Abstract Background The clinical implications of high-sensitivity cardiac troponin T (hs-cTnT) measurements in all-comers with AKI in the emergency department (ED) is largely unknown. Purpose To investigate the prevalence of myocardial injury, associations between serum-creatinine concentrations (SCC) and hs-cTnT kinetics, and the accuracy of established hs-cTnT-based myocardial infarction (MI) risk stratification algorithm criteria in patients with AKI. Methods This observational cohort study was based on patient visits to 7 EDs in Sweden from December 9, 2010 to August 31, 2017. All visits by patients ≥18 years fulfilling criteria for AKI with ≥1 hs-cTnT measurement and the exposure was dynamic change in SCC (ΔSCC) during the visits. Linear mixed models were used to estimate linear predictors of kinetic change in hs-cTnT (Δhs-cTnT). Logistic regression models were applied to calculate odds ratios (ORs) for Δhs-cTnT indicative of acute myocardial injury (Δhs-cTnT >20% and elevated hs-cTnT) in relation to ΔSCC among patients without MI, and to assess the diagnostic performance of hs-cTnT for MI in a subgroup of patients presenting with chest pain. Results A total of 15,211 visits by patients with AKI were included, of whom 1174 (8%) had an MI. Four of five (81%) patients without MI had myocardial injury (hs-cTnT >14 ng/l), and almost one of three (31%) had acute myocardial injury. These entities were common in both cardiac and non-cardiac conditions. The Δhs-cTnT in non-MI patients was 1.8-fold (β: 1.78, 95% CI: 1.62-1.96) in the highest quartile of ΔSCC and was paralleled by a 2-fold risk of acute myocardial injury (OR: 2.32, 95% CI: 2.08-2.59), compared with the lowest ΔSCC quartile. Using a 0 h hs-cTnT cut-off ≥52 ng/l assigned 627 (26%) patients to a high-risk group in whom the specificity and positive predictive value (PPV) for MI was low (78.5% (95% CI: 76.7-80.2) and 27.6% (95% CI: 24.1-31.3), respectively). Conclusions Hs-cTnT concentrations indicative of acute myocardial injury are frequently observed among all-comers with AKI in the ED and are associated with dynamic changes in SCC. These observations are accompanied by poor performance of recommended hs-cTnT-based algorithms for MI risk stratification in patients with chest pain.
Read full abstract