This retrospective chart review was conducted in an outpatient human immunodeficiency virus (HIV) clinic to determine if macrocytosis can be used as a clinical indicator of zidovudine adherence in HIV-infected outpatients. This study included 71 HIV-positive patients who were taking zidovudine and 93 HIV-positive controls who were not taking zidovudine, for 8 weeks or longer. One control and 16 subjects were excluded secondary to insufficient laboratory data or preexistence of other macrocytosis-inducing etiologies. The incidence of macrocytosis (mean corpuscular volume [MCV] >/= 100 fL) was significantly different among subjects and controls: 78% versus 32.6% (p < 0.001), respectively. Adherence (determined by documentation from a physician/pharmacist) was assessed among subjects for whom zidovudine was prescribed, comparing those with and without macrocytosis. Adherence was observed in 77% and 18% of the macrocytosis-positive and macrocytosis-negative subjects, respectively (p < 0.001). Macrocytosis was also observed in patients receiving stavudine (another thymidine analogue, in the class of nucleoside reverse transcriptase inhibitors). Twenty-seven of the 41 patients for whom stavudine was prescribed developed macrocytosis (65.8%). These results indicate that macrocytosis may be useful in assessing adherence to zidovudine-containing antiretroviral regimens and may also have a role in assessing stavudine adherence.