Abstract Aim To produce a systematic review and meta-analysis to identify radiological features that predict shunt response in normal pressure hydrocephalus (NPH), evaluate their effectiveness and therefore recommend the most predictive feature. Method The authors searched MEDLINE, Embase, Scopus, PubMed, Google scholar and JSTOR databases for all original studies investigating NPH radiological features that predict shunt response. Selected studies were assessed with ROBINS-1 tool and included studies were evaluated by a univariate meta-analysis. Results 301 full text papers were selected from screening, 28 studies were included, and 26 features were identified, of these 5 were selected for the meta-analysis based off the inclusion criteria: disproportionately enlarged subarachnoid space (DESH), callosal angle, periventricular white matter changes, cerebral blood flow (CBF), and computerized tomography cisternography. The diagnostic odds ratios of colossal angle (1.88) and periventricular white matter changes (1.01) showed they were significant differentiators between shunt responders from non-responders. None of the other features could differentiate between the two. Conclusions Colossal angle and periventricular white matter changes are the only statistically effective features which can identify shunt responders. Their role, however, may be overstated as their DORs approximate 1, therefore it is advised they be used in combination with other diagnostic predictors, such as clinical tests, of shunt responders. More research is required to evaluate the combined use of multiple radiological features together.
Read full abstract