Prevalence and impact of cardiac injury on COVID‐19: A systematic review and meta‐analysis

Abstract Background The exact prevalence and impact of cardiac injury in hospitalized patients with coronavirus disease 2019 (COVID‐19) is still controversial. Hence, we aim to investigate prevalence of cardiac injury and its impact on the outcomes in patients with COVID‐19. Hypothesis Cardiac injury is common and associated with higher risk of death. Methods We searched the Cochrane Library, PubMed, MedRxiv, and EMBASE databases from December 2019 to July 15, 2020 for studies that evaluated the prevalence and impact of cardiac injury on COVID‐19. This study has been registered with PROSPERO (International prospective register of systematic reviews)‐registration number‐CRD‐42020186120. Results Twenty‐one studies including 6297 participants were identified. The proportions of cardiac injury were 22%, 28% among hospitalized patients with COVID‐19 or severe COVID‐19 patients, respectively. The incidences of cardiac injury in advance age (>60 years) (30%) was about two‐fold than young patients (<60 years) (15%) with COVID‐19. Severe cases (42%) have seven‐fold prevalence cardiac injury than in their non‐ severe counterparts (6%). Furthermore, cardiac injury is associated with an increased risk of all‐cause mortality in patients with COVID‐19 (OR 10.11, 95% CI 4.49–22.77). In patients with severe COVID‐19, cardiac injury is associated with an increased risk of all‐cause mortality (OR: 16.79, 95% CI: 5.52–51.02). Conclusions This was the first meta‐analysis exploring the prevalence and impact of cardiac injury on COVID‐19. Cardiac injury is common in hospitalized patients and advanced age and severe COVID‐19 patients prone to experience more risk of cardiac injury. Furthermore, cardiac injury is associated with increased risk of all‐cause mortality.


| INTRODUCTION
Coronavirus disease 2019  pandemic is an ongoing global public health emergency, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). As of July 11, 2020, there are 12 322 395 confirmed cases with 556 335 deaths reported in 216 countries. 1 COVID-19 is associated with an increased risk of acute respiratory distress syndrome and has adverse effects on other organ systems, including the heart, kidney, and liver. 2 Cardiovascular disease is a common comorbidity in patients with COVID-19. 3 COVID-19 also lead to many cardiovascular diseases, such as cardiac arrhythmias, myocardial infarction, cardiomyopathy, shock, and cardiac arrest. 4 Cardiac injury is defined as the level of serum troponin with at least one value was above the 99th percentile upper reference according to latest guideline. 5 Recently, a body of the literatures 4, [6][7][8][9][10][11][12][13][14][15][16][17][18] shown that acute cardiac injury occurs in patients with  However, the reported prevalence of cardiac injury on COVID-19 varies from one to another. In a cohort of 138 hospitalized patients with COVID-19 from Wuhan, 4 cardiac injury was reported in 11.2% of hospitalized patients and 22% of intensive care unit (ICU) patients.
Another study from New York City has reported the existence of cardiac injury in 36% of patients hospitalized with COVID-19. 14 The exact prevalence of cardiac injury in hospitalized patients with COVID-19 is still not clear.
Moreover, accumulating evidence 3,4,6-24 supports the notion that acute cardiac injury leads to other poor outcomes patients with cardiac injury had a significantly higher mortality risk than those without cardiac injury. We therefore conducted a systematic review and meta-analysis to investigate the reported prevalence and impact of cardiac injury on COVID-19.

| METHODS
This meta-analysis was performed based on the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines for studies that evaluate healthcare interventions (http://www. prisma-statement.org). 25 This study has been registered with PROS-PERO (International prospective register of systematic reviews)registration number-CRD 42018090474. Severe COVID-19 patients is defined as patients with COVID-19 who are admitted to an ICU patients, while the non-severe COVID-19 patients is defined as hospital patients with COVID-19 who are not admitted to an ICU. Ethical approval is not applicable for this study.

| Literature search
We conducted computerized searches of the Cochrane Library, PubMed, MedRxiv (https://www.medrxiv.org/), and Embase databases from December 2019 to July 2020 (Table S1 in Supplemental material). To identify studies involving relevant COVID-19, we performed the search using the following terms: 2019-novel coronavirus, SARS-CoV-2, COVID-19, and 2019-nCoV. To identify studies involving outcomes, we performed the search using the following terms: cardiac injury, myocardial injury, and cardiac troponin. The definition of cardiac injury was those adopted by the original studies (Table S2 in Supplemental material).
Two groups of keywords were combined using the Boolean operator "and." No language restrictions were applied for the literature search. We also reviewed reference lists, relevant journals, and conference abstracts to identify relevant studies. Additionally, we searched ClinicalTrial.gov (https://www.clinicaltrials.gov/) to obtain information on studies that were terminated before being published.

| Study selection
Studies were considered eligible if they (1) were cohort or nested case-control studies; (2) reported the prevalence of cardiac injury; (3) reported the association between cardiac injury and outcomes (e.g. all-cause death) in this illness. Certain publication types (e.g., reviews, editorials, and animal studies) or studies with insufficient data were excluded from this analysis.

| Data extraction and quality assessment
Two researchers (Linghua Fu and Yuhao Su) extracted the basic characteristics, including the first author, publication year, geographical location, study type, participants (sex, age, and sample size), and duration of follow-up. For multiple reports using the same data, we included the articles with the longest follow-up or the largest numbers of participants. The quality of all prevalence studies was independently assessed by the Joanna Briggs Institute critical appraisal checklist. In addition, we assessing the quality of included studies involved with the impact of cardiac injury by Newcastle-Ottawa quality scale (NOS), which entailed evaluations of the selection of cohorts, the comparability of cohorts, and the assessment of outcome. In this meta-analysis, we defined studies with an NOS of ≥6 stars as moderate to high-quality studies; studies which lower scores were defined as low-quality studies. 26,27 Two researchers discussed the topics during several face-to-face and web-based meetings. ware. The exact binomial (Clopper-Pearson) method was used to calculate 95% confidence intervals (CIs). Estimates were normalized using the Freeman-Tukey double arcsine transformation. We selected a random effects model to evaluate summary prevalence. The effect measurement estimate of mortality risk chosen was the odds ratio (ORs) in our study. Risk ratio and hazard ratio in other studies were considered to be OR, as OR was shown to be a more effective measure. 28 If not available, the ORs were calculated by events and total numbers of patients in two groups. The natural logarithm of the OR (log [OR]) and its standard error (SElog [OR]) were calculated and then pooled using statistical software. Cochran's chi-square test and I 2 statistic were used to evaluate the heterogeneity among the included studies. I 2 values of 25%, 50%, and 75% were considered to represent low, moderate, and high heterogeneity, respectively. 29,30 The statistical significance threshold was set at p < .05.

| Study selection
We screened 811 potentially relevant articles in the Cochrane Library (n = 19), PubMed (n = 393), MedRxiv (n = 286), and Embase (n = 113) databases ( Figure S1 in Supplemental Material). We excluded 91 studies after screening the titles and abstracts, and the full texts of the remaining 720 studies were reviewed. After a quick screening of the full-text articles, 30 records were received full-test review, and 9 excluded for the following reasons: (1) seven studies did not have relevant results, and (2) two studies 31,32 were based on same population. Finally, 21 studies 3,4,6-24 were included in the present metaanalysis.

| Study characteristics and quality
The study characteristics were shown in Table 1  studies, and three 10,12,15 were prospective observational (cohort) studies. 3.3 | The prevalence of cardiac injury in hospitalized patients with COVID-19

| The impact of cardiac injury on all-cause death in hospitalized patients with COVID-19
There were 10 publications 3,4,12,13,19-24 reported the association between cardiac injury and death. As shown in Figure 3,

| The impact of cardiac injury on all-cause death in hospitalized patients with severe COVID-19
There were two studies 4,7,12,15 reported the association between cardiac injury and death in patients with severe COVID-19. None of F I G U R E 2 Meta-analysis for the proportion of cardiac injury in patients hospitalized with severe and non-severe COVID-19

F I G U R E 3 Meta-analysis for association between cardiac injury and all-cause death in patients hospitalized with COVID-19
included studies reported the adjusted results, thus, forced us to perform a univariate analysis. As presented in Figure 4, cardiac injury is

| Study limitations
Some limitations may influence the validity of this meta-analysis. First, all the included studies were observational studies and some studies did not adjust for the clinical confounding in the outcome of death and these biases might influence on our results. For example, ACEI/ ARB, as we describe previously, was showed be associated with decreased risk of death in COVID-19 42 . However, although with number changes, the positive association between cardiac injury and death was still persisted when excluding the unadjusted studies, which showed the robustness of our conclusion. Second, the definition of the cardiac injury differed across the included trials and potentially affected the findings. Third, the majority of studies are from China, the studies involved with the prevalence and impact of cardiac injury on COVID-19 is needing to confirm this conclusion. Fourth, high heterogeneity existed across studies in some comparisons, limiting the interpreting on the results, however, we attempted to account for this by using a random effect model to make our results more conservable. The precise number for the type of cardiac injury is lacking, therefore, we cannot evaluate the association between type of cardiac involvement and adverse prognosis. Finally, lots of studies did not provide the adjusted results, thus, further studies with welldesigned are needed to confirm our results.

| CONCLUSIONS
This was the first meta-analysis exploring the prevalence and impact of cardiac injury on COVID-19. Cardiac injury is common in hospitalized patients and advanced age and severe COVID-19 patients prone to experience more risk of cardiac injury. Furthermore, cardiac injury is associated with increased risk of all-cause mortality.