Positive end-expiratory pressure (PEEP) has been used to prevent and treat venous air embolism in patients in the seated position undergoing neurosurgical operations. However, the safety of PEEP has recently been questioned, because of concern that PEEP might increase right atrial pressure (RAP) more than left atrial pressure, thereby predisposing patients with a probe-patent foramen ovale to paradoxical air embolism. In a prior study in dogs, the authors found that to up 10 cm H2O PEEP did not affect the interatrial pressure difference. In the present study, the authors examined the effects of 0, 5, and 10 cm H2O PEEP in 12 anesthetized neurosurgical patients positioned both supine and seated prior to operation. Measurements were made of systemic arterial pressure, RAP, mean pulmonary artery pressure (PAP), pulmonary artery wedge pressure (PAWP), and cardiac output. PAWP was higher (average 2 mmHg) than RAP in all patients. PEEP increased RAP and PAWP in patients, both seated and supine (mean 3 mmHg at 10 cm H2O), but did not affect the PAWP-RAP difference. In an additional eight patients in the seated position, the authors examined the effects of 0, 10, and 20 cm H2O PEEP during operation. PEEP again increased PAWP and RAP, but did not significantly affect the PAWP-RAP difference. The PAWP-RAP difference became negative (-1 mmHg) in one patient with 20 cm H2O PEEP. The authors conclude that levels of PEEP up to 10 cm H2O do not alter the interatrial pressure difference in seated neurosurgical patients, and, therefore, would not predispose these patients to paradoxical air embolism.