Prospective evaluation of emergency department patients with potential coronary syndromes using initial absolute CK-MB vs. CK-MB relative index.

Publication/Presentation Date

5-1-2003

Abstract

We compared the predictive properties of an initial absolute creatine kinase-MB (CK-MB) to creatine kinase-MB relative index (CK-MB RI) for detecting acute myocardial infarction (AMI), acute coronary syndromes (ACS), and serious cardiac events (SCE). Consecutive patients > 24 years of age with chest pain who received an electrocardiogram (EKG) as part of their Emergency Department (ED) evaluation had CK and CK-MB drawn at presentation. Patients were followed prospectively during their hospital course. The main outcome was AMI, ACS or SCE (death, AMI, dysrhythmias, CHF, PTCA/stent, CABG) within 30 days. The sensitivity, specificity, PPV and NPV of CK-MB and CK-MB RI to predict AMI, ACS, and SCE were calculated with 95% CIs. We enrolled 2028 patients. There were 105 patients (5.2%) with AMI, 266 (13.1%) with ACS, and 150 with SCE (7.4%). Absolute CK-MB had a higher sensitivity than CK-MB RI for AMI (52.0 vs. 46.9, respectively), ACS (23.5 vs. 20.8, respectively), and SCE (39.6 vs. 36.0, respectively), but a lower specificity than CK-MB RI for AMI (93.2 vs. 96.1, respectively), ACS (93.1 vs. 96.1, respectively) and SCE (93.3 vs. 96.3, respectively); and lower PPV for AMI (35.7 vs. 46.5, respectively), ACS (42.0 vs. 53.4, respectively) and SCE (38.5 vs. 50.5, respectively). The negative predictive values were similar for all outcomes. We conclude that the risk stratification of ED chest pain patients by absolute CK-MB has higher sensitivity, similar NPV, but a lower specificity and PPV than CK-MB relative index for detection of AMI, ACS, and SCE. The optimal test depends upon the relative importance of the sensitivity or specificity for clinical decision-making in an individual patient.

Volume

24

Issue

4

First Page

361

Last Page

367

ISSN

0736-4679

Disciplines

Business Administration, Management, and Operations | Health and Medical Administration | Management Sciences and Quantitative Methods

PubMedID

12745035

Department(s)

Administration and Leadership

Document Type

Article

Share

COinS