High sensitivity troponin as a 60-day mortality marker in patients with COVID-19 in intensive care at a Latin American hospital

Abstract

Abstract Introduction Coronavirus disease (COVID) - 19 presents with periodic outbreaks with a large number of critical cases that overwhelm the health system. One of the objectives of screening and management is to have markers that can efficiently stratify the individual risk of experiencing adverse results. High-sensitivity troponin has been proposed as one of the possible prognostic markers. Purpose To evaluate the prevalence of troponin elevation and its performance for predicting 60-days mortality in patients in ICU due to COVID - 19 in a Latin American population. Methodology A single-center retrospective cohort study was carried out in which patients in ICU for COVID-19 confirmed by RT-PC between May and December 2020 were included. Patients with previous conditions that determined a limitation of therapeutic effort and pregnant women were excluded. The outcome was 60-days mortality. Kaplan Meier survival curves were constructed, a ROC curve was generated for the prediction of death according to troponin levels by calculating the area under the curve. Finally, a Cox proportional hazards model was generated. Results 296 patients were included, 110 (37.2%) female, the mean age was 60 (SD 14) years and the duration of symptoms before admission to the ICU was 8.4 (SD 4.2) days. 81.4% of the patients had at least one comorbidity, 117 (39.5%) had hypertension, 65 (21.9%) had diabetes, and 63 (21.3%) had chronic lung disease. Mortality occurred in 151 cases (51%), 225 (76%) presented shock, 242 (81.8%) severe ARDS (PaO2 / FiO2 <100 mmHg), 232 (78.4%) required invasive ventilatory support, and 153 (5.7%) developed acute kidney injury. The APACHE II average was 13.1 and SOFA 4.8. Troponin was positive in 118 (39.9%) patients, 29.6% of the survivors and 49.7% in non survivors. The AUC was 0.65 (95% CI 0.59–0.71) for predicting mortality (Figure 1). The Kaplan Meier curves are shown in Figure 2 with a p less than 0.001 by achieving the test. In the bivariate analysis, troponin obtained a HR of 1.94 and p<0.001, but in the multivariate model it was 1.39 (95% CI of 0.21–1.56) and p of 0.12. Conclusions High sensitivity troponin was found to be elevated in patients hospitalized for COVID-19 in the intensive care unit who died, with regular discrimination (AUC: 0.65), even though the univariate analysis by the Cox and Kaplan Meier methods showed association with 60-days mortality, this was not maintained in the multivariate model. Funding Acknowledgement Type of funding sources: None. Figure 1. ROC curveFigure 2. Kaplan Meier survival curves