New metabolic health definition might not be a reliable predictor for diabetes in the nonobese Chinese population.

<bold>Aim: </bold>To investigate the predictive values of the new metabolic health (MH) definition for future diabetes in a nonobese Chinese population, compared with the MH definition from metabolic syndrome (MetS).<bold>Methods: </bold>The data were collected in 1992 and then again in 2007 from the same group of 653 participants. The risk assessment of the new MH definition and the MH definition from MetS for future diabetes was performed by Cox regression analysis with overlap weighting as the primary analysis.<bold>Results: </bold>During the follow-up, 62 participants were diagnosed with diabetes. In the primary analysis with overlap weighting, there was no significant association between new MH and diabetes (HR: 1.12; 95% CI: 0.45-2.78, p = 0.803); conversely, based on the MH definition from MetS, the participants with MH were less likely to have had diabetes than the participants with MUHs (HR: 0.41; 95% CI: 0.22-0.78, p = 0.007). Furthermore, other analysis methods also confirmed the reproducibility of abovementioned results. In addition, sensitivity analysis excluding participants with prediabetes also demonstrated similar results with the primary analysis.<bold>Conclusion: </bold>In contrast to the previous MH definition from MetS, the new MH definition was not a reliable predictor for future diabetes in the nonobese Chinese population. [ABSTRACT FROM AUTHOR]