A simple clinical score accurately predicts outcome in a community-based population undergoing stress testing

Scoring systems based on clinical variables are available but not widely applied for evaluating patients with chronic coronary artery disease. The purpose of this study was to validate the prognostic value of a simple clinical scoring system, originally developed in patients referred for a nuclear stress test at a tertiary-care medical center, in a less-selected, community-based population undergoing stress testing for known or suspected coronary artery disease.

Over a 4-year period, 3546 residents of Olmsted County, Minn, underwent stress testing. A previously developed clinical score was calculated for every patient by assigning 1 point each for: male sex, history of myocardial infarction, typical angina, diabetes, insulin use, and each decade of age beginning at age 40. The associations between the assigned score and clinical endpoints were tested using logistic regression. A previously established cutoff point of 5 was used to establish risk groups.

During follow-up (7.6 +/- 2.7 years) there were 363 total deaths, 109 cardiac deaths, and 132 nonfatal myocardial infarctions. The clinical score was strongly associated with overall mortality, cardiac death, and cardiac death/myocardial infarction (P <0.001 for all 3 endpoints). Annual mortality was .6% for the 3076 patients (86%) with a score < or =4, 2.4% for 275 patients (8%) with a score = 5 and 6.2% for the 215 patients (6%) with a score > or =6.

This study enhances the generalizability of this simple clinical score, which was highly effective for risk-stratifying this community-based population undergoing evaluation of chronic coronary artery disease.