OBJECTIVE To examine whether a transmantle pressure gradient exists in adult patients with communicating and noncommunicating hydrocephalus. METHODS Ten patients participated in the study. The mean patient age was 57 +/- 18 years (range, 20-80 yr); seven patients had communicating hydrocephalus, and three had noncommunicating hydrocephalus. Microsensors were used to measure the intracranial pressure (ICP), for 17 to 24 hours during sleeping and waking periods, in the right lateral ventricle (ICP(IV)) and in the subarachnoid space (ICP(SAS)) over the right cerebral convexity simultaneously. Patient activities and body positions were documented. The hydrostatic pressure difference between the two sensors was calculated from cranial x-rays for four basic body positions and compared with the actual body positions of the patients and the measured difference between the two sensors. For three 10-minute periods, the exact transmantle pressure gradient was calculated for each patient as ICP(IV) - ICP(SAS), adjusted for the hydrostatic pressure difference. RESULTS The measured pressure difference between the two sensors was always within the limits of the maximal possible hydrostatic pressure difference, and it correlated well with the expected difference for the various body positions: mean correlation coefficient, 0.79 +/- 0.10 (range, 0.65-0.92). The exact mean transmantle pressure was -0.01 +/- 0.24 mmHg (range, -0.4 to 0.4 mmHg). ICP waves caused by cardiac pulse, respiration, and B waves were identical in both spaces. CONCLUSIONS This study demonstrates no factual support for existence of a transmantle pressure gradient in nonacute communicating or noncommunicating hydrocephalus.