Predictive role of high sensitivity troponin T within four hours from presentation of acute coronary syndrome in elderly patients

BACKGROUND: Previous studies indicate that the introduction of high-sensitivity troponin T (HsTnT) as a diagnostic tool for chest pain patients in the emergency department (ED) creates a high rate of false-positive tests. In the present study, we aimed to evaluate if the diagnostic performance of HsTnT for acute coronary syndrome (ACS) up to 3–4 h after presentation in elderly patients can be improved. METHODS: A total of 477 consecutive patients ≥ 75 years, admitted to in-hospital care for chest pain suspicious of ACS, were retrospectively included. HsTnT values at presentation (0 h) and at 3–4 h were analysed. Receiver operating characteristic (ROC) curves were created for absolute and relative changes from 0 to 3–4 h. ACS, non-elective percutaneous coronary intervention, coronary artery bypass grafting and death of all causes were recorded for all patients during a follow-up of 60 days. Sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV) were analysed for different HsTnT cut-off values at 0 and 3–4 h and for the combination of a HsTnT at presentation and an absolute change from 0 to 3–4 h. RESULTS: Twenty-seven percent of the patients had ACS and 21 % acute myocardial infarction (AMI) during the hospital stay. The standard cut-off 14 ng/L gave sensitivity and NPV for ACS of 88 and 90 % at 3–4 h. Specificity and PPV was 38 and 32 % respectively. Analysing for non-ST elevation myocardial infarction (NSTEMI) alone gave a sensitivity and NPV of 100 % but did not improve specificity and PPV. The area under the ROC-curve was larger for absolute than relative HsTnT changes from 0 to 3–4 h. A combination of HsTnT at presentation > 30 ng/L and/or a change > 5 ng/L up to 3–4 h gave a 63 % specificity and a PPV of 46 %, a 99 % sensitivity and a NPV of 99 % for NSTEMI. CONCLUSION: Our study indicates that HsTnT can neither exclude nor confirm ACS within 3–4 h from presentation in patients ≥ 75 years. NSTEMI can be excluded with HsTnT within 3–4 h, but HsTnT cannot be used to rule in NSTEMI during the first 3–4 h, not even by using a combination of the initial HsTnT result and the change from 0 to 3–4 h. With combined criteria, the majority of the positive tests were still false positive. Our results indicate that in patients > 75 years, HsTnT should be used primarily as an early rule-out tool for AMI.