Objective To investigate the risk factors of anemia in HIV-infected patients receiving highly active antiretroviral therapy （HAART） in the Pudong New Area.

Methods A retrospective cohort study was conducted among HIV-infected patients who started HAART from 2005 to 2020 in Pudong New Area. Cox proportional hazards model was used to analyze the risk factors of anemia， moderate or severe anemia， and chronic anemia. The piecewise linear mixed-effect model was used to analyze the association between initial HAART classes and hemoglobin change in the follow-up.

Results A total of 2 403 HIV-infected patients were included in the analysis. Among them， there were 357 cases of new onset anemia， 86 cases of chronic anemia and 102 cases of moderate or severe anemia， with the incidence density of 4.41/100 person years， 0.89/100 person years and 0.96/100 person years respectively. Multifactorial Cox regression analysis results showed that female， age >45 years， baseline CD4⁺ T lymphocyte count （CD4） <200 cells‧μL^-1， opportunistic infections， glomerular filtration rate （eGFR） <60 mL‧min^-1‧（1.73 m²）^-1， and zidovudine （AZT） or protease inhibitor （PIs） based regimens were associated factors for the development of anemia. Female， age >45 years， CD4 <200 cells‧μL^-1， opportunistic infections， and AZT-based regimens were associated with the development of chronic anemia. Mild anemia at baseline and AZT-based regimens were associated with the development of moderate or severe anemia. Linear mixed-effects model showed that the use of AZT （-7.87 g‧L^-1， 95%CI： -9.42 to -6.32） or PIs （-3.43 g‧L^-1， 95%CI： -5.57 to -1.30） was associated with lower Hb at follow-up.

Conclusion Initial use of AZT and PIs is associated with progression to anemia and a lower follow-up hemoglobin level. Increased hemoglobin monitoring in users of AZT and PIs may be beneficial， especially during the first 6 months after initiation of HAART.