Idiopathic normal, pressure hydrocephalus: A systematic review of diagnosis and outcome

OBJECTIVE: Patient selection for cerebrospinal fluid diversion is difficult, because idiopathic normal pressure hydrocephalus (INPH) mimics other neurodegenerative disorders and no findings reliably predict outcome. The literature was reviewed to identify diagnostic criteria that predict shunt response and to formulate prognostic expectations. METHODS: MEDLINE was searched, and 44 articles meeting predetermined criteria were included. RESULTS: Clinical series were frequently retrospective with small patient numbers and unstandardized outcome evaluation. Clinical findings suggestive of shunt responsiveness were the complete triad (gait disturbance, urinary incontinence, and dementia) with early gait disturbance. Degree of hydrocephalus was not correlated with clinical improvement. Reduction of the subcortical low-blood flow area was correlated with improvement in three small studies. Clinical response to prolonged cerebrospinal fluid drainage predicted shunt outcome in all cases in two small series. Overall, 59% (range, 24-100%) of patients improved after shunting, and 29% (range, 10-100%) of patients experienced prolonged improvement. Complications occurred in 38% (range, 5-100%) of patients, additional surgery was required in 22% (range, 0-47%) of patients, and there was a 6% (range, 0-35%) combined rate of permanent neurological deficit and death. CONCLUSION: Shunting INPH is associated with an approximately 29% rate of significant improvement and a 6% significant complication rate. Enlargement of the subcortical low-flow area and clinical improvement secondary to prolonged lumbar drainage may provide additive predictive value above clinical and computed tomographic criteria. A multicenter clinical trial that focuses on the value of ancillary tests, defines the clinical course of a patient with a ventriculoperitoneal shunt, and evaluates the cost effectiveness of shunting INPH is needed to better describe outcome from shunting in INPH.