Pediatric endoscopic third ventriculostomy: a population-based study

Clinical article

Full access

Object

Endoscopic third ventriculostomy (ETV) is an alternative to ventriculoperitoneal shunting for hydrocephalus treatment. Choice of treatment options raises questions about which patients are likely to benefit from ETV. The authors performed a population-based analysis using an administrative claims database, examining current practice and outcomes for pediatric patients in the US.

Methods

The authors queried the MarketScan (Truven Health Analytics) database for Current Procedural Terminology codes corresponding to ETV and ventriculoperitoneal shunting from 2003 to 2011; they included patients 19 years or younger and extracted data from initial and subsequent hospitalizations. Hydrocephalus etiology was classified with ICD-9-CM coding. ETV failure was defined as any subsequent ETV or shunt procedure.

Results

Five hundred one patients underwent ETV. Of these, 46% were female. The mean age was 8.7 ± 6.4 years (± SD). The mean follow-up was 1.9 ± 1.8 years. Etiology of hydrocephalus was primarily tumor (41.7%) and congenital/aqueductal stenosis (24.4%). ETV was successful in 354 patients (71%). The mean time to failure was 109.9 ± 233 days. Of the 147 patients with ETV failure, 35 (24%) underwent repeat ETV and 112 (76%) had shunt placement. Patients in age groups 0 to < 6 months and 6 months to < 1 year had a significantly higher rate of ETV failure than those 10–19 years (HR 2.9, p = 0.05; and HR 2.3, p = 0.001, respectively). History of prior shunt was associated with higher risk of failure (HR 2.5, p < 0.001). There were no significant associations between hydrocephalus etiology and risk of failure. A second wave of failures occurred at 2.5–3.5 years postoperative in tumor and congenital/aqueductal stenosis patients; this was not observed in other etiology groups.

Conclusions

This study represents a cross-section of nationwide ETV practice over 9 years. ETV success was more likely among children 1 year and older and those with no history of prior shunt.

Abbreviations used in this paper:CPT = Current Procedural Terminology; ETV = endoscopic third ventriculostomy; IVH = intraventricular hemorrhage; LOS = length of stay; MMC = myelomeningocele; VP = ventriculoperitoneal.

Abstract

Object

Endoscopic third ventriculostomy (ETV) is an alternative to ventriculoperitoneal shunting for hydrocephalus treatment. Choice of treatment options raises questions about which patients are likely to benefit from ETV. The authors performed a population-based analysis using an administrative claims database, examining current practice and outcomes for pediatric patients in the US.

Methods

The authors queried the MarketScan (Truven Health Analytics) database for Current Procedural Terminology codes corresponding to ETV and ventriculoperitoneal shunting from 2003 to 2011; they included patients 19 years or younger and extracted data from initial and subsequent hospitalizations. Hydrocephalus etiology was classified with ICD-9-CM coding. ETV failure was defined as any subsequent ETV or shunt procedure.

Results

Five hundred one patients underwent ETV. Of these, 46% were female. The mean age was 8.7 ± 6.4 years (± SD). The mean follow-up was 1.9 ± 1.8 years. Etiology of hydrocephalus was primarily tumor (41.7%) and congenital/aqueductal stenosis (24.4%). ETV was successful in 354 patients (71%). The mean time to failure was 109.9 ± 233 days. Of the 147 patients with ETV failure, 35 (24%) underwent repeat ETV and 112 (76%) had shunt placement. Patients in age groups 0 to < 6 months and 6 months to < 1 year had a significantly higher rate of ETV failure than those 10–19 years (HR 2.9, p = 0.05; and HR 2.3, p = 0.001, respectively). History of prior shunt was associated with higher risk of failure (HR 2.5, p < 0.001). There were no significant associations between hydrocephalus etiology and risk of failure. A second wave of failures occurred at 2.5–3.5 years postoperative in tumor and congenital/aqueductal stenosis patients; this was not observed in other etiology groups.

Conclusions

This study represents a cross-section of nationwide ETV practice over 9 years. ETV success was more likely among children 1 year and older and those with no history of prior shunt.

The prevalence of congenital/infantile hydrocephalus in the US and Europe has been estimated to range from 0.5 to 0.8 per 1000 live births.5–7,16 Hydrocephalus represents a high health care burden in the US, with ventriculoperitoneal (VP) shunting and its associated hospital charges totaling almost $2 billion annually in children 0–18 years.18 The burden of hydrocephalus is reported to be even higher in developing nations.23

Potential treatments for hydrocephalus include CSF diversion with VP shunt placement and endoscopic third ventriculostomy (ETV). The lack of implanted hardware in the case of ETV makes it an attractive alternative to shunting, but only for the subset of patients in whom ETV is an adequate and durable treatment for hydrocephalus. ETV has been used as a treatment of hydrocephalus secondary to different etiologies, including tumor-related obstructive hydrocephalus, aqueductal stenosis, myelomeningocele (MMC), postinfectious hydrocephalus, and intraventricular hemorrhage (IVH).3,8,20,23,24 Multicenter studies have characterized outcomes after ETV in different pediatric populations. Drake, in a large (n = 368) multicenter Canadian study, found that younger age is a significant predictor of increased failure after ETV, but that hydrocephalus etiology and surgeon/center volume were not.3 Warf et al. found that both older age and certain etiologies (postinfectious and MMC) were associated with a higher chance of ETV success in the Ugandan population. 25 Other smaller studies have reported conflicting information on the role of age, etiology, and other factors on the impact of success after ETV.1,8,9,15,19 Kulkarni et al. described multinational series; predictive strategies have been used to create success scores for various populations. 10,11,25 However, it remains to be seen if a national sample has similar ETV success compared with these site-specific studies.

The availability of large secondary administrative databases allows for the opportunity to examine practices in a wide variety of academic and community settings across the US. While administrative claims data have been used to explore topics in adult neurosurgery,13 it has not been used to investigate ETV in children. This study aims to use a large administrative database to describe the current practice, correlates, and effectiveness of ETV in pediatric patients in the US.

Methods

Data Source

The MarketScan database (Truven Health Analytics) is a collection of health insurance claims for working adults and early retirees with employer-sponsored health insurance and their dependents. For the current project, we used the MarketScan Commercial Claims and Encounters database, constructed from paid claims for employee-sponsored health insurance for 2003 to 2011, representing 17 million enrollees in 2003 to 52 million enrollees in 2011. The Commercial Claims and Encounters database contains 7 tables, including inpatient admission, facility header, inpatient service, outpatient service, population, outpatient pharmaceutical claims, and enrollment. In this study, we used the inpatient admission, inpatient service, outpatient service, and enrollment tables. Within these tables, we used records from January 1, 2003, to December 31, 2011. All pediatric patients were covered under their parents' employer-sponsored health insurance. The study received exempt status from the University of Chicago and Baylor College of Medicine institutional review boards.

Patient Selection

For the initial procedure hospitalization, we queried inpatient service tables for all hospitalizations for patients 0 through 19 years of age in which the following Current Procedural Terminology (CPT) procedure codes were recorded: 1) ETV codes: 62200, 62201; or 2) shunt placement codes: 62220, 62223. These codes and associated dates of procedures were used to determine the type and timing of index and follow-up surgeries. For each patient, the first occurrence of ETV in the database at 19 years or younger was considered the index procedure.

Follow-Up Data

For the included patients, we examined details of any subsequent hospitalization in the inpatient service and inpatient admission tables from the date of the index procedure until the end of the records in 2011. Analysis was then based on initial and subsequent hospitalizations for ETV or shunt.

ETV failure was defined as any subsequent surgery for definitive treatment of hydrocephalus after the initial ETV surgery (i.e., repeat ETV or placement of a VP shunt). Time to failure was computed from the date of initial ETV surgery and date of subsequent procedures. Temporary CSF diversion measures such as ventricular puncture, lumbar puncture, and external ventricular drain placement were not included. Death was not included as an end point as noninpatient deaths are not captured in MarketScan. However, inpatient deaths were examined.

To determine duration of postoperative follow-up, we used the enrollment table to obtain the final month of insurance enrollment for each patient. End of follow-up was the last day of the final month of enrollment, typically because of change in annual election of health insurance plan or change of employment. Those with continuing coverage at the conclusion of the study period were assigned end dates of December 31, 2011, the last date in the database. For inpatient deaths, the date of hospital discharge for any reason with death as the discharge status was treated as the end of follow-up. Thus, follow-up ended on the earlier of the date of discharge status of deceased or last day of the final month of enrollment. Postoperative follow-up time was calculated as the difference between the end of follow-up and the date of the initial procedure. Two patients were excluded because their index surgeries occurred during long hospital admissions with surgery and discharge after insurance enrollment ended.

Covariates

Age in years on the date of index admission was reported in the inpatient admission tables; date of birth was not available in MarketScan databases due to their de-identified nature. For those 0 years of age a variable based on the Major Diagnostic Category and Diagnosis Related Group of the admission indicated whether the admission was for maternity/neonate or other reason. Patients with this indicator were considered to be neonates, and age was estimated from the earliest date of admission and date of ETV. We assigned patients estimated to be younger than 182 days old to age group 0 to < 6 months, and the remaining patients 0 years of age were assigned to age group 6 months to < 1 year. Sex, geographic region, year, and length of stay (LOS) were reported in the inpatient admission tables.

We classified patients by hydrocephalus etiology based on International Classification of Diseases, Ninth Edition—Clinical Modification (ICD-9-CM) diagnosis codes reported for the index hospitalization in the inpatient admission tables. Indication for ETV surgery was determined based on previously published methodology using administrative data to examine pediatric hydrocephalus and CSF shunts.17 We reviewed the ICD-9-CM diagnosis codes, focusing on those that occurred at a frequency of ≥ 1% of the study population. We assigned etiology at the time of ETV with the concurrent assignment of one of the following groups of diagnosis codes: IVH (772.1x), MMC (653.7, 655.0, 741.x), CNS tumor (191–194, 198.3–4, 225.0–2, 225.8–9, 227.4, 237.0–1, 237.5–7, 239.6–7), meningitis (320–322, 326), trauma (767.4, 851. xx–854.xx, 995.55), and congenital hydrocephalus/aqueductal stenosis (742.3). Those with no indications or multiple indications were classified as other etiology.

History of previous CSF shunt was determined from diagnosis codes of inpatient and outpatient encounters from initial enrollment until the day prior to index ETV surgery. Codes considered to indicate history of CSF shunt include ICD-9-CM procedure codes of shunt surgery (02.32–35, 02.42, and 02.43), ICD-9-CM codes indicating presence of or complications of shunt device (V45.2, V53.01, 996.2, 996.63), and CPT codes for shunt surgery (62190, 62192, 62194, 62220, 62223, 62225, 62230, 62180, 62256, 62258). Patients were classified either as having a history of prior shunt or no history of prior shunt.

Statistical Analysis

We summarized the distribution of patient demographics and surgery-related characteristics using descriptive analyses. Mean values are presented as the mean ± SD. We performed chi-square tests for bivariate analyses of categorical variables and the Mann-Whitney U-test for continuous variables. Kaplan-Meier plots were used to estimate overall survival as well as 1-, 2-, and 5-year survival rates. The association of age, sex, year, region, etiology, index LOS, and history of prior shunt with survival was quantified using hazard ratios estimated from multivariate Cox proportional hazards models. Multivariate logistic regression was used to examine factors associated with second ETV versus shunt among ETV failures. We conducted a sensitivity analysis to examine the effects of mortality and changes in insurance coverage before the end of the study period. We used 2-sided tests, with p values < 0.05 considered to be statistically significant. All statistical analyses were performed with Stata (version 12, StataCorp).

Results

Cohort

Five hundred one patients ranging from 0 to 19 years of age were identified as having undergone an index ETV procedure from the years 2003 to 2011, inclusive. The sample represents 4.1 (95% CI 2.3–5.8) to 7.3 (95% CI 5.9–8.7) index ETV procedures per million enrollees and dependents of enrollees 0 to 19 years of age in 2003 and 2011, respectively. Forty-six percent were female. The mean age ± SD was 8.8 ± 6.4 years (Table 1). Six patients were 0 to < 6 months, 84 patients were 6 months to < 1 year, 168 patients were 1 to < 10 years, and 243 patients were 10–19 years. The distribution of etiology during index surgery was 41.7% tumor, 24.4% congenital/aqueductal stenosis, 3.6% MMC, 1.4% IVH, 2.2% meningitis, 1.4% trauma, and 25.4% other. The study period was divided into 3 epochs: Years 2003–2005, 2006–2008, and 2009–2011. There were no significant differences in hydrocephalus etiology composition or in ETV success rates by epoch (log-rank test, p = 0.378).

TABLE 1:

Characteristics of the study sample (n = 501)

CharacteristicGroup*p Value
OverallSuccessFailure
no. of patients501354147
mean age in yrs8.8 ± 6.4
age group<0.001
 0 to <6 mos6 (1.2)2 (0.6)4 (2.7)
 6 mos to <1 yr84 (16.8)44 (12.4)40 (27.2)
 1 to <10 yrs168 (33.5)126 (35.6)42 (28.6)
 10–19 yrs243 (48.5)182 (51.4)61 (41.5)
sex0.35
 male270 (53.9)186 (52.5)84 (57.1)
 female231 (46.1)168 (47.5)63 (42.9)
years0.74
 2003–200583 (16.6)56 (15.8)27 (18.4)
 2006–2008148 (29.5)107 (30.2)41 (27.9)
 2009–2011270 (53.9)191 (54.0)79 (53.7)
region0.07
 Northeast90 (18.0)61 (17.2)29 (19.7)
 Midwest128 (25.6)80 (22.6)48 (32.7)
 South185 (36.3)143 (40.4)42 (28.6)
 West89 (17.8)64 (18.1)25 (17.0)
etiology<0.001
 congenital122 (24.4)76 (21.5)46 (31.3)
 tumor209 (41.7)157 (44.4)52 (35.4)
 MMC18 (3.6)12 (3.4)6 (4.1)
 IVH7 (1.4)1 (0.3)6 (4.1)
 meningitis11 (2.2)4 (1.1)7 (4.8)
 trauma7 (1.4)6 (1.7)1 (0.7)
 other127 (25.4)98 (27.7)29 (19.7)
history of previous shunt108 (21.6)59 (16.7)49 (33.3)<0.001
mean index LOS in days10.3 ± 22.98.2 ± 20.315.4 ± 27.50.001
mean no. of total admissions3.7 ± 3.82.7 ± 2.86.1 ± 4.60.001
mean no. of surgeries1.5 ± 1.21.0 ± 0.02.8 ± 1.7
median follow-up time in yrs1.3 (0.5–2.6)1.2 (0.5–2.5)1.3 (0.6–2.6)0.28
failure147 (29.3)0 (0.0)147 (100.0)
mean time to failure in days109.9 ± 233
mortality16 (3.2)8 (2.3)8 (5.4)0.07

Values are presented as the number of patients (%) unless stated otherwise. Mean values are presented as the mean ± SD. The median value is presented as the median (interquartile range).

The mean LOS for the entire cohort was 10.3 ± 22.9 days. The LOS encompassed a wide range, and the influence of outliers is reflected in the mean (range 0–279 days, 10th percentile 1.0 day, 25th percentile 2.0 days, 50th percentile 4.0 days, 75th percentile 10.0 days, 90th percentile 24.0 days). Length of index hospitalization was longer in the patients with ETV failure, with a mean of 8.2 ± 20.3 inpatient days for the ETV success group and 15.4 ± 27.5 inpatient days for the ETV failure group (p = 0.001). The failure group had more total all-cause admissions subsequent to index ETV surgery than the success group.

Mortality rates did not differ between the groups. Sixteen inpatient deaths (3.19%) were associated with this pediatric ETV cohort. The median number of days to death was 157 days (interquartile range 55–297 days, range 13–587 days) following the index surgery. Half of these patients had 1–4 subsequent surgeries or admissions between the index procedure and death. Seven died during an admission for an ETV-related surgery (median 35 days, range 13–339 days following the index procedure) at ages 0–16 years. Fourteen of the 16 inpatient deaths occurred in patients with tumor diagnoses; the other 2 cases included 1 patient in the IVH category and 1 in the “other” category.

Patients underwent on average 1.9 ± 1.8 years of follow-up after the index surgery. Follow-up duration ranged from 1 day to 8.9 years with a median of 1.3 years. Sensitivity analysis showed that results were not changed when the sample was restricted to minimum 1-month and minimum 3-month postoperative follow-up times. The ETV success and ETV failure groups did not differ in duration of follow-up (p = 0.28).

End Points

Endoscopic third ventriculostomy was successful in 71% of the cohort as a whole, with 354 of the 501 patients not having a subsequent surgery. Overall success rates for the entire population of patients is shown as survival in Fig. 1. By Kaplan-Meier estimates, the 6-month, 1-year, 2-year, and 5-year ETV overall success rates were 73.5%, 70.8%, 68.9%, and 62.8%, respectively. The mean time to failure was 109.9 ± 233 days. The median time to failure was 24 days with a range of 1–1330 days. Of those in whom failure occurred, the failure occurred in 57.1% in the first 30 days, in 75.5% in the first 90 days, and in 85.0% in the first 6 months.

Fig. 1.
Fig. 1.

Overall Kaplan-Meier estimate of ETV time to failure.

Kaplan-Meier estimates of ETV success rate stratified by age showed increasing success with older age (Fig. 2). These curves were significantly different by log-rank analysis (p = 0.001). The 1-year ETV success rates for the 0-to < 6-month, 6-month to < 1-year, 1- to < 10-year, and 10- to 19-year age groups were 50.0%, 48.7%, 76.3%, and 75.1%, respectively. Up to 3 years after the index procedure, success rates did not differ between the 1- to < 10- year and 10- to 19-year age groups. While the majority of failures occurred within the first 6 months across all age groups, those occurring at 1–19 years of age showed a bimodal distribution of failures with a dip in the Kaplan-Meier estimates between 2.5 and 3.5 years after the index ETV surgery. This second failure time point was more pronounced in the 1- to < 10-year age group such that 5-year success rates were significantly higher in the 10- to 19-year group compared with the 1- to < 10-year age group and also the < 1-year age group. The 5-year success rates for the 6-month to < 1-year, 1- to < 10-year, and 10- to 19-year age groups were 48.7%, 58.7%, and 69.7%, respectively. All remaining patients in the 0- to < 6-month age group were lost to follow-up after 2 years.

Fig. 2.
Fig. 2.

Kaplan-Meier estimates of ETV time to failure stratified by age group.

Kaplan-Meier estimates of ETV success over time stratified by hydrocephalus etiology are shown in Fig. 3. The 1-year ETV success rates for congenital, tumor, MMC, trauma, meningitis, and other were 64.0%, 75.9%, 65.5%, 85.7%, 36.4%, and 75.3%, respectively. The 6-month ETV success rate for IVH was 14.3% (all remaining patients with IVH were lost to follow-up by 1 year). The 2-year success rates were similar to the 1-year rates. The congenital/ aqueductal stenosis patients had a second round of failures between 2.5 and 3.0 years following ETV, and the tumor group had a similar round of failures at about the same time. The 5-year ETV success rates for congenital, tumor, MMC, and other were 52.3%, 61.4%, 65.5%, and 75.3%, respectively. All remaining patients who suffered trauma were lost to follow-up after 3 years.

Fig. 3.
Fig. 3.

Kaplan-Meier estimates of ETV success stratified by diagnosis at index admission.

In the multivariate regression model for factors associated with ETV failures, the 0- to < 6-month age group had an HR of 2.91 (95% CI 1.00–8.49, p = 0.051) and the 6- to 12-month group had an HR of 2.31 (95% CI 1.41–3.76, p = 0.001) compared with those 10–19 years of age (Table 2). Sex, region, and year of ETV did not impact ETV success. Controlling for the other factors, LOS for index surgery also had no impact on ETV success (HR 1.00 [95% CI 1.00–1.01], p = 0.061). Similarly, etiology was not significantly associated with ETV failure. However, history of prior CSF shunt was significantly associated with ETV failure with an HR of 2.47 (95% CI 1.70–3.59, p < 0.001) compared with no history of prior shunt (Table 2).

TABLE 2:

Multivariate Cox regression of factors associated with ETV failures

CharacteristicHR95% CIp Value
age group
 0 to <6 mos2.911.00–8.490.05
 6 mos to <1 yr2.311.41–3.760.001
 1 to <10 yrs1.160.77–1.740.47
 10–19 yrs1.00referencereference
sex
 male1.00referencereference
 female0.910.65–1.270.58
year
 2003–20051.00referencereference
 2006–20080.650.39–1.080.29
 2009–20110.820.51–1.330.43
region
 Northeast1.00referencereference
 Midwest1.080.66–1.770.75
 South0.700.43–1.140.15
 West0.830.47–1.440.50
 unknown0.740.21–2.520.63
etiology
 tumor1.00referencereference
 congenital1.170.75–1.840.49
 MMC0.760.32–1.840.55
 IVH2.170.84–5.600.11
 meningitis2.030.88–4.650.10
 trauma0.430.06–3.170.41
 other0.750.47–1.210.24
history of prior shunt2.471.70–3.59<0.001
index LOS (days)1.001.00–1.010.06

Second Treatment

Table 3 summarizes the characteristics of the patients who experienced an ETV failure and underwent repeat ETV versus shunt placement. This describes the pattern of surgeons' choice and does not reflect eventual outcome of hydrocephalus treatment. The median age of the 23.8% of ETV failure patients who underwent repeat ETV was 8.0 years compared with a median age of 6.0 years for children who received shunts. Younger patients, longer mean hospital LOS (18.2 vs 6.5 days, p = 0.03), and shorter time to failure (86 vs 186 days, p = 0.03) were associated with shunt placement versus repeat ETV. Sex, year of surgery, and etiology of hydrocephalus were not related to the type of surgery received at the time of ETV failure. In multivariate analysis, hospital LOS for the index surgery was the only factor significantly associated with an increased likelihood of receiving a shunt (OR 0.93 [95% CI 0.87–0.99], p = 0.02; Table 4). The outcomes in follow-up after repeat ETV are as follows: among those with a second ETV, 34.3% had a second failure. The range of time to the second failure after repeat ETV was 2–784 days, median 10 days, mean 94.8 ± 228.6 days.

TABLE 3:

Characteristics of patients with ETV failure, by second treatment (n = 147)

CharacteristicGroup*p Value
ShuntRepeat ETV
no. of patients11235
median age in yrs68
mean age in yrs7.0 ± 7.19.3 ± 6.30.08
age group0.01
 0 to <6 mos2 (1.8)2 (5.7)
 6 mos to <1 yr37 (33.0)3 (8.6)
 1 to <10 yrs27 (24.1)15 (42.9)
 10–19 yrs46 (41.1)15 (42.9)
sex0.43
 male62 (55.4)22 (62.9)
 female50 (44.6)13 (37.1)
year0.63
 2003–200521 (18.8)6 (17.1)
 2006–200829 (25.9)12 (34.3)
 2009–201162 (55.4)17 (48.6)
region0.73
 Northeast23 (20.5)6 (17.1)
 Midwest34 (30.4)14 (40.0)
 South32 (28.6)10 (28.6)
 West20 (17.9)5 (14.3)
etiology0.57
 congenital37 (33.0)9 (25.7)
 tumor36 (32.1)16 (45.7)
 MMC4 (3.6)2 (5.7)
 IVH5 (4.5)1 (2.9)
 meningitis7 (6.2)0 (0.0)
 trauma1 (0.9)0 (0.0)
 other22 (19.6)7 (20.0)
prior shunt40 (35.7)9 (25.7)0.27
mean LOS in days18.2 ± 30.76.5 ± 7.00.03
mean time to failure in days86 ± 211186 ± 2830.03

Values are the number of patients (%) unless indicated otherwise. Mean values are presented as the mean ± SD.

TABLE 4:

Logistic regression for second ETV vs shunt after index ETV failure (n = 147)

CharacteristicOR*95% CIp Value
age group
 0 to <6 mos3.520.22–56.50.37
 6 mos to <1 yr0.290.06–1.430.13
 1 to <10 yrs1.720.62–4.830.30
 10–19 yrs1.00referencereference
sex
 male1.00referencereference
 female0.780.30–2.000.60
year
 2003–20051.00referencereference
 2006–20081.130.31–4.190.85
 2009–20111.190.34–4.180.79
region
 Northeast1.00referencereference
 Midwest0.950.26–3.430.94
 South0.640.16–2.540.53
 West0.490.10–2.330.37
etiology
 tumor1.00referencereference
 congenital0.660.19–2.260.51
 MMC1.140.11–11.630.92
 IVH1.420.06–33.740.83
 meningitis
 trauma
 other1.570.42–5.800.50
history of prior shunt0.820.26–2.560.73
index LOS (days)0.930.87–0.990.02
time to failure (mos)1.040.98–1.090.20
constant0.790.13–4.900.80

OR > 1.0 favors second ETV.

Effects of Censoring

The sensitivity analysis tested the effect of mortality and changes in health insurance prior to the end of the study period on the Kaplan-Meier estimates of median time to failure overall and by age group. Change in health insurance causes loss to follow-up in MarketScan claims data. We wanted to examine the effect of these activities on the study results. Although 28% of the sample was censored for these reasons, the median time to failure was similar when this group was excluded or was assumed to have failure the day after changing insurance. The 1-, 2-, and 5-year ETV survival was slightly better for patients younger than 1 year old when the informatively censored patients were excluded; other age and etiology groups did not differ. This suggests that the changes in insurance did not substantially affect the conclusions of the study.

Discussion

This study examined a large administrative claims data set representing up to 52 million persons per year to identify ETV procedures and outcomes of 501 children across the US over 9 years from 2003 to 2011. The overall ETV success rate at 1 year was 70.7%. We found that age younger than 1 year and patients with a history of prior shunt were more likely to have ETV failure when adjusting for etiology and other factors. This is the first study to use population-based administrative claims data to examine ETV outcomes.

Our findings are consistent with other reports of ETV success rates. For example, Drake reported 1- and 5-year success rates of 65% and 52%, respectively, based on 368 Canadian patients.3 Kulkarni et al. reported a 6-month success rate of 66.3% in 618 patients collected from 3 countries.10 Kadrian et al. reported an 89% overall longterm success rate in a patient population that included both children and adults.8 Due to the national population- based nature of this administrative claims data set, the results are more generalizable than institution-based studies.

We found a second round of failures around 3 years after the index ETV procedure in the children 1 to < 10 years of age and in tumor and aqueductal stenosis cases. Early ETV failures are more common than late ETV failures, although many of the largest ETV studies to date do not include this length of follow-up.11 Late rapid deterioration has been described with a mean of 2.5 years (as late as 7.8 years) after index ETV,2 although late failure events may not be captured in this data set. The use of administrative data limits our long-term evaluation in the present study, as subjects who change insurance provider are lost to follow-up. However, the suggestion of a second distinct period of risk for ETV failure is consistent with previous reports of late failures and highlights the importance of long follow-up in subsequent large clinical series that study ETV.

Age and history of prior shunt were the only significant predictors of ETV success in the present study, with index LOS approaching significance. We found that older children had a higher chance of avoiding subsequent surgery for hydrocephalus treatment. Success at 1 year for the < 1-year age group was 48.7%, while it was 75.1% for the 10- to 19-year age group. Drake found that only age had a significant effect on outcome with shunts failing in younger patients at higher rates. Factors included in their model were age, sex, etiology of hydrocephalus, previous surgery, center volume, and surgeon volume.3 The ETV success score paradigm described by Kulkarni et al. has over 50% of the weighted score attributable to age alone. In that study, the model for prediction of ETV success found that age, cause of hydrocephalus, and previous shunt were all important and independent predictors of ETV success, with age being the strongest predictor.10,11 This algorithm for estimating ETV success has been validated by review of select published case series12 and by other groups in single-institution retrospective studies.4,14

None of the hydrocephalus-related diagnoses in these data had a significant association with ETV success. While some authors reported no correlation of hydrocephalus etiology with ETV outcomes,3,8,19,21 others have found etiology to be predictive of ETV success.1,9,10 Etiology of hydrocephalus accounts for up to 33% of the possible points in the aforementioned ETV success score paradigm.10 A notable limitation of administrative databases, the MarketScan database included, is that ICD-9 coding with insurance claims may not be as reliable in classifying diagnoses as clinical studies. Therefore, it is unclear if we can confidently delineate the impact of hydrocephalus etiology on ETV success.

One-quarter of patients in whom ETV failed underwent repeat ETV and 76% received shunt placement. Those who had repeat ETV tended to be older (mean 9.3 ± 6.3 years) and had a shorter index length of stay than those who had later shunt placement (mean 7.0 ± 7.1 years, p = 0.08). Time to ETV failure was also longer in the repeat ETV group. Taken together, this may reflect surgeons' willingness to choose repeat ETV in children who may be older, with fewer comorbidities requiring shorter index hospital stay, and who had relatively better response to the initial ETV. Among those who underwent repeat ETV, 34.3% of patients had a second failure, giving a 65.7% overall rate of success in this group.

Our study has 2 notable strengths. First, administrative claims data including MarketScan and Medicare are routinely used for population-based comparative effectiveness research.13 This data source reduces the common biases that result from selecting patients from particular surgeons, institutions, or regions, and reflects a nationally representative range of skill level and institutional size and volume. Although the data are not weighted to the national population, this reflects ETV outcomes in the real-world setting of up to one-sixth of the US population. Thus, the results are roughly generalizable to the pediatric neurosurgery community. Administrative data are useful for additional research questions including those concerned with care across multiple settings and total resource utilization and cost. Second, this methodology allows us to evaluate a relatively large number of children (n = 501).

Despite these strengths, a few notable limitations are present in this study. Various patient and clinical details cannot be discerned. Specific anatomy and imaging characteristics are not available. However, these are not suspected to be correlates of ETV failure and choice of second surgery. As patients and procedures were selected for this cohort using ICD-9-CM diagnosis and CPT procedure codes, we are cognizant that miscoding may be present. While CPT codes for billing associated with surgeries are less ambiguous, ICD-9 diagnosis codes may vary in clinical precision and quality control. Simon et al. used the same ICD-9 diagnosis codes to classify etiology using the Kids' Inpatient Database and found only 47% of patients to be classifiable.18 Here we are able to classify 75%. Third, there is no separate CPT code for choroid plexus cauterization, so the possibility exists that the effects of choroid plexus cauterization, if done, may confound outcomes results.22 Non–hospital death is not captured as an end point. Finally, this study is retrospective and uses only the private payer category of health insurance coverage in the US, and this is not representative of all demographic strata.

Patients had on average 1.9 ± 1.8 years of follow-up after the index surgery, which may seem short for a cohort spanning from 2003 to 2011. The follow-up duration for this study was stopped at the time of a change in insurance enrollment or end of data availability, which occurred in almost 30% of patients. In MarketScan databases, patients were lost to follow-up after changing health insurance policies, parental job change, or outpatient death. Additionally, the largest sample was for 2011, with the end of the database at the end of the year. Nevertheless, 85% of failures were within 6 months of the index ETV. Moreover, the ETV success and failure groups did not differ in follow-up duration (p = 0.28). The sensitivity analysis did not indicate that the loss to follow-up biased the results. Our results should be interpreted cautiously and with the understanding of the inherent limitations of the data source. We do not represent these data as a replacement for clinical retrospective data, prospectively collected data, or randomized data. This study aims to leverage the strengths of the large reach of administrative data to generate a large sample size and to show patterns in care and utilization across the nation.

Conclusions

Use of a US population-based administrative claims data set shows ETV success in a cohort of 501 pediatric patients to be 70.8% at 1 year. Of the ETV failures, average time to failure was 109.9 days. ETV success was more likely among children 1 year and older and those with no history of prior shunt. One-quarter of these patients who experienced ETV failure underwent repeat ETV and 76% underwent shunt placement. This study reflects ETV practice in the US from 2003 to 2011. Further studies to examine outcomes, longitudinal costs, and health care utilization in this population are warranted.

Disclosure

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author contributions to the study and manuscript preparation include the following. Conception and design: Lam. Acquisition of data: Lam, Ham. Analysis and interpretation of data: Lam, Harris, Ham. Drafting the article: Lam, Rocque. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Lam. Statistical analysis: Harris, Ham. Administrative/technical/material support: Lam. Study supervision: Lam.

This article contains some figures that are displayed in color online but in black-and-white in the print edition.

References

  • 1

    Baldauf JOertel JGaab MRSchroeder HW: Endoscopic third ventriculostomy in children younger than 2 years of age. Childs Nerv Syst 23:6236262007

  • 2

    Drake JChumas PKestle JPierre-Kahn AVinchon MBrown J: Late rapid deterioration after endoscopic third ventriculostomy: additional cases and review of the literature. J Neurosurg 105 :2 Suppl1181262006

  • 3

    Drake JM: Canadian Pediatric Neurosurgery Study Group: Endoscopic third ventriculostomy in pediatric patients: the Canadian experience. Neurosurgery 60:8818862007

  • 4

    Durnford AJKirkham FJMathad NSparrow OC: Endoscopic third ventriculostomy in the treatment of childhood hydrocephalus: validation of a success score that predicts long-term outcome. Clinical article. J Neurosurg Pediatr 8:4894932011

  • 5

    Fernell EHagberg GHagberg B: Infantile hydrocephalus epidemiology: an indicator of enhanced survival. Arch Dis Child Fetal Neonatal Ed 70:F123F1281994

  • 6

    Garne ELoane MAddor MCBoyd PABarisic IDolk H: Congenital hydrocephalus—prevalence, prenatal diagnosis and outcome of pregnancy in four European regions. Eur J Paediatr Neurol 14:1501552010

  • 7

    Jeng SGupta NWrensch MZhao SWu YW: Prevalence of congenital hydrocephalus in California, 1991-2000. Pediatr Neurol 45:67712011

  • 8

    Kadrian Dvan Gelder JFlorida DJones RVonau MTeo C: Long-term reliability of endoscopic third ventriculostomy. Neurosurgery 62 :Suppl 26146212008

  • 9

    Koch-Wiewrodt DWagner W: Success and failure of endoscopic third ventriculostomy in young infants: are there different age distributions?. Childs Nerv Syst 22:153715412006

  • 10

    Kulkarni AVDrake JMKestle JRMallucci CLSgouros SConstantini S: Predicting who will benefit from endoscopic third ventriculostomy compared with shunt insertion in childhood hydrocephalus using the ETV Success Score. Clinical article. J Neurosurg Pediatr 6:3103152010. (Erratum in J Neurosurg Pediatr 7: 221 2011)

  • 11

    Kulkarni AVDrake JMMallucci CLSgouros SRoth JConstantini S: Endoscopic third ventriculostomy in the treatment of childhood hydrocephalus. J Pediatr 155:254259.e12009

  • 12

    Kulkarni AVRiva-Cambrin JBrowd SR: Use of the ETV Success Score to explain the variation in reported endoscopic third ventriculostomy success rates among published case series of childhood hydrocephalus. Clinical article. J Neurosurg Pediatr 7:1431462011

  • 13

    Lad SPBabu RRhee MSFranklin RLUgiliweneza BHodes J: Long-term economic impact of coiling vs clipping for unruptured intracranial aneurysms. Neurosurgery 72:100010132013

  • 14

    Naftel RPReed GTKulkarni AVWellons JC: Evaluating the Children's Hospital of Alabama endoscopic third ventriculostomy experience using the Endoscopic Third Ventriculostomy Success Score: an external validation study. Clinical article. J Neurosurg Pediatr 8:4945012011

  • 15

    O'Brien DFHayhurst CPizer BMallucci CL: Outcomes in patients undergoing single-trajectory endoscopic third ventriculostomy and endoscopic biopsy for midline tumors presenting with obstructive hydrocephalus. J Neurosurg 105 :3 Suppl2192262006

  • 16

    Persson EKAnderson SWiklund LMUvebrant P: Hydrocephalus in children born in 1999-2002: epidemiology, outcome and ophthalmological findings. Childs Nerv Syst 23:111111182007

  • 17

    Simon TDHall MRiva-Cambrin JAlbert JEJeffries HELafleur B: Infection rates following initial cerebrospinal fluid shunt placement across pediatric hospitals in the United States. Clinical article. J Neurosurg Pediatr 4:1561652009

  • 18

    Simon TDRiva-Cambrin JSrivastava RBratton SLDean JMKestle JR: Hospital care for children with hydrocephalus in the United States: utilization, charges, comorbidities, and deaths. J Neurosurg Pediatr 1:1311372008

  • 19

    Sufianov AASufianova GZIakimov IA: Endoscopic third ventriculostomy in patients younger than 2 years: outcome analysis of 41 hydrocephalus cases. Clinical article. J Neurosurg Pediatr 5:3924012010

  • 20

    Vogel TWBahuleyan BRobinson SCohen AR: The role of endoscopic third ventriculostomy in the treatment of hydrocephalus. Clinical article. J Neurosurg Pediatr 12:54612013

  • 21

    Wagner WKoch D: Mechanisms of failure after endoscopic third ventriculostomy in young infants. J Neurosurg 103 :1 Suppl43492005

  • 22

    Warf BC: Comparison of endoscopic third ventriculostomy alone and combined with choroid plexus cauterization in infants younger than 1 year of age: a prospective study in 550 African children. J Neurosurg 103 :6 Suppl4754812005

  • 23

    Warf BC: Hydrocephalus in Uganda: the predominance of infectious origin and primary management with endoscopic third ventriculostomy. J Neurosurg 102 :1 Suppl1152005

  • 24

    Warf BC: The impact of combined endoscopic third ventriculostomy and choroid plexus cauterization on the management of pediatric hydrocephalus in developing countries. World Neurosurg 79:2 SupplS23.e13S23.e152013

  • 25

    Warf BCMugamba JKulkarni AV: Endoscopic third ventriculostomy in the treatment of childhood hydrocephalus in Uganda: report of a scoring system that predicts success. Clinical article. J Neurosurg Pediatr 5:1431482010

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Address correspondence to: Sandi Lam, M.D., M.B.A., Department of Neurosurgery, Baylor College of Medicine, Texas Children's Hospital, 6701 Fannin St., CCC Ste. 1230, Houston, TX 77030. email: sklam@texaschildrens.org.

Please include this information when citing this paper: published online September 19, 2014; DOI: 10.3171/2014.8.PEDS13680.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Overall Kaplan-Meier estimate of ETV time to failure.

  • View in gallery

    Kaplan-Meier estimates of ETV time to failure stratified by age group.

  • View in gallery

    Kaplan-Meier estimates of ETV success stratified by diagnosis at index admission.

References

1

Baldauf JOertel JGaab MRSchroeder HW: Endoscopic third ventriculostomy in children younger than 2 years of age. Childs Nerv Syst 23:6236262007

2

Drake JChumas PKestle JPierre-Kahn AVinchon MBrown J: Late rapid deterioration after endoscopic third ventriculostomy: additional cases and review of the literature. J Neurosurg 105 :2 Suppl1181262006

3

Drake JM: Canadian Pediatric Neurosurgery Study Group: Endoscopic third ventriculostomy in pediatric patients: the Canadian experience. Neurosurgery 60:8818862007

4

Durnford AJKirkham FJMathad NSparrow OC: Endoscopic third ventriculostomy in the treatment of childhood hydrocephalus: validation of a success score that predicts long-term outcome. Clinical article. J Neurosurg Pediatr 8:4894932011

5

Fernell EHagberg GHagberg B: Infantile hydrocephalus epidemiology: an indicator of enhanced survival. Arch Dis Child Fetal Neonatal Ed 70:F123F1281994

6

Garne ELoane MAddor MCBoyd PABarisic IDolk H: Congenital hydrocephalus—prevalence, prenatal diagnosis and outcome of pregnancy in four European regions. Eur J Paediatr Neurol 14:1501552010

7

Jeng SGupta NWrensch MZhao SWu YW: Prevalence of congenital hydrocephalus in California, 1991-2000. Pediatr Neurol 45:67712011

8

Kadrian Dvan Gelder JFlorida DJones RVonau MTeo C: Long-term reliability of endoscopic third ventriculostomy. Neurosurgery 62 :Suppl 26146212008

9

Koch-Wiewrodt DWagner W: Success and failure of endoscopic third ventriculostomy in young infants: are there different age distributions?. Childs Nerv Syst 22:153715412006

10

Kulkarni AVDrake JMKestle JRMallucci CLSgouros SConstantini S: Predicting who will benefit from endoscopic third ventriculostomy compared with shunt insertion in childhood hydrocephalus using the ETV Success Score. Clinical article. J Neurosurg Pediatr 6:3103152010. (Erratum in J Neurosurg Pediatr 7: 221 2011)

11

Kulkarni AVDrake JMMallucci CLSgouros SRoth JConstantini S: Endoscopic third ventriculostomy in the treatment of childhood hydrocephalus. J Pediatr 155:254259.e12009

12

Kulkarni AVRiva-Cambrin JBrowd SR: Use of the ETV Success Score to explain the variation in reported endoscopic third ventriculostomy success rates among published case series of childhood hydrocephalus. Clinical article. J Neurosurg Pediatr 7:1431462011

13

Lad SPBabu RRhee MSFranklin RLUgiliweneza BHodes J: Long-term economic impact of coiling vs clipping for unruptured intracranial aneurysms. Neurosurgery 72:100010132013

14

Naftel RPReed GTKulkarni AVWellons JC: Evaluating the Children's Hospital of Alabama endoscopic third ventriculostomy experience using the Endoscopic Third Ventriculostomy Success Score: an external validation study. Clinical article. J Neurosurg Pediatr 8:4945012011

15

O'Brien DFHayhurst CPizer BMallucci CL: Outcomes in patients undergoing single-trajectory endoscopic third ventriculostomy and endoscopic biopsy for midline tumors presenting with obstructive hydrocephalus. J Neurosurg 105 :3 Suppl2192262006

16

Persson EKAnderson SWiklund LMUvebrant P: Hydrocephalus in children born in 1999-2002: epidemiology, outcome and ophthalmological findings. Childs Nerv Syst 23:111111182007

17

Simon TDHall MRiva-Cambrin JAlbert JEJeffries HELafleur B: Infection rates following initial cerebrospinal fluid shunt placement across pediatric hospitals in the United States. Clinical article. J Neurosurg Pediatr 4:1561652009

18

Simon TDRiva-Cambrin JSrivastava RBratton SLDean JMKestle JR: Hospital care for children with hydrocephalus in the United States: utilization, charges, comorbidities, and deaths. J Neurosurg Pediatr 1:1311372008

19

Sufianov AASufianova GZIakimov IA: Endoscopic third ventriculostomy in patients younger than 2 years: outcome analysis of 41 hydrocephalus cases. Clinical article. J Neurosurg Pediatr 5:3924012010

20

Vogel TWBahuleyan BRobinson SCohen AR: The role of endoscopic third ventriculostomy in the treatment of hydrocephalus. Clinical article. J Neurosurg Pediatr 12:54612013

21

Wagner WKoch D: Mechanisms of failure after endoscopic third ventriculostomy in young infants. J Neurosurg 103 :1 Suppl43492005

22

Warf BC: Comparison of endoscopic third ventriculostomy alone and combined with choroid plexus cauterization in infants younger than 1 year of age: a prospective study in 550 African children. J Neurosurg 103 :6 Suppl4754812005

23

Warf BC: Hydrocephalus in Uganda: the predominance of infectious origin and primary management with endoscopic third ventriculostomy. J Neurosurg 102 :1 Suppl1152005

24

Warf BC: The impact of combined endoscopic third ventriculostomy and choroid plexus cauterization on the management of pediatric hydrocephalus in developing countries. World Neurosurg 79:2 SupplS23.e13S23.e152013

25

Warf BCMugamba JKulkarni AV: Endoscopic third ventriculostomy in the treatment of childhood hydrocephalus in Uganda: report of a scoring system that predicts success. Clinical article. J Neurosurg Pediatr 5:1431482010

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 256 256 59
PDF Downloads 101 101 14
EPUB Downloads 0 0 0

PubMed

Google Scholar