Weight change at 1 mo of antiretroviral therapy and its association with subsequent mortality, morbidity, and CD4 T cell reconstitution in a Tanzanian HIV-infected adult cohort

Abstract

Background: The development of low-cost point-of-care technologies to improve HIV treatment is a major focus of current research in resource-limited settings.

Objective: We assessed associations of body mass index (BMI; in kg/m(2)) at antiretroviral therapy (ART) initiation and weight change after 1 mo of treatment with mortality, morbidity, and CD4 T cell reconstitution.

Design: A prospective cohort of 3389 Tanzanian adults initiating ART enrolled in a multivitamin trial was followed at monthly clinic visits (median: 19.7 mo). Proportional hazard models were used to analyze mortality and morbidity associations, whereas generalized estimating equations were used for CD4 T cell counts.

Results: The median weight change at 1 mo of ART was +2.0% (IQR: -0.4% to +4.6%). The association of weight loss at 1 mo with subsequent mortality varied significantly by baseline BMI (P = 0.011). Participants with ≥2.5% weight loss had 6.43 times (95% CI: 3.78, 10.93 times) the hazard of mortality compared with that of participants with weight gains ≥2.5%, if their baseline BMI was <18.5 but only 2.73 times (95% CI: 1.49, 5.00 times) the hazard of mortality if their baseline BMI was ≥18.5 and <25.0. Weight loss at 1 mo was also associated with incident pneumonia (P = 0.002), oral thrush (P = 0.007), and pulmonary tuberculosis (P < 0.001) but not change in CD4 T cell counts (P > 0.05).

Conclusions: Weight loss as early as 1 mo after ART initiation can identify adults at high risk of adverse outcomes. Studies identifying reasons for and managing early weight loss are needed to improve HIV treatment, with particular urgency for malnourished adults initiating ART. The parent trial was registered at clinicaltrials.gov as NCT00383669.

FIGURE 1.

Relation between the percentage of weight change at 1 mo of ART and mortality for individuals with baseline BMI (in kg/m²) <18.5 (left graph) and ≥18.5 and <25.0 (right graph) after multivariate adjustment (adjusted for randomized regimen, sex, age, district, educational attainment, household assets, hemoglobin, WHO HIV stage, oral candidiasis, tuberculosis treatment and diagnosis, chronic diarrhea, ART regimen, and season of ART initiation). The solid line shows the estimated HR of mortality for the percentage of weight change relative to the reference weight change of +2.5%, and the horizontal dotted line designates the null HR of 1.0. Dashed lines denote 95% CIs of the HR. The relation was significantly nonlinear for individuals with a baseline BMI <18.5 (P-nonlinear relation = 0.008) and was significantly linear for patients with a baseline BMI ≥18.5 and <25.0 (P-linear relation < 0.001). ART, antiretroviral therapy.

FIGURE 2.

Multivariate (adjusted for randomized regimen, sex, age, district, educational attainment, household assets, hemoglobin, WHO HIV stage, oral candidiasis, tuberculosis treatment and diagnosis, chronic diarrhea, ART regimen, and season of ART initiation) joint association of baseline BMI (in kg/m²) and weight change after 1 mo of ART on the hazard of mortality with a BMI of 18.5–25.0 and ≥2.5% weight gain as the reference. ART, antiretroviral therapy.