Accuracy of CSF Tap Test and Lumbar Infusion Test in Predicting Shunt Response in Idiopathic Normal Pressure Hydrocephalus: A Systematic Review and Meta-Analysis
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Idiopathic normal pressure hydrocephalus (iNPH) presents with gait disturbance, cognitive impairment, and urinary incontinence. The cerebrospinal fluid tap test (CSF-TT) and lumbar infusion test (LIT) are commonly used to predict postoperative improvement after shunt surgery; however, their validity remains debated.
Methods
We performed a systematic review and meta-analysis to assess the sensitivity and specificity of CSF-TT and LIT. The protocol was registered in PROSPERO (CRD42023454502). Reporting followed the PRISMA guidelines, and certainty of evidence was assessed using the GRADE approach.
Results
From 1762 studies, 14 were included, reporting 697 CSF-TT and 393 LIT patients with shunt surgery as the reference standard. Considerable heterogeneity existed in test protocols, timing of assessment, and outcome measures. Several studies were retrospective with a high risk of bias. Using a bivariate random-effects model, pooled sensitivity and specificity of CSF-TT were 67.5% (95% CI 52.2–79.8, I 2 82.3%) and 53.3% (40.7-65.4, I 2 49.4%), respectively. The certainty of evidence was very low for sensitivity due to bias, inconsistency, and imprecision, and low for poor specificity due to bias and inconsistency. LIT showed a pooled sensitivity of 81% (70.3–88.3, I 2 = 27.6%) and specificity of 42.8% (20.8–68.1, I 2 = 60.8%), with moderate certainty for sensitivity and poor specificity.
Conclusion
Both CSF-TT and LIT demonstrate only modest accuracy in predicting shunt outcomes. The pooled specificity of the CSF-TT is similar to a coin toss, limiting its standalone clinical utility. The moderate certainty regarding the poor specificity of the LIT highlights the need for improved prognostic models. These tests should be interpreted in conjunction with other clinical and imaging findings, rather than in isolation. We need standardized, high-quality studies to better define their diagnostic value and support shared decision-making.
