by Rui Zhang, Liqiang Zheng, Zhaoqing Sun, Xingang Zhang, Jue Li, Dayi Hu, Yingxian Sun
Few studies reported the associations between decreased glomerular filtration rate (GFR) and mortality, coronary heart disease (CHD), and stroke in hypertensive patients. We aim to assess the associations between GFR and mortality, CHD, and stroke in hypertensive patients and to evaluate whether low GFR can improve the prediction of these outcomes in addition to conventional cardiovascular risk factors. Methods and Findings
This is an observational prospective study and 3,711 eligible hypertensive patients aged =5 years from rural areas of China were used for the present analysis. The associations between eGFR and outcomes, followed by a median of 4.9 years, were evaluated using Cox proportional hazards models adjusting for other potential confounders. Low eGFR was independently associated with risk of all-cause mortality, cardiovascular mortality, and incident stroke [multivariable adjusted hazard ratios (95% confidence intervals) for eGFR <60 ml/min/1.73 m2 relative to eGFR =90 ml/min/1.73 m2 were 1.824 (1.047–3.365), 2.371 (1.109–5.068), and 2.493 (1.193–5.212), respectively]. We found no independent association between eGFR and the risk of CHD. For 4-year all-cause and cardiovascular mortality, integrated discrimination improvement (IDI) was positive when eGFR were added to traditional risk factors (1.51%, P?=?0.016, and 1.99%, P?=?0.017, respectively). For stroke and CHD events, net reclassification improvements (NRI) were 5.9% (P?=?0.012) and 1.8% (P?=?0.083) for eGFR, respectively. Conclusions
We have established an inversely independent association between eGFR and all-cause mortality, cardiovascular mortality, and stroke in hypertensive patients in rural areas of China. Further, addition of eGFR significantly improved the prediction of 4-year mortality and stroke over and above that of conventional risk factors. We recommend that eGFR be incorporated into prognostic assessment for patients with hypertension in rural areas of China. Limitations
We did not have sufficient information on atrial fibrillation to control for the potential covariate. These associations should be further confirmed in future.