Non-traditional biomarkers and incident diabetes in the Diabetes Prevention Program: comparative effects of lifestyle and metformin interventions

Higher baseline adiponectin was associated with lower risk of incident diabetes in the placebo group of the DPP, independent of traditional diabetes risk factors including BMI, fasting glucose, insuli

Effect: improvement; HR 0.84; CI: 95% CI 0.71, 0.99

Higher baseline adiponectin was more strongly associated with lower incident diabetes risk in the lifestyle group than in the placebo group, and this association persisted after full adjustment for ba

Effect: improvement; HR 0.69; CI: 95% CI 0.52, 0.92

Higher baseline adiponectin was associated with lower incident diabetes risk in the metformin group, persisting after full adjustment for baseline and 1-year changes in diabetes risk factors and after

Effect: improvement; HR 0.79; CI: 95% CI 0.66, 0.94

Higher baseline sE-selectin was the most robust biomarker predictor of incident diabetes in the placebo group, remaining significant after adjustment for traditional diabetes risk factors, 1-year chan

Effect: decline; HR 1.19; CI: 95% CI 1.06, 1.34

The strong association between baseline sE-selectin and incident diabetes seen in the placebo group was completely absent in both the lifestyle and metformin intervention groups, suggesting that both

Effect: null; non-significant in both ILS and metformin groups

Higher baseline tPA was associated with increased incident diabetes risk in the placebo group, independent of traditional diabetes risk factors and their 1-year changes, and independent of all other b

Effect: decline; HR 1.13; CI: 95% CI 1.03, 1.24

Higher baseline leptin was directly and robustly associated with incident diabetes in the lifestyle group only, independent of diabetes risk factors including BMI, a novel association not seen in plac

Effect: decline; HR 1.31; CI: 95% CI 1.06, 1.63

Change in IL-6 at 1 year independently predicted incident diabetes in the metformin group only, surviving adjustment for 1-year changes in diabetes risk factors, a novel association not seen in placeb

Effect: decline; HR 1.09; CI: 95% CI 1.021, 1.173

Baseline CRP and sICAM-1 were associated with incident diabetes in the placebo group after demographic adjustment but lost significance after adjustment for traditional diabetes risk factors, indicati

Effect: null; CRP and sICAM-1 significant only in demographics-adjusted models, attenuated to non-significance after DRF adjustment

Neither fibrinogen nor MCP-1 showed any significant association with incident diabetes in any intervention group under any level of adjustment.

Effect: null; non-significant in all groups and all models