1. Introduction
In December 2019, an outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was reported in Wuhan, China [
1]. Within just a few weeks, the disease, COVID-19, spread throughout the world with a high mortality rate [
2].
Various published studies have shown that myocardial damage, defined as increased plasma cardiac troponin levels, is a common finding in affected patients, and that it is associated with a poorer outcome during admission and high mortality [
3,
4]. However, little is known about its real effect on the heart and whether increased levels of specific biomarkers reported in studies simply reflect a more systemic disease in the most severely ill patients [
5]. Given the difficulty of performing an echocardiography study under strict isolation conditions, in addition to the potential risk of contagion for health care staff, the exact prevalence and nature of cardiac involvement using systematic cardiac imaging in patients with COVID-19 has been difficult to establish but of great interest. In order to analyze the cardiac involvement of SAR-CoV-2, multicenter alliances have emerged, like the ECHO_COVID research group. They assessed cardiac disorders in six hundred and seventy-seven severely ill patients affected by COVID-19 in the ICU and found that almost one-third presented abnormal LV and/or RV systolic function. Additionally, RV involvement was prevalent, with different phenotypes of RV involvement leading to different ICU mortality rates; acute cor pulmonale had the worst outcome [
6,
7].
Determination of the left ventricular ejection fraction using echocardiography is the most commonly used technique for the assessment of ventricular function in clinical practice. However, it may fail to detect early or subclinical abnormalities of systolic function. Measurement of the global longitudinal strain (GLS) using two-dimensional speckle-tracking echocardiography is a sensitive parameter of systolic function. Cardiomyopathy, heart failure, and evaluation of chemotherapy-induced cardiac toxicity are common clinical conditions that use left ventricular GLS to identify subclinical left ventricular dysfunction [
8]. Data regarding the use of left ventricular GLS in patients with COVID-19 are limited. Accordingly, the objectives of the present study were to evaluate the characteristics and severity of myocardial injury using left ventricular GLS in patients hospitalized with COVID-19 outside the ICU, to investigate its association with plasma cardiac troponin levels, and to assess the evolution of GLS in the short term.
3. Results
Figure 1 shows the flowchart for patient recruitment. Briefly, 4 of the initial 36 patients were excluded because of a previous history of heart disease, and 1 was excluded due to inadequate image quality. The final study population comprised 31 patients hospitalized with confirmed COVID-19. The mean age was 65 ± 15.18 years, and 61.3% were men. Myocardial damage in the form of increased troponins was detected in nine patients (29%).
Table 1 shows the patients’ clinical and laboratory data.
All patients had a normal left ventricular ejection fraction. However, after assessing GLS, we observed that this parameter was reduced in 11 of the 31 patients (35.5%) (
Figure 2). Comorbidities such as hypertension and chronic kidney disease were more common in patients with left ventricular dysfunction determined by GLS. These patients also had higher plasma hs-cTnI levels (23.7 vs. 3.2 ng/L;
p < 0.05) and NT-proBNP levels (753 vs. 81 pg/mL;
p < 0. 05). Moreover, 45.5% of patients with reduced GSL had elevated troponin I plasma levels. No differences were observed for plasma markers of inflammation (interleukin 6 and C-reactive protein) (
Table 1 and
Table 2).
Panel A shows an example of a 65-year-old patient with normal left ventricular global longitudinal strain (LV GLS) and 1.5 ng/L of high-sensitivity troponin I (hs-TnI); a longitudinal strain of −22.4% was obtained from the bull’s eye plot derived from two-dimensional speckle tracking imaging. Panel B shows results from a 72-year-old patient with impaired LV GLS and 121.4 ng/L of hs-TnI; a longitudinal strain of −11.6% was obtained from the bull’s eye plot derived from two-dimensional speckle tracking imaging.
Table 3 shows the patients’ echocardiographic characteristics according to the presence or absence of altered GLS. Patients with reduced left ventricle GLS had also lower right ventricle free wall strain. There were no differences in left ventricular filling pressure (E/e’).
A significant and moderate positive correlation was found between GLS and plasma hs-cTnI and NT-proBNP values (
Figure 3).
The multivariate analysis showed that the presence of myocardial damage, defined as an increase in hs-cTnI levels, was significantly associated with left ventricular GLS values, and patients with myocardial damage presented an abnormal 5% reduction in GLS (
Table 4).
None of the patients evaluated required transfer to the intensive care unit (ICU) or died during their admission.
At follow-up after one month, one patient from the reduced-GLS group had died, and there were no rehospitalizations. In the echocardiography performed 30 days after discharge (30 patients), there was a significant increase in GLS values in the reduced-GLS group (−16.4 ± 2.07% vs. −13.2 ± 2.40%;
p < 0.01) (
Table 5). Four of the surviving patients of this group had normalized GLS values (>−15.9%). There were no significant differences in the evolution of GLS in the group with normal systolic function during the acute phase of infection (−19.9 ± 2.73% vs. −19.8 ± 3.28%;
p = 0.959) (
Figure 4).
However, because this was an observational study on a population of a small sample of patients with COVID-19 in an internal medicine ward, our results may not reflect those for the wider population infected with the virus.
4. Discussion
To the best of our knowledge, this is one of the first studies to evaluate cardiac involvement based on left ventricular GLS using 2-D speckle tracking echocardiography in noncritical patients hospitalized with COVID-19. Our findings illustrate the high frequency of subclinical left ventricular dysfunction in the early stages of COVID-19 in patients with no previous history of heart disease and a preserved ejection fraction (one in three patients). Our results show that this dysfunction may be reversible on clinical recovery, since GLS had improved 1 month after the first echocardiogram in 40% of patients with impaired left ventricular function. Our data also show that lower GLS values are correlated with the magnitude of increased hs-cTnI plasma levels.
Numerous studies have reported acute cardiac injury as an important manifestation of COVID-19. In studies published to date, acute cardiac injury, defined as an increase in cardiac troponin to the >99th percentile, affects 7% to 28% of hospitalized patients. However, this number may depend in part on the definition used and the severity of cases at the hospital where the data were recorded [
3,
13]. In our study, 29% of all hospitalized patients had acute cardiac injury, even though the patients’ condition was not severe and did not require admission to the intensive care unit. The percentage we report is higher, probably because, given the dynamic changes in troponin levels, we made several determinations in plasma during hospitalization and selected the highest value. Patients with high troponin levels usually present a poor clinical profile, are generally older, mainly men, and have a higher prevalence of comorbidities, including arterial hypertension, diabetes, coronary heart disease, and chronic kidney disease [
14,
15]. Furthermore, these values increase significantly in patients with more severe infection compared with those who have milder forms of the disease and do not require admission to the ICU [
4,
16].
Acute cardiac injury has consistently proven to be a strong negative prognostic marker in patients with COVID-19. Various studies have shown a poorer outcome in patients with increased troponin I or T: greater risk of being admitted to the ICU [
17], mechanical ventilation, and more severe complications during admission, as well as respiratory distress, malignant ventricular arrhythmias, acute kidney failure, and clotting disorders, has been reported [
3,
13]. In a sample of 416 cases, Shi et al. [
3] showed higher mortality in patients with myocardial damage (51.2% vs. 4.5%,
p < 0.001), with an association between the magnitude of the increase in hs-cTnI levels and mortality.
Importantly, many aspects of this increase in troponins remain undefined, including the frequency and severity of associated cardiac structural abnormalities. To date, few data have been published about how the virus affects the heart. Furthermore, not many systematic echocardiography-based studies report the exact prevalence and nature of cardiac involvement in COVID-19. In some cases, it could go undetected. The few published studies report right ventricular dysfunction and dilatation as the most common echocardiographic abnormality, with a considerable impact on prognosis of the disease [
18], whereas systolic function of the left ventricle is rarely affected in patients with COVID-19.
In their retrospective study of 74 patients with COVID-19 pneumonia, Mahmoud-Elsayed et al. [
19] reported frequent right ventricular abnormalities in the form of dilatation (41%) and dysfunction (27%), whereas left ventricular systolic function was normal or hyperdynamic in most patients (89%). The study, however, was subject to major limitations, mainly that the left ventricular systolic function was determined visually. A recent study of 120 patients hospitalized with COVID-19 showed that the left ventricular ejection fraction—assessed according to the Simpson biplane method—was impaired to some extent in 11% of patients, with no repercussion on patient progression during admission [
20]. Similar findings were reported by Szekely et al. [
21] in a series of 100 consecutive patients, among whom only 10% had a depressed ejection fraction. Among these patients, two cases were due to the presence of previous coronary disease. No differences were found in terms of cardiac injury prevalence depending on the cardiac troponin increase. The results of these studies suggest that ventricular systolic dysfunction in patients with acute COVID-19 infection is not very common, while ventricular diastolic dysfunction is more prevalent, as has been reported in other coronavirus infections [
22].
In a multicentric study by Lassen et al. [
23] of 214 patients with COVID-19, both left and right ventricular strain were reduced in comparison with those of control subjects, and left ventricular strain reduction was associated with poorer outcome. In this regard, our study provides consistent information, considering that using left ventricular GLS is an effective and reproducible assessment to quantify global systolic function of the left ventricle in patients with acute COVID-19 infection. This approach is more sensitive than the ejection fraction as a marker of cardiac dysfunction. GLS helps in the diagnosis of early and subclinical stages of myocardial involvement in different conditions such as cardiomyopathy and chemotherapy-induced cardiotoxicity before significant changes occur in ventricular function [
8]. Thus, we have been able to show for the first time that asymptomatic involvement of left ventricular function is very common in a population of stable hospitalized patients (35.5%), even though these patients have a preserved left ventricular ejection fraction. The degree of ventricular dysfunction is associated with the magnitude of the increase in plasma troponin levels, as seen by the fact that patients with lower strain values have higher hs-cTnI levels.
Previous studies suggest that myocardial tissue remodeling could precede functional remodeling, and our results point in this direction by showing early dysfunction without ejection fraction abnormalities. Among the prior studies, a small-scale study based on cardiac magnetic resonance in patients who had recovered from COVID-19 showed that 54% had myocardial edema affecting 33% of left ventricular segments and late gadolinium uptake, with no alterations in contractility [
24]. Our study shows a similar left ventricular dysfunction persistence: 60% of patients still had reduced GLS at the one-month follow-up. Previous experiences with SARS showed that 40% of patients who had recovered from the disease had cardiovascular abnormalities at a 12-year follow-up, in addition to lipid abnormalities and glucose metabolism [
25].
In contrast with other studies in which different degrees of right ventricular abnormalities such as dilatation or systolic dysfunction have been reported in up to 40% of patients [
18,
26], we found light involvement of the right ventricle during infection. This difference is mainly because these previous studies included more critical patients with acute respiratory failure who required mechanical ventilation. However, focusing on the right free wall strain, 11 of our patients (35.5%) had reduced strain values of <–19%.
Regarding the cardiac biomarkers, it has been reported that less than 50% of patients with increased levels of troponins had left ventricular systolic or diastolic dysfunction with abnormal right ventricular function. The authors of [
21] suggest that the more common mechanism underlying elevated troponins in patients with COVID-19 is an acute right ventricular dysfunction due to vascular or parenchymal involvement in the lung. These findings are quite controversial since left ventricular GLS was not determined in their study, as it was in ours. We observed that patients with increased troponins have significantly more marked impairment of GLS, whereas no differences were observed in right ventricular function parameters between the two study populations. Therefore, subclinical left ventricular dysfunction may be the common mechanism in the increase in troponins in this infection.
Although the exact pathophysiology of myocardial damage in COVID-19 is not completely known, various potential mechanisms could play a role in this phenomenon. On the one hand, there is a direct invasion of cardiomyocytes by the virus, in which expression of angiotensin-converting enzyme 2 by myocytes is essential. In the same way, severe hypoxia caused by respiratory damage can lead to ischemia and oxidative stress due to the increased oxygen demand in the myocardium. On the other hand, cardiac microvascular damage can be caused by intravascular coagulation and/or by an excessive immune and inflammatory response leading to a cytokine storm and myocardial dysfunction [
26,
27]. It has been recently pointed out that people with chronic diseases have a higher risk of developing more severe forms of the disease. In such more ill people, lipid peroxidation is catalyzed by iron released from hemoglobin, transferrin, and ferritin, which are released by tissue acidosis and free oxygen radicals. Such ferroptosis-inducing factors can directly or indirectly affect glutathione peroxidase, resulting in a decrease in the antioxidant capacity and accumulation of lipid reactive oxygen species (ROS) in the cells, ultimately leading to oxidative cell stress and, finally, death [
28].
In our study, SAR-CoV-2 infection was associated with a high incidence of subclinical cardiac involvement, highlighted by impairment in GLS and elevated hs-cTnI; this was reversible in most patients but it could affect prognosis as it has been seen in other myocarditis. This could mean that patients with cardiac dysfunction should have a longer follow-up to exclude former cardiac abnormalities in the long term. The application of these findings to clinical practice would involve the hypothesis that in other kinds of myocarditis with normal left ventricular ejection fraction but with subclinical left ventricular dysfunction shown by GLS and elevated hs-cTnI plasma levels, the impaired ventricular function might have a high recovery rate, as we have seen in this study with the SAR-CoV-2 virus.
Despite being based on a consecutive and homogeneous cohort of patients hospitalized for COVID-19, our main limitation is that this was a single-center observational study and was performed in a relatively small sample; because of this, the results may vary in a wider population. Furthermore, we included patients with mild disease and not with severe conditions (respiratory distress or multiorgan failure), where the systemic inflammatory response is more marked, as are its cardiac implications. Given the broad clinical variety of SARS-CoV-2 infection, which includes asymptomatic patients who are not hospitalized, our findings cannot be applied to the general population affected by COVID-19. Finally, our study does not enable us to evaluate whether these echocardiographic abnormalities can lead to chronic consequences in the long term. Therefore, careful follow-up of patients who have recovered from COVID-19 should be recommended, in order to identify the long-term implications of the cardiovascular damage.