Risk factors for COVID-19 patients with cardiac injury: pulmonary ventilation dysfunction and oxygen inhalation insufficiency are not the direct causes

Background: Cardiac injury in patients with coronavirus disease 2019 (COVID-19) has been reported in recent studies. However, reports on the risk factors for cardiac injury and their prognostic value are limited. Results: In total, 15.9% of all cases were defined as cardiac injury in our study. Patients with severe COVID-19 were significantly associated with older age and higher respiratory rates, Sequential Organ Failure Assessment (SOFA) scores, cardiac injury biomarkers and PaO2/FiO2 ratios. Male patients with chest distress and dyspnea were more likely to have severe disease. Patients with cardiac injury were significantly more likely to have a severe condition and have an outcome of death. However, no significant difference was found in respiratory rates, dyspnea or PaO2/FiO2 ratio between patients with or without cardiac injury. In the logistic regression model, pre-existing hypertension and higher SOFA score were independent risk factors for patients with COVID-19 developing cardiac injury. Conclusions: Our study revealed that cardiac injury was an important predictor for patients having a severe or fatal outcome. Patients with pre-existing hypertension and higher SOFA scores upon admission were more likely to develop cardiac injury. Nevertheless, pulmonary ventilation dysfunction and oxygen inhalation insufficiency were not the main causes of cardiac injury in patients with COVID-19. Methods: A total of 113 confirmed cases were included in our study. Severe patients were defined according to American Thoracic Society guidelines for community-acquired pneumonia. Cardiac injury was defined as a serum cTnI above the 99th-percentile of the upper reference limit. Patient characteristics, clinical laboratory data and treatment details were collected and analyzed. The risk factors for patients with and without cardiac injury were analyzed.


INTRODUCTION
First reported in Wuhan, Hubei Province, China, coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has now caused considerable morbidity and mortality in almost all countries [1,2], with an overall mortality rate of approximately 3.4% [3]. The common clinical A total of 140 adult patients with COVID-19 confirmed by SARS-CoV-2 RNA detection in Renmin Hospital of Wuhan University between February 16 and March 21, 2020 were enrolled in this retrospective observational cohort study. After excluding seven patients who were previously diagnosed with coronary heart disease and 20 patients without available basic examinations in their medical records, we included 113 inpatients in the final analysis ( Figure 1). The median age of the 113 patients in this study was 63.00 years (IQR 49.50-70.00), ranging from 23 years to 87 years, and 54.9% of them were male. Comorbidities were present in over half of the patients, and the most common comorbidity was hypertension, followed by diabetes (Table 1). Fever (78.8%) was the most common symptom on admission, with a median duration of 10 days (IQR 4.50-15.00). Cough (60.2%) was the second most common symptom, followed by myalgia or fatigue (43.4%). Among all 113 patients, lymphocytopenia occurred in 63 (55.8%) patients, and 18 (15.9%) patients were confirmed to have cardiac injury upon admission. During the hospital treatment period, 92.9% patients were provided with oxygen supplementation, and 14 patients required invasive mechanical ventilation, of whom 9 (64.3%) died. All patients who died were in the severe group.  The severe patients presented higher neutrophils, lower lymphocytes and lower monocytes in both proportion and numbers than patients with non-severe disease ( Table 2). In addition, the levels of albumin, C-reactive protein (CRP), total bilirubin (TB), Ddimer, serum IL6, and IL10 were significantly higher in the severe patients compared with the non-severe patients, whereas the level of sodium and calcium were lower in severe patients. It is worth noting that all the biomarkers related to cardiac injury in the severe patients showed dramatically significant differences from those in the non-severe patients (P < 0.001, Table 2), indicating that the severe patients were accompanied with relatively severe cardiac dysfunction.

Kinetic analysis of serum cardiac injury biomarkers in COVID-19 patients and correlation analysis between cardiac injury biomarkers and other parameters
Serum cardiac injury biomarkers, including CK, CKMB, LDH, cTnI and BNP, were observed as significant predictors for severe patients at the time of hospital admission. Next, we analyzed the kinetic alterations of these five biomarkers over at least 14 days. The significantly increased serum CK in the severe group, compared with that in non-severe patients, was only observed at the onset (within 3 days) but not during the following period of disease progression ( Figure 2A). The other four biomarkers were increased significantly in the severe patients, not only on the first day but also during two weeks after admission compared with the non-severe patients ( Figure 1B-1E).
Due to the high sensitivity and specificity of cTnI in distinguishing cardiac injury and the widespread application of BNP in distinguishing heart failure in clinical diagnoses, we further analyzed the correlation of cTnI or BNP with other parameters, which showed significant differences between the two groups in Tables 1-3. As shown in Table 4, the level of serum cTnI was positively correlated with age (R = 0.208, P = 0.027), CRP (R = 0.273, P = 0.008), CREA (R = 0.258, P = 0.009), serum IL6 level (R = 0.302, P = 0.037), pneumonia severity index (PSI) score (R = 0.210, P = 0.025), APACHE II score (R = 0.296, P = 0.001) and SOFA score (R = 0.323, P < 0.001) and negatively correlated with lymphocyte counts (R = -0.245, P = 0.012) in all patients. The level of serum BNP was only positively correlated with creatinine (CREA) (R = 0.957, P < 0.001), CURB65 score (R = 0.252, P = 0.032), APACHEII score (R = 0.241, P = 0.042) and SOFA score (R = 0.326, P = 0.005). It is worth noting that serum cTnI showed no correlation with CURB65 score (R = 0.163, P = 0.085), and serum BNP showed no correlation with PSI score (R = 0.224, P = 0.059). Both biomarkers showed no correlation with the P/F ratio (Table 4). In addition, we found that the serum level of IL6 showed a significantly positive correlation with serum cTnI level (P = 0.008 on day 3, P < 0.001 on day 7 and P < 0.001 on day 14), and these two indicators presented similar variation tendency over time ( Figure 3).

Dysfunction of pulmonary ventilation in COVID-19 patients was not the main cause of cardiac injury
In total, 113 patients were confirmed to have COVID-19 in our study, 18 patients (15.9%) were defined as having cardiac injury, and 95 patients (84.1%) were defined as having no cardiac injury. As indicated in Table 5, patients with cardiac injury showed significantly higher proportion in severe patients (13/18, 72.2%) than patients without cardiac injury (40/95, 42.1%) (P = 0.018). Compared with patients without cardiac injury, patients with cardiac injury were older (median 69.50, IQR 66.50-78.00 years). Moreover, hypertension (P = 0.011) as well as a higher heart rate (P = 0.011) and systolic pressure (P = 0.002) were more common among patients with cardiac injury. No significant difference was observed in respiratory rate or in the proportion of patients with chest distress and dyspnea among patients with cardiac injury, which were dramatically different between severe patients and non-severe patients. The days from fever onset to patient admission showed no significant difference between patients with or without cardiac injury. st and P/F ratio, showed no significant differences between the two groups (Table 7).

AGING
In patients with cardiac injury, the PSI (89.22±16.01 vs 64.78±32.58, P < 0.001), CURB65 (0.72±0.78 vs AGING 1.33±0.59, P < 0.001), APACHEII (9.56±2.83 vs 5.56±4.31, P < 0.001) and SOFA scores (3.00, IQR 2.00-4.00 vs 1.00, IQR 1.00-2.00, P < 0.001) were all higher than in those without cardiac injury. As APACHEII score was calculated based on age and other scores (PSI, SOFA, CURB65) were related to the rest of the laboratory indicators, we chose age, hypertension, PSI, CURB65 score and SOFA score as the five variables for our multivariable logistic regression model. A significant difference was observed in the logistic model with χ 2 (5) = 21.998, P <0.0005. The predictive model was able to classify 86.7% of the cardiac injury patients among patients with COVID-19, with a sensitivity of 33.3% and specificity of 96.8%. Under this hazard regression model, the variables age, CURB65 score and APACHEII score showed no significant difference between groups, and they were not independent risk factors for cardiac injury of patients with COVID-19. Regarding the independent risk factors, patients with pre-existing hypertension were 3.2 times more likely to have cardiac injury than those without hypertension (OR 3.28, 95% CI 1.02-10.61), and the risk of cardiac injury rose 66% with a one-score increase of the SOFA score on the hospital admission day.

DISCUSSION
In the present study, our findings indicated that pulmonary ventilation dysfunction were not directly associated with cardiac injury in patients with COVID-19. In contrast, patients with pre-existing hypertension and elevated SOFA scores upon admission, which were regarded as independent risk factors in our study, were more likely to progress to cardiac injury.
The elevated biomarker levels in severe patients and cardiac injury patients indicated that the myocardial damage in COVID-19 patients were not random. This factor might be associated with the outcome with the patients. A previous study pointed out that in hospitalized patients with COVID-19, cardiac injury is a common condition in disease progression, and it is tightly associated with a higher risk of in-hospital mortality [8]. The elevated levels of biomarkers, such as BNP and troponin, were regarded as the prominent features in COVID-19 patients and reported to be associated with ICU admission and mortality [6,[9][10][11]. Cardiac involvement is of great importance in determining the prevalence and prognosis of COVID-19 patients. In our current study, the proportion of cardiac injuries in the severe patients or non-survival patients was significantly higher than that in non-severe patients or survival patients, respectively, which was similar with previous study.
Exaggerated systemic inflammation, lymphocytopenia, hypoxemia and cardiovascular stress might be the hallmarks of severe patients with COVID-19 [10]. In our current study, lymphocytopenia was more common in severe patients in terms of the numbers of lymphocytes and monocytes, as well as in patients with cardiac injury, than in patients without cardiac injury. The inflammatory factor IL6 showed similar alterations of CK, CKMB, LDH, cTnI and BNP in the two groups. "*" means significant difference between the two groups. *, P < 0.05, **, P < 0.01, ***, P < 0.001.  AGING and was significantly higher in both patients with severe disease and patients with cardiac damage than in their corresponding comparison groups. Compared with IL6, some other cytokines, such as IL2 and IL10, are elevated in cardiac injury patients and severe patients, respectively, activating the pathways leading to the differentiation of immune cells, stimulating the leukocytes to the infection sites and promoting the proliferation of hematopoietic progenitor cells after viral infection. We also found that the serum level of IL6 was tightly associated with the biomarker cTnI in all patients with COVID-19 at different time points, which supports the point of view that the inflammation starting in and propagating from the lung or other initial organ injuries probably resulted in some bystander effects on other organs, such as the heart, due to amplifying inflammatory responses [9,12].

AGING
Dysfunction of pulmonary ventilation, hypoxemia and other related symptoms in COVID-19 patients were indicators for disease progression to severe conditions or death in previous studies [13][14][15]. In some critical cases, patients with respiratory failure might develop ARDS, sepsis, multiorgan dysfunction or even septic shock [6]. In our study, the respiratory rate, chest distress and dyspnea were more frequent in severe patients. However, all the clinical symptoms related to pulmonary ventilation dysfunction and hypoxemia showed no difference no matter COVID-19 patients with or without cardiac injury. These findings were consistent with previous data, although these conclusions were not given enough attention in their article [8]. In addition, the PaO 2 on admission and the P/F ratio after oxygen inhalation were not improved in cardiac damaged patients. Moreover, the biomarkers cTnI and BNP showed no correlation with these hypoxemia-related indicators. Therefore, we concluded that pulmonary ventilation dysfunction and oxygen inhalation insufficiency were not direct causes of cardiac injury or myocardial ischemia in patients with COVID-19.
Although the PSI, CURB65, APACHEII and SOFA scores were significantly higher in both severe patients and patients with cardiac damage, the pre-existing hypertension and SOFA scores on admission to the hospital in patients with COVID-19 were independent risk factors for patients progressing to cardiac injury ( Table 8). Gu J and his colleagues had pointed out that immune and lung damage were the key features of coronavirus infection, accompanied by lymphocytopenia and inflammatory cytokine storms, which led to multiple organ infections or injuries, including in respiratory epithelial cells, the intestinal mucosa, the renal distal tubule epithelium and cerebral neurons [16,17]. Human coronaviruses (SARS-CoV and SARS-CoV-2) target epithelial cells or immune cells by binding their spike protein to angiotensin-converting enzyme 2 (ACE2), which is expressed by epithelial cells of the lung, intestine, kidney, heart and blood vessels, thus priming the serine protease TMPRSS2 for S protein [18,19]. Single-cell RNAseq data analysis of receptor ACE2 expression demonstrated that more than 7.5% myocardial cells had positive ACE2 expression, indicating that the heart or vessels could be directly infected and damaged in overloaded SARS-CoV-2 patients [20]. As patients with hypertension or diabetes were commonly treated with ACE inhibitors and angiotensin II type-I receptor blockers, an upregulation in the expression of ACE2 was mentioned in previous studies [18,21,22], which perhaps led to the heart being more vulnerable to SARS-CoV-2 infection. In our current study, pre-existing hypertension was observed in a significantly higher proportion in cardiac damaged patients, and it was an independent risk factor for COVID-19 patients developing myocardial damage in our multivariate logistic regression model. Furthermore, COVID-19 patients with cardiac injury were more likely to die in our study. However, some researchers pointed out that no difference in ACE2 expression or activity was found after antihypertensive calcium channel blocker treatment [23].
A study including 50 hospitalized hypertensive patients with laboratory-confirmed COVID-19 in Wuhan revealed no obvious difference in clinical characteristics between RAS blockers and non-RAS blockers groups [24]. Therefore, we inferred that SARS-CoV-2 might invade vessels by some uncertain pathways to reach the heart and damage myocardial cells by binding to upregulated ACE2, directly to lead to cardiac injury.  Our study has several inevitable limitations. First, the number of patients with cardiac injury and mortality was not very high in our study, but the proportions both were in reasonable ranges. Furthermore, the drugs that the patients with preexisting hypertension took before SARS-CoV-2 infection were not analyzed in our study due to the incomplete medical history collected from the patients.

CONCLUSIONS
In conclusion, our study indicated that cardiac injury was an important indicator for patients with severe or fatal disease, and patients with preexisting hypertension and higher SOFA scores upon admission were more likely to progress to cardiac injury. Nevertheless, pulmonary ventilation dysfunction and oxygen inhalation insufficiency were not the main causes of cardiac injury in patients with COVID-19.

Study participants
This retrospective cohort study included adult patients (≥18 years old) admitted to Renmin Hospital of Wuhan University from February 16 to March 21, 2020. All inpatients were confirmed to have COVID-19 by SARS-CoV-2 RNA detection. As this study focused on patients with or without cardiac injury, participants with preexisting coronary heart disease or other myocardial AGING  diseases were excluded. This study was approved by the Ethics Committee of Renmin Hospital of Wuhan University.

Data collection
Epidemiological data, demographic data, laboratory indicators, treatment details and outcome data were collected from electronic medical records and confidentially protected by assigning a deidentified ID to each patient. Laboratory indicators and treatment details were collected for at least 14 days on days 1, 3, 7 and 14.
The degree of severity of COVID-19 patients (severe vs. non-severe) was defined at the time of admission, according to the American Thoracic Society guidelines for community-acquired pneumonia [25]. Cardiac injury was defined by cardiac biomarker (cardiac troponin I, [cTnI]) levels in the blood above the 99 th -percentile of the upper reference limit. Patients with unavailable key information were excluded from our study.

Statistical analysis
Categorical variables are presented as N, % and were compared using Fisher's exact test or χ2 test. Continuous variables are presented as the mean ± SEM or median (interquartile range [IQR]) values and were compared using Student's t test or the Mann-Whitney U test, as appropriate. To explore the risk factors associated with cardiac injury in patients with COVID-19, univariable and multivariate logistic regression models were used. Correlation analysis between the two parameters was performed using the Pearson correlation coefficient. Data were analyzed using SPSS version 22.0 (IBM), and statistical charts were generated using Prism 7.0 (GraphPad 7.0). For all the statistical analyses, P < 0.05 was considered significant.