OBJECTIVE To demonstrate that pentoxifylline (PTX) and not placebo improves oxygen consumption (VO2) in critically ill patients with severe sepsis. METHODS Multidisciplinary intensive care unit in a university affiliated hospital. METHODS A randomized, double blinded clinical trial comparing 300 mg of PTX administered in a 120 min iv infusion with an identically looking placebo. METHODS 13 patients (9 men and 4 women) average age 39 (24-62) years old received PTX, and 12 patients (5 men and 7 women) average age 38 (21-83) years old received placebo. All satisfied ACCP/SCCM criteria for severe sepsis. METHODS Patients fulfilling criteria for severe sepsis was identified on admission, cardiac output, DO2 and VO2 were measured by thermodilution and standard oximetric technics after adequate volume replacement at baseline, 60 and 120 during infusion. F-test of analysis of variance was used to test hypothesis about differences of DO2 and VO2 by group, by time and for the interaction terms, a "p" value < 0.05 was considered significant. RESULTS Evaluation of baseline measurements of both groups revealed no significant difference in any haemodynamic function or oxygen transport variables. The average VO2 difference (0-120 min) between groups was 21 mL/min.m2 and it was higher in the experimental group, however, this difference was non significant. CONCLUSIONS We think that the trend in VO2 followed by the experimental group is clinically important. If this difference is sustained we will probably be able to demonstrate our hypothesis.