Department of Geriatrics, The First Affiliated Hospital of China Medical University, No.155 Nanjing North Street, Shenyang, 110001, China.
BMC Geriatr. 2022 May 4;22(1):392. doi: 10.1186/s12877-022-03087-3.
Hypertension-related mortality has been increasing in older adults, resulting in serious burden to society and individual. However, how to identify older adults with hypertension at high-risk mortality remains a great challenge. The purpose of this study is to develop and validate the prediction nomogram for 5-year all-cause mortality in older adults with hypertension.
Data were extracted from National Health and Nutrition Examination Survey (NHANES). We recruited 2691 participants aged 65 years and over with hypertension in the NHANES 1999-2006 cycles (training cohort) and 1737 participants in the NHANES 2007-2010 cycles (validation cohort). The cohorts were selected to provide at least 5 years follow-up for evaluating all-cause mortality by linking National Death Index through December 31, 2015. We developed a web-based dynamic nomogram for predicting 5-year risk of all-cause mortality based on a logistic regression model in training cohort. We conducted internal validation by 1000 bootstrapping resamples and external validation in validation cohort. The discrimination and calibration of nomogram were evaluated using concordance index (C-index) and calibration curves.
The final model included eleven independent predictors: age, sex, diabetes, cardiovascular disease, body mass index, smoking, lipid-lowering drugs, systolic blood pressure, hemoglobin, albumin, and blood urea nitrogen. The C-index of model in training and validation cohort were 0.759 (bootstrap-corrected C-index 0.750) and 0.740, respectively. The calibration curves also indicated that the model had satisfactory consistence in two cohorts. A web-based nomogram was established ( https://hrzhang1993.shinyapps.io/dynnomapp ).
The novel developed nomogram is a useful tool to accurately predict 5-year all-cause mortality in older adults with hypertension, and can provide valuable information to make individualized intervention.
与高血压相关的死亡率在老年人中不断增加，给社会和个人带来了严重的负担。然而，如何识别高血压高危死亡的老年人仍然是一个巨大的挑战。本研究旨在开发和验证预测高血压老年患者 5 年全因死亡率的预测列线图。
数据来自国家健康和营养检查调查（NHANES）。我们招募了 NHANES 1999-2006 周期（训练队列）中年龄 65 岁及以上的 2691 名高血压患者和 NHANES 2007-2010 周期（验证队列）中 1737 名高血压患者。这些队列的选择是为了通过链接国家死亡指数，至少提供 5 年的全因死亡率随访，截至 2015 年 12 月 31 日。我们在训练队列中基于逻辑回归模型开发了一个用于预测 5 年全因死亡率风险的基于网络的动态列线图。我们通过 1000 次自举重采样进行内部验证，并在验证队列中进行外部验证。通过一致性指数（C 指数）和校准曲线评估列线图的区分度和校准度。
最终模型包括 11 个独立预测因素：年龄、性别、糖尿病、心血管疾病、体重指数、吸烟、降脂药物、收缩压、血红蛋白、白蛋白和血尿素氮。模型在训练和验证队列中的 C 指数分别为 0.759（自举校正的 C 指数为 0.750）和 0.740。校准曲线也表明该模型在两个队列中具有较好的一致性。建立了一个基于网络的列线图（https://hrzhang1993.shinyapps.io/dynnomapp）。
新开发的列线图是一种准确预测高血压老年患者 5 年全因死亡率的有用工具，可以为制定个体化干预措施提供有价值的信息。
