3394 Alpha-1-acid glycoprotein as outcome, independent predictor, and effect modifier in a randomized, placebo-controlled, factorial trial of recombinant human growth hormone and rosiglitazone in people living with HIV

OBJECTIVES/SPECIFIC AIMS: In a randomized controlled trial in participants with HIV infection, recombinant human growth hormone (rhGH) reduced visceral adipose tissue (VAT); addition of rosiglitazone to rhGH prevented the accompanying decline in insulin sensitivity (SI). Within this parent RCT, we sought to determine the effect of rosiglitazone and rhGH intervention on alpha-1-acid glycoprotein (AGP), a biomarker of inflammation. We also investigated AGP as an independent risk factor for SI and VAT changes along with any potential effect modification by AGP of the intervention. METHODS/STUDY POPULATION: Participants with HIV-infection (n = 72) with abdominal adiposity and insulin resistance were randomized to rosiglitazone, rhGH, combination, or placebo for 12 weeks (NCT00130286). SI was determined by frequently sampled intravenous glucose tolerance test, and VAT by whole body MRI. AGP concentrations were determined by immunoturbidimetric assay in available serum samples at baseline (time 0), 4, and 12 weeks (n = 41 participants with samples at all 3 time points). A linear mixed model was used to assess the impact of intervention over time on AGP concentrations. General linear models were used to assess baseline AGP concentrations as an independent predictor of SI and VAT changes by treatment group with the model initially including age quartile, gender, race, ethnicity, BMI, HIV RNA <400 copies/mL, antiretroviral regimen, CD4 count, Stavudine use, and zidovudine use with step-by-step removal of least significant predictors. Effect modification was assessed by adding an interaction term between AGP and assigned intervention. RESULTS/ANTICIPATED RESULTS: AGP did not differ among treatment groups at baseline; overall median (Q1, Q3): 0.608 (.526,.727) g/L, P = 0.92. Treatment with rosiglitazone, rhGH, or the combination significantly reduced AGP concentrations from baseline to week 12, compared to placebo (time by treatment interaction, P = 0.0038). Baseline AGP was not a significant predictor or effect modifier of SI change in response to treatment (P ≥ 0.50). Baseline AGP (g/L) was an independent predictor of VAT change (L) (β = 1.91, SE = 0.89, P = 0.038) in addition to a treatment effect (P < 0.001) and age quartile effect (P < 0.001). No other predictors or interactions were significant, including effect modification of AGP (AGP by treatment interaction P = 0.50). DISCUSSION/SIGNIFICANCE OF IMPACT: It is known that immune and metabolic pathways are highly integrated, and biomarkers of inflammation have predictive abilities for cardiovascular and metabolic disease outcomes. This analysis provides data showing that treatment with rosiglitazone or rhGH in the context of HIV reduces AGP concentrations, indicating efficacy in reducing systemic inflammation. Baseline AGP was an independent risk factor for VAT changes as those with lower AGP at baseline showed a greater reduction in VAT in response to treatment. Biomarkers of inflammation may provide prognostic information for individualized patient outcomes to help guide treatment and follow-up.