ObjectivesTo evaluate the differences in efficacy and in clinical and biochemical tolerance to amphotericin B administered in a lipid emulsion compared with amphotericin B administered in 5% dextrose in water in the treatment of Candida albicans infection in intensive care unit (ICU) patients.DesignProspective, controlled, randomized study, conducted during a 2.5-yr period, comparing the two treatment protocols.SettingGeneral ICU of a university-affiliated municipal hospital.PatientsSixty consecutive critically ill patients with confirmed or suspected Candida infection.InterventionsPatients received amphotericin B (1 mg/kg/24 hrs), administered randomly in 5% dextrose in water (group A), or in lipid emulsion (20% Intralipid Registered Trademark) (group B).Measurements and Main ResultsClinical tolerance (fever, chills, hemodynamics), hepatorenal tolerance, and biological tolerance (serum electrolytes and coagulation profile) were evaluated. Patients receiving amphotericin B in lipid emulsion experienced a lower frequency rate of drug-associated fever (61.4% vs. 5.8%, p < .003) rigors (54% vs. 8.5%, p < .004), hypotension (17% vs. 0%), and nephrotoxicity (increase of serum creatinine concentration, 66.7% vs. 20%, p < .0002). Significant (264,500 +/- 71,460 to 163,570 +/- 34,450 mm3, p < .01) thrombocytopenia, not associated with active bleeding, occurred in patients receiving amphotericin B lipid in emulsion but not in patients receiving the drug in dextrose.ConclusionsTreatment with amphotericin B in a lipid emulsion when given to critically ill patients with Candida sepsis seems to be safer and as effective as the conventional mode of administration.(Crit Care Med 1996; 24:1311-1315)