Effect of Diabetes on the Performance of Algorithms for the Detection of AMI Without ST-Elevation

Heart attack
Heart attack, conceptual computer artwork.
The performance of a European Society of Cardiology algorithm in ruling out acute myocardial infarction without ST-elevation was comparable in patients with and without diabetes mellitus.

The performance of the European Society of Cardiology (ESC) algorithm, ESC 0/1-h, in ruling out acute myocardial infarction (AMI) without ST-elevation was comparable in patients with and without diabetes mellitus (DM), according to a study published in Diabetes Care.

The ESC 0/1-h and 0/3-h algorithms are used to diagnose patients with suspected acute non-ST-elevation myocardial infarction (NSTEMI). The levels of high-sensitivity cardiac troponin (hs-cTn) are often chronically elevated in individuals with DM, rendering it difficult to identify NSTEMI in this patient population. Investigators sought to assess whether the presence of DM affects the diagnostic abilities of 2 ESC algorithms in patients presenting to the emergency department with symptoms indicative of AMI.

In this secondary analysis of 2 studies, the Biomarkers in Acute Cardiac Care (BACC) and stenoCardia trials (ClinicalTrials.gov identifiers NCT02355457 and NCT03227159, respectively), 3,681 patients (mean age, 64.0 years; 64.2% men) with prospectively evaluated suspected acute NSTEMI with (n=563) and without DM, were enrolled. Data from the Advantageous Predictors of Acute Coronary Syndromes study (APACE; n=2895; ClinicalTrials.gov identifier NCT00470587) were used to calculate and externally validate alternative cutoffs for the algorithms.

The levels of hs-cTn were measured at admission, 1 hour (only in the BACC study), and 3 hours (in both studies). Negative and positive predictive values (NPV and PPV, respectively) for NSTEMI were calculated for both algorithms. The study’s primary safety outcome was the NPV for NSTEMI (ie, for ruling out the condition), and the primary efficacy outcome was the PPV for ruling in NSTEMI. The sensitivity and specificity of both algorithms were the study’s secondary endpoints.

Of 563 participants with DM, 137 (24.3%) had comorbid acute NSTEMI, compared with 15.9% of patients without DM (P <.001). Participants with DM were older and had more cardiovascular risk factors and comorbidities.

Related Articles

The ESC 0/1-h algorithm had a comparable NPV for NSTEMI in patients with and without DM (absolute difference [AD], -1.50; 95% CI, -5.95 to 2.96; P =.54), but the ESC 0/3-h algorithm had a lower NPV in patients with vs without DM (AD, -2.27; 95% CI, -4.47 to -0.07; P =.004). The diagnostic performance to rule-in NSTEMI was comparable for patients with vs without DM with both algorithms: ESC 0/1-h (AD, -6.59; 95% CI, -19.53 to 6.35; P =.34) and ESC 0/3-h (AD, 1.03; 95% CI, -7.63 to 9.7; P =.88).

The sensitivity for ruling out NSTEMI was comparable in patients with vs without DM with both ESC0/1-h (AD, -0.9; 95% CI, -5.1 to 3.3; P =1.00) and ESC 0/3-h (AD, -4.0; 95% CI, -10.4 to 2.4; P =.19) algorithms. The specificity for ruling in NSTEMI was higher for patients without vs with DM when using both the ESC 0/1-h (AD, -6.9; 95% CI -12.5 to -1.2; P =.0035) and ESC 0/3-h (AD, -4.4; 95% CI, -8.2 to 0.6; P =.01) algorithms. The use of alternative cutoffs improved the PPV of both algorithms.

Study strengths include large sample sizes and external validation of proposed alternative cutoffs. Study limitations include the sole use of data from the BACC study to evaluate the 0/1-h algorithm, possible misclassification of AMI and DM, and a lack of accounting for disease duration.

“Although alternative cutoffs might be helpful, patients with DM remain a high-risk population in whom identification of AMI is challenging and who require careful clinical evaluation,” noted the authors.

Follow @EndoAdvisor


Haller PM, Boeddinghaus J, Neumann JT, et al. Performance of the ESC 0/1-h and 0/3-h algorithm for the rapid identification of myocardial infarction without ST-elevation in patients with diabetes. Diabetes Care. 2019;43(2):460-467. doi: 10.2337/dc19-1327

This article originally appeared on The Cardiology Advisor