Animal experiments suggest that the application of positive end-expiratory pressure (PEEP) levels > or = 10 cm H2O increase intraocular pressure (IOP), probably through ophthalmic and episcleral venous outflow obstruction secondary to PEEP-induced increase in central venous pressure (CVP). To evaluate whether a similar response occurs in humans, we studied the effects of varying levels of PEEP on IOP in 11 critically ill sedated and mechanically ventilated patients, aged 35 to 88 yrs (mean: 68 yrs), without evidence of ocular disease. Measured variables included PaCO2, PaO2, mean arterial pressure (MAP), CVP and IOP, and were recorded at zero end-expiratory pressure and at 5, 10 and 15 cm H2O PEEP, applied in random order. IOP was measured by the same investigator at the bedside, using a portable and battery-operated tonometer (Tono-Pen). As expected, PaO2 increased significantly from baseline at all PEEP levels. At the three levels of PEEP, no significant change of PaCO2 from its corresponding baseline values was observed. At 5 cm H2O PEEP neither CVP nor IOP raised significantly from baseline. IOP increased significantly (p < 0.01) from 12 +/- 4 to 14 +/- 4 mmHg at 10 cm H2O PEEP and from 13 +/- 4 to 16 +/- 5 mmHg at 15 cm H2O PEEP. CVP also increased significantly (p < 0.01) and in parallel with IOP at 10 and 15 cm H2O PEEP. Since IOP tends to rise significantly when PEEP levels exceed 10 cm H2O, PEEP-ventilated patients in whom such increase is undesirable might benefit from regular bedside IOP monitoring.