Journal of Clinical Neuroscience 24 (2016) 94–98
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Proximal ventricular shunt malfunctions in children: Factors associated with failure Bryan E. Buster, Phillip A. Bonney, Ahmed A. Cheema, Chad A. Glenn ⇑, Andrew K. Conner, Sam Safavi-Abbasi, Mason B. Andrews, Naina L. Gross, Timothy B. Mapstone Department of Neurosurgery, University of Oklahoma Health Sciences Center, 1000 N. Lincoln Boulevard, Suite 400, Oklahoma City, OK 73104, USA
a r t i c l e
i n f o
Article history: Received 23 August 2015 Accepted 29 August 2015
Keywords: Hydrocephalus Foramen of Monro Proximal failure Shunt
a b s t r a c t Ventricular shunt failures and subsequent revisions are a significant source of patient morbidity. We conducted a review of pediatric patients undergoing placement or revision of ventricular shunts at our institution between January 2007 and December 2008. Patients were followed through to July 2014. Data collected included patient demographics, shunt history and indication for procedure, approach taken for shunt placement, and location of shunt tip in relation to the foramen of Monro. Univariate and multivariate analyses were conducted to identify factors associated with proximal failure. A total of 87 procedures were identified in 40 patients, consisting of 23 initial placements and 64 revisions. Thirty-nine proximal catheter malfunctions were identified. Indications for shunt placement included Chiari II malformation (33%) and intraventricular hemorrhage (33%). Mean follow-up period was 5.5 years. Median time to shunt failure was 1.57 years. In the multivariate model, younger age at placement was associated with decreased time to proximal failure (hazard ratio [HR] = 0.80 per increasing year of age, 95% confidence interval [CI] 0.64–0.98). Both anterior approach (HR = 0.39, 95% CI 0.23–0.67) and farther distance to foramen of Monro (HR = 0.02 per increasing 10 mm, 95% CI 0.00–0.22) were associated with increased time to proximal failure when the catheter tip was located within the contralateral lateral ventricle. Optimizing outcomes in patients with shunt-dependent hydrocephalus continues to be a challenge. Despite unsatisfactory outcomes, particularly in the pediatric population, few conclusions can be drawn from studies assessing operative variables. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Shunt failures are an important cause of patient morbidity and healthcare costs. Despite technological advances in shunt valves and image guidance, overall shunt survival has increased only modestly in patients with shunt-dependent hydrocephalus over the last several decades [1,2] and these surgical adjuvants have not been shown to be effective in randomized, controlled trials [3–5]. Proximal malfunction is an important cause of shunt failure in the pediatric population. This typically occurs as a result of choroid plexus, glial tissue, connective tissues, and other substances both native and pathologic obstructing the catheter tip [6]. It is thought that catheter tip placement in the frontal horn of the lateral ventricle (LV) and anterior to the foramen of Monro decreases the likelihood of obstruction and subsequent failure. However, despite use
⇑ Corresponding author. Tel.: +1 405 271 4912. E-mail address:
[email protected] (C.A. Glenn). http://dx.doi.org/10.1016/j.jocn.2015.08.024 0967-5868/Ó 2015 Elsevier Ltd. All rights reserved.
of advanced imaging techniques, including endoscopy, ultrasound, and image guidance, the proximal failure rate for shunts within several years of the procedure exceeds 30% in the pediatric population [5,7,8]. The best practice for placing ventricular shunts remains to be elucidated. To better understand the optimal treatment of shuntdependent children, we reviewed our institution’s experience with shunt procedures in a cohort of pediatric patients. Here we present the results of our analysis and place our findings in the context of the published literature.
2. Methods A prospective and retrospective cohort study of pediatric patients admitted to The Children’s Hospital in Oklahoma City, USA, for cerebral ventricular shunt placement or revision was conducted with approval from our Institutional Review Board. Data including age, sex, etiology of hydrocephalus, etiology of shunt failure and post-operative CT imaging were collected on a cohort of
95
B.E. Buster et al. / Journal of Clinical Neuroscience 24 (2016) 94–98
pediatric patients (<18 years old at time of shunt placement) evaluated from January 2007 to December 2008. Complete records were gathered with respect to prior and subsequent shunt procedures through to July 2014. Multiple observations were included for a single patient if they underwent additional shunt revision or revisions during the period of the study. Patients with more than one concurrent intraventricular catheter were excluded. 2.1. Studied variables Surgical technique, use of image guidance, type of ventricular catheter, valve type, and timing of follow-up imaging varied by attending surgeon. Etiology of hydrocephalus was recorded as myelomeningocele/Chiari II malformation, intraventricular hemorrhage (IVH), aqueductal stenosis, encephalocele, meningitis, Chiari I malformation, other congenital, or unknown if the etiology was not clear. Shunt revisions were classified as due to proximal malfunction, distal malfunction, valve malfunction, or infection. Shunt failure was determined at time of surgery or externalization. Time to shunt failure was determined as the time between shunt placement and shunt revision or externalization. Follow-up time for each patient was computed as the time from initial shunt placement to the patient’s most recent visit to The Children’s Hospital. For the survival analysis, observations were censored at the date of most recent visit of the Children’s Hospital, for any reason, that was captured in the electronic medical record system. Distance of the tip of the ventricular shunt catheter from the foramen of Monro was determined by post-operative CT imaging. Horizontal and vertical distances were measured using 5 mm axial cuts. Absolute distance was determined using the Pythagorean theorem. 2.2. Statistical analysis Time-to-failure analyses examining factors associated with proximal shunt failure were conducted using gap-time conditional proportional hazards models for recurrent events. The data analysis was conducted with SAS version 9.3 (SAS Institute, Cary, NC, USA). For univariate and multivariate modeling, etiology of hydrocephalus was classified as myelomeningocele, IVH, or other. For the multivariate model, all covariates were initially included along with clinically reasonable two-way interactions. A variable was considered to be a confounder if its removal resulted in a clinically significant difference in interpretation of another variable. A clinically significant difference was defined a priori as P20% change in the estimate of the variable effect, or loss of statistical significance. All statistically significant covariates, interaction terms, and confounder variables were included in the final model. Only hierarchical multivariate models were considered, and any covariate involved in a significant interaction was left in the model. Only stratified results were reported for these covariates. All results for univariate and multivariate Cox regression modeling are reported with point estimates of the hazard ratio (HR) and 95% confidence intervals (CI) and p values determined using robust sandwich variance estimates. A p value 6 0.05 was considered statistically significant. 3. Results Forty patients with proximal catheter ventriculoperitoneal shunt malfunctions were included in the analysis. Patient demographics are displayed in Table 1. The most prevalent etiologies of hydrocephalus were myelomeningocele/Chiari II malformation and IVH. The patients were followed for a mean of 5.5 years after first shunt procedure.
Table 1 Characteristics of patients with proximal catheter ventriculoperitoneal shunt malfunction Characteristic
Value (%)
Patients Sex Female Male Etiology of hydrocephalus Chiari II malformation Intraventricular hemorrhage Congenital Aqueductal stenosis Encephalocele Fungal meningitis Chiari I malformation Unknown Follow-up, years (mean ± SD)
40 17 (42.5) 23 (57.5) 14 (35.0) 12 (30.0) 5 (12.5) 3 (7.5) 2 (5.0) 1 (2.5) 1 (2.5) 2 (5.0) 5.5 ± 2.8
SD = standard deviation.
The 40 patients contributed a total of 87 observations over the period of the study, which consisted of 23 initial placements and 64 revisions. The 1 year failure rate for first-time shunt placements (N = 23) was 51.3%. Of the revisions, 53 had time-to-failure data. Characteristics of the observations are displayed in Table 2. The mean age for the observations was 2.7 years (standard deviation [SD] = 3.1). The etiology in 36 of the observations (41.4%) was myelomeningocele/Chiari II malformation, followed by 24 for IVH (27.6%). Of the 53 shunt failures and subsequent revisions observed, 39 (73.6%) were attributed to proximal shunt malfunctions. Kaplan–Meier survival plot of the observations is shown in Figure 1. Median and mean time to failure were 1.57 and 2.18 years, respectively. The placement characteristics of the shunts including approach of the catheter (frontal versus occipital-parietal), distance of shunt catheter tip from the foramen of Monro ipsilateral to the entry point of the shunt, and shunt catheter tip with respect to the side of entry the shunt catheter are shown in Table 3. The average catheter tip distance from the foramen of Monro was 27.7 mm (SD = 17.1 mm). The 39 proximal failures with complete records were included along with the 34 shunts that remained failure free in univariate and multivariate Cox regression analyses. Results of univariate analysis are displayed in Table 4. Younger age of patient was Table 2 Characteristics of observations of ventriculoperitoneal shunt malfunction Characteristic
Value (%)
Observations Sex Female Male Age, years (mean ± SD) Etiology of hydrocephalus Chiari II malformation Intraventricular hemorrhage Congenital Aqueductal stenosis Chiari I malformation Fungal meningitis Encephalocele Unknown Initial placements Revisions with time to survival data Proximal malfunction Infection Distal malfunction Valve malfunction Revisions without time-to-survival data
87
SD = standard deviation.
40 (46.0) 47 (54.0) 2.7 ± 3.1 36 (41.4) 24 (27.6) 10 (11.5) 4 (4.6) 4 (4.6) 3 (3.5) 1 (1.2) 5 (5.8) 23 (26.4) 53 (60.9) 39 (73.6) 7 (13.2) 5 (9.4) 2 (3.8) 11 (12.6)
96
B.E. Buster et al. / Journal of Clinical Neuroscience 24 (2016) 94–98
Fig. 1. Kaplan–Meier analysis of shunt survival for all pediatric patients, considering all causes of shunt failures.
Table 4 Univariate survival analysis for shunts undergoing proximal failure or no failure (N = 73)
Table 3 Placement characteristics of ventriculoperitoneal shunts
y
Covariate
Value (%)
Observations Distance of catheter tip to foramen of Monro 0–10 mm 10–20 mm 20–30 mm 30–40 mm 40–50 mm 50–60 mm 60–70 mm 70–80 mm Approach of catheter Frontal Occipital-parietal Shunt tip placementy Ipsilateral lateral ventricle Contralateral lateral ventricle Third ventricle
87 16 (18.4) 18 (20.7) 14 (16.1) 18 (20.7) 11 (12.6) 5 (5.8) 4 (4.6) 1 (1.2) 48 (55.2) 39 (44.8) 65 (74.7) 14 (16.1) 8 (9.2)
With respect to approach side.
associated with proximal failure, with each increasing year of age seen to be protective (HR = 0.84 per increasing year, 95% CI 0.71– 0.98, p = 0.031). Frontal approach was associated with improved outcomes compared to occipital-parietal approach (HR = 0.48, 95% CI 0.24–0.99, p = 0.048). Etiology of hydrocephalus was also statistically significant, with non-IVH, non-Chiari II malformation etiologies associated with improved outcomes compared to IVH (HR = 0.31, 95% CI 0.13–0.76, p = 0.0133). Outcomes were similar between myelomeningocele/Chiari II malformation and IVH etiologies. Distance of shunt catheter tip from foramen of Monro, catheter tip location (with respect to side of entry of the shunt catheter), and sex were not associated with time to shunt failure. For the multivariate analysis, the following significant interactions were identified: (1) catheter entry approach and catheter tip distance from foramen of Monro; and (2) catheter entry approach and catheter tip location. There was no interaction
Covariate
Hazard ratio (95% CI)
p valuey
Distance from foramen of Monro (per 10 mm) Age at shunt placement (per year of age) Ventricular placement Ipsilateral Contralateral Third ventricle Approach of catheter Frontal Occipital-parietal Sex Female Male Etiology of hydrocephalus Intraventricular hemorrhage Chiari II malformation Other
1.13 (0.96–1.32) 0.84 (0.71–0.98)
0.1521 0.0310 0.6735
1 0.90 (0.35–2.32) 0.39 (0.05–3.09) 0.48 (0.24–0.99) 1
0.0481
0.86 (0.45–1.62) 1
0.6318 0.0133
1 1.28 (0.71–2.31) 0.31 (0.13–0.76)
y Type 3 analysis with sandwich variance estimate; values in bold are statistically significant. CI = confidence interval.
between distance and approach or distance and age. Sex and etiology did not contribute significantly to the model, nor did they confound any significant covariates or interactions in the model. As such, they were excluded. Findings for distance of catheter tip from foramen of Monro and approach of the catheter were stratified by catheter tip location (with respect to side of entry of the catheter). Results of the final model are shown in Table 5. Increasing age was again found to be protective, as in the univariate analysis. Accounting for distance, catheter approach, and ventricular placement, each additional year of age was associated with a 20% lower risk of shunt failure (HR = 0.80, 95% CI 0.64–0.98). Interestingly, increasing distance of the catheter tip from the foramen of Monro appeared to be protective when the catheter tip was in the contralateral LV. Accounting for age and catheter approach, each additional 10 mm of distance resulted in a 60%
B.E. Buster et al. / Journal of Clinical Neuroscience 24 (2016) 94–98 Table 5 Multivariate survival analysis for shunts undergoing proximal failure or no failure (N = 73) Covariate Age at shunt placement (per year of age) Distance from foramen of Monro (per 10 mm) by catheter tip location Ipsilateral lateral ventricle Third ventricle Contralateral lateral ventricle Anterior approach (vs. posterior) by catheter tip location Ipsilateral lateral ventricle Third ventricle Contralateral lateral ventricle
Hazard ratio (95% CI)
p valuey
0.80 (0.64–0.98)
0.0330 0.0004
1.01 (0.81–1.27) 0.02 (0.00–2.61) 0.39 (0.23–0.67) 0.0156 0.62 (0.25–1.53) 0.62 (0.25–1.53) 0.02 (0.00–0.22)
y Type 3 analysis with sandwich variance estimate; values in bold are statistically significant. CI = confidence interval, vs. = versus.
reduction in the risk of shunt failure (HR = 0.39, 95% CI 0.23–0.67). In our dataset, there appeared to be no relation between the catheter tip distance and the time to failure when the catheter tip was in the ipsilateral LV or the third ventricle when accounting for age and catheter approach. In the univariate analysis, frontal approach was associated with a nearly 50% reduction in the risk of shunt failure. In the multivariate analysis, when accounting for patient age and catheter tip distance from foramen of Monro, a frontal approach is quite protective in the case where the catheter tip is in the contralateral LV (HR = 0.02, 95% CI 0.00–0.22), but not in the cases where the catheter tip is in the ipsilateral LV or the third ventricle. 4. Discussion In a group of 40 pediatric patients, median time to shunt failure was 18.8 months. Twelve month shunt survival of patients receiving an initial shunt operation was 48.7%. The mean follow-up for all patients was 5.5 years. Nearly three-quarters of revisions were due to proximal malfunctions. Assessing variables associated with time to proximal failure, younger age, occipital-parietal approach, and both IVH and Chiari II malformation were associated with proximal malfunctions on univariate analysis. On multivariate analysis, age remained significant, while etiology of hydrocephalus dropped out of the model. Approach remained significant, but only for catheter tips located in the contralateral LV. Distance to foramen of Monro, not statistically significant in the univariate model, was significant within the contralateral LV. Surprisingly, farther distance from the foramen of Monro was seen to be protective. 4.1. Patient factors The association of decreased shunt survival in younger pediatric patients has been identified previously. In a large review, age less than 40 weeks gestation and age between 40 weeks gestation and 1 year were significant risk factors for failure compared to age older than 1 year (HR 2.49 and 1.77, respectively) [9]. Similarly, others have found age less than 40 weeks and decreasing age were significant risk factors for failure (HR = 2.15 and 1.04/year, respectively) [10]. In another large retrospective review, however, age groups younger than 9 years were no less likely to fail than children 9 years and older [11]. Each of these studies considered all failures in their analyses, rather than only proximal failures. Studies have also assessed the role of etiology of hydrocephalus in shunt outcomes. All non-congenital hydrocephalus, including IVH, myelomeningocele, and others have previously been identified as predictors of poorer shunt survival, with HR ranging from 1.56 to 2.80 [9]. Despite this, others have noted no association
97
between etiology and shunt failure [10]. Lastly, IVH has been identified as a risk factor for repeat surgery within a year of initial placement (adjusted odds ratio = 2.2) [11]. 4.2. Surgical factors Unfortunately, most studied factors affecting shunt outcomes are patient dependent and are thus determined at the time of presentation. Only a few surgical factors have been shown to improve outcomes in any study. While a better understanding of the patient population at risk for shunt failures is helpful, this ultimately does not dictate how to improve care of these patients, beyond a more accurate depiction of the risk profile for each patient. Because of this, a better understanding of how to prolong shunt life with surgical techniques is needed. Regarding technique-dependent factors, a recent systematic review [12] analyzed the effect of catheter entry point on shunt outcomes. They identified five relevant articles [13–17] and concluded that there was insufficient evidence to recommend one approach over the other, based on the inconsistent findings and the fact that all but one study was retrospective. Four of the articles suggested that the parietal-occipital approach was superior to the frontal approach, and one concluded the opposite. While several authors have concluded that catheter tips surrounded by cerebrospinal fluid (CSF) are superior to non-CSF surrounded catheter tips, [17] no study to our knowledge has specifically addressed the effect of approach stratified by the ventricle in which the tip rests. Interestingly, we found that the frontal approach is optimal compared the parietal-occipital approach when the catheter tip is contralateral to the entry side, while there was no difference based on approach when the tip was in the ipsilateral LV. There was no difference in outcome based solely on which ventricle the tip ended up in, contralateral versus ipsilateral. 4.3. Limitations An important limitation of our study is the retrospective nature of the analysis and the modest sample size. Our results must be interpreted accordingly. In addition, we did not study several variables that have been identified by others, including the effect of the surgeon, [18,19] the time of year of the surgery and the effect of increasing shunt procedures [10]. 5. Conclusions Optimizing outcomes in patients with shunt-dependent hydrocephalus continues to be a challenge. Our finding that increasing distance of the ventricular catheter tip from the foramen of Monro in the contralateral ventricle is protective from shunt malfunction requires additional study as the reasons for this are unclear. Despite unsatisfactory outcomes, particularly in the pediatric population, few conclusions can be drawn from studies assessing operative variables. More studies and better techniques are needed. Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. References [1] Kulkarni AV, Riva-Cambrin J, Butler J, et al. Outcomes of CSF shunting in children: comparison of Hydrocephalus Clinical Research Network cohort with historical controls: clinical article. J Neurosurg Pediatr 2013;12:334–8. [2] Stein SC, Guo W. Have we made progress in preventing shunt failure? A critical analysis. J Neurosurg Pediatr 2008;1:40–7.
98
B.E. Buster et al. / Journal of Clinical Neuroscience 24 (2016) 94–98
[3] Flannery AM, Duhaime AC, Tamber MS, et al. Pediatric hydrocephalus: systematic literature review and evidence-based guidelines. Part 3: endoscopic computer-assisted electromagnetic navigation and ultrasonography as technical adjuvants for shunt placement. J Neurosurg Pediatr 2014;14:24–9. [4] Drake JM, Kestle JR, Milner R, et al. Randomized trial of cerebrospinal fluid shunt valve design in pediatric hydrocephalus. Neurosurgery 1998;43:294–303 [discussion 303–5]. [5] Kestle JR, Drake JM, Cochrane DD, et al. Lack of benefit of endoscopic ventriculoperitoneal shunt insertion: a multicenter randomized trial. J Neurosurg 2003;98:284–90. [6] Sekhar LN, Moossy J, Guthkelch AN. Malfunctioning ventriculoperitoneal shunts. Clinical and pathological features. J Neurosurg 1982;56:411–6. [7] Levitt MR, O’Neill BR, Ishak GE, et al. Image-guided cerebrospinal fluid shunting in children: catheter accuracy and shunt survival. J Neurosurg Pediatr 2012;10:112–7. [8] Villavicencio AT, Leveque JC, McGirt MJ, et al. Comparison of revision rates following endoscopically versus nonendoscopically placed ventricular shunt catheters. Surg Neurol 2003;59:375–9 [discussion 379–80]. [9] Tuli S, Drake J, Lawless J, et al. Risk factors for repeated cerebrospinal shunt failures in pediatric patients with hydrocephalus. J Neurosurg 2000;92: 31–8. [10] McGirt MJ, Leveque JC, Wellons 3rd JC, et al. Cerebrospinal fluid shunt survival and etiology of failures: a seven-year institutional experience. Pediatr Neurosurg 2002;36:248–55.
[11] Simon TD, Whitlock KB, Riva-Cambrin J, et al. Association of intraventricular hemorrhage secondary to prematurity with cerebrospinal fluid shunt surgery in the first year following initial shunt placement. J Neurosurg Pediatr 2012;9:54–63. [12] Kemp J, Flannery AM, Tamber MS, et al. Pediatric hydrocephalus: systematic literature review and evidence-based guidelines. Part 9: Effect of ventricular catheter entry point and position. J Neurosurg Pediatr 2014;14:72–6. [13] Albright AL, Haines SJ, Taylor FH. Function of parietal and frontal shunts in childhood hydrocephalus. J Neurosurg 1988;69:883–6. [14] Bierbrauer KS, Storrs BB, McLone DG, et al. A prospective, randomized study of shunt function and infections as a function of shunt placement. Pediatr Neurosurg 1990;16:287–91. [15] Nakahara K, Shimizu S, Utsuki S, et al. Shortening of ventricular shunt catheter associated with cranial growth: effect of the frontal and parieto-occipital access route on long-term shunt patency. Childs Nerv Syst 2009;25:91–4. [16] Sainte-Rose C, Piatt JH, Renier D, et al. Mechanical complications in shunts. Pediatr Neurosurg 1991–1992;17:2–9. [17] Tuli S, O’Hayon B, Drake J, et al. Change in ventricular size and effect of ventricular catheter placement in pediatric patients with shunted hydrocephalus. Neurosurgery 1999;45:1329–33 [discussion 1333–5]. [18] Korinek AM, Fulla-Oller L, Boch AL, et al. Morbidity of ventricular cerebrospinal fluid shunt surgery in adults: an 8-year study. Neurosurgery 2011;68:985–94 [discussion 994–5]. [19] Liptak GS, McDonald JV. Ventriculoperitoneal shunts in children: factors affecting shunt survival. Pediatr Neurosci 1985–1986;12:289–93.