The Authors Reply.

We appreciate the comments by Dr. Fukui and colleagues about our paper (1), and we fully agree with some of their views. However, there are some issues that need to be clarified and discussed.

We acknowledge and agree that right ventricular (RV) function is closely linked with the afterload. In our cohort study, 40 (34.2%) patients developed acute respiratory distress syndrome (ARDS), which contributed to increase RV afterload. Furthermore, both ARDS and right ventricular longitudinal strain (RVLS) were found to be associated with mortality, consistent with the previous study (2).

It is difficult for us to evaluate the right ventricular to pulmonary artery (RV-PA) coupling by measuring a ratio between tricuspid annular plane systolic excursion and pulmonary artery systolic pressure (PASP), as suggested by Cavalcante (3). Noninvasive PASP measurement depends on the presence of tricuspid regurgitation (TR), and there were only 61 patients (50.8%) had interpretable TR jet signal in our study. Among them, the majority was mild or trivial TR, which may affect the accuracy of PASP assessment. As the right heart catheterization was not available in our designated treatment hospital, PASP derived from echocardiography may be less accurate to evaluate the RV-PA coupling.

COVID-19 patients with underlying cardiovascular disease are more likely to have more severe clinical course and represent a higher proportion of the patients who are at greater risk of mortality. Although cardiovascular disease has a marginal significance in the univariate analysis (p = 0.058), it was not included in the multivariate analysis due to the smaller number of endpoints.

Speckle-tracking echocardiography is dependent on image quality, which may relatively preclude its wide application in clinical practice, especially in patients who had a poor echocardiographic window. Therefore, we had to exclude 24 patients due to suboptimal image quality. Except for RVLS, RV size and function parameters were measured in these excluded cases. Our study showed that there were no significant differences in RV structure and function parameters between patients included for final analysis and those who were excluded. Therefore, we think the effect of excluding cases on our findings may be negligible.

We agree with the comments of Dr. Fukui and colleagues regarding the need for exploring the incremental prognostic value of RVLS over comprehensive clinical assessment. However, the small number of observed events does not enable us to explore the additive value of RVLS over the known risk factors. Future studies with larger simple sizes are needed to determine the additional prognostic value of RVLS over the other clinical and echocardiographic parameters in COVID-19 patients.

We are also delighted to respond to Dr. Schiavon and colleagues, who commented on the relatively high prevalence of deep vein thrombosis (DVT) and no cases of pulmonary embolization (PE) in our study. Each patient in the present study received both bedside echocardiography and the lower-extremity venous ultrasound. The levels of DVT included the bilateral common femoral, deep, and superficial femoral, and the popliteal veins as well as the posterior tibial, peroneal, and calf muscle veins. Moreover, if there was any clinical suspicion of PE, computed tomographic angiography of the pulmonary arteries (CTPA) was considered (4) and obtained, if possible. In the present study, CTPA was performed in 2 patients with DVT, but neither of them was diagnosed with PE. Although previous studies demonstrated an increased risk of DVT and potential PE in respiratory and other intensive care settings (5), our data suggested that CTPA-confirmed PE was less frequent in this group of patients. Nevertheless, the finding was limited because we did not perform screening CTPA for every patient with DVT.

We agree with the comments by Dr. Schiavon and colleagues about the effect of the timing of echocardiographic study and ventilation on RV echo parameters. In our study, there were 105 patients (87.5%) with oxygen therapy, 6 patients (5%) with noninvasive mechanical ventilation, and 15 patients (12.5%) with invasive mechanical ventilation at the time of echocardiographic examinations. Indeed, two-thirds of patients with invasive mechanical ventilation had the worst RV longitudinal strain.

We acknowledge that the proposed cutoff values for RV function are largely within the normal range. Except for the thoughtful explanation proposed by Schiavon and colleagues, another important reason may be the small proportion of patients (22%) admitted to the intensive care unit as well as those (12.5%) receiving invasive mechanical ventilation.