Since macrolide antibiotics inhibit the oxidative hepatic metabolism of various drugs, including midazolam, the present double blind studies were conducted to find out if azithromycin, a new macrolide of the azalide type, would inhibit the metabolism of midazolam and enhance the effects of midazolam on human performance. In Study I, 64 healthy medical students, divided in four parallel groups received placebo, midazolam (10 mg or 15 mg), and midazolam 10 mg combined with azithromycin (500 mg + 250 mg). In Study II, three males received oral midazolam 10 mg in combination with placebo, azithromycin or erythromycin 750 mg (as a positive control) in a cross-over trial. Objective and subjective tests were done before the intake of midazolam and 30 and 90 min after it, and venous blood was sampled for the assay of midazolam. In the placebo group in Study I, the mean numbers of letters cancelled (LC) at baseline, 30 min and 90 min were 21, 20 and 20, respectively, and the corresponding mean numbers of correct digit symbol substitutions (DSS) were 126, 137 and 140, indicating a practice effect. Midazolam 10 mg impaired these performances (21, 13 and 12 for LC, and 127, 113 and 111 for DSS). Either dose of midazolam produced clumsiness, mental slowness and poor subjective performance, midazolam 15 mg being slightly more active. The corresponding, scores in the azithromycin+midazolam group were 21, 16, 16 for LC, and 132, 121 and 119 for DSS, the only significant difference from placebo being the impairment of DSS at 90 min.(ABSTRACT TRUNCATED AT 250 WORDS)