Objectives: Some authors have recommended troponin measurement to stratify patient mortality risk, but it is unclear whether troponin values add to age and routine admission laboratory tests in the prediction of in-hospital mortality of older adult patients without suspected acute coronary syndrome (ACS). The aim of our study was to determine whether troponin testing adds significantly to routine admission laboratory testing in predicting in-hospital mortality in patients without a suspected ACS.
Methods: In 2018-2019, we reviewed all acutely admitted patients aged 60 years or older to Internal Medicine wards of a regional hospital after excluding those admitted to intensive care or with chest pain. The independent variables were troponin, age, sex, and routine admission laboratory tests. The outcome measure was in-hospital mortality. We compared c-statistics and the observed 10% to 90% risk gradients using logistic regression models for age and routine laboratory testing before and after the addition of troponin.
Results: The mortality risk gradient for age and admission laboratory tests was 0.2% to 29.5%. Adding troponin did not increase the gradient significantly (0.2%-34.6%, P = 0.170), and the 95% confidence intervals for the c-statistics overlapped, increasing from 0.845 (0.818-0.876) to 0.866 (0.839-0.892).
Conclusions: In older adult patients without suspected ACS, troponin testing did not improve the prediction of hospital mortality above that of a model including age and common admission blood tests. In the absence of suspected ACS, troponin testing is not needed to predict the hospital mortality of older adult patients.