Queensland Centre for Gynaecological Cancer Research, UQ Centre for Clinical Research, The University of Queensland, Brisbane, Australia; Cancer Research Centre, Cancer Council Queensland, Brisbane, Australia.
Menzies School of Health Research, Darwin, Australia.
Nutr Metab Cardiovasc Dis. 2020 Oct 30;30(11):2008-2016. doi: 10.1016/j.numecd.2020.06.017. Epub 2020 Jun 26.
High blood pressure (BP) is a significant predictor for cardiovascular disease (CVD) mortality. The 2017 American College of Cardiology/American Heart Association guideline reclassified the BP categories; however, its impact on CVD mortality prediction is still unclear. Our study aimed to examine whether the application of new BP cut-offs could improve 10-year CVD mortality prediction among US adults.
This population-based cohort study linked data from the US National Health and Nutrition Examination Survey (1988-1994 and 1999-2004) and National Death Index (up to December 31, 2015). We constructed original and modified, using new BP cut-offs, Systematic COronary Risk Evaluation models to predict 10-year CVD mortality. We measured model discrimination and calibration using the Harrell's C statistic and calibration plots, respectively. We calculated the net reclassification index to evaluate the reclassification. In addition, we compared the sensitivity, specificity, predictive values (PVs), and likelihood ratios (LRs). Among 28,964 adults (aged ≥ 20 years), 1493 have died of CVD within ten years of follow-up. The modified models had improvements in calibration and reclassification instead of discrimination compared to the original models. The modified models have higher sensitivity and negative PV; however, they have lower specificity, positive PV, positive LR, and negative LR.
The modified models failed to improve the discrimination of 10-year CVD mortality. However, they could increase the calibration and reclassification and capture more participants with high CVD risk. More studies are needed on the potential use of the new BP cut-offs in the CVD primary prevention.
高血压（BP）是心血管疾病（CVD）死亡率的重要预测因素。2017 年美国心脏病学会/美国心脏协会指南重新分类了 BP 类别；然而，其对 CVD 死亡率预测的影响尚不清楚。我们的研究旨在检验新的 BP 切点的应用是否能提高美国成年人的 10 年 CVD 死亡率预测。
本基于人群的队列研究将来自美国国家健康和营养检查调查（1988-1994 年和 1999-2004 年）和国家死亡指数（截至 2015 年 12 月 31 日）的数据进行了关联。我们构建了原始和修改后的使用新 BP 切点的系统冠状动脉风险评估模型，以预测 10 年 CVD 死亡率。我们分别使用 Harrell 的 C 统计量和校准图来衡量模型的区分度和校准度。我们计算了净再分类指数来评估再分类。此外，我们比较了敏感性、特异性、预测值（PV）和似然比（LR）。在 28964 名成年人（年龄≥20 岁）中，有 1493 人在随访的 10 年内死于 CVD。与原始模型相比，修改后的模型在校准和再分类方面有所改善，而在区分度方面则没有改善。修改后的模型具有更高的敏感性和阴性 PV；然而，它们的特异性、阳性 PV、阳性 LR 和阴性 LR 较低。
修改后的模型未能提高 10 年 CVD 死亡率的区分度。然而，它们可以提高校准度和再分类度，并捕捉到更多具有高 CVD 风险的参与者。需要更多的研究来探讨新的 BP 切点在 CVD 一级预防中的潜在应用。
