The ESC algorithm for serial high-sensitivity troponin T changes and long-term outcomes in patients with suspected acute coronary syndrome

Abstract

Abstract Background The fourth universal definition of myocardial infarction (MI) consensus paper suggests that patients with changing troponins not reaching concentrations greater than the 99th percentile may be at high risk and deserve close scrutiny. Purpose To determine long-term prognostic implications of high-sensitivity troponin T (hs-TnT) levels and their relative change (Δ) from baseline in subjects with suspected acute coronary syndrome (ACS). Methods We conducted a retrospective cohort study through individual participant-level linkage between Danish national registries, including subjects with a final discharge diagnosis of acute MI, unstable angina, suspected MI, or chest pain from March 2013 through December 2016 who had a record of at least two serial hs-TnT measurements during hospitalization. Individuals were followed for 12 months, until the occurrence of an event, or censoring due to emigration. Kaplan-Meier analysis and Cox regression, incorporating the competing risk of death, were used to examine the prognostic implications of serial hs-TnT. Subjects were categorized according to whether their first and second hs-TnT were normal/elevated as well as Δhs-TnT and its direction, the latter employing a modified version of the 0/3-hour diagnostic algorithm proposed by ESC, i.e., using cut-offs for Δhs-TnT of 20% and 50%. The primary outcome was a composite of presumed death from cardiovascular causes, recurrent MI, or repeat revascularization (i.e., not including the index event unless the patient died) within 12 months. Results A total of 13,494 individuals (mean age 63.4 years, 39.5% women) were included. Of these, 6129 (45.4%) had a final diagnosis of MI, 941 (7.0%) of unstable angina, and 6414 (47.5%) of either suspected MI or chest pain. Median baseline hs-TnT was 20 ng/l (72.1% elevated), second hs-TnT 27 ng/l (74.6% elevated), Δhs-TnT 4.8%, and time between samples 5.4 hours. At 12 months, 1055 (7.8%) had experienced a primary event. Baseline hs-TnT and Δhs-TnT both displayed a significant association with the primary outcome (P<0.001 for both overall trends and for non-linearity vs. linearity). The Figure shows the prognostic implications of serial hs-TnT. Overall, subjects with two consecutively elevated hs-TnT had the highest 12-month event risk (10.0%), followed by those who went from an elevated to a normal hs-TnT (8.6%), those who went from a normal to an elevated hs-TnT (6.3%), and those with two normal hs-TnT levels (1.6%). The majority either had non-significant Δhs-TnT (−20% to 20%: 56.8%) or a large positive Δhs-TnT (>50%: 30.6%). Individuals with a positive Δhs-TnT (>20%) had a worse prognosis than those without. Conclusions An elevated hs-TnT at any time and Δhs-TnT were both determinants of poorer prognosis in subjects with suspected ACS. Individuals with two normal hs-TnT had a good prognosis, irrespective of their Δhs-TnT. Figure 1 Funding Acknowledgement Type of funding source: None