基于马尔可夫模型的社区人群糖尿病筛查预防心血管病的效果评价

目的: 在中国鄞州电子健康档案研究(Chinese electronic health records research in Yinzhou, CHERRY)的队列人群中，评估国内外不同指南最新推荐的2型糖尿病筛查策略在我国发达地区人群中预防心血管病的效果。方法: 采用马尔可夫(Markov)模型模拟并比较的系统性筛查策略包括：(1) 根据《中国2型糖尿病防治指南(2020年版)》的推荐，在40~70岁人群中筛查(策略1);(2) 根据2022年美国糖尿病学会《糖尿病医学诊疗标准》的推荐，在35~70岁人群中筛查(策略2);(3)根据美国预防服务工作组2021年更新的《2型糖尿病的筛查建议声明》的推荐，在35~70岁且超重或肥胖(体重指数24 kg/m²及以上)的人群中进行筛查(策略3)。根据指南推荐，对筛查阳性(空腹血糖7.0 mmol/L及以上)的人群强化控制血糖以达到目标值(糖化血红蛋白控制在7.0%以下)。马尔可夫模型循环周期设为1年，研究期限设为10年，模拟10个周期，计算的结局指标包括心血管病事件发病数和全因死亡数等结局事件数，以及每预防一例心血管病事件或全因死亡需筛查人数等效果评价指标。马尔可夫模型的参数主要来源于CHERRY队列人群和公开发表的文献。采用单因素敏感性分析探讨筛查方法的灵敏度变化对结果的影响，采用概率敏感性分析探讨糖尿病发病率、筛查方法的灵敏度、强化干预措施的效应强度等参数的不确定性。结果: 研究纳入的289 245名基线无心血管病且未诊断糖尿病的35~70岁人群中，与机会性筛查相比，在40~70岁人群中进行系统性筛查的策略1可预防的心血管病发病数为22295%不确定性区间(uncertainty interval, UI)：180~264]例，在35~70岁人群中筛查的策略2为227(95%UI: 185~271)例，在35~70岁且超重或肥胖人群中筛查的策略3为131(95%UI: 98~164)例。每预防一例心血管病发病数的需筛查人数在策略1、2和3分别为1 184(95%UI: 994~1 456)人、1 274(95%UI: 1 067~1 564)人和814(95%UI: 649~1 091)人。策略2相比策略1每预防一例心血管病的需筛查人数增加90(95%UI: -197~381)人，但心血管病预防效果相似; 策略3相比策略2的需筛查人数减少460(95%UI: 185~724)人，筛查效率更高。单因素敏感性分析和概率敏感性分析的结果与主分析结果一致。结论: 在我国发达地区人群中，根据现有的最新指南开展糖尿病系统性筛查能够减少心血管病发病和全因死亡，但仅降低筛查起始年龄从40岁到35岁对预防心血管病效果的增益并不明显，如果降低筛查的起始年龄到35岁需要同时考虑超重或肥胖的危险因素以便提高筛查效率。

Effectiveness of different screening strategies for type 2 diabete on preventing cardio-vascular diseases in a community-based Chinese population using a decision-analytic Markov model

Objective: To evaluate the effectiveness of different screening strategies for type 2 diabetes to prevent cardiovascular disease in a community-based Chinese population from economically developed areas based on the Chinese electronic health records research in Yinzhou (CHERRY) study. Methods: A Markov model was used to simulate different systematic diabetes screening strategies, including: (1) screening among Chinese adults aged 40-70 years recommended by the 2020 Chinese Guideline for the prevention and Treatment of Type 2 Diabetes (Strategy 1); (2) screening among Chinese adults aged 35 to 70 years recommended by the 2022 American Diabetes Association Standard of Medical Care in Diabetes (Strategy 2); and (3) screening among Chinese adults aged 35-70 years with overweight or obesity recommended by the 2021 United States Preventive Services Task Force Recommendation Statement on Screening for Prediabetes and Type 2 Diabetes (Strategy 3). According to the guidelines, individuals who were screened positively (fasting plasma glucose ≥ 7.0 mmol/L) would be introduced to intensive glycemic targets management (glycated hemoglobin < 7.0%).The Markov model simulated different screening scenarios for ten years (cycles) with parameters mainly from the CHERRY study or published literature. Number of cardiovascular disease events or deaths could be prevented and number needed to screen (NNS) were calculated to compare the effectiveness of the different strategies. One-way sensitivity analysis on the sensitivity of screening methods and probabilistic sensitivity analysis on uncertainties of diabetes incidence, the sensitivity of screening methods, and intensive glycemic management effects were conducted. Results: Totally 289 245 Chinese adults aged 35-70 years without cardiovascular diseases or diagnosed diabetes at baseline were enrolled. In terms of the number of cardiovascular disease events could be prevented, Strategy 1 for systematic diabetes screening among the adults aged 35-70 years was 222 (95%UI: 180-264), Strategy 2 for systematic diabetes screening among the adults aged 40-70 years was 227 (95%UI: 185-271), and Strategy 3 for systematic diabetes screening among the adults aged 35-70 years with obesity or overweight (body mass index ≥ 24 kg/m²) was 131 (95%UI: 98-164), compared with opportunistic screening. NNS per cardiovascular disease event for the strategies 1, 2 and 3 were 1 184 (95%UI: 994-1 456), 1 274 (95%UI: 1 067-1 564) and 814 (95%UI: 649-1 091), respectively. Compared with Strategy 1, NNS per cardiovascular disease event for Strategy 2 increased by 90 (95%UI: -197-381) with similar effectiveness of cardiovascular prevention; however, NNS per cardiovascular disease event for Strategy 3 was reduced by 460 (95%UI: 185-724) in contrast to the Strategy 2, suggesting that the Strategy 3 was more efficient. The results were consistent in multiple sensitivity analyses. Conclusion: Systematic screening for diabetes based on the latest guidelines in economically developed areas of China can reduce cardiovascular events and deaths. However, merely lowering the starting age of screening from 40 to 35 years seems ineffective for preventing cardiovascular disease, while screening strategy for Chinese adults aged 35-70 years with overweight or obesity is recommended to improve efficiency.