Anemia and survival in HIV infection.

Several clinical studies have suggested that anemia is an independent risk factor for dying in patients with HIV disease. We analyzed data from a large urban HIV clinical practice in Baltimore to assess the annual incidence of anemia, the risk of dying in patients who develop anemia, and the association between recombinant human erythropoietin use to treat anemia and subsequent survival. In 2348 patients observed between 1989 and 1996, 498 (21%) developed at least grade 1 anemia (hemoglobin <9.4 g/dl); 95 (4%) developed grade 4 anemia (hemoglobin <6.9 g/dl). Development of anemia was associated with decreased survival, independent of other prognostic factors. Use of erythropoietin was more likely in patients of nonminority race, those who did not inject drugs, those with a lower CD4 count or AIDS, and those being treated for cytomegalovirus disease (p < .05). Adjusting for these factors as well as severity of anemia, age, diagnosis of opportunistic disease, blood transfusion, and antiretroviral therapy in a time-dependent Cox proportional hazards analysis, erythropoietin use (n=91) was associated with a decreased hazard of dying (relative hazard [RH]=0.57; 95% confidence interval [CII, 0.40-0.81; p=.002). Although we cannot rule out a treatment selection bias, adjusting for available prognostic factors and factors potentially associated with a decision to use erythropoietin suggests that erythropoietin for treatment of anemia is associated with improved survival in HIV disease.