Key-words:
Hydrocephalus - infection - malfunction - pediatric - shunt
Introduction
Hydrocephalus (HCP) is a common neurosurgical disorder with an estimated prevalence
and incidence of 0.9–1.2/1000 and 0.2–0.6/1000, respectively, in developed world.[[1]] Although a reliable estimate is not available, the incidence is likely to be higher
in developing countries.[[2]]
Cerebrospinal fluid (CSF) shunt surgeries have dramatically reduced morbidity and
mortality of children with HCP. However, these procedures are associated with potential
complications such as infection and malfunction. In pediatric surgical series, shunt
failures occur in 14% of patients just within the 1st month after shunt placement[[3]] and 40%–50% of the shunts will fail within the 1st year.[[4]],[[5]] Shunt revision surgeries account for 48% of all shunt-related procedures performed
in the United States.[[6]]
Lately, endoscopic third ventriculostomy (ETV) has emerged as an alternative to ventriculoperitoneal
(VP) shunt for HCP. Absence of hardware makes ETV a viable alternative to shunting
to reduce the complications of VP shunt. However, effectiveness of ETV in all etiologies
and in all age groups has yet to be proven. Pan et al. found that the early risk of
failure was significantly higher in ETV compared to shunt. The relative risk of ETV
failure compared to shunt at 3 months of follow-up was 40% higher than shunt (P =
0.011).[[7]] Further, due to availability of expertise, technology, infrastructure, and cost-related
issues, VP shunt is still the most commonly performed and only available option to
treat HCP in low-resource setting like Nepal.
Early identification and management of risk factors may help in improving patient
outcome. Here, we report our 4-year experience of management of pediatric HCP with
VP shunts, their complications, and analysis of risk factors for shunt failure.
Materials and Methods
Study population
A retrospective analytical study of prospectively collected data was designed and
conducted in the Department of Neurosurgery, National Academy of Medical Sciences
(NAMS), Bir Hospital, Kathmandu, Nepal. All pediatric patients with HCP aged below
or equal to 15 years who underwent first VP shunting and consented for participation
in the study were included. Patients with prior VP shunt or those who had other diversion
procedures (ventriculoatrial shunts, ventriculopleural shunts, or external ventricular
drainage) than VP shunt were excluded.
Surgical procedure
VP shunt surgery was performed under general anesthesia (with endotracheal intubation)
in an operation theater dedicated only for neurosurgical procedures. All elective
VP shunt procedures were done as the first case of the day. The operative area was
first cleaned with savlon solution, and then, the skin was meticulously prepped with
povidone-iodine solution from scalp, neck, and upper chest down to the abdomen. After
this, draping was done exposing the operative field. The number of scrubbing personnel
in the operating room was limited to an operative surgeon, an assistant, a scrub nurse,
a circulating nurse, and an anesthesiologist. Normal Chhabra shunt (nonantimicrobial
impregnated) was used in all cases, which was assembled, immersed, and rinsed with
80 mg of gentamycin diluted in 100 ml of normal saline. Nontouch technique was used
to avoid contact between the shunt and patient's skin, as well as the surgeon's gloves,
using instruments, and clean gauze to handle the shunt components. Intravenous antibiotics
were given before induction and continued for 1 week postoperatively.
Data collection and management
Ethical approval was taken from the Institutional Review Board of NAMS. Demographic
details, clinical findings, and preoperative radiological imaging were obtained in
all cases from medical records. Surgical variables were documented on a data sheet
designed for the study. Computed tomography scan of head, clinical status, and CSF
analysis reports were documented before discharging the patient. All the patients
were evaluated for complications at the time of admission, at discharge, at routine
outpatient department follow-up, and at emergency department visits for shunt complications.
All cases were consecutively identified between April 15, 2014, and May 6, 2017, and
were followed up prospectively until the study closure on May 6, 2017. Any complication
(either shunt infection and/or shunt malfunction) developed during the study period
was documented by the investigation team members or neurosurgical residents.
Primary study outcome
The primary outcome was shunt infection or malfunction. As per previous literature,[[8]] shunt infection was defined as the identification of a bacterial pathogen from
the reservoir CSF or reservoir CSF pleocytosis (more than 50 leukocytes/cumm) in association
with positive blood culture when a patient presents with any of the following after
surgery: (1) fever – body temperature >100.1°F, (2) neurologic symptoms – headache
or vomiting or Glasgow Coma Scale ≤14, (3) abdominal symptoms – pain abdomen or abdominal
distension or abdominal tenderness, and (4) shunt malfunction. Any obstruction, overdrainage,
or underdrainage of CSF and/or malposition and migration of shunt catheter after surgery
were defined as malfunction.
Demographic variables
Age ≤15 years (dichotomized as infants versus other than infants) and sex.
Clinical variables
Etiology of HCP (categorized into four groups: congenital aqueductal stenosis, tumor,
infection, and inflammatory and others. Others include benign cysts, Dandy–Walker
malformation, neural tube defects, and posttraumatic HCP) and duration of illness
(dichotomized to less than or more than 1 month).
Surgical variables
Surgeon's experience was defined as consultant versus resident on training, duration
of surgery (dichotomized as less than or more than 1 h from skin incision to closure),
and case priority (dichotomized as elective or an emergency surgery).
Statistical analyses
Dichotomous and categorical variables were reported using frequencies and percentages,
while mean was reported in continuous variables. Stratification with respect to age,
gender, duration of illness, etiology, duration of surgery, priority of case, and
the surgeon's experience was done. Association of independent variables with the primary
outcome variable (shunt complication) was analyzed using Chi-square test. Statistical
significance was determined at P < 0.05. Bivariate logistic regression was performed
to identify unadjusted odds ratio (OR) with 95% confidence interval (CI). Then, multivariate
logistic regression model was designed for those variables with P < 0.1 at bivariate
level to calculate adjusted OR with 95% CI. Analysis was performed in SPSS 17 (IBM,
Chicago, IL, USA).
Results
Demographic and clinical variables
Among 133 patients with first VP shunting followed up in the study period, 13 were
lost to follow-up, so 120 patients were included in the analysis. Mean age was 62.97
months, with infants being the most common age group (30.8%). Males were more common
than females. Congenital aqueductal stenosis (40.8%) was the most common cause of
HCP followed by tumors and inflammatory causes. Majority of the procedures were performed
by consultants (71.7%), duration of surgery was <1 h in 45%, and 61.7% of cases were
done in elective basis. The mean length of follow-up was 16 months. Overall shunt
complication was seen in 26.7% (95% CI 19.0%–35.5%). Five percent had shunt infection
while shunt malfunction without infection was found in 21.7% [[Table 1]].
Table 1: Distribution of complications according to patient characteristics
Distribution of complications according to demographic and clinical characteristics
More complications were seen among male (29.9%), among children other than infants
(31.3%), and among tumor etiology (34.3%). Similarly, failure rates were higher in
cases done by resident doctors (41.2%), in cases who had duration of surgery more
than 1 h (34.8%), and among emergency procedures (43.5%) [[Table 1]].
Risk factors analysis for ventriculoperitoneal complications
On bivariate logistic regression analysis, duration of surgery (P = 0.028), priority
of case (P = 0.001), and surgeon experience (P = 0.026) were significantly correlated
with outcome. On multivariate analysis, urgency of the case was only significantly
associated with higher complications (P = 0.025) [[Table 2]].
Table 2: Association of complication with demographic and clinical variables
Discussion
This study showed that one-fourth of the pediatric patients (26.7%) developed complications
after VP shunting within 30 months of surgery. Shunt malfunction was more common than
infection (21.7% vs. 5%). Complications were found to be more frequent among male,
older children other than infants, tumor etiology, operation time more than 1 h, cases
operated by resident doctors, and emergency procedures. Of them, surgeon experience,
duration of surgery, and priority of case were significantly associated with failure.
Shunt failure occurred as early as 1 week up to 30 months after surgery.
The overall shunt complication rate in our study (26.7%) was comparable or less than
reported from the literature including that from developed world. Previous reported
pediatric shunt failure rate was 30%–40% at 1 year and approximately 50% at 2 years.[[9]],[[10]],[[11]] Shannon et al. found that almost half of the patients (49%) experienced one or
more shunt failures within 2 years of follow-up.[[12]] A multicentered prospective cohort study analyzing the risk factors for shunt malfunction
in pediatric HCP reported 33.2% failure rate, 23% of them were due to infection.[[10]] Another study from Pakistan reported shunt failure rate of 23% in mean follow-up
of 11 months.[[13]] In our study, shunt complication was 26.7% at mean follow-up of 16 months, of which
18.75% were due to infection.
The overall infection rate reported in the literature ranges from 3% to 15%.[[3]],[[14]],[[15]],[[16]],[[17]] Shunt infection rate of 5% in our study was at internationally accepted level.
HCP Clinical Research Network Quality Improvement Initiative reported an infection
rate of 5.7% in 2011.[[14]] The same research group introduced the new protocol to use an antibiotic-impregnated
catheter (AIC) and found a new infection rate of 6% which was not different than their
previous infection rate (5.7%) with or without AIC.[[18]] In our series, normal catheter was used in all cases due to cost factors.
Majority of shunt infection occurs early after surgery. Choux et al. found that 90%
of the infections occurred within the first 6 months.[[19]] Another study by McGirt et al. found that infection was responsible for 45% of
failures within the 1st month.[[3]] In our case, 87.5% of the infections occurred within first 2 months of surgery,
50% of them occurred in the 1st month. The most common organism was coagulase-negative
Staphylococcus (50%), followed by Staphylococcus aureus (17%) which was similar as
reported by previous studies.[[8]]
We evaluated demographic, clinical, and surgical variables to find their association
with shunt complications. Of these, surgeon experience, duration of surgery, and the
priority of case were found to be significantly associated with shunt failure. Tuli
et al. examined the effects of patient characteristics, shunt hardware, and surgical
details in pediatric patients.[[5]] Their findings suggested that age (prematurity and infants), certain etiology of
HCP (intraventricular hemorrhage, postmeningitis, and tumor), and the concurrent other
surgical procedures were significantly associated with increased risk of failure.
Their study did not show any association with gender, American Society of Anesthesiologists
class, emergency surgery, duration of surgery, use of prophylactic antibiotic agents,
and shunt-related factors. Simon et al. found age (6–12 months), prior neurosurgery,
and the presence of gastrostomy tube as independent risk factors for infection.[[20]] Similarly, a multicenter prospective cohort study found that age younger than 6
months, cardiac comorbidity, and endoscopic placement were independent risk factors
for failures, while other factors such as etiology, payer, center, valve design, valve
programmability, the use of ultrasound or stereotactic guidance, and surgeon experience
and volume were not associated with risk for failure.[[10]] Our study could not detect the association of age with shunt survival. This could
be because of low sample size and the fact that majority of the patients were other
than infants. As reported by Riva-Cambrin et al. and in contrast to the findings of
Tuli et al., our findings did not show any relationship between etiology and the risk
of shunt failure.[[5]],[[10]]
Although our study found surgeon experience, duration of surgery, and case priority
as significant factors of failure, there is no consistency in these associations in
the reported literature. Omrani et al. and Griebel et al. found surgeon experience
as a statistically significant predictor of complication,[[15]],[[21]] while others failed to find this association.[[10]],[[22]] Similarly, Tuli et al. and Griebel et al.'s study could not found the duration
of surgery and emergency surgery as the significant predictors of failure.[[5]],[[15]]
Being a retrospective study, it has its own limitations. We did not evaluate the impact
of comorbidity as well as any other concurrent operation or neurosurgical intervention
before shunt which could have influenced the outcome. Some patients were lost to follow-up
and status of their shunt could not be assessed. Although this is the first and largest
study in Nepal to analyze risk factors of VP shunt complication, the small sample
size may have affected in power of study to detect all statistical significance.
Conclusion
The overall shunt complication rate among pediatric patients with HCP who had their
first VP shunt surgery was 26.7% (95% CI 19.0%–35.5%). Malfunction was more common
than infection (21.6% and 5%, respectively). Logistic regression analysis showed that
longer duration of surgery, less experience of the surgeon, and the priority of the
case (emergency) were independent risk factors for shunt complication. A prospective
cohort study with larger sample size is recommended for further analysis of risk factors
which will help reduce morbidity and mortality associated with shunt failure.
