Comparative durability and costs analysis of ventricular shunts

View More View Less
  • Department of Neurological Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
Full access

OBJECTIVE

Ventricular shunt (VS) durability has been well studied in the pediatric population and in patients with normal pressure hydrocephalus; however, further evaluation in a more heterogeneous adult population is needed. This study aims to evaluate the effect of diagnosis and valve type—fixed versus programmable—on shunt durability and cost for placement of shunts in adult patients.

METHODS

The authors retrospectively reviewed the medical records of all patients who underwent implantation of a VS for hydrocephalus at their institution over a 3-year period between August 2013 and October 2016 with a minimum postoperative follow-up of 6 months. The primary outcome was shunt revision, which was defined as reoperation for any indication after the initial procedure. Supply costs, shunt durability, and hydrocephalus etiologies were compared between fixed and programmable valves.

RESULTS

A total of 417 patients underwent shunt placement during the index time frame, consisting of 62 fixed shunts (15%) and 355 programmable shunts (85%). The mean follow-up was 30 ± 12 (SD) months. The shunt revision rate was 22% for programmable pressure valves and 21% for fixed pressure valves (HR 1.1 [95% CI 0.6–1.8]). Shunt complications, such as valve failure, infection, and overdrainage, occurred with similar frequency across valve types. Kaplan-Meier survival curve analysis showed no difference in durability between fixed (mean 39 months) and programmable (mean 40 months) shunts (p = 0.980, log-rank test). The median shunt supply cost per index case and accounting for subsequent revisions was $3438 (interquartile range $2938–$3876) and $1504 (interquartile range $753–$1584) for programmable and fixed shunts, respectively (p < 0.001, Wilcoxon rank-sum test). Of all hydrocephalus etiologies, pseudotumor cerebri (HR 1.9 [95% CI 1.2–3.1]) and previous shunt malfunction (HR 1.8 [95% CI 1.2–2.7]) were found to significantly increase the risk of shunt revision. Within each diagnosis, there were no significant differences in revision rates between shunts with a fixed valve and shunts with a programmable valve.

CONCLUSIONS

Long-term shunt revision rates are similar for fixed and programmable shunt pressure valves in adult patients. Hydrocephalus etiology may play a significant role in predicting shunt revision, although programmable valves incur higher supply costs regardless of initial diagnosis. Utilization of fixed pressure valves versus programmable pressure valves may reduce supply costs while maintaining similar revision rates. Given the importance of developing cost-effective management protocols, this study highlights the critical need for large-scale prospective observational studies and randomized clinical trials of ventricular shunt valve revisions and additional patient-centered outcomes.

ABBREVIATIONS NPH = normal pressure hydrocephalus; VS = ventricular shunt.

OBJECTIVE

Ventricular shunt (VS) durability has been well studied in the pediatric population and in patients with normal pressure hydrocephalus; however, further evaluation in a more heterogeneous adult population is needed. This study aims to evaluate the effect of diagnosis and valve type—fixed versus programmable—on shunt durability and cost for placement of shunts in adult patients.

METHODS

The authors retrospectively reviewed the medical records of all patients who underwent implantation of a VS for hydrocephalus at their institution over a 3-year period between August 2013 and October 2016 with a minimum postoperative follow-up of 6 months. The primary outcome was shunt revision, which was defined as reoperation for any indication after the initial procedure. Supply costs, shunt durability, and hydrocephalus etiologies were compared between fixed and programmable valves.

RESULTS

A total of 417 patients underwent shunt placement during the index time frame, consisting of 62 fixed shunts (15%) and 355 programmable shunts (85%). The mean follow-up was 30 ± 12 (SD) months. The shunt revision rate was 22% for programmable pressure valves and 21% for fixed pressure valves (HR 1.1 [95% CI 0.6–1.8]). Shunt complications, such as valve failure, infection, and overdrainage, occurred with similar frequency across valve types. Kaplan-Meier survival curve analysis showed no difference in durability between fixed (mean 39 months) and programmable (mean 40 months) shunts (p = 0.980, log-rank test). The median shunt supply cost per index case and accounting for subsequent revisions was $3438 (interquartile range $2938–$3876) and $1504 (interquartile range $753–$1584) for programmable and fixed shunts, respectively (p < 0.001, Wilcoxon rank-sum test). Of all hydrocephalus etiologies, pseudotumor cerebri (HR 1.9 [95% CI 1.2–3.1]) and previous shunt malfunction (HR 1.8 [95% CI 1.2–2.7]) were found to significantly increase the risk of shunt revision. Within each diagnosis, there were no significant differences in revision rates between shunts with a fixed valve and shunts with a programmable valve.

CONCLUSIONS

Long-term shunt revision rates are similar for fixed and programmable shunt pressure valves in adult patients. Hydrocephalus etiology may play a significant role in predicting shunt revision, although programmable valves incur higher supply costs regardless of initial diagnosis. Utilization of fixed pressure valves versus programmable pressure valves may reduce supply costs while maintaining similar revision rates. Given the importance of developing cost-effective management protocols, this study highlights the critical need for large-scale prospective observational studies and randomized clinical trials of ventricular shunt valve revisions and additional patient-centered outcomes.

ABBREVIATIONS NPH = normal pressure hydrocephalus; VS = ventricular shunt.

Hydrocephalus is a broad condition encompassing the extraneous entrapment of CSF within the ventricular system of the brain. It is most commonly described in the pediatric population, although hydrocephalus in adults is common as well, with an incidence of approximately 17 patients per year per 100,000 and a prevalence of approximately 1%–1.5%.2,3 Approximately 5.5 patients per 100,000 undergo ventricular shunt (VS) implantation for the treatment of hydrocephalus annually, representing one of the most regularly performed neurosurgical procedures.4,17,19 Unfortunately, postoperative complications such as subdural hematoma, infection, and mechanical obstruction are common after shunt implantation. Revision rates vary across studies, ranging from as little as 13% up to 40% in adult patients at 1 year.9,13,15,17,18 Multiple valve types, including fixed pressure valves and programmable differential pressure-regulated valves, have been designed to regulate the amount of CSF drainage. Programmable valves offer the ability to noninvasively regulate CSF drainage to readjust flow after initial implantation, while fixed pressure valves require reoperation to readjust flow with a new valve.

Despite innovation in valve design, there is no clear clinical superiority between fixed and programmable pressure valves. The lack of clinical evidence makes it difficult to justify the use of one valve type over the other in the adult hydrocephalus population. A significant disadvantage of programmable valves is the increased supply cost relative to that of fixed pressure valves. Given an estimated $1.1 billion annual cost of ventricular shunt procedures in the United States,11 cost-effectiveness is an important consideration in the evaluation of treatment options for hydrocephalus. Thus, the primary aim of the present study was to analyze revision rates of fixed and programmable pressure valves in adult patients with shunt-dependent hydrocephalus with a specific focus on cost analysis. Although we do not have specific outcome measures, shunt revision is a surrogate marker for shunt malfunction as well as a requirement for a different valve. Therefore, shunt revision was used in this study as a proxy for shunt efficacy. As hydrocephalus etiology may influence shunt outcomes, the secondary aim was to examine the role of valve type on long-term failure rates across various diagnoses. The results presented herein may help improve the quality of care while decreasing overall hospital costs associated with VS implantation.

Methods

Patient Selection

We performed a retrospective review of all patients with hydrocephalus treated with a VS at our institution over a 3-year period between April 2013 and October 2016. All patients included in this study had a minimum follow-up of 6 months. As part of a departmental infection prevention and cost containment physician-awareness program, all attending physicians were informed of their personal perioperative infection rates as well as the costs of various VS brands from May 2015 to November 2015.1 However, we had no role in designating shunt types used on a case-by-case basis, and the shunts were selected at the discretion of the attending neurosurgeon. Upon review of the medical records, we were unable to discern the rationale of attending physicians for choosing one VS valve type over the other. Two different programmable and 2 different fixed shunt brands were used in this study. All fixed shunts used in this study were composed of differential pressure valves.

Data Collection

The first shunt placed for each patient within the study timeframe was termed the “index shunt.” Information regarding initial diagnosis was collected from the medical record and classified into one of the following categories: aqueductal stenosis, Chiari malformation, cerebral cyst, hemorrhage, central nervous system infection, normal pressure hydrocephalus (NPH), pseudotumor cerebri, shunt malfunction, trauma, or tumor. The “hemorrhage” category consisted of any nontraumatic intracranial hemorrhage such as subarachnoid, intraventricular, and intraparenchymal hemorrhage. The “shunt malfunction” category consisted of shunts placed within the index timeframe due to failure of a shunt placed prior to April 2013 and served solely to delineate which patients from our cohort had a prior shunt history. Furthermore, the type of shunt placed (fixed vs programmable) and the shunt cost (direct supply cost only) was obtained for each procedure. Costs consisted only of direct implant costs, such as shunt valves, catheters, and surgical tools, such as the shunt passer. Hospitalization and other operative expenses, such as operating room time and anesthesia, were excluded. Additionally, we collected the rate of subdural hematoma and hygroma formation, including the size, midline shift, course of action, and radiographic outcome.

Definitions

Each shunt operation was categorized for placement of either a fixed or programmable valve type (categorical, independent variable). Primary outcomes included shunt revision (categorical, dependent variable), the number of shunt revisions (categorical, dependent variable), and cost (continuous, dependent variable) per valve type.

Median shunt supply costs were calculated on a per-patient basis. For each patient, we initially calculated the cost of all index shunts, which were classified as either fixed or programmable. Then, both catheter and valve replacement costs were calculated for each patient requiring shunt revision. Total cost was calculated using an intention-to-treat analysis, such that the cost per patient was categorized as fixed or programmable based on the index shunt placed. To calculate total cost per patient, we added the costs from all procedures, including the index shunt, catheter revision, and valve replacement costs. Revision costs included in the total cost per patient were differentiated based on the equipment specifically used in each revision surgery, rather than utilizing an estimated cost for all revision surgeries for each valve type.

Shunt revision was defined as reoperation for any indication for shunts placed between April 2013 and November 2016. Importantly, the patient count categorized in the “shunt malfunction” group, which consisted of revisions on shunts placed prior to April 2013, was not counted toward the primary outcome (shunt revision) and was not used to compare shunt revision rates between fixed and programmable valves. Comparative analysis of revision rates between valve types on shunts placed prior to April 2013 was not performed, as the number of shunts without revision was unknown. A patient selection flowchart summarizing the classification of patients by prior shunt history, valve type, and shunt revision is illustrated in Fig. 1.

Fig. 1.
Fig. 1.

Flowchart demonstrating prior shunt history, total patient cohort, and primary outcomes including the time range of data collection. Figure is available in color online only.

Statistical Analysis

To evaluate differences in shunt revision rates across fixed and programmable pressure valves, hazard ratios were calculated using a Cox proportional hazards regression model. Hazard ratios were used given the variable follow-up times observed across patients included in this study. Kaplan-Meier survival curve analysis was used to evaluate the probability of shunt revision as a function of time, with the date of shunt revision serving as the event marker. For comparisons of continuous variables such as cost, a Wilcoxon rank-sum test was used. Continuous variables are reported as the median (interquartile range [IQR]) unless otherwise mentioned. The number of revisions per valve type was analyzed utilizing a chi-square test.

We then used a multivariate Cox proportional hazards regression model to evaluate the combinatorial effects of valve type and hydrocephalus etiology on revision rate to address the potential influence of confounding variables. In an initial univariate analysis, we evaluated the rates of various hydrocephalus etiologies between fixed and programmable valves. In a second univariate analysis, we evaluated the effects of hydrocephalus etiology on shunt revision. All significant etiologies from both univariate analyses were then used as covariates with shunt type (fixed or programmable) as predictors for shunt revision in the multivariate regression model.

Three subgroup analyses were performed in this study. In the first, we compared shunt revision rates between valve types without the shunt malfunction category, as shunt failure is known to predict future revision. The second subgroup analysis examined revision rates after excluding patients with shunt infection, as these outcomes are likely not related to valve type. Lastly, the rates of revision in fixed and programmable valves were compared within each diagnosis category to elicit any valve advantages that might be etiology specific. Statistical analysis was performed using IBM SPSS (version 24, IBM Corp.) and Matlab R2016a (MathWorks).

Results

Baseline Information

During the index time frame, 417 patients underwent initial shunt placement, consisting of 62 fixed shunts (15%) and 355 programmable shunts (85%) (Table 1). The most common indication for shunt placement was NPH (35%, 144 of 417), followed by shunt malfunction of a shunt placed prior to April 2013 (18%, 73 of 417). Within the shunt malfunction category, 29 cases were due to failure of a previous fixed shunt (40%, 29 of 73) and 41 were due to failure of a programmable shunt (56%, 41 of 73); 4% of cases had an unknown prior valve type. There was a statistically significant difference between the rates of fixed and programmable shunts implanted across etiologies such that patients with NPH were more likely to receive programmable pressure valves (p < 0.0001). Patients with previous shunt malfunction (p = 0.026) or intracranial tumors (p = 0.043) were more likely to receive fixed pressure valves.

TABLE 1.

Indications for initial shunt placement in 417 patients

DiagnosisNo. of Shunts (%)p Value
Programmable (n = 355)Fixed (n = 62) 
Aqueductal stenosis (n = 3)3 (100)0 (0)>0.99
Chiari malformation (n = 10)8 (80)2 (20)0.647
CSF leak (n = 7)4 (57)3 (43)0.069
Cysts (n = 9)7 (78)2 (22)0.627
Hemorrhage (n = 32)27 (84)5 (16)0.881
Primary CNS infection (n = 18)15 (83)3 (17)0.737
NPH (n = 144)136 (94)8 (6)<0.0001*
Other (n = 4)3 (75)1 (25)0.476
Pseudotumor cerebri (n = 53)45 (85)8 (15)0.910
Shunt malfunction (n = 73)56 (77)17 (23)0.026*
 Fixed (n = 29)16 (55)13 (45)<0.001*
 Programmable (n = 41)37 (90)4 (10)0.486
Stroke (n = 2)1 (50)1 (50)0.278
Trauma (n = 23)21 (91)2 (9)0.553
Tumor (n = 39)29 (74)10 (26)0.043*

Statistically significant at p < 0.05.

Two cases of postoperative hydrocephalus, 1 case of congenital hydrocephalus, and 1 case of syringomyelia.

Shunt Revisions

A total of 509 procedures were performed (431 programmable, 78 fixed), with a 6-month minimum follow-up. The shunt revision rate was 22% (94 of 431) for programmable pressure valves and 21% (16 of 78) for fixed pressure valves (HR 1.1 [95% CI 0.6–1.8]; Table 2). Indications for shunt revision included 37 valve failures (7%), 34 catheter failures (7%), 28 shunt infections (6%), 2 cases of shuntalgia (0.4%), 2 cases of integumentary shunt erosion (0.4%), and 5 cases of overshunting (1%). Two shunts (0.4%) were removed because of persistent symptoms, although no specific malfunction could be identified. No significant differences in valve failures (HR 1.0 [95% CI 0.4–2.5]), catheter failures (HR 1.1 [95% CI 0.4–2.8]), or any of the failure modes between valve types could be appreciated. Overshunting was experienced in 1% of both programmable shunts (4 of 431) and fixed shunts (1 of 78) (HR 1.3 [95% CI 0.2–11.3]). Shunt infection was observed in 5% of programmable shunts (23 of 431) and in 6% of fixed pressure shunts (5 of 78) (HR 1.4 [95% CI 0.5–3.6]).

TABLE 2.

Summary outcomes for 509 shunt procedures during study timeframe

OutcomeNo. of Shunts (%)HR (95% CI)
Programmable (n = 431)Fixed (n = 78) 
Shunt revision94 (21.8)16 (20.5)1.1 (0.6–1.8)
Reop
 Replacement45 (10.4)9 (11.5)1.2 (0.6–2.5)
 Removal15 (3.5)2 (2.6)0.8 (0.2–3.6)
 Catheter revision29 (6.7)5 (6.4)1.1 (0.4–2.8)
 Ligation2 (0.5)0 (0.0)0.4 (0.0–3.4E+6)
 Add antisiphon device3 (0.7)0 (0.0)0.4 (0.0–3.7E+4)
Condition requiring revision
 Valve failure32 (7.4)5 (6.4)1.0 (0.4–2.5)
 Catheter failure29 (6.7)5 (6.4)1.1 (0.4–2.8)
 Infection/wound dehiscence23 (5.3)5 (6.4)1.4 (0.5–3.6)
 Shuntalgia2 (0.5)0 (0.0)0.0 (0.0–5.7E+6)
 Erosion2 (0.5)0 (0.0)0.0 (0.0–2.1+E6)
 Overshunting4 (0.9)1 (1.3)1.3 (0.2–11.3)
 Other*2 (0.5)0 (0.0)0.0 (0.0–9.4+E5)

Two patients had headache and abdominal pain.

Total revisions consisted of 54 valve replacements (11%, 54 of 509), 17 shunt removals (3%, 17 of 509), 34 catheter revisions (7%, 34 of 509), 2 shunt ligations (0.5%, 2 of 509), and 3 to add an antisiphon device (0.6%, 3 of 509). Valve replacements (n = 54) were performed due to a variety of failure modes, including infection (n = 14), shuntalgia (n = 2), erosion (n = 2), and overshunting (n = 1), in addition to valve failures (n = 35). Revision rates were also found to be similar between valve types after excluding patients in the shunt malfunction category (Table 3). The majority of patients who experienced shunt failure required only one revision, although 17 of the total 417 patients (4%) required more than one revision (Table 4). There was no significant difference in the number of revisions between fixed and programmable pressure valves (p = 0.935, chi-square test). Kaplan-Meier survival curve analysis showed no difference in durability between fixed shunts (mean 39 months) and programmable shunts (mean 40 months) in all patients (p = 0.980, log-rank test) and in those without shunt infection (p = 0.989, log-rank test) (Fig. 2).

TABLE 3.

Summary outcomes for 406 shunt procedures excluding shunt malfunction category

OutcomeNo. of Shunts (%)HR (95% CI)
Programmable (n = 350)Fixed (n = 56) 
Shunt revision66 (18.9)11 (19.6)1.2 (0.6–2.3)
Reop
 Replacement28 (8.0)6 (10.7)1.6 (0.7–3.9)
 Removal13 (3.7)1 (1.8)0.6 (0.1–4.3)
 Catheter revision21 (6.0)4 (7.1)1.4 (0.5–4.0)
 Ligation2 (0.6)0 (0.0)0.0 (0.0–9.2E+6)
 Add antisiphon device2 (0.6)0 (0.0)0.0 (0.0–2.3E+6)
Condition requiring revision
 Valve failure20 (5.7)4 (7.1)1.5 (0.5–4.3)
 Catheter failure21 (6.0)4 (7.1)1.4 (0.5–4.0)
 Infection/wound dehiscence16 (4.6)2 (3.6)1.0 (0.2–4.2)
 Shuntalgia1 (0.3)0 (0.0)0.0 (0.0–1.7E+11)
 Erosion2 (0.6)0 (0.0)0.0 (0.0–7.6E+6)
 Overshunting4 (1.1)1 (1.8)1.5 (0.2–13.0)
 Other*2 (0.6)0 (0.0)0.4 (0.0–3.3E+6)

Two patients had headache and abdominal pain.

TABLE 4.

Number of shunt revisions per patient by valve type

No. of FailuresNo. of Shunts (%)p Value
Programmable (n = 355)Fixed (n = 62) 
140 (11)8 (13)0.935
211 (3)2 (3)
31 (0.3)0 (0)
53 (1)0 (0)
Fig. 2.
Fig. 2.

Probability of shunt survival with time in months as estimated by Kaplan-Meier survival curve analysis for fixed and programmable valve types.

Subdural Hygroma and Hematoma Formation

The rate of subdural fluid collection, including all hygromas and hematomas, was 12% (51 of 431) for programmable valves and 5% (4 of 78) for fixed pressure valves (p = 0.110) (Table 5). Subdural hygromas were observed in 7% (29 of 431) of programmable valves and 4% (3 of 78) of fixed pressure valves (p = 0.451). The subdural hematoma rate was 6% (28 of 431) and 1% (1 of 78) for programmable and fixed pressure valves, respectively (p = 0.105). There were no significant differences in the time to formation (p = 0.906), size of collection (p = 0.349), or midline shift (p = 0.569) between valve types. Sixteen patients with programmable valves (4%, 16 of 431) were treated by adjusting the valve pressure. Nine patients with programmable valves (2%, 9 of 431) and 1 patient with a fixed pressure valve (1%, 1 of 78) required surgical intervention for a subdural hygroma or hematoma, either in the form of evacuation or shunt revision (p > 0.99).

TABLE 5.

New-onset subdural hematoma and hygroma formation after shunting

VariableProgrammable (n = 431)Fixed (n = 78)p Value
Any subdural collection51 (12)4 (5)0.110
Subdural hygroma29 (7)3 (4)0.451
 After craniotomy5 (1)0 (0)>0.99
 Symptomatic6 (1)2 (3)0.620
Subdural hematoma28 (6)1 (1)0.105
 Nontraumatic17 (4)0 (0)0.703
 Symptomatic17 (4)1 (1)>0.99
Median time to formation in mos2.0 (0.6–9.1)1.6 (0.6–22.1)0.906
Median size in mm7.0 (5.0–14.0)5.0 (3.0–9.5)0.349
Median midline shift in mm0.0 (0.0–2.0)0.5 (0.0–3.0)0.569
Course of action
 Observation19 (4)3 (4)0.320
 Reprogram valve16 (4)NANA
 Surgery*9 (2)1 (1)>0.99
 Unknown7 (2)0 (0)0.633
Outcome
 Stable15 (3)2 (3)0.327
 Resolution32 (7)2 (3)>0.99
 Unknown4 (1)0 (0)>0.99

NA = not applicable.

Values are presented as the number of shunts (%) unless stated otherwise. Median values are presented as the median (IQR).

Evacuation or revision.

Shunt Supply Costs

The difference in median index shunt implant cost between fixed ($1492) and programmable ($3320) shunts was highly statistically significant (p < 0.001) (Table 6). The cost of shunt replacement was also higher for programmable shunts (median $3206) than for fixed shunts (median $1601) (p = 0.002), suggesting that surgeons tended to replace failed shunts with shunts of the same valve type. The median sum total of all supply costs after combining index and revision supply costs was estimated to be $3438 (IQR $2938–$3876) and $1504 (IQR $753–$1584) for programmable and fixed shunts, respectively (p < 0.001). Cost savings of fixed shunt implantation are thus estimated to be $1934 (56%) per patient. This figure is an underestimation, given the increased maintenance costs of programmable shunts. We did not take into account personnel costs related to checking programmable shunts after any MRI session or costs associated with reprogramming the shunts. Some of the programmable shunts also require plain radiographs of the skull, which is an uncalculated added expense. Fixed pressure valves do not require any of these additional costs.

TABLE 6.

Shunt supply costs with valve type

CostProgrammableFixedp Value
Index shunt$3320 (2907–3663)$1492 (718–1508)<0.001*
Shunt replacement$3206 (2906–3578)$1601 (1435–2186)0.002*
Catheter revision$158 (81–215)$372 (129–570)0.222
Total cost per patient$3438 (2938–3876)$1504 (753–1584)<0.001*

Costs were calculated on a per-patient basis. Values are presented as the medians (IQRs) unless stated otherwise.

Statistically significant at p < 0.05.

Risk Factor Analysis

Hydrocephalus etiologies were then assessed with respect to the frequency of shunt revision. Of all diagnosis categories, pseudotumor cerebri (HR 1.9 [95% CI 1.2–3.1]) and shunt malfunction (HR 1.8 [95% CI 1.2–2.7]) significantly increased the likelihood of requiring a shunt revision. A comparison of shunt revision rates between fixed and programmable shunts was then performed according to hydrocephalus etiology subgroups (Tables 7 and 8); however, no significant differences in revision rates were detected between valve types across all etiologies.

TABLE 7.

Frequency of shunt revisions with shunt indication and valve type

DiagnosisShunt Revision SurgeryHR (95% CI)
Fixed (% of fixed)Programmable (% of programmable) 
Aqueductal stenosis (n = 4)0/0 (0)1/4 (25)NA
Chiari malformation (n = 13)2/3 (67)3/10 (30)5.1 (0.7–37.3)
CSF leak (n = 9)0/3 (0)1/6 (17)0.0 (0.0–9.4E+5)
Cyst (n = 9)0/2 (0)0/7 (0)NA
Hemorrhage (n = 36)0/7 (0)2/29 (7)0.0 (0.0–8.5E+5)
NPH (n = 159)2/10 (20)25/149 (1)1.2 (0.3–5.1)
Primary CNS infection (n = 32)8/27 (30)1/5 (20)1.1 (0.1–8.6)
Pseudotumor cerebri (n = 71)18/55 (33)4/16 (25)0.9 (0.3–3.2)
Shunt malfunction (n = 103)5/22 (23)28/83 (34)0.7 (0.3–1.7)
Stroke (n = 2)0/1 (0)0/1 (0)NA
Trauma (n = 24)0/2 (0)4/22 (18)0.0 (0.0–4.6E+5)
Tumor (n = 42)2/31 (6)2/11 (18)3.2 (0.4–22.4)
TABLE 8.

Frequency of shunt revisions with shunt indication and valve type

DiagnosisRelative Risk of Fixed Valve*% Confidence
Aqueductal stenosis (n = 4)NANA
Chiari malformation (n = 13)2.236
CSF leak (n = 9)0.00
Cyst (n = 9)NANA
Hemorrhage (n = 36)0.00
NPH (n = 159)1.20
Primary CNS infection (n = 32)1.50
Pseudotumor cerebri (n = 71)1.30
Shunt malfunction (n = 103)0.723
Stroke (n = 2)NANA
Trauma (n = 24)0.04
Tumor (n = 42)0.40

Frequency of fixed shunt revisions divided by frequency of programmable shunt revisions.

NPH, shunt malfunction, tumor, and pseudotumor cerebri were then all used as covariates to shunt type (fixed or programmable) as inputs into a multivariate Cox proportional hazards regression model to evaluate for the combined effects of hydrocephalus etiology and valve type on shunt revision. NPH and tumor were used due to their significant association with valve type, while shunt malfunction and pseudotumor cerebri were used due to the significant effect on shunt revision. In the combined model, only previous shunt malfunction (p = 0.014, HR 1.9 [95% CI 1.1–3.3]) and pseudotumor cerebri (p = 0.038, HR 1.8 [95% CI 1.0–3.3]) exhibited a significant effect on shunt revision rate. Shunt revision rates were not significantly associated with valve type (p = 0.919, HR 1.0 [95% CI 0.6–1.7]), NPH (p = 0.771, HR 0.9 [95% CI 0.5–1.6]), or tumor (p = 0.227, HR 0.5 [95% CI 0.2–1.5]).

Discussion

We analyzed the costs and reoperation rates per diagnostic category associated with fixed and programmable pressure valves in adult patients with shunt-dependent hydrocephalus. The primary finding of this analysis was that revision rates between fixed and programmable pressure valves were similar, and this was robust across various hydrocephalus etiologies. Programmable pressure valves were observed to be substantially more expensive than fixed pressure valves; however, they did not offer any advantage from a durability perspective.

To date, there have been mixed results from studies comparing the two VS options, with most reports unable to demonstrate a difference in revision rates between valve types.5,9,10,12,14,16,19 A recent meta-analysis by Li et al. concluded that there was no difference in the 1-year shunt survival rate, overall complications rate, infection rate, or catheter-related complications rate.8 They observed that programmable pressure valves were associated with fewer revision surgeries at 1 year; however, this effect was only substantiated in patients younger than 18 years. In that report, only 3 of the 11 studies evaluated included adult patients6,7,14 and only one of the studies included adult hydrocephalus etiologies other than NPH.14

A prospective study by Farahmand et al. reported that nonprogrammable pressure valves were a risk factor for shunt revision within 6 months of initial implantation with an HR of 1.86 (95% CI 1.18–2.93).5 The study included 450 adults with hydrocephalus due to idiopathic causes (n = 125), infection (n = 19), subarachnoid hemorrhage (n = 113), cerebrovascular disease (n = 24), trauma (n = 47), tumor (n = 65), and miscellaneous causes (n = 57). In another study, Ringel et al. demonstrated that programmable pressure valves were associated with a higher rate of overshunting and subsequent subdural hematoma formation compared with fixed pressure valves in patients with communicating hydrocephalus.14 However, shunt revision rates were observed to be similar in both groups, as overshunting can be nonoperatively corrected by increasing the valve opening pressure. Weiner et al. also found no significant differences in reoperation rates, but found no association between valve type and overshunting, in contrast to the study by Ringel et al.14,16 Our study similarly showed no difference in subdural hematoma or hygroma formation in patients with programmable and nonprogrammable shunts. Given the variation in results across studies, randomized control trials that specifically compare the effectiveness and complications of shunt treatments in adult patients are needed to guide clinical decision-making regarding shunt-dependent hydrocephalus.

The role of hydrocephalus etiology in the choice of valve type may play a significant role in shunt durability. In the present study, it was observed that hydrocephalus caused by pseudotumor cerebri and prior shunt malfunction significantly increased the risk of shunt revision. However, within each of these diagnoses, as for the group as a whole, no significant difference in shunt durability between fixed and programmable pressure valves could be elicited. These findings may provide insights into the ideal treatment paradigm for increased cost-effectiveness. Other previously described risk factors for shunt revision include occipital placement, increased valve pressure (decreased flow), and ventricular catheter length.5 Furthermore, we have shown that physician awareness of operative outcomes and costs can decrease postoperative ventricular shunt infections while simultaneously decreasing costs.1

There is a substantial cost difference between implantation of the two valve types, and thus the cost of programmable pressure valve implantation should be predicated on improved clinical benefit in terms of efficacy and complications. In this study, approximately $1934 of cost savings (56%) was associated with implanting a fixed pressure valve compared with a programmable pressure valve. These results are in contrast to those of the theoretical cost analysis reported by Zemack and Romner in 2001.18 In that study, 541 patients underwent implantation of programmable pressure valves and were observed for a mean of 23 months. During follow-up, 107 nonoperative valve adjustments were made (20%) by a magnitude of 50 mm H2O. The authors concluded that programmable pressure valves are economical because the cost savings (£341,437) of nonoperative valve adjustments outweighed the increased cost (£337,608) of implantation. Notably, this effect was most profound in patients with NPH.18 The cost-effectiveness of programmable valves was contingent on the assumption that programmable valves requiring noninvasive pressure adjustment by 50 mm H2O or more would have incurred operative costs had they been fixed pressure valves. In fact, when adjustment magnitudes of 60 mm H2O were considered, programmable valves were observed to be cost-ineffective.

Several factors limit the conclusions of the present study. First, there was a preponderance of programmable shunts implanted early in the study time frame, whereas fixed shunts were predominantly implanted later. Although it is possible that the subsequent difference in follow-up duration may have allowed for a higher proportion of shunt revision in the programmable shunt group, our results suggest that it is unlikely to play a major role, as the Kaplan-Meier analysis, which estimates survival despite varying follow-up in the study cohort, could not elicit a difference in durability across valves.

The small number of patients in some of the etiological subgroups also limits the power of the analysis. Furthermore, clinical information such as baseline and postoperative neurological status was not included in this study and may further delineate the efficacy, effectiveness, and complications associated with each valve type. Shunt revision, albeit a proxy for shunt malfunction, provides some information pertaining to the durability and the effectiveness of the shunt in addressing the primary symptomatology. Additionally, shunt supply costs associated with revision surgery may be overestimated, as these included costs associated with catheter revision and ancillary surgical tools, which may or may not be related to valve failure.

Finally, the scope of this study is limited by its retrospective nature. Primarily, we were unable to discern the rationale for valve choice from the medical records. The lack of defined indications or an institutional protocol makes it difficult to identify potential subgroups that might differentially respond to various shunt types. The potential for confounding bias further limits the reliability of our conclusions. For instance, there was a preponderance of NPH patients treated with programmable shunts, which may have influenced revision rates between the two valve types. Although we used potential confounders such as NPH and prior shunt malfunction in our multivariate regression analysis, confounding biases may be better addressed via randomization of patients. Further research including prospective, protocol-driven observational cohort studies as well as multicenter, randomized controlled trials of adult hydrocephalus interventions may further improve risk stratification of patients, ultimately improving shunt revision rates, hospital costs, and clinical outcomes.

Conclusions

In a heterogeneous adult population, long-term shunt revision rates are similar across fixed and programmable shunt pressure valves. Hydrocephalus etiology may play a role in predicting requirement for shunt revision, with previous shunt revision and pseudotumor cerebri serving as unfavorable factors. With the detection of revision rates similar between valve types, it is difficult to justify election of implantation of the more expensive programmable valve without additional studies of clinical effectiveness. Utilization of fixed pressure valves versus programmable pressure valves may reduce supply costs while maintaining similar revision rates and effectiveness across various hydrocephalus etiologies. Our study provides some insight into an analysis of comparative cost and effectiveness of fixed pressure and programmable VSs. Given the limitations of this study and the significant cost differential between shunt types, further research, including prospective, observational studies and multiinstitutional randomized trials, should be performed.

Disclosures

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

Conception and design: Friedlander, Agarwal. Acquisition of data: Friedlander, Agarwal, Kashkoush, Ismail. Analysis and interpretation of data: Friedlander, Agarwal, Kashkoush, Ismail. Drafting the article: Agarwal, Kashkoush, Ismail. 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: Friedlander. Statistical analysis: Kashkoush. Administrative/technical/material support: Friedlander, Ismail. Study supervision: Friedlander.

References

  • 1

    Agarwal N, Agarwal P, Querry A, Mazurkiewicz A, Whiteside B, Marroquin OC, : Reducing surgical infections and implant costs via a novel paradigm of enhanced physician awareness. Neurosurgery [epub ahead of print], 2017

    • Search Google Scholar
    • Export Citation
  • 2

    Bergsneider M, Miller C, Vespa PM, Hu X: Surgical management of adult hydrocephalus. Neurosurgery 62 (Suppl 2):643660, 2008

  • 3

    Bir SC, Patra DP, Maiti TK, Sun H, Guthikonda B, Notarianni C, : Epidemiology of adult-onset hydrocephalus: institutional experience with 2001 patients. Neurosurg Focus 41(3):E5, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Brean A, Eide PK: Prevalence of probable idiopathic normal pressure hydrocephalus in a Norwegian population. Acta Neurol Scand 118:4853, 2008

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Farahmand D, Hilmarsson H, Högfeldt M, Tisell M: Perioperative risk factors for short term shunt revisions in adult hydrocephalus patients. J Neurol Neurosurg Psychiatry 80:12481253, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Farahmand D, Sæhle T, Eide PK, Tisell M, Hellström P, Wikkelsö C: A double-blind randomized trial on the clinical effect of different shunt valve settings in idiopathic normal pressure hydrocephalus. J Neurosurg 124:359367, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Lemcke J, Meier U, Müller C, Fritsch MJ, Kehler U, Langer N, : Safety and efficacy of gravitational shunt valves in patients with idiopathic normal pressure hydrocephalus: a pragmatic, randomised, open label, multicentre trial (SVASONA). J Neurol Neurosurg Psychiatry 84:850857, 2013

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8

    Li M, Wang H, Ouyang Y, Yin M, Yin X: Efficacy and safety of programmable shunt valves for hydrocephalus: a meta-analysis. Int J Surg 44:139146, 2017

  • 9

    Lund-Johansen M, Svendsen F, Wester K: Shunt failures and complications in adults as related to shunt type, diagnosis, and the experience of the surgeon. Neurosurgery 35:839844, 1994

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    McGirt MJ, Buck DW II, Sciubba D, Woodworth GF, Carson B, Weingart J, : Adjustable vs set-pressure valves decrease the risk of proximal shunt obstruction in the treatment of pediatric hydrocephalus. Childs Nerv Syst 23:289295, 2007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Patwardhan RV, Nanda A: Implanted ventricular shunts in the United States: the billion-dollar-a-year cost of hydrocephalus treatment. Neurosurgery 56:139145, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Pollack IF, Albright AL, Adelson PD: A randomized, controlled study of a programmable shunt valve versus a conventional valve for patients with hydrocephalus. Neurosurgery 45:13991411, 1999

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13

    Reddy GK, Shi R, Nanda A, Guthikonda B: Obstructive hydrocephalus in adult patients: the Louisiana State University Health Sciences Center-Shreveport experience with ventriculoperitoneal shunts. World Neurosurg 76:176182, 2011 (Erratum in World Neurosurg 78:e1, 2012)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Ringel F, Schramm J, Meyer B: Comparison of programmable shunt valves vs standard valves for communicating hydrocephalus of adults: a retrospective analysis of 407 patients. Surg Neurol 63:3641, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 15

    Sotelo J, Arriada N, López MA: Ventriculoperitoneal shunt of continuous flow vs valvular shunt for treatment of hydrocephalus in adults. Surg Neurol 63:197203, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    Weiner HL, Constantini S, Cohen H, Wisoff JH: Current treatment of normal-pressure hydrocephalus: comparison of flow-regulated and differential-pressure shunt valves. Neurosurgery 37:877884, 1995

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Wu Y, Green NL, Wrensch MR, Zhao S, Gupta N: Ventriculoperitoneal shunt complications in California: 1990 to 2000. Neurosurgery 61:557563, 2007

  • 18

    Zemack G, Romner B: Do adjustable shunt valves pressure our budget? A retrospective analysis of 541 implanted Codman Hakim programmable valves. Br J Neurosurg 15:221227, 2001

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19

    Ziebell M, Wetterslev J, Tisell M, Gluud C, Juhler M: Flow-regulated versus differential pressure-regulated shunt valves for adult patients with normal pressure hydrocephalus. Cochrane Database Syst Rev (5):CD009706, 2013

    • PubMed
    • Search Google Scholar
    • Export Citation

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

Contributor Notes

Correspondence Robert M. Friedlander: University of Pittsburgh Medical Center, Pittsburgh, PA. friedlanderr@upmc.edu.

INCLUDE WHEN CITING Published online May 11, 2018; DOI: 10.3171/2017.11.JNS172212.

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

  • View in gallery

    Flowchart demonstrating prior shunt history, total patient cohort, and primary outcomes including the time range of data collection. Figure is available in color online only.

  • View in gallery

    Probability of shunt survival with time in months as estimated by Kaplan-Meier survival curve analysis for fixed and programmable valve types.

  • 1

    Agarwal N, Agarwal P, Querry A, Mazurkiewicz A, Whiteside B, Marroquin OC, : Reducing surgical infections and implant costs via a novel paradigm of enhanced physician awareness. Neurosurgery [epub ahead of print], 2017

    • Search Google Scholar
    • Export Citation
  • 2

    Bergsneider M, Miller C, Vespa PM, Hu X: Surgical management of adult hydrocephalus. Neurosurgery 62 (Suppl 2):643660, 2008

  • 3

    Bir SC, Patra DP, Maiti TK, Sun H, Guthikonda B, Notarianni C, : Epidemiology of adult-onset hydrocephalus: institutional experience with 2001 patients. Neurosurg Focus 41(3):E5, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Brean A, Eide PK: Prevalence of probable idiopathic normal pressure hydrocephalus in a Norwegian population. Acta Neurol Scand 118:4853, 2008

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Farahmand D, Hilmarsson H, Högfeldt M, Tisell M: Perioperative risk factors for short term shunt revisions in adult hydrocephalus patients. J Neurol Neurosurg Psychiatry 80:12481253, 2009

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Farahmand D, Sæhle T, Eide PK, Tisell M, Hellström P, Wikkelsö C: A double-blind randomized trial on the clinical effect of different shunt valve settings in idiopathic normal pressure hydrocephalus. J Neurosurg 124:359367, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Lemcke J, Meier U, Müller C, Fritsch MJ, Kehler U, Langer N, : Safety and efficacy of gravitational shunt valves in patients with idiopathic normal pressure hydrocephalus: a pragmatic, randomised, open label, multicentre trial (SVASONA). J Neurol Neurosurg Psychiatry 84:850857, 2013

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8

    Li M, Wang H, Ouyang Y, Yin M, Yin X: Efficacy and safety of programmable shunt valves for hydrocephalus: a meta-analysis. Int J Surg 44:139146, 2017

  • 9

    Lund-Johansen M, Svendsen F, Wester K: Shunt failures and complications in adults as related to shunt type, diagnosis, and the experience of the surgeon. Neurosurgery 35:839844, 1994

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    McGirt MJ, Buck DW II, Sciubba D, Woodworth GF, Carson B, Weingart J, : Adjustable vs set-pressure valves decrease the risk of proximal shunt obstruction in the treatment of pediatric hydrocephalus. Childs Nerv Syst 23:289295, 2007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Patwardhan RV, Nanda A: Implanted ventricular shunts in the United States: the billion-dollar-a-year cost of hydrocephalus treatment. Neurosurgery 56:139145, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Pollack IF, Albright AL, Adelson PD: A randomized, controlled study of a programmable shunt valve versus a conventional valve for patients with hydrocephalus. Neurosurgery 45:13991411, 1999

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13

    Reddy GK, Shi R, Nanda A, Guthikonda B: Obstructive hydrocephalus in adult patients: the Louisiana State University Health Sciences Center-Shreveport experience with ventriculoperitoneal shunts. World Neurosurg 76:176182, 2011 (Erratum in World Neurosurg 78:e1, 2012)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Ringel F, Schramm J, Meyer B: Comparison of programmable shunt valves vs standard valves for communicating hydrocephalus of adults: a retrospective analysis of 407 patients. Surg Neurol 63:3641, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 15

    Sotelo J, Arriada N, López MA: Ventriculoperitoneal shunt of continuous flow vs valvular shunt for treatment of hydrocephalus in adults. Surg Neurol 63:197203, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    Weiner HL, Constantini S, Cohen H, Wisoff JH: Current treatment of normal-pressure hydrocephalus: comparison of flow-regulated and differential-pressure shunt valves. Neurosurgery 37:877884, 1995

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Wu Y, Green NL, Wrensch MR, Zhao S, Gupta N: Ventriculoperitoneal shunt complications in California: 1990 to 2000. Neurosurgery 61:557563, 2007

  • 18

    Zemack G, Romner B: Do adjustable shunt valves pressure our budget? A retrospective analysis of 541 implanted Codman Hakim programmable valves. Br J Neurosurg 15:221227, 2001

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19

    Ziebell M, Wetterslev J, Tisell M, Gluud C, Juhler M: Flow-regulated versus differential pressure-regulated shunt valves for adult patients with normal pressure hydrocephalus. Cochrane Database Syst Rev (5):CD009706, 2013

    • PubMed
    • Search Google Scholar
    • Export Citation

Metrics

All Time Past Year Past 30 Days
Abstract Views 439 170 0
Full Text Views 470 191 32
PDF Downloads 163 51 10
EPUB Downloads 0 0 0