Early vs. Late Onset Cardiac Injury and Mortality in Hospitalized COVID-19 Patients in Wuhan

Background: Increasing evidence points to cardiac injury (CI) as a common coronavirus disease 2019 (COVID-19) related complication. The characteristics of early CI (occurred within 72 h of admission) and late CI (occurred after 72 h of admission) and its association with mortality in COVID-19 patients is unknown. Methods: This retrospective study analyzed patients confirmed with COVID-19 in Union Hospital (Wuhan, China) from Jan 29th to Mar 15th, 2020. Clinical outcomes (discharge, or death) were monitored to April 15, 2020, the latest date of follow-up. Demographic, clinical, laboratory, as well as treatment and prognosis were collected and analyzed in patients with early, late CI and without CI. Results: A total of 196 COVID-19 patients were included for analysis. The median age was 65 years [interquartile range (IQR) 56–73 years], and 112 (57.1%) were male. Of the 196 COVID-19 patients, 49 (25.0%) patients had early and 20 (10.2%) patients had late CI, 56.6% developed Acute-Respiratory-Distress-Syndrome (ARDS) and 43 (21.9%) patients died. Patients with any CI were more likely to have developed ARDS (87.0 vs. 40.2%) and had a higher in-hospital mortality than those without (52.2 vs. 5.5%, P < 0.001). Among CI subtypes, a significantly higher risk of in-hospital death was found in patients with early CI with recurrence [19/49 patients, adjusted odds ratio (OR) = 7.184, 95% CI 1.472–35.071] and patients with late CI (adjusted OR = 5.019, 95% CI 1.125–22.388) compared to patients with early CI but no recurrence. Conclusions: CI can occur early on or late after, the initial 72 h of admission and is associated with ARDS and an increased risk of in-hospital mortality. Both late CI and recurrent CI after the initial episode were associated with worse outcomes than patients with early CI alone. This study highlights the importance of early examination and periodical monitoring of cardiac biomarkers, especially for patients with early CI or at risk of clinical deterioration.


INTRODUCTION
Coronavirus disease 2019  caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has affected over 200 countries (1). With the increasing number of confirmed cases, the cardiovascular manifestations associated by this highly contagious viral infection have gained more and more attention. Several observational studies have found that between 7.2 and 37.5% of COVID-19 patients had cardiac injury (CI) which was associated with higher mortality in COVID-19 patients (2)(3)(4)(5)(6)(7)(8)(9)(10)(11). However, a portion of CI does not occur on admission and the association of the timing of CI with prognosis is unknown. Furthermore, the clinical features and risk factors associated with early or late onset CI in COVID-19 patients have not been formally evaluated. The clinical sequence preceding and following CI at the time of admission may provide additional understanding of the pathogenesis associated with CI in COVID-19. Therefore, this study compared the clinical characteristics, risk factors and prognostic value of early vs. late onset of CI in COVID-19 patients.

Study Design and Participants
We performed this retrospective study at Union Hospital (Affiliated Tongji Medical College, Huazhong University of Science and Technology) Wuhan, China. The West Branch of Union Hospital was one of the major designated hospitals for critically ill COVID-19 patients. We enrolled 429 consecutive patients with confirmed COVID-19, according to the WHO interim guidance criteria (12), who were either discharged alive or died during hospitalization from Jan 29th to April 15th, 2020. Only participants who had high-sensitivity troponin I (hs-TNI) measured before and after 72 h from admission during their hospitalization were included in the study (233 patients excluded). The study was approved by the ethics committee of the Union hospital, Tongji Medical College, Huazhong University of Science and Technology. Per institutional policy, written informed consent was waived for all participants with emerging infectious diseases.

Data Collection
Data were extracted from the electronic medical records including demographic information and clinical characteristics (i.e., vital signs, symptoms, laboratory findings, medical history, underlying comorbidities, treatments, complications, and outcomes) of the participants on admission and during hospitalization. The date of illness onset was defined as the day when symptoms of COVID-19 as defined by the World Health Organization (12) were appreciated. Laboratory measurements within and after 72 h of admission were collected. If multiple measurements were available, the patient's first abnormal measurements, both within and after 72 h, were recorded for the determination of the timing of the CI. The duration from the onset of admission to the onset of clinical complications and death in the hospital were also recorded. Clinical outcomes (discharge and mortality) were monitored up to April 15, 2020, the last date of follow-up. Complete hospitalization data was available in all patients included in the study.

Timing of CI
The hs-TNI data for each patient were collected from admission to discharge or death. COVID-19 related CI was defined as the serum levels of cardiac high-sensitivity troponin I (hs-TNI) above the 99th percentile upper reference limit in a patient diagnosed with COVID-19 per Huang et al. and Shi et al. (2,(7)(8)(9)(10). Early CI was defined as CI that occurred within 72 h of admission, whereas late CI was defined as occurring after 72 h of admission. We also defined a subgroup of recurrent CI within the early CI group as a second rise of hs-TNI value of >20% from its previous value after 72 h of admission.

Non-cardiac Complications
Acute respiratory distress syndrome (ARDS) was defined according to the World Health Organization interim guidance criteria (13). Acute kidney injury was identified according to the KDIGO clinical practice guidelines as an increase in serum creatinine by ≥0.3 mg/dl (≥26.5 µmol/l) within 48 h or by 1.5 times of the baseline values (14). Coagulation dysfunction was defined as a >3-s prolongation of prothrombin time (PT) or a 5s prolongation of activated partial thromboplastin time (APTT). Thrombocytopenia was characterized by a platelet count <125 × 10 9 /L (15).

Statistical Analysis
Categorical variables were expressed as number (%), and continuous variables were expressed as mean ± standard deviation (SD) or median [interquartile range (IQR)]. The normality of the distribution was tested with the Shapiro-Wilk normality test. Differences among the three groups (without CI, early CI or late CI) were assessed by ANOVA for normally distributed and Kruskal-Wallis H-test for nonnormally distributed continuous variables. Categorical variables were compared by Chi-square or Fisher exact test where applicable. In-hospital survival curves of four groups of patients with early but no recurrent CI, early with recurrent CI, late CI, and no CI were estimated with the Kaplan-Meier method and the groups compared with the log-rank test. Univariate and multivariate logistic regression analyses were used to determine the independent risk associated with each of the four groups (early but no recurrent CI, early with recurrent CI, late CI, and no CI) of patients with in-hospital death, adjusted by known risk factors of COVID-19 mortality in the literature [age, sex, respiratory rate, heart rate, SpO 2 , temperature, mean arterial pressure, coma, hypertension history, Lymphocyte count, Creactive protein (CRP), and lactate dehydrogenase (LDH)] (16)(17)(18). All statistical analyses were performed with SPSS version 24.0 (Statistical Package for the Social Sciences, Chicago, Illinois), and STATA software version 10 (StataCorp, Texas, USA). A twotailed P-value of <0.05 was considered statistically significant.

Clinical Characteristics, Laboratory Findings, and Treatments Within 72 h of Admission
Among the 429 patients, 100 patients were excluded due to missing hs-TNI data during the entire hospitalization, 133 patients were excluded due to missing hs-TNI data within 72 h of admission, the remaining 196 patients were included for analysis. Compared with non-CI group, patients with early and late CI presented with more abnormal laboratory findings within 72 h of admission including lower lymphocyte and platelet counts, higher inflammation-related indices [CRP, procalcitonin (PCT)] and further elevations in liver and renal function indices.
Concerning the treatment of the 196 patients within 72 h of admission, there was no difference in the antiviral (P = 0.551) or antibiotic therapy (P = 0.235) among these three groups. However, compared with the non-CI group, more patients with early and late CI received glucocorticoid therapy (18.9 vs. 38.8% and 30.0%, P = 0.021), high-flow oxygen (18.1 vs. 69.4% and 80.0%, P <0.001), invasive mechanical ventilation (0.8 vs. 4.1% and 15.0%; P = 0.005), noninvasive mechanical ventilation (0.0 vs. 8.2% and 15.0%; P <0.001), and more subjects were transferred to the intensive care unit (ICU) (0.8 vs. 10.2% and 5.0%; P = 0.007) ( Table 1). Patients with early and late CI were more likely to have developed ARDS, ICU transfer and receive invasive mechanical ventilation (IMV) during their hospitalization compared to the non-CI group ( Table 2). A majority, 87.0% (60/69) of patients with CI vs. 40.2% (51/127) without CI, developed ARDS. Overall, the median time from admission to ARDS was 4 days, to acute kidney injury was 7 days, to late CI was 11 days, and to coagulation dysfunction was 11 days for all patients.
Using the onset of ARDS as the temporal reference point, 15/20 (75%) patients with late CI had ARDS before their CI (Figure 2), with a median time between ARDS and CI of 4 days (IQR 3-17 days).

Cardiac Injury and Mortality
As of April 15, 2020, 153 patients (78.1%) were discharged and 43 patients (21.9%) died in the hospital. Patients with any CI had significantly higher in-hospital mortality than those without (52.2 vs. 5.5%, P < 0.001) (Figure 3). Among CI subtypes, multivariable regression modeling showed that compared to patients with early CI but no recurrence, a significantly higher risk of in-hospital death was found in patients with early CI with recurrence [odds ratio (OR) = 7.184, P = 0.015] and patients with late CI (OR = 5.019, P = 0.034) ( Table 3).

DISCUSSION
In this cohort of 196 hospitalized COVID-19 patients from Wuhan, China, we found 35.2% had evidence of CI during hospitalization that included 25.0% of patients with early CI and 10.2% patients with late CI. In addition, there were 38.8% of patients with early CI who also had recurrent CI. Patients with any CI had significantly higher incidence of ARDS and in-hospital mortality than those without. Moreover, among CI subtypes, a significantly higher risk of in-hospital death was found in patients with early CI with recurrence and patients with late CI compared to patients with early CI and no recurrence.
Increasingly, researchers are reporting CI in COVID-19 patients, with the prevalence varying from 7.2 to 37.5% (4,7,8,11). In our study, a remarkable 35.2% of patients had any CI. The pathogenesis of CI associated with COVID-19 is still unknown partly due to a dearth of autopsy or biopsy reported in these patients. The following potential mechanisms for CI have been proposed. The first possibility is direct myocardial damage by the virus, because angiotensin-converting enzyme 2 has been identified as a functional receptor for coronaviruses, which is also expressed abundantly in the myocardium (19,20). Mixed literature from autopsy and biopsy case series were reported without conclusive evidence yet (21)(22)(23)(24)(25). The second mechanism is presumably the systemic inflammatory cytokine response, namely cytokine storm (26,27). It can cause the proliferation of highly pro-inflammatory CCR4+CCR6+Th17 cells amongst the CD4+T cells, the expression of high concentrations of cytotoxic particles in CD8+T cells and over activation of T cells in general, all of which lead to a stronger inflammatory response in return (3,21). Inflammation can in turn lead to thromboembolic complications (28). Our study showed that inflammation-related indices (CRP, PCT) were higher      in the CI group compared to the non-CI group. Another mechanism is CI related to hypoxia. The balance between the oxygen demand and supply of the myocardium is disrupted during acute hypoxia. A cascade of cellular, biochemical and inflammatory reactions can occur during hypoxia, eventually causing myocardial apoptosis (29). Acute severe hypoxia can   (30). In this study, patients with early or late CI had lower level of SpO 2 on admission and higher incidence of ARDS compared to the non-CI group. There was a large timing overlap between CI and ARDS in the early CI group and 75% of patients with late CI were preceded by ARDS which supports a strong relationship between CI and hypoxia. Lastly, antiviral drugs can cause cardiac insufficiency, arrhythmia or other cardiovascular disorders with variable individual susceptibility (4,31). In the present study, almost all patients (93.4% of COVID-19 patients) were administrated with antiviral drugs either on admission or after admission for which the opportunity of individual susceptibility and/or interaction with the underlying comorbid conditions could contribute to onset of late CI or recurrence of the initial one. In addition, patients in the CI group were more likely to have received glucocorticoid therapy. The relationship between glucocorticoid therapy and cardiac injury remains controversial and is under investigation (32). On the other hand, it is also likely that patients with any CI had higher disease severity for which more treatment was administered. This may suggest that the CI after admission was a sign of disease severity and/or progression.
Recent studies have demonstrated that CI was associated with increased mortality in COVID-19 patients (7,8,33). However, to the best of our knowledge, this is the first study to depict the clinical characteristics of both early CI and late CI and their association with in-hospital mortality. Our study showed that early CI was an independent predictor of death in COVID-19 patients even after accounting for variables that have proven prognostic value in the risk stratification of acutely ill patients, such as respiratory rate, heart rate, SpO 2 , temperature, mean arterial pressure, and coma (16,34). Cardiac injury remained significant after accounting for potential confounding laboratory variables with proven prognostic value in COVID hospitalizations such as lymphocyte count, LDH, CRP (17). This is important because clinicians should not consider patients   being out of danger despite absence of CI in the first 72 h since it can occur late. Similarly, a fall after the early rise in Trop I should not translate into a lower risk status since recurrence of CI can occur. Conversely, we also found that CI is closely related to the occurrence of ARDS as opposed to a primary cardiac event. Accordingly, our study suggested the need for a more systematic assessment of cardiac troponins for risk stratification of COVID-19 patients on admission, with repeat measures for those who already have initial CI or who are at risk for clinical deterioration. With the advent of multiple therapies now to reduce the morbidity and mortality of COVID-19 (35)(36)(37), early identification of higher risk patients with cardiac biomarkers may be helpful to balance the cost vs. the effect of therapy.

LIMITATIONS
Our study has some limitations. First, this was a relatively small sample size and single-center retrospective observational study for which residual confounding cannot be excluded and we did not include all the potential factors associated with mortality in our study. Second, only 196 patients both had the data of hs-TNI within and after 72 h of admission, which may have underestimated the true incidence of CI in our study. Therefore, future studies should be multi-centered, with larger sample size and systematic data collection, in order to promote a more comprehensive understanding of the association between cardiac injury and mortality in hospitalized patients with COVID-19.

CONCLUSIONS
CI is a common condition that can occur early on or late after, the initial 72 h of admission and is associated with ARDS and an increased risk of in-hospital mortality. Both late CI and recurrent CI after the initial episode were associated with worse outcomes than patients with early CI alone. This study highlights the importance of early examination and periodical monitoring of cardiac biomarkers to identify predictors and markers of clinical deterioration in COVID-19 patients to guide intervention.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology. Written informed consent was waived for all participants with emerging infectious diseases. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.