Alzheimer’s disease biomarkers, DESH-like imaging phenotype, and tap test outcomes in idiopathic normal pressure hydrocephalus: a retrospective observational study

Idiopathic normal pressure hydrocephalus is clinically heterogeneous, and Alzheimer’s disease-related pathology is common in this syndrome. It remains unclear whether amyloid/tau (A/T) biomarker stratification mainly reflects baseline clinical and imaging heterogeneity or also predicts short-term cerebrospinal fluid tap-test responsiveness. This single-center retrospective study included 63 patients with possible idiopathic normal pressure hydrocephalus from the Peking Union Medical College Hospital dementia cohort. Patients were classified into A − T−, A + T−, and A + T+ groups using available CSF or positron emission tomography biomarkers. Baseline clinical features, disproportionately enlarged subarachnoid-space hydrocephalus-like imaging features, and 72-hour tap-test outcomes, including gait and executive-domain measures, responder rates, and response patterns, were compared across groups. A/T stratification was associated with distinct baseline phenotypes. Compared with the A − T− group, the A + T+ group showed greater cognitive impairment and a higher burden of psychiatric symptoms, whereas the A − T− group more closely resembled a typical hydrocephalus phenotype. Imaging features also differed substantially: the A − T− group had the highest disproportionately enlarged subarachnoid-space hydrocephalus burden, whereas the A + T+ group had the lowest. In contrast, 72-hour tap-test outcomes were not clearly stratified by A/T status. Gait-related outcomes showed only directional trends favoring the A − T− group, and neither A/T status nor imaging burden showed robust independent associations with short-term benefit in exploratory analyses. A/T biomarker stratification in idiopathic normal pressure hydrocephalus primarily reflects baseline clinical and imaging heterogeneity rather than clearly predicting short-term tap-test responsiveness. AD-related pathology may be associated with a modified hydrocephalus phenotype but should not be used alone to exclude potentially responsive patients.