Prognostic implications of biventricular strain measurement in COVID‐19 patients by speckle‐tracking echocardiography

Abstract Background Recent reports have indicated the beneficial role of strain measurement in COVID‐19 patients. Hypothesis To determine the association between right and left global longitudinal strain (RVGLS, LVGLS) and COVID‐19 patients' outcomes. Methods Hospitalized COVID‐19 patients between June and August 2020 were included. Two‐dimensional echocardiography and biventricular global longitudinal strain measurement were performed. The outcome measure was defined as mortality, ICU admission, and need for intubation. Appropriate statistical tests were used to compare different groups. Results In this study 207 patients (88 females) were enrolled. During 64 ± 4 days of follow‐up, 22 (10.6%) patients died. Mortality, ICU admission, and intubation were significantly associated with LVGLS and RVGLS tertiles. LVGLS tertiles could predict poor outcome with significant odds ratios in the total population (OR = 0.203, 95% CI: 0.088–0.465; OR = 0.350, 95% CI: 0.210–0.585; OR = 0.354, 95% CI: 0.170–0.736 for mortality, ICU admission, and intubation). Although odds ratios of LVGLS for the prediction of outcome were statistically significant among hypertensive patients, these odds ratios did not reach significance among non‐hypertensive patients. RVGLS tertiles revealed significant odds ratios for the prediction of mortality (OR = 0.322, 95% CI: 0.162–0.640), ICU admission (OR = 0.287, 95% CI: 0.166–0.495), and need for intubation (OR = 0.360, 95% CI: 0.174–0.744). Odds ratios of RVGLS remained significant even after adjusting for hypertension when considering mortality and ICU admission. Conclusion RVGLS and LVGLS can be acceptable prognostic factors to predict mortality, ICU admission, and intubation in hospitalized COVID‐19 patients. However, RVGLS seems more reliable, as it is not confounded by hypertension.


| Study design
In this prognostic prospective cohort study, 225 consecutive patients with established COVID-19 were included; the diagnosis of COVID-19 was made by real-time RT-PCR. Patients were consecutively enrolled from a general educational hospital (Shahid Modarres Hospital, Tehran, Iran) between June 21 and August 24, 2020. A consecutive series of COVID-19 patients were included and underwent echocardiography in the first 12-24 h of admission. Furthermore, both deceased and survived cases were selected from a single center to reduce selection bias.
The flow chart of patients' enrollment is depicted in Figure 1. Finally, 207 patients were included in our analyses. Clinical data were gathered by a predefined checklist including age, sex, outcome (mortality, ICU admission, and intubation), and background diseases (ischemic heart disease, chronic obstructive pulmonary disease, hypertension, asthma, hyperlipidemia, smoking, and diabetes mellitus). It should be maintained that all the study protocols were approved by the local ethics committee of Shahid Beheshti University of Medical Sciences.   Figure S1). RVGLS was also calculated based on the 4-chamber view, and RVGLS was automatically calculated by the device ( Figure S2). Both measurements were performed in a 16-segment model. It should be noted that the margin of the endocardium should be clearly defined for these evaluations and patients who did not meet this criterion were excluded from the study. All the measurements were performed by a single cardiologist specialized in this

| Statistical analysis
Data analysis were performed using the SPSS version 20.0 software.  Tables 3 and 4, mortality (p = .001, 0.001), ICU admission (p = .001, 0.001), and need for intubation (p = .006, 0.022) were significantly associated with LVGLS and RVGLS tertiles, respectively (lower measurement was associated with poor outcomes, ROC curves are presented in Figure 2, sensitivity and specificity analysis is summarized in Table 5). LVGLS tertiles could predict poor out-

| DISCUSSION
This is one of the largest COVID-19 populations studied by 2D strain echocardiography. In this cohort, it was found that the majority of findings in two-dimensional echocardiography are related to the outcomes in COVID-19 patients. Both LVGLS and RVGLS could predict all outcomes including ICU admission, need for intubation, and mortality in admitted COVID-19 cases, even when adjusted for hypertension and ischemic heart diseases.
Our findings for RVGLS were congruent with the study of Krishnamoorthy et al. 13  We also found no association between the troponin level and any of RVGLS, LVGLS, or patients' outcomes. Consistent with this finding, Kocas et al. 19 reported no significant association between troponin level and RVGLS. It may demonstrate the potential indirect effect of SARS-CoV-2 on cardiac function.
Finally, there is still a question: Why to use GLS measurements over routine echocardiographic parameters such as LVEF? First, GLS is considered to be a more reproducible method regardless of echocardiographic training as it has better inter-and intra-observer reproducibility. 20 Also, Strain echocardiography may reveal myocardial dysfunction before significant changes in ejection fraction and cardiac output. 21 Strain measurement can also be valuable in detecting acute myocarditis, 22 which may increase COVID-19 mortality. 23 Considering all the above-mentioned rationales and the findings of our study, strain echocardiography can be a valuable tool for the prediction of outcome in hospitalized COVID-19 patients.
This study was limited by the fact that only hospitalized COVID-19 patients were included. Also, patients' pharmacologic therapy was not included in our analysis. Thus, further prospective studies are needed to take all confounding factors into account.
In conclusion, RVGLS and LVGLS can be both acceptable prognostic factors to predict mortality, ICU admission, and intubation in hospitalized COVID-19 patients. However, RVGLS seems more reliable, as it is not confounded by hypertension. Also, other echocardiographic findings, such as LVEF, E', LAVI, FAC, TAPSE, and systolic PAP may be used as prognostic factors, if speckle tracking echocardiography is not available.