The effects on shunt flow from the position of an antisiphon device (ASD) and from changes in posture in hydrocephalic patients were examined. Fifty patients with hydrocephalus (including 36 with normal pressure hydrocephalus) were investigated, using quantitative radionuclide shuntography (99mtechnetium-pertechnetate) in the supine, sitting, and standing positions. The types of shunt valve used were as follows: Mishler dual chamber low pressure without ASD (16 cases), with ASD 40 cm below the level of the foramen of Monro (three cases), and with ASD 10 cm below the level of the foramen of Monro (12 cases); low pressure with integral ASD (14 cases); and medium pressure with integral ASD (five cases). In patients with a low pressure valve without ASD, shunt flow was least in the supine position (0.0011 ml/min) but increased significantly in the sitting position (0.4381 ml/min, P less than 0.001) because of the siphon effect. Conversely, in patients with a low pressure valve with integral ASD, shunt flow was maximal in the supine position (0.1056 ml/min) and decreased significantly in the sitting position (0.0017 ml/min, P less than 0.001), indicating overfunction of the ASD. Intracranial pressure (ICP) in the supine position increased significantly compared with patients with a low pressure valve without ASD (93.6 and 20.7 mm H2O, respectively, P less than 0.01). Intermediate values for shunt flow in the supine and sitting positions (0.0279 and 0.0896 ml/min, respectively) and for ICP (55.8 mm H2O) were obtained with patients with a low pressure valve with the ASD 10 cm below the level of the foramen of Monro (P less than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)