Predictive value of glomerular filtration rate estimated by two formulas for all-cause death after interventional coronary artery disease

Bolin Li,Xiyang Li,Bo Wang,Huihui Guo,Yue Wu,Xiaolin Xue,Hongbing Li

DOI: https://doi.org/10.3969/j.issn.1674-4055.2019.10.05

2019-01-01

Abstract:compare the predictive value of glomerular filtration rate (eGFR) levels calculated by the two formulas for all-cause mortality in patients with chronic total occlusive disease (CTO) undergoing percutaneous coronary intervention (PCI). Methods 671 patients with CTO who underwent PCI from June 2013 to October 2017. Preoperative baseline eGFR was calculated using the Chinese modified modification of diet in renal disease (cMDRD) equation and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, and patients were grouped (group A eGFR≥90 ml/min·1.73 m2; Group B 90ml/min·1.73 m2＞eGFR≥60 ml/min·1.73 m2; Group C eGFR＜60 ml/min·1.73 m2). Follow-up was conducted by telephone questionnaire, etc. The main study endpoint was all-cause death. The Cox regression risk model was used to estimate the risk ratio of eGFR, including eGFR univariate analysis and multiple-correction model analysis (corrected age, gender, smoking, body mass index, diabetes, low-density lipoprotein and other common risk factors for coronary heart disease). Results Of the 671 CTO patients who underwent PCI, 471 (70.2%) were successful, 200 (29.8%) failed, and 64 (9.5%) died. Univariate analysis of eGFR-cys showed that the difference in HR values between the three groups was statistically significant; the trend test was P<0.001, and the difference was statistically significant. eGFR-cys as a continuous variable (per 10 ml/min·1.73 m2) had HR=0.810, 95%CI: 0.374~0.894, P＜0.01. In combination with the trend test, there was a linear relationship between eGFR-cys level and all-cause mortality risk. For every 10 ml/min·1.73 m2 of eGFR, the risk of all-cause mortality decreased by 19%. The difference was statistically significant (P＜0.05), after the model correction, the above differences are still statistically significant. The difference in HR and trend test between the three groups was statistically significant by eGFR-cMDRD single factor analysis, but the difference was not statistically significant after model correction. Conclusions In patients with CTO undergoing PCI, eGFR-cys may be a stronger predictor of all-cause mortality than eGFR-cMDRD.