Increased Concentrations of Cardiac Troponin I Are Equivalent to Increased Cardiac Troponin T in Identifying Chest Pain Patients at Short-Term Risk of Myocardial Infarction

A recent paper by Christenson et al. (1) presented a comparative analysis of cardiac troponin I (cTnI) and T (cTnT). Their objective was to stratify patients for short-term risk of an acute coronary event on the basis of initial marker concentration as determined within 3.5 h of ischemic symptoms. The authors suggest that cTnT is a better predictor than cTnI on the basis of comparisons of the area under the curve (AUC) of ROC curves for cTnT (AUC = 0.68) and cTnI (AUC = 0.64). However, although weakly significant (P = 0.0375), the 95% confidence interval (CI) for the two areas overlap almost completely (95% CI for cTnI, 0.56–0.72; 95% CI for cTnT, 0.6–0.75). The fact that the samples were not tested for the two analytes simultaneously, but >1 year apart, could easily account for this small difference. Using cutoffs of >0.1 ng/mL (0.1 μg/L) for cTnT and >1.5 ng/mL (1.5 μg/L) for cTnI, the authors found 90.4% concordance between the two assays. After correction for samples falling within the 95% CI around the cutoffs, the overall concordance rose to 93%. Of the 66 cTnT-positive/cTnI-negative patients, 7.5% died and 70% suffered from an acute myocardial infarction (AMI). The authors use this as evidence that cTnT is more sensitive than cTnI for risk stratification but neglect to present data recalculating the cTnI sensitivity/specificity profile using a cutoff of 0.4 ng/mL (0.4 μg/L), which has been demonstrated to be the more appropriate concentration for cTnI in risk stratification of chest pain patients (2). The authors cite that the cutoffs used are in accordance with the manufacturer’s specifications; however, the Dade Stratus® Cardiac Troponin-I package insert (3) describes in detail the multicenter, outcomes-based study of Antman et al. (2), where 0.4 ng/mL (0.4 μg/L) was used as a cutoff. The insert also states that in two separate studies, the 97.5 percentile distribution of ostensibly healthy individuals (n = 156) and patients with chest pain but confirmed not to have AMI (n = 149) was 0.4–0.6 ng/mL (0.4–0.6 μg/L).