Abstract
OBJECTIVE: Time-dependent covariates are generally available as longitudinal data were collected periodically in the cohort study. To examine whether time-dependent triglyceride to high-density lipoprotein cholesterol (TG/HDL-C) ratio could predict the future risk of type 2 diabetes mellitus (T2DM) and assess its potential impact on the risk of T2DM incidence. RESEARCH DESIGN AND METHODS: This study enrolled 1460 participants without T2DM aged 55 or above in 1992 in the Beijing Longitudinal Study of Aging during 25 years. The questionnaire data were collected in nine surveys from 1992 to 2017. Physical examination and blood laboratory tests including TG and HDL-C concentrations were measured in five surveys. Incident T2DM cases were confirmed via a self-reported history of T2DM or the fasting plasma glucose level. RESULTS: 119 new cases of T2DM were identified. In the Cox regression analysis with time-dependent TG/HDL-C ratios and covariates, the adjusted hazard ratios (95% confidence interval) of T2DM incidence were 1.90 (1.12 to 3.23), 2.75 (1.58 to 4.80) and 2.84 (1.69 to 4.77), respectively, for those with TG/HDL-C ratios (both TG and HDL-C were expressed in millimole per liter) in the ranges of 0.87-1.30, 1.31-1.74 and ≥1.75, compared with individuals with TG/HDL-C ratios <0.87. The similar results of subdistribution hazard ratios were obtained by performing the Fine-Gray model with time-dependent TG/HDL-C ratios. This positive association and the statistically significant trend with increased risk of T2DM incidence in the three categories of elevated TG/HDL-C ratio was confirmed by multiple sensitivity analyses. Furthermore, the T2DM discriminatory power of TG/HDL-C ratio combining with other risk factors was moderately high. CONCLUSIONS: We found that time-dependent TG/HDL-C ratios were positively associated with the risk of T2DM risk. The elevated TG/HDL-C ratios increased the future risk of T2DM incidence. Lowering the TG/HDL-C ratio could assist in the prevention of diabetes for older adults.