It has been postulated that increased cardiac surface pressure with continuous positive airway pressure (CPAP) results in decreased left-ventricular (LV) transmural pressure. We tested this hypothesis in seven sedated, unanesthetized, and previously instrumented pigs. We measured pericardial (Pperi), LV, airway (P(aw)), and esophageal (Pes) pressures at CPAP values of 0, 4, 8, and 12 cm H2O before and after blood-volume expansion. With normovolemia, CPAP resulted in an increase in Pperi (from -2.0 +/- 7.6 mm Hg at CPAP 0 to 2.3 +/- 5.6 mmHg at CPAP 12, p < 0.05). Baseline end-diastolic Pperi rose with volume expansion from -2.0 +/- 7.6 mm Hg to 5.4 +/- 5.4 mm Hg, p < 0.05. With hypervolemia, CPAP was associated with a decrease in Pperi (from 5.42 +/- 5.4 mm Hg to 2.3 +/- 5.6 mm Hg, p < 0.05). By contrast, Pes rose equally under both conditions with CPAP. LV transmural end-diastolic pressure (TMEDP) fell significantly under normovolemic conditions (from 16.5 +/- 7.4 mm Hg at CPAP 0 to 13.6 +/- 9.0 mm Hg at CPAP 12). The changes in FRC (pneumotachometry) with CPAP were similar under both conditions. We conclude that the CPAP-induced decrease in LV volume under hypervolemic conditions cannot be explained by an increase in cardiac surface pressure. We present a model to explain the decrease in cardiac surface pressure with CPAP.