Abstract 115: Predictors Of A Positive Stress Test Among Patient Presenting To The Emergency Department With Chest Pain

Abstract

Objectives: Patients frequently present to the hospital with chest pain. Once myocardial infarction is ruled out based on EKG and cardiac enzymes, most patients undergo stress testing, but only few patients have a positive test. In ambulatory practice, age, sex and symptomatology can establish pretest probability of the coronary disease. However, there are no studies evaluating the predictors of a positive stress test in the emergency department (ED). We assessed predictors for a positive stress test in patients presenting to our hospital with chest pain. Methods: This is a case-control study conducted on a subset of patients admitted to our tertiary care center with chest pain between 2007 and 2009, and who had an inpatient stress test (n=1474). Using chart review, we identified 87 patients, whose stress tests were positive (abnormals), defined as presence of ischemia on EKG and/or imaging modalities. We then used a pseudorandom number generator to select 194 patients whose stress test results were normal (normals) for comparison. Clinical features of chest pain and CAD risk factors were abstracted from the medical record for comparison. A bivariable screening process was used to identify characteristics for inclusion in a multivariable predictive model. Sex and age were maintained in the model for face validity, and remaining covariates were removed in ascending order of their z-statistics until only those with a two-sided p-value of <0.10 remained. Stata 12.1 (Copyright 2011, StataCorp LP) was used for all analyses. Results: Patients with an abnormal stress test were older and more likely to be male and to have a history of vascular disease. Although patients with abnormal stress test were more likely to have history of hypertension, hyperlipidemia and current or ex-smoking, this difference was not statistically significant. Over half of the patients presented with non-cardiac chest pain and there was no significant difference in the chest pain characteristics between patients who had a normal and an abnormal stress test result. In the final multivariable model, when compared to the normals, abnormals were four times as likely to have a history of revascularization (OR 4.13, 95% CI 2.11, 8.09) and twice as likely to have a history of hyperlipidemia (OR 2.1, 95% CI 1.18, 3.79). They were also more likely to have an EKG suggestive of ischemia at presentation (OR 1.90, 95% CI 1.03, 3.53). Specificity of the model was 89%; sensitivity was 43%, and the c-statistic for the final multivariable model was 0.76, suggesting fair to good discrimination. Conclusions: Among patients presenting to the ED with chest pain, a past history of revascularization and hyperlipidemia and an EKG suggestive of ischemia may independently predict the likelihood of an abnormal stress test. Further validation of this model on an external dataset is necessary.