A prediction score for significant coronary artery disease in Chinese patients ≥50 years old referred for rheumatic valvular heart disease surgery. (18th December 2017)

• Record Type: Journal Article
• Authors: Xu, Zhenjun; Pan, Jun; Chen, Tao; Zhou, Qing; Wang, Qiang; Cao, Hailong; Fan, Fudong; Luo, Xuan; Ge, Min; Wang, Dongjin
• Is Part Of: Interactive cardiovascular and thoracic surgery. Volume 26:Number 4(2018)
• Page Start: 623
• Page End: 630
• Publication Date: 2017-12-18
• Subjects: Rheumatic valvular heart disease -- Coronary artery disease -- Coronary angiography -- Logistic regression model; Chest -- Surgery -- Periodicals; Cardiovascular system -- Diseases -- Periodicals; 616.1
• DOI: 10.1093/icvts/ivx408
• Languages: English
• ISSNs: 1569-9293

Abstract

Abstract: OBJECTIVES: Our goal was to establish a prediction score and protocol for the preoperative prediction of significant coronary artery disease (CAD) in patients with rheumatic valvular heart disease. METHODS: Using multivariate logistic regression analysis, we validated the model based on 490 patients without a history of myocardial infarction and who underwent preoperative screening coronary angiography. Significant CAD was defined as ≥50% narrowing of the diameter of the lumen of the left main coronary artery or ≥70% narrowing of the diameter of the lumen of the left anterior descending coronary artery, left circumflex artery or right coronary artery. RESULTS: Significant CAD was present in 9.8% of patients. Age, smoking, diabetes mellitus, diastolic blood pressure, low-density lipoprotein cholesterol and ischaemia evident on an electrocardiogram were independently associated with significant CAD and were entered into the multivariate model. According to the logistic regression predictive risk score, preoperative coronary angiography is recommended in (i) postmenopausal women between 50 and 59 years of age with ≥9.1% logistic regression predictive risk score; (ii) postmenopausal women who are ≥60 years old with a logistic regression predictive risk score ≥6.6% and (iii) men ≥50 years old whose logistic regression predictive risk score was ≥2.8%. Based on this predictive model, 246 (50.2%) preoperative coronary angiograms could be safely avoided. The negative predictive value of the model was 98.8% (246 of 249). CONCLUSIONS: This model was accurate for the preoperative prediction of significant CAD in patients with rheumatic valvular heart disease. This model must be validated in larger cohorts and various populations.