We tested the hypothesis that the concentration of atrial natriuretic factor in the cerebrospinal fluid is an indicator of brain injury in patients with intracranial disease. Atrial natriuretic factor concentration was measured in 72 samples of cerebrospinal fluid from 28 patients with intraventricular drains and in nine samples from outpatient controls undergoing diagnostic lumbar puncture. Levels were correlated with diagnosis; systemic fluid administration; concentration of atrial natriuretic factor in the plasma; intracranial pressure; sodium, glucose, and protein concentrations, osmolality, and cell count in the cerebrospinal fluid; sodium concentration in the serum; and hemodynamics. Atrial natriuretic factor concentration was highest in cerebrospinal fluid from patients with intracerebral hematoma, followed by those with obstructive hydrocephalus and subarachnoid hemorrhage (19 +/- 2, 13 +/- 3, and 8 +/- 2 pg/ml, respectively); atrial natriuretic factor concentration was less than 4 pg/ml in the controls. Patients treated with fluid restriction had significantly higher atrial natriuretic factor levels than those receiving maintenance or high-volume fluids (16 +/- 3, 8 +/- 2, 10 +/- 1 pg/ml, respectively). The concentration of atrial natriuretic factor in the plasma was significantly elevated in patients with intracerebral hematoma and subarachnoid hemorrhage (155 +/- 38 and 92 +/- 20 pg/ml, respectively) and did not correlate with fluid administration or the concentration of atrial natriuretic factor in the cerebrospinal fluid. Neither cerebrospinal fluid nor plasma concentrations of atrial natriuretic factor correlated with intracranial pressure; cerebrospinal fluid sodium, glucose, or protein concentrations, osmolality, or cell count; serum sodium concentration; or hemodynamics. We conclude that the concentration of atrial natriuretic factor in the cerebrospinal fluid is a nonspecific indicator of brain injury.(ABSTRACT TRUNCATED AT 250 WORDS)