Possible Misdiagnosis of Myocardial Infarction Using Regulatory-Approved and Close-to-Bioequivalent Upper Limits of Normal for Cardiac Troponin.

BACKGROUND

Possible misdiagnosis of acute myocardial infarction (AMI) may occur due to inappropriate upper limit of normal (ULN) for cardiac troponin and has the potential to harm patients. In this observational international multicenter study, we aimed to assess to what extent the novel hs-cTn-assays are affected.

METHODS

A total of 6646 patients presenting with suspected AMI to the emergency department were enrolled. All level pairs (n=18 732) of 4 widely used high-sensitivity cardiac troponin T/I (hs-cTnT/I) assays using (1) the regulatory-approved uniform and sex-specific clinical ULN and (2) mathematically derived close-to-bioequivalent ULNs were assessed. The primary outcome was the quantification of the incidence of inconsistencies in the diagnosis of AMI. Inconsistency was defined as hs-cTnT/I concentration above the recommended ULN in one but not the other assay: for example, hs-cTnT-Elecsys+/hs-cTnI-Architect- or hs-cTnT-Elecsys-/hs-cTnI-Architect+.

RESULTS

AMI was the adjudicated diagnosis in 1422 patients (21.4%). When the regulatory-approved uniform ULN was used, the rate of inconsistent AMI diagnoses was 17.6% (Elecsys/Architect), 18.8% (Elecsys/Centaur), 14.2% (Elecsys/Access), 4.9% (Architect/Centaur), 8.3% (Architect/Access), and 7.4% (Access/Centaur), respectively. Overall, diagnostic mismatches were not decreased, but in fact increased using regulatory-approved sex-specific ULNs. In women as compared with men, they were 23.8% versus 17.6% (Elecsys/Architect), 30.1% versus 19.1% (Elecsys/Centaur), 23.2% versus 15% (Elecsys/Access), 7.2% versus 4.5% (Architect/Centaur), 8.3% versus 8.7% (Architect/Access) and 7.8% versus 8.2% (Access/Centaur), respectively. Using close-to-bioequivalent ULNs reduced inconsistencies by 15% to 20% (<0.001). Findings were confirmed in a sensitivity analysis among all level pairs with final diagnosis of AMI (mismatches in 7.3%-20.5%).

CONCLUSIONS

Current regulatory-approved uniform and sex-specific ULNs for hs-cTnT/I result in discordances in binary assay results, possibly impacting the diagnosis of AMI. A regulatory process that defines bioequivalent ULNs could reduce inconsistencies significantly.

REGISTRATION

URL: https://www.clinicaltrials.gov; Unique identifier: NCT00470587.