非高密度与高密度脂蛋白胆固醇比值及其动态变化和2型糖尿病发病风险的前瞻性研究

目的 探索非高密度脂蛋白胆固醇（Non-high-density lipoprotein cholesterol, non-HDL-c)/高密度脂蛋白胆固醇（HDL-c）及其变化与2型糖尿病发病风险的关联，为2型糖尿病防控提供依据。方法 2010年采用多阶段分层整群抽样抽取9 280人建立贵州自然人群队列，并于2016—2020年，对所有队列人群进行随访，随访到8 163人，剔除基线时患有糖尿病者、服用降血脂药物、基线总胆固醇、HDL-c缺失者或极值者后，最终有5 078 人纳入分析。non-HDL-c为总胆固醇与HDL-c差值。动态变化由随访non-HDL-c/HDL-c值减去基线non-HDL-c/HDL-c值得出。采用Cox回归分析non-HDL-c/HDL-c值与糖尿病发病的关联及调整风险比(adjusted hazard ratio, aHR )和其95%可信限(confidential interval, CI)。采用限制性立方样条进行剂量反应关系描述。结果 中位随访时间为6.58年，共新发2型糖尿病820例。与位于non-HDL-c/HDL-c第1四分位数区间对象相比，位于第2四分位数，第3四分位数和第4四分位数区间对象的2型糖尿病发病风险 aHR 值分别为1.065(95%CI:0.855~1.325)、1.272(95%CI:1.037~1.562)、1.266(95%CI:1.022~1.569)。亚组分析中，在男性、年龄大于45岁和超重肥胖人群中，与位于non-HDL-c/HDL-c第1四分位数区间对象相比、第3四分位数和第4四分位数区间对象的2型糖尿病发病风险 aHR 值分别为1.529（95%CI:1.130~2.067）、1.748(95%CI:1.279~2.388) 、1.389(95%CI:1.046~1.844) 、1.394(95%CI:1.049~1.854) 、1.380(95%CI:0.955~1.994) 、1.462(95%CI:1.014~2.107),且趋势性检验P＜0.05。与non-HDL-c/HDL-c动态变化值的第2四分位数区间对象相比，第1四分位数，第3四分位数和第4四分位数区间对象的2型糖尿病发病风险aHR值分别为1.258(95%CI:0.999~1.582)、1.390(95%CI:1.115~1.731)、1.530(95%CI:1.228~1.906)。限制性立方样条显示non-HDL-c/HDL-c与糖尿病发病风险呈正相关。结论 non-HDL-c/HDL-c值的增高将增加糖尿病发病的风险，尤其在男性，年龄大于45岁和超重肥胖人群中更为明显。

Associations between non-HDL-c/HDL-c,its dynamics changes and incident type 2 diabetes mellitus in a prospective cohort study

Objective To explore associations between the ratio of non-high density lipoprotein cholesterol (HDL-c) to HDL-c (non-HDL-c /HDL-c) and the incident type 2 diabetes mellitus(T2DM) so as to provide evidences for the prevention and control of T2DM. Methods A multi-stage stratified cluster sampling method was used to target 9 280 adults in a natural population cohort in 2010. From 2016 to 2020, a total of 8 163 subjects were followed up. After exclusion of those with diabetes at baseline, taking lipid lowering drugs, with missing or extreme values of total cholesterol and HDL-c baseline, 5 078 subjects were included in this analysis finally. Non-HDL-c was derived from the difference between total cholesterol and HDL-c. Dynamic changes were calculated from non-HDL-c/HDL-c at follow-up minus non-HDL-c/ HDL-c at baseline. Cox model was occupied to explore associations between non-HDL-c/HDL-c, its changes and the incident T2DM, and to estimate the adjusted hazard ratio ( aHR ) and its 95% confidential interval (CI). The dose-response relationship was presented by using restricted cubic spline. Results With the median follow-up of 6.58 years, a total of 820 new cases of T2DM were identified. Compared with the first quantile of non-HDL-c/HDL-c, the aHR values of diabetes in the second, third and fourth quantiles were 1.065(95%CI: 0.855-1.325), 1.272(95%CI: 1.037-1.562), 1.266 (95% CI : 1.022 -1.569) respectively. In subgroup analysis, among men, those older than 45, and those who were overweight and obese, compared with the first quartile of non-HDL-C/HDL-C, The risk of type 2 diabetes in the third and fourth quartile of non-HDL-c/HDL-c adjusted hazard ratio were 1.529(95%CI: 1.130-2.067), 1.748(95%CI: 1.279-2.388), 1.389(95%CI: 1.046-1.844), 1.394(95%CI: 1.049-1.854), 1.380(95%CI: 0.955-1.994), 1.462(95%CI: 1.014-2.107), respectively, and the trend test P <0.05. Compared with the second quartile of non-HDL-c/HDL-c dynamic variation, the aHR values of diabetes in the first, third and fourth quantiles were 1.258(95%CI: 0.999-1.582) ,1.390(95%CI: 1.115-1.731),1.530(95%CI: 1.228-1.906) respectively. Restricted cubic splines showed a positive correlation between non-HDL-c/HDL-c and the risk of diabetes. Conclusion The increase in non-HDL-c/HDL-c is associated with an increased risk of diabetes, especially in men, older than 45 years and overweight and obese individuals.