Abstract Background High-sensitivity troponin assays allow for accurate and rapid rule-in or rule-out of myocardial infarction (MI) among patients with acute-onset chest pain. However, prognostic implications of serial high-sensitivity troponin concentrations are unknown. Purpose To determine short- and long-term prognostic implications of high-sensitivity troponin T (hsTnT) concentrations and their changes from baseline, in patients with suspected acute coronary syndrome. Methods Retrospective cohort study based on Danish national registries. We identified all patients discharged from the hospital with either MI, unstable angina, suspected MI, or chest pain from January 2012 through December 2019 and merged these individuals with all records of two serial hsTnT measurements obtained ≤7 hours apart during the same hospitalization. The primary outcome was death at days 0–30 and 31–365. Prognostic implications of serial hsTnT were examined in accordance with the 2012 ESC algorithm stratifying patients for normal baseline concentrations and relative changes of 20% and 50% from baseline. In case of a normal baseline concentration, 20% and 50% of the upper reference level (14 ng/l) were used as thresholds instead, i.e., 3 ng/l and 7 ng/l, respectively. Absolute risks were calculated through multivariable logistic regression with average treatment effect modeling (G-formula). Results Complete data were available in 28,902 individuals (median age [25th-75th percentile] 65.2 [53.4–75.4] years, 11,632 [40.2%] women). Of these, 11,116 (38.5%) had a final diagnosis of MI, 1518 (5.3%) of unstable angina, and 16,268 (56.3%) of either suspected MI or chest pain. Median baseline hsTnT was 18 ng/l (25th-75th percentile, 10–69), second hsTnT 21 ng/l (25th-75th percentile, 10–248), relative hsTnT change 3.6% (25th-75th percentile, 0–66.7), and time between samples 4.0 hours (25th-75th percentile, 3.2–5.4). Most patients had either two normal hsTnT concentrations (9483, 32.8%) or two elevated hsTnT concentrations (18,235, 63.1%). At 30 days, 796 (2.8%) individuals had died, while an additional 1287 (4.6% of 30-day survivors) died between days 31–365. Baseline hsTnT and the relative hsTnT change both displayed a significant, non-linear association with death and interacted with each other (P<0.001). Tables 1 and 2 show the standardized, absolute risks of death (with 95% confidence intervals) from days 0–30 and from days 31–365, respectively. Patients with two normal hsTnT concentrations had very low mortality rates, irrespective of the magnitude of relative change. Conversely, patients with two elevated hsTnT concentrations consistently had high mortality rates. Conclusions This is the first study to assess both short- and long-term outcomes as a function of both baseline hsTnT and its change from first to second measurement. In general, patients with two normal hsTnT concentrations have an excellent prognosis while those with two elevated concentrations require scrutiny. Funding Acknowledgement Type of funding sources: None.
Read full abstract