ABSTRACT
OBJECTIVE: To estimate the prevalence of HIV-associated weight loss among HIV patients in a US managed care population, and compare demographic and clinical characteristics of HIV patients with and without evidence of HIV-associated weight loss. RESEARCH DESIGN AND METHODS: A retrospective observational study was conducted using a large, geographically diverse US managed care population to identify commercial enrollees with HIV/AIDS from 1/1/2005-7/31/2007, based on a combination of HIV/AIDS diagnosis codes or antiretroviral treatment. HIV-associated weight loss status was defined according to an algorithm combining evidence for weight loss-associated conditions, anorexia symptoms, and various treatments for weight loss or wasting. Among HIV patients continuously enrolled in the health plan for one year, patient demographics, treatments, and comorbidities were compared between patients with and without evidence for weight loss. RESULTS: A total of 22,535 patients with HIV/AIDS were identified, including 2098 who met the criteria for weight loss (estimated prevalence 9.3%; 95% CI: 8.9% - 9.7%). Among 12,187 continuously enrolled patients with HIV, 1006 (8.3%) had evidence of HIV-associated weight loss. Patients with HIV-associated weight loss were older (44.1 vs. 42.6 years), and more men had HIV-associated weight loss than women (8.8% vs. 5.3%). A number of comorbidities were more common among patients with HIV-associated weight loss. On average, these patients also had more ambulatory (24.0 vs. 13.4), ER (1.4 vs. 0.8), and inpatient visits (0.5 vs. 0.1). Total annual health care costs for patients with HIV-associated weight loss were more than double (mean $45,686 vs. $19,960) the costs for HIV patients without weight loss. CONCLUSIONS: Despite the availability of effective antiretroviral therapy, weight loss remains a problem among patients with HIV. Based on this analysis, almost 1 in 10 managed care patients with HIV have evidence of HIV-associated weight loss. These patients tend to have more comorbidities, use more health care resources, and incur greater costs compared to patients without HIV-associated weight loss. Patients with HIV-associated weight loss were generally sicker than the non-weight loss cohort; thus, the increased costs observed in this population may not be directly or wholly attributable to HIV-associated weight loss. In addition, limitations common to analyses of administrative claims data should be considered when interpreting these results.
