Recurrent myocardial injury is a strong predictor of ICU mortality in COVID-19

Abstract Funding Acknowledgements Type of funding sources: None. Background Myocardial injury is common in COVID-19, and has been linked to disease severity and mortality.1 Its pathophysiology is not fully elucidated and suggested mechanisms include cytokine damage, endotheliitis, myocarditis, ferroptosis, acute coronary syndrome, oxygen supply-demand mismatch and pulmonary embolism.2-3 Elevated serum cardiac troponin (cTn) – as surrogate marker of myocardial injury – at admission and during hospitalisation is associated with poor outcome.4-5 Comparison of available trial data, however, is unfeasible, because of variable definition and sampling timing/intervals/assays.6 Peculiarly, prognostic cut-off cTn values are often – retrospectively – set much lower than recommended by the guidelines.6-8 Furthermore, robust intensive care unit (ICU) outcome data regarding cTn evolution over time is lacking. Purpose The aim of this study is to critically explore the correlation between myocardial injury and mortality in a tertiary ICU population, using both standard and high cTn cut-off values, and to scrutinise the impact of timing and recurrence. Methods All adult COVID-19 patients admitted to our tertiary ICU between 06/03/2020 and 27/09/2021 were enrolled in this large, retrospective, single centre, cohort study. Patients requiring extracorporeal membrane oxygenation (n=29) were excluded. Myocardial injury was defined – as per ESC guidelines - as high-sensitive cardiac troponin I (hs-cTnI) peak >45ng/L (99th percentile upper reference limit) and ‘major’ myocardial injury was defined as hs-cTnI >135ng/L (3x upper limit of normal).6-7 Recurrent myocardial injury was considered if a new hs-cTnI peak above cut-off point occurred after normalisation. Patients were categorised according to peak (‘standard’ >45ng/L vs ‘high’ >135ng/L), onset (‘on admission’ vs ‘during hospitalisation’) and recurrence (‘single’ vs ‘multiple’) of hs-cTnI rise. Statistical analysis was done by logistic regression, using IBM SPSS version 28. Results A total of 193 COVID-19 ICU patients were included. Patient characteristics, disease severity indices and treatment details are summarised in Table 1. The mean age was 62 years and ICU mortality was 22%. Myocardial injury overall occurred in 51% (n=99) and major myocardial injury occurred in 39% (n=75) of patients. Obtained results were consistent across standard (>45ng/L) and high (>135ng/L) hs-cTnI cut-off groups. Myocardial injury only present at ICU admission was not correlated with mortality in our cohort, contradicting prior (mostly non-ICU) trials. Single hs-cTnI peak during ICU hospitalisation, however, and, especially recurrent hs-cTnI peaks were strong predictors of ICU mortality. The respective incidence, odds ratio, 95% confidence intervals and p-values are displayed in Table 2. Conclusion Myocardial injury occurring during ICU hospitalisation and, especially, recurrent myocardial injury is strongly correlated with ICU mortality in COVID-19 patients.