Elliott Paul, Muller David C, Schneider-Luftman Deborah, Pazoki Raha, Evangelou Evangelos, Dehghan Abbas, Neal Bruce, Tzoulaki Ioanna
From the Department of Epidemiology and Biostatistics, School of Public Health (P.E., D.C.M., D.S.-L., R.P., E.E., A.D., B.N., I.T.), Imperial College London, United Kingdom.
MRC Centre for Environment and Health, School of Public Health (P.E., D.S.-L., R.P., E.E., A.D., I.T.), Imperial College London, United Kingdom.
Hypertension. 2020 Sep;76(3):683-691. doi: 10.1161/HYPERTENSIONAHA.119.14302. Epub 2020 Jul 6.
We report on an analysis to explore the association between estimated 24-hour urinary sodium excretion (surrogate for sodium intake) and incident cardiovascular disease (CVD) and mortality. Data were obtained from 398 628 UK Biobank prospective cohort study participants (40-69 years) recruited between 2006 and 2010, with no history of CVD, renal disease, diabetes mellitus or cancer, and cardiovascular events and mortality recorded during follow-up. Hazard ratios between 24-hour sodium excretion were estimated from spot urinary sodium concentrations across incident CVD and its components and all-cause and cause-specific mortality. In restricted cubic splines analyses, there was little evidence for an association between estimated 24-hour sodium excretion and CVD, coronary heart disease, or stroke; hazard ratios for CVD (95% CIs) for the 15th and 85th percentiles (2.5 and 4.2 g/day, respectively) compared with the 50th percentile of estimated sodium excretion (3.2 g/day) were 1.05 (1.01-1.10) and 0.96 (0.92-1.00), respectively. An inverse association was observed with heart failure, but that was no longer apparent in sensitivity analysis. A J-shaped association was observed between estimated sodium excretion and mortality. Our findings do not support a J-shaped association of estimated sodium excretion with CVD, although such an association was apparent for all-cause and cause-specific mortality across a wide range of diseases. Reasons for these differences are unclear; methodological limitations, including the use of estimating equations based on spot urinary data, need to be considered in interpreting our findings.
我们报告了一项分析，以探讨估计的24小时尿钠排泄量（钠摄入量的替代指标）与心血管疾病（CVD）发病及死亡率之间的关联。数据来自于2006年至2010年间招募的398628名英国生物银行前瞻性队列研究参与者（年龄在40 - 69岁之间），这些参与者无CVD、肾病、糖尿病或癌症病史，并在随访期间记录了心血管事件和死亡率。根据即时尿钠浓度估计24小时钠排泄量与CVD及其组成部分、全因死亡率和特定病因死亡率之间的风险比。在受限立方样条分析中，几乎没有证据表明估计的24小时钠排泄量与CVD、冠心病或中风之间存在关联；与估计钠排泄量的第50百分位数（3.2克/天）相比，第15和85百分位数（分别为2.5和4.2克/天）的CVD风险比（95%置信区间）分别为1.05（1.01 - 1.10）和0.96（0.92 - 1.00）。观察到与心力衰竭呈负相关，但在敏感性分析中这种关联不再明显。在估计的钠排泄量与死亡率之间观察到J形关联。我们的研究结果不支持估计的钠排泄量与CVD呈J形关联，尽管在广泛的疾病中，这种关联在全因死亡率和特定病因死亡率中很明显。这些差异的原因尚不清楚；在解释我们的研究结果时，需要考虑方法学上的局限性，包括使用基于即时尿样数据的估算方程。
