Diagnosing Cerebrospinal Fluid Shunt Failure: Systematic Review and Meta-Analysis of Noninvasive and Minimally Invasive Tests in Children and Adults.

BACKGROUND AND OBJECTIVES: The goal of this study was to assess the diagnostic performance of noninvasive and minimally invasive tools for the diagnosis of shunt failure in pediatric and adult patients. METHODS: Eligible studies included patients with permanent cerebrospinal fluid shunt systems who underwent diagnostic testing (X-ray shunt series, computed tomography [CT], MRI, nuclear scintigraphy, or thermal flow evaluation) for possible shunt failure. A systematic review and meta-analysis was conducted (PROSPERO registration number CRD42024596142). Performance measures (accuracy, sensitivity, and specificity) were analyzed using a meta-regression model. The model incorporated index test type, age group, follow-up time frame, and shunt type. Odds ratios were calculated for each level of the analyzed variables. A bias assessment of each study was conducted using Quality Assessment Tool for Diagnostic Accuracy Studies-2 and Quality Assessment Tool for Diagnostic Accuracy Studies-Comparative, when applicable. RESULTS: In total, 465 articles were reviewed (PubMed search October 22, 2024) with 33 included in the final analysis. There was a significant effect of index test type on sensitivity ( P = .013). In the subgroup analysis, X-ray shunt series performed worse than CT (reference group) (odds ratio 0.106, 95% CI [0.056-0.201]). There was no overall effect of index test type on specificity ( P = .192), but, in the subgroup analysis, X-ray shunt series performed better than CT (odds ratio 4.04, 95% CI [1.938-8.424]). There were no significant effects of age group, follow-up time frame, or shunt type on any of the performance measures. CONCLUSION: This is the largest systematic review and meta-analysis conducted to date to analyze the performance of minimally invasive tools to diagnose shunt failure. None of the tests performed well enough to be considered a stand-alone test for the diagnosis of shunt failure, and only X-ray shunt series performed significantly different than CT. Limitations include the high risk of bias in the reference standard domain for most studies as identified using Quality Assessment Tool for Diagnostic Accuracy Studies-2.

Duke Scholars

Author Gerald Arthur Grant Neurosurgery