tion rates, we believe that a shunt procedure should be
carried out only by an experienced neurosurgeon. In
contrast, Renter, et al., 29 reported that the experience
of the surgeon does not result in a statistically lower
rate of infection.
The majority (44%) of infections in our study were

878 J. Neurosurg. / Volume 7 7 / D e c e m b e r , 1992

Reduction of infection in shunt implantation

caused by S. epidermidis: in the literature this inci- TABLE 3

dence varies between 50% and 75%.~4.",2~,24~-~,39Staph- Use of proplo,lactic antibiotic medications in previously
yh)coccus aureus was implicated in 20% to 25% of reported series
reported cases, 4"5'2839 and we observed the same inci-
dence (23%). In our study a variety of organisms were No. No. % of
Authors & Year of of Medication* Infeclions
cultured; in nine (19%) of the 47 patients with shunt
Cases Ops Cases Ops
infection, more than one organism was grown from the
same culture, a slightly higher incidence of polymicro- Venes, 1976 54 150 O 18.5 6.6
bial infection than reported by Shapiro, et al." Staph- George, et al., 1979 - - 388 M 26.58 12.97
Welch, 1979 404 624 O/C 3.2 2.91
ylococci are a frequent cause of shunt infection; they McCullough, el al., 1980 257 223 M 8.91 2.62
are thought to form discrete colonies within the silicone Ajir, et aL, 1981 73 66 M 8.82 4.54
shunt tubing and have been shown to secrete a muco- Alvares-Garijo& Mengual, 90 - - M 17.77 --
polysaccharide that protects the colony and ensures 1982
Duret, et al., 1983 48 56 O/C 4.16 3.54
adherence to the shunt material? ~4 Other organisms Fitzgerald& Connolly, 43 82 O/C 4.65 2.43
isolated include Gram-negative bacilli4 and Gram-pos- 1984
itive bacilliJ 33
* Abbreviations: O = oxacillin; O/C = oxacillin or cloxacillin;
M = methicillin.
Prophylactic Antibiotics
There is no clearly defined role for prophylactic
antibiotic medications in the prevention of shunt infec-
tion (Table 3). Despite the number of investigators who rate of shunt infection is to be avoided. This has im-
have compared the prevalence of infection in patients portant implications for the obvious and hidden costs
with and without antibiotic coverage, no study with- of repeated shunt infections in our patients.
stands scientific scrutiny and most of the early trials
were reported as sequential studies. 1,21,n_3,31,34,38Accord- Conclusions
ing to Haines, 16'17 an infection rate between 0.8% and An effective protocol for the prevention of shunt
5.7% in clean neurosurgical operations is acceptable; infection involving meticulous surgical technique and
more recently, Tenney, el al., 37 and Young and Law- modifications to the pre-, intra-, and postoperative care
ner 4~ have reported infection rates of 2.6% and 3.6%, of patients undergoing a shunt procedure has been
respectively, in neurosurgical procedures performed devised. Clinical audit has proved to be a valuable tool
without antibiotic prophylaxis. In studies involving in monitoring the results of introducing of a new
shunt implantation, it has been estimated that perhaps scheme of patient management. A clear reduction in
as many as 1600 patients would require antibiotic med- the incidence of shunt infection, a known life-threat-
ications to achieve a 50% reduction in the shunt infec- ening complication of shunt implantation, has been
tion rate. 37 Haines and Taylor,~9 despite demonstrating shown.
reduction in shunt infection rates (five of 39 patients
with placebo and two of 39 treated with antibiotic References
medications) were unable to show a statistically signif- 1. Ajir F, Levin AB, Duff TA: Effect of prophylactic meth-
icant reduction. Other reports in which infection rates icillin on cerebrospinal fluid infections in children. Neu-
are disproportionately high and which have demon- rosurgery 9:6-8, 1981
strated no apparent benefit from antibiotic prophylaxis 2. Alvares-GarijoJA, Mengual MV: Infection rate with and
may be negative because of an unduly high burden. 6'32'42 without prophylactic antibiotic therapy after shunt infec-
tion. Monogr Neurol Sci 8:66-68, 1982
We chose to use prophylactic antibiotic medications
3. Ammirati M, Raimondi AJ: Cerebrospinal fluid shunt
guided by our findings in the cases studied between infections in children. A study on the relationship between
1978 and 1982. Cloxacillin and oxacillin are in our the etiology of hydrocephalus, age at the time of shunt
opinion the most appropriate antibiotic medications placement, and infection rate. Childs Nerv Syst 3:
because of their narrow spectrum, bactericidal proper- 106-109, 1987
ties, and high affinity for plasma proteins which facili- 4. Bayston R: Hydrocephalus Shunt Infections. London:
tates antibacterial activity throughout the operation. In Chapman Hall, 1987, pp 1-157
5. Bayston R, Penny SR: Excessive production of mucoid
agreement with the experimental work of Burke, 9 we substance in Staphylococcus SIIA: a possible factor in
gave the antibiotic intravenously before the operation colonisation of Holter shunts. Dev Med Child Neurol 14
was begun. Although inconclusive results have been (Suppl 27):25-28, 1972
obtained in studies where topical antibiotics are used, TM 6. Blomstedt GC: Results of trimelhoprim-sulfamethox-
we soaked the shunt material in gentamicin prior to azole prophylaxis in ventriculostomy and shunting proce-
implantation; in our view, this reduced the electrostatic dures. A double-blind randomized trial. J Neurosurg 62:
694-697, 1985
properties of the silicone tubing which favor adherence 7. Borges LF: Cerebrospinal fluid shunts interfere with host
of bacteria. defenses. Neurosurgery 10:55-60, 1982
We have demonstrated that attention to detail and 8. Boynton BR, Boynton CA, Merritt TA, etal: Ventricu-
meticulous surgical technique are important if a high loperitoneal shunts in low birth weight infants with intra-

J. Neurosurg. / Volume 77/December, 1992 879


cranial hemorrhage: neurodevelopmental outcome. Neu-
rosurgery 18:141-145, 1986
9. Burke JF: The effective period of preventive antibiotic
action in experimental incisions and dermal lesions. Sur-
gery 50:161-168, 1961
10. Duret N, Derex MC, Chacornac R: Essai de prevention
du risque infectieux au cours des traitements de l'hy-
drocephalie de l'enfant par drrivation du liquide c~ph-
alo-rachidien. Agressologie 24:239-242, 1983
11. Fitzgerald R, Connolly B: An operative technique to
reduce valve colonisation. Z Kinderehir 39 (Suppl 2):
107-108, 1984
12. Gardner P, Leipzig T, Phillips P: Infections of central ner-
vous system shunts. Med Clin North Am 69:297-314,
1985
13. George R, Leibrock L, Epstein M: Long-term analysis of
cerebrospinal fluid shunt infections. A 25-year experience.
d Neurosurl~51:804-811, 1979
14. Guevara JA, Zficcaro G, Trevis~n A, et al: Bacterial
adhesion to cerebrospinal fluid shunts. J Neurosurg 67:
438-445, 1987
15. Haase J, Bang F, Tango M: Danish experience with the
one-piece shunt. A long-term follow-up. Chiids Nerv Syst
3:93-96, 1987
16. Haines SJ: Antibiotic prophylaxis in neurosurgery. Clin
Nenrosurg33:633-642, 1986
17. Haines SJ: Systemic antibiotic prophylaxis in neurological
surgery. Neurosurgery 6:355-361, 1980
18. Haines SJ: Topical antibiotic prophylaxis in neurosurgery.
Neurosurgery 11:250-253, 1982
19. Haines SJ, Taylor F: Prophylactic methicillin for shunt
operations: effects on incidence of shunt malfunction and
infection. Childs Brain 9:10-22, 1982
20. Keucher TR, Mealey J Jr: Long-term results after ven-
triculoatrial and ventriculoperitoneal shunting for in-
fantile hydrocephalus. J Neurosurg50:179-186, 1979
21. Klein DM: Comparison of antibiotic methods in the
prophylaxis of operative shunt infections, in Humphreys
RP (ed): Concepts in Pediatric Neurosurgery 4. Basel: S
Karger, 1983, pp 131-141
22. Leggate JR, Baxter P, Minns RA, et al: Role of separate
subcutaneous cerebro-spinal fluid reservoir in the man-
agement of hydrocephalus. Br J Neurosurg 2:327-337,
1988
23. McCullough DC, Kane JG, Prosper JH, et al: Antibiotic
prophylaxis in ventricular shunt surgery. Part I. Reduc-
tion of operative infection rates with methicillin. Childs
Brain 7:182-189, 1980
24. McLaurin RL, Frame PT: Treatment of infections of ce-
rebrospinal fluid shunts. Rev Infect Dis 9:595-603, 1987
25. Meirovitch J, Kitai-Cohen Y, Keren G, et al: Cerebrospi-
nal fluid shunt infections in children. Pediatr Infect Dis
J 6:921-924, 1987
26. Noetzel MJ, Baker RP: Shunt fluid examination: risks
and benefits in the evaluation of shunt malfunction and
infection. J Neurnsurg 61:328-332, 1984
27. O'Brien M, Parent A, Davis B: Management of ventric-
880
M. Choux, et al.
ular shunt infections. Childs Brain 5:304-309, 1979
28. Raimondi AJ, Robinson JS, Kuwamura K: Complica-
tions of ventriculo-peritoneal shunting and a critical com-
parison of the three-piece and one-piece systems. Childs
Brain 3:321-342, 1977
29. Renier D, Lacombe J, Pierre-Kahn A, et al: Factors
causing acute shunt infection. Computer analysis of 1174
operations. J Neurnsurg 61:1072-1078, 1984
30. Salmon JH: Adult hydrocephalus. Evaluation of shunt
therapy in 80 patients. J Neurosurg 37:423-428, 1972
31. Schmidt K, Gjerris F, Osgaard O, et al: Antibiotic pro-
phylaxis in cerebrospinal fluid shunting: a prospective
randomized trial in 152 hydrocephalic patients. Neuro-
surgery 17:1-5, 1985
32. Schoenbaum SC, Gardner P, Shillito J: Infections of
cerebrospinal fluid shunts: epidemiology, clinical mani-
festations, and therapy. J Infect Dis 131:543-552, 1975
33. Shapiro S, Boaz J, Kleiman M, et al: Origin of organisms
infecting ventricular shunts. Neurosurgery 22:868-872,
1988
34. Shimoji T, Sato K, Ishii S: Measures taken to control
shunt infections. Childs Brain 9:370-371, 1982 (Abstract)
35. Shurtleff DB, Foltz EL, Loeser JD: Hydrocephalus. A
definition of its progression and relationship to intellec-
tual function, diagnosis, and complications. Am J Dis
Child 125:688-693, 1973
36. Shurtleff DB, Stuntz JT, Hayden PW: Experience with
1201 cerebrospinal fluid shunt procedures. Pediatr Neu-
rosci 12:49-57, 1985/86
37. Tenney JH, Vlahov D, Salcman M, et al: Wide variation
in risk of wound infection following clean neurosurgery.
Implication for perioperative antibiotic prophylaxis. J
Neurosurg62:243-247, 1985
38. Venes JL: Control of shunt infection. Report of 150
consecutive cases. J Neurosurg 45:31 I-314, 1976
39. Waiters BC, Hoffman HJ, Hendrick EB, et al: Cerebro-
spinal fluid shunt infection. Influences on initial manage-
ment and subsequent outcome. J Neurosurg 60:
1014-1021, 1984
40. Welch K: Residual shunt infection in a program aimed
at its prevention. Z Kinderchir 28:374-377, 1979
41. Young RF, Lawner PM: Perioperative antibiotic prophy-
laxis for prevention of postoperative neurosurgical infec-
tions. A randomized clinical trial. J Neurosurg 66:
701-705, 1987
42. Younger JJ, Simmons JC, Barnett FF: Failure of single-
dose intraventricular vancomycin for cerebrospinal fluid
shunt surgery prophylaxis. Pediatr Infect Dis J 6:
212-213, 1987
Manuscript received January 28, 1992.
Accepted in final form May 1, 1992.
Address reprint requests to: Maurice Choux, M.D., Service
de Neurochirurgie Prdiatrique, Hrpital des Enfants de la
Timone, Marseille 13005, France.
J. Neurosurg. / Volume 77/December, 1992