Impact of randomized antiretroviral therapy initiation on glucose metabolism

Abstract

Objective: Prior studies have found that early HIV protease inhibitors contribute to glucose dysregulation. Few randomized trials have evaluated glucose indices in antiretroviral-naive individuals on newer antiretroviral therapy (ART).

Methods: A5224s was a substudy of A5202, a prospective trial of 1857 ART-naive participants randomized to blinded abacavir-lamivudine (ABC/3TC) or tenofovir DF-emtricitabine (TDF/FTC) with open-label efavirenz (EFV) or atazanavir-ritonavir (ATV/r). Analyses used two-sample t-tests, Spearman correlation coefficients and linear regression.

Results: A5224s included 269 nondiabetic individuals: 85% men, 47% white non-Hispanic, baseline median age 38 years, HIV-1 RNA 4.6 log10 copies/ml and CD4 cell count 233 cells/μl. Overall, significant 96-week increases occurred in fasting glucose, insulin and the homeostatic model assessment of insulin resistance (HOMA-IR), P ≤ 0.004. Assignment to EFV (versus ATV/r) resulted in significantly greater glucose increase [mean difference 4.4; 95% confidence interval (CI) 1.3, 7.5 mg/dl; P = 0.006] but not insulin or HOMA-IR (P ≥ 0.72). Glucose indices were not significantly different between ABC/3TC and TDF/FTC arms, P ≥ 0.18. Significant correlations were detected between changes in glucose indices and changes in BMI; all r ≥ 0.23, P ≤ 0.001. In multivariable analyses, in addition to the EFV effect, higher baseline HIV-1 RNA and greater BMI change were significant independent factors associated with greater glucose increase.

Conclusion: Changes in glucose metabolism were not significantly different between TDF/FTC and ABC/3TC-based regimens. A small but significantly greater increase in glucose was observed in those assigned to EFV. As glucose dysregulation may increase with time on ART, longer term studies will be needed to further clarify the clinical significance of these findings.

Trial registration: ClinicalTrials.gov NCT00118898.

Figure 1

Changes in measures of glucose metabolism by treatment arms (intent-to-treat). Mean and 95% confidence intervals are represented by symbols and error bars; P value from comparison between arms at 96 weeks; TDF/FTC, tenofovir-emtricitabine; ABC/3TC, abacavir-lamivudine; EFV, efavirenz; ATV/r, atazanavir-ritonavir. A. Changes in glucose between the nucleoside reverse transcriptase inhibitor (NRTI) components and non-nucleoside reverse transcriptase inhibitor/protease inhibitor (NNRTI/PI) components. B. Changes in insulin between NRTI components and NNRTI/PI components. C. Changes in homeostatic model assessment of insulin resistance (HOMA-IR) between NRTI components and NNRTI/PI components.

Figure 2

Correlation of changes in glucose (top row), insulin (middle row), homeostatic model assessment of insulin resistance (HOMA-IR; bottom row) with body mass index (BMI), lean body mass (LBM), limb fat, and trunk fat.