Eleven HIV-infected men participated in a randomized, two-treatment, two-period crossover study to determine the effect of a 25 g protein meal on zidovudine pharmacokinetics. On two separate occasions, 1 week apart, each patient received 200 mg zidovudine in a fasting state or immediately following the protein meal. A protein meal significantly decreased Cmax [532 (228 s.d.) vs 802 (452 s.d.) ng ml-1, P = 0.004] and increased mean residence time (138 (26 s.d.) vs 114 (26 s.d.) min, corrected for lag times, P = 0.001). However, AUC, tmax, terminal half-life and renal clearance were not significantly altered (P greater than 0.05). The power to detect a 20% change in AUC was 98% at the 5% significance level. In contrast to fat-containing foods, protein-based meals may not alter the extent of zidovudine absorption.