The prevalence of obesity is increasing among HIVpositive patients, with current estimates—20% to 30%— rivaling estimates for the general US population [1-3]. Obesity is associated with a chronic, systemic state of inflammation that may contribute to the development of many obesity-related comorbidities. Obesity increases the risk of hypertension, diabetes, myocardial infarction, heart failure, stroke, and some malignancies [4-6]. Because adipocytes are metabolically active, obesity also increases the production of a number of substances with proinflammatory or immune-modulating functions, which might mediate the adverse health effects: adipokines, such as leptin and adiponectin; C-reactive protein; interleukin (IL)-6; tumor necrosis factor–α; and IL-1-R agonist [4,7,8] .I n a large cohort study of subjects without HIV, obesity was associated with higher CD3, CD4, and CD8 counts [9]. The treatment of morbid obesity with gastric bypass surgery in HIV-negative adults has also been reported in small studies to change cytokine levels and a number of immune markers [10,11]. In the first study, abnormal CD95 antigen expression on Tcells was reversed with surgical weight loss [10]; and in the second study, C-reactive protein levels decreased significantly postsurgery as did lymphocyte subsets CD4 and CD8 [11]. Little is known about specific effects of obesity on the immune system in HIV-positive patients. The aim of our study was to explore the effects of obesity on lymphocyte subsets in HIV-positive patients with diabetes. 2. Methods We conducted a cross-sectional study in a cohort of 216 HIV-positive patients with diabetes at the CORE Center in Chicago. The study was approved by the Cook County Bureau of Health Services institutional review board. Patients were eligible for study if they were taking highly active antiretroviral therapy (HAART) and had undetectable HIV RNA for at least 6 months. We chose to include only patients who had well-controlled HIV infection, as ongoing HIV viremia is known to cause elevations in CD8 levels. The measure of obesity was based on body mass index (BMI), calculated as weight in kilograms divided by height in meters squared: normal weight, BMI 18.5 to 24.9; overweight, BMI 25 to 29.9; obese, BMI 30 to 34.9; and morbidly obese, BMI of at least 35. Our dependent variables were the white blood cell count and the following lymphocyte subgroups, defined by surface antigens: CD3 (Tcells), CD4 (helper Tcells), CD8 (cytotoxic or suppressor T cells), CD19 (B cells), and CD56 (natural killer cells). We used linear regression models to test bivariate relationships between BMI and the concentrations of the different cell types, after visually inspecting locally weighted scatterplot smoothercurvestoensurethatlinearmodelswereappropriate. Because of consistently nonlinear relationships at BMI greater than 40, we restricted the analysis to patients with a BMI of 40 or less. We then used multivariable linear models to test the same relationships after adjusting for age, sex, ethnic group, renal function (inverse of serum creatinine), hemoglobin A1c (HbA1c) (measureofaverageglucosecontrol),andthenumber of years since HIV diagnosis. To meet model distribution assumptions,wetransformeddependent,predictor,andcontrol variables to best approximate the normal distribution (square root transformations for HIV duration and CD3, CD4, CD8, and total lymphocyte counts; log transformations for BMI and