新疆维吾尔族农村居民肾功能轨迹与心血管疾病发病关系前瞻性队列研究

Association between kidney function trajectories and cardiovascular disease incidence among rural Uyghur residents in Xinjiang:a prospec-tive cohort study

目的了解新疆维吾尔族农村居民肾功能轨迹与心血管疾病(CVD)发病的关系,为CVD的早期识别和预防提供参考依据。方法采用前瞻性队列研究方法,于2016年9月采用典型抽样方法在新疆生产建设兵团第三师51团招募6417名≥18岁新疆维吾尔族农村居民进行基线调查,分别于2019年4月、2020年6月、2021年7月和2022年6月对5585名基本信息完整且基线调查无慢性肾脏病和CVD的非频繁流动居民进行了4次随访调查,基于群组轨迹模型构建肾功能轨迹确定轨迹的组数和趋势,采用log-rank检验比较不同肾功能轨迹组居民的CVD累积发病率,应用Cox比例风险回归模型分析肾功能轨迹组与CVD发病的关系,并采用亚组分析和敏感性分析对结果进行验证。结果截至2022年6月,5585名新疆维吾尔族农村居民中失访414人,失访率为7.41%;排除血肌酐测定<3次的429人后最终纳入分析的4742名新疆维吾尔族农村居民共随访27018.48人年,平均随访(5.70±0.96)人年,随访期间CVD共发病563例,CVD累积发病率为11.87%;基于群组轨迹模型构建了肾小球滤过率(e GFR)高水平稳定进展组(2128人,44.88%)、eGFR逐渐上升组(1207人,25.45%)、eGFR逐渐下降组(988人,20.84%)和eGFR快速下降组(419人,8.83%)共4种肾功能轨迹,其CVD累积发病数分别为232、87、156和88例,CVD累积发病率分别为10.90%、7.21%、15.79%和21.00%,不同肾功能轨迹的新疆维吾尔族农村居民CVD累积发病率差异均有统计学意义(χ^(2)=74.882,P<0.001);在调整了性别、年龄、吸烟情况、饮酒情况、是否高血压、是否糖尿病、是否超重肥胖、是否血脂异常、有无高血压家族史、有无糖尿病家族史、有无冠心病家族史和有无脑卒中家族史等混杂因素后,Cox比例风险回归模型分析结果显示,eGFR逐渐上升组、e GFR逐渐下降组和e GFR快速下降组维吾尔族农村居民CVD发病风险分别为eGFR高水平稳定进展组维吾尔族农村居民的0.766倍(HR=0.766,95%CI=0.598~0.983)、1.297倍(HR=1.297,95%CI=1.059~1.590)和1.865倍(HR=1.865,95%CI=1.458~2.385);亚组分析结果显示,男性、女性、<45岁和≥45岁e GFR快速下降组维吾尔族农村居民的CVD发病风险分别为eGFR高水平稳定进展组维吾尔族农村居民的2.577倍(HR=2.577,95%CI=1.698~3.909)、1.629倍(HR=1.629,95%CI=1.196~2.220)、1.729倍(HR=1.729,95%CI=1.078~2.774)和1.867倍(HR=1.867,95%CI=1.399~2.492);依次排除基线调查患高血压、糖尿病和血脂异常者后的敏感性分析结果显示,eGFR快速下降组维吾尔族农村居民的CVD发病风险分别为eGFR高水平稳定进展组维吾尔族农村居民的1.638倍(HR=1.638,95%CI=1.157~2.319)、1.854倍(HR=1.854,95%CI=1.426~2.410)和1.667倍(HR=1.667,95%CI=1.227~2.264),结果稳健性较好。结论肾功能轨迹eGFR上升和下降均与新疆维吾尔族农村居民的CVD发病风险相关,e GFR上升组居民的CVD发病风险较低,eGFR下降组居民的CVD发病风险较高。

Objective To understand the relationship between kidney function trajectories and cardiovascular disease(CVD)incidence among adult rural Uyghurs in Xinjiang Uyghur Autonomous Region(Xinjiang),and to provide a reference for early detection and prevention of CVD in the population.Methods A total of 6417 rural Uyghur residents aged≥18 years in Regiment 51,Third Division of the Xinjiang Production and Construction Corps were recruited for a baseline survey in September 2016 using typical sampling.Follow-up surveys were conducted in April 2019,June 2020,July 2021 and June 2022 among 5585 non-frequent migrants with complete basic information and no chronic kidney disease or CVD at baseline.Kidney function trajectories were constructed using a group-based trajectory model to determine the number and trend of trajectories.The log-rank test was used to compare the cumulative incidence of CVD among residents with different renal function trajectories.The Cox proportional hazards regression model was used to analyze the relationship between renal function trajectory groups and CVD incidence,and subgroup analysis and sensitivity analysis were used to validate the results.Results As of June 2022,414(7.41%)of the 5585 participants were lost to follow-up;after excluding 429 participants with<3 serum creatinine measurements,a total of 4742 participants were finally included in the analysis,with a total follow-up of 27018.48 person-years and a mean follow-up of 5.70±0.96 person-years.A total of 563 cases of CVD were observed during the follow-up period,with a cumulative incidence of 11.87%.Based on the group-based trajectory model analysis,the 4742 participants were classified into four groups with different kidney function trajectories:high-level stable progression of estimated glomerular filtration rate(eGFR)group(2128 individuals,44.88%of the participants),gradually increasing eGFR group(1207,25.45%),gradually decreasing eGFR group(988,20.84%),and rapidly decreasing eGFR group(419,8.83%),respectively;the cumulative incidence of CVD observed at follow-up in these groups was 232,87,156,and 88,with significantly different cumulative incidence rates of 10.90%,7.21%,15.79%,and 21.00%(x2=74.882,P<0.001).After adjustment for sex,age,smoking status,alcohol consumption,hypertension,diabetes,overweight/obesity,dyslipidaemia/hypertension/diabetes/CVD/stroke,the Cox proportional hazards regression model analysis showed that,compared with those in the high stable eGFR group,participants in the gradually increasing eGFR group had a reduced risk of CVD incidence(hazard ratio[HR]=0.766,95%confidence interval[95%CI]:0.598-0.983),whereas participants in the gradually decreasing eGFR group and the rapidly decreasing eGFR group had an increased risk of CVD incidence,with HR(95%CI)of 1.297(1.059-1.590)and 1.865(1.458-2.385),respectively;further subgroup analysis revealed HR(95%CI)of 2.577(1.698-3.909),1.629(1.196-2.220),1.729(1.078-2.774),and 1.867(1.399-2.492)for participants who were male,female,<45 years of age,and≥45 years of age in the rapidly declining eGFR group,respectively.Sensitivity analysis results after sequential exclusion of those with hypertension,diabetes and dyslipidaemia at baseline showed increased risks of CVD incidence in participants in the rapidly declining eGFR group,with HR(95%CI)of 1.638(1.157-2.319),1.854(1.426-2.410)and 1.667(1.227-2.264)compared with those in the high-grade stable eGFR group,indicating good robustness of the results.Conclusion Both increasing and decreasing eGFR trajectories are associated with the risk of CVD incidence among rural Uyghur residents in Xinjiang;those with increasing eGFR have a decreased risk of CVD incidence,whereas those with decreasing eGFR have an increased risk of CVD incidence.