Clinical risk scoring beyond initial troponin values: results from a large, prospective, unselected acute chest pain population.

BACKGROUND: Risk stratifying the diverse group of patients who present to hospital with chest discomfort remains challenging. Current clinical risk models, typically derived from selected populations, are limited by their relative complexity and the absence of a well-defined role of troponin. OBJECTIVE: To derive a simple clinical risk score from a large, unselected population of patients with chest discomfort and to delineate the prognostic value of an initial troponin measurement. METHODS: Prospective, consecutive data were collected from patients who presented to a tertiary care hospital. Multivariate analysis was used to identify variables predictive of the primary end point: death, nonfatal myocardial infarction or revascularization at 30 days. Integer values were assigned, generating a risk score to quantify individual patient risk. RESULTS: Among 1054 patients, predictor variables included ST-segment deviation (strongest predictor -- assigned two points), male sex, prior congestive heart failure, three or more cardiac risk factors and prior acetylsalicylic acid use (one point each). There was a progressive increase in events with increasing total score (P<0.0001), with a 15-fold gradient from scores of 0 to 4 and greater. Although a negative troponin measurement was associated with fewer events for all scores, patients with higher scores remained exposed to substantial risk. A negative initial troponin measurement conferred a negative predictive value of 97.3% (95% CI 93.7% to 99.1%) among patients with a risk score of 0. CONCLUSION: Significant 30-day events occurred in patients with elevated risk scores, despite negative initial troponin measurements, emphasizing the importance of clinical risk stratification. This simple clinical risk score, in conjunction with a single troponin I measurement, facilitates triage of patients who present to hospital with chest discomfort.