Serum albumin as a predictor of survival in HIV-infected women in the Women's Interagency HIV Study

Background: The level of serum albumin is associated with mortality in a wi de variety of chronic diseases. However, few studies have examined the rela tionship between serum albumin and survival in HIV-1 infection. Objectives: To determine whether the serum albumin level is associated with survival in HIV-1 infected women. Design: Prospective cohort study. Patients were interviewed and examined at 6 month intervals. Setting: A North American multi-institutional cohort of HIV-infected women from five geographical areas. Participants: A total of 2056 HIV-infected women at various stages of disea se. Measurements: Mortality during the first 3 years of follow-up. The relative risk of death by serum albumin level was estimated using a proportional ha zards ratio adjusted for CD4 cell count, HIV-1-RNA level and other relevant covariates. Result: Three year mortality for women in the lowest serum albumin category (< 35/l) was 48% compared with 11% in the highest category (greater than o r equal to 42 g/l; P<0.001). The adjusted relative hazard (RH) of death was 3.1 times greater for those in the lowest albumin category (P < 0.01). The excess risk associated with lower serum albumin levels remained when subje cts with moderate to severe immunosuppression and abnormal kidney and liver function were excluded (P < 0.01). Conclusion: The baseline serum albumin level is an independent predictor of mortality in HIV-l-infected women. The serum albumin level may be a useful additional marker of HIV-1 disease progression, particularly among asympto matic women with little or no evidence of immunosuppression. (C) 2000 Lippi ncott Williams & Wilkins.