Objective: To describe anemia prevalence and correlates with biomarkers and antiretroviral therapy (ART) in HIV/AIDS.

Methods: Multicenter, cross-sectional study; clinical laboratory data collected at single visits, including hemoglobin (Hb), CD4+ count, HIV-1 RNA. Patients receiving care at US physician offices during the year 2000. Main outcome measure was anemia (Hb < 14 g/dL [men]; < 12 g/dL [women]) and marked anemia (Hb < 11 g/dL [men]; < 10 g/dL [women]) prevalence. Multivariable models examined association of anemia prevalence with HIV-1 biomarkers and ART.

Results: Among 9690 patients, prevalence of anemia and marked anemia was 36% and 5%, respectively. Among 1721 patients receiving no ART, 39.7% were anemic; among 7252 receiving highly active antiretroviral therapy (HAART), 35.5% were anemic (p = 0.001). Anemia was most prevalent among men (37.3 vs. 32.3%; p = 0.0008), blacks (49 vs. 26% [whites]; p < 0.0001), patients with CD4+ < 200 cells/mm(3) (57 vs. 23% [> or = 500 CD4+]; p < 0.00001), and HIV-1 RNA > 30 000 copies/ml (53 vs. 30% [< 500 copies/ml]; p < 0.00001). Marked anemia was more common in women (6.8 vs. 4.3%; p < 0.0001). Among treated patients, logistic regression analysis controlling for CD4+, HIV-1 RNA, sex, and ethnicity, zidovudine (ZDV)-containing regimens (except combination with saquinavir/ZDV/lamivudine) were associated with increased overall anemia risk (odds ratio, 1.39 : 1.74). No regimen was associated with increased risk for marked anemia. Multivariable logistic regression showed CD4+, sex, and ethnicity more strongly associated with anemia than any ART regimen.

Conclusion: This large, single-visit, cross-sectional, US-based study shows that anemia remains highly prevalent in HIV-infected patients. Data from this analysis suggest low CD4+ count, black ethnicity, and male sex are consistently strongest correlates of overall anemia; women are significantly more likely to have marked anemia.