An important goal in managing patients with respiratory failure using mechanical ventilatory support and positive end-expiratory pressure (PEEP) is to optimize tissue oxygen delivery relative to oxygen consumption. To this end, systemic hypothermia has been reported to reduce oxygen consumption. Cooling, however, may antagonize hypoxic pulmonary vasoconstriction and depress cardiac output. To determine whether these potentially adverse effects of cooling on tissue oxygen delivery would outweigh any potential benefits, we studied the effects of systemic hypothermia and end-expiratory pressure on venous admixture, intrapulmonary blood distribution, and oxygenation variables in 40 dogs with oleic acid-induced pulmonary edema of the right lung. The dogs were randomly assigned to four treatment groups of 10 dogs each: normothermia and zero end-expiratory pressure (ZEEP); normothermia and 10 cm H2O PEEP; hypothermia and ZEEP; hypothermia and PEEP. Hypothermia to 31.9 +/- 0.1 degree C (mean +/- SEM) caused no adverse effects on intrapulmonary blood flow distribution (measured by radioactive microspheres) or on venous admixture. Tissue oxygen delivery and arterial oxygenation did not improve with hypothermia, the latter being 109 +/- 13 mm Hg and 70 +/- 8 mm Hg with PEEP and ZEEP, respectively. However, hypothermia significantly reduced oxygen consumption, so that the coefficient of oxygen delivery (i.e., the ratio of oxygen supply to consumption) increased from 2.5 +/- 0.1 to 3.2 +/- 0.2 (p less than 0.01) with ZEEP and from 2.0 +/- 0.1 to 2.6 +/- 0.3 with PEEP (p = 0.016). Thus, although systemic hypothermia failed to improve arterial oxygenation and tissue oxygen delivery, it decreased systemic oxygen demands, thereby improving the oxygen supply-demand balance.