Influence of Thrombolysis and Mechanical Ventilation on Echocardiographic Predictors of Survival after Acute Pulmonary Embolism

To the Editor: Khemasuwan et al. retrospectively evaluated a cohort of 211 patients admitted to intensive care units (ICUs) with acute pulmonary embolism (PE). The investigators examined the correlation between echocardiographic parameters and clinical outcomes, in particular ICU, hospital, and long-termmortality. Several simple echocardiographic parameters were associated with ICU, hospital, and long-termmortality in this group of patients, while right ventricular (RV) strain analysis was not, although optimal data were available for only 54% of the patients. However, the study findings are partly flawed. The investigators performed adjusted analyses for Acute Physiology and Chronic Health Evaluation score (ICU mortality), Pulmonary Embolism Severity Index score (hospital mortality), or age and gender (longterm mortality), but they did not consider two important issues. First, intravenous thrombolytic therapy was administered to 9% of the study population. Thrombolytic therapy has been previously associated with significant early improvements in echocardiographic parameters of RV function compared with heparin alone in patients with acute massive or submassive PE. Therefore, patients receiving thrombolysis may have had more pronounced improvement in all or some of the parameters that were positively correlated with clinical outcomes in this study. Unfortunately, the investigators did not present a subgroup analysis of the echocardiographic findings in the subgroup of patients who received thrombolysis, and thrombolysis was not considered in the regression analysis. Second, absence of $50% inferior vena cava (IVC) collapsibility was correlated with ICU and hospital mortality. Nonetheless, 26% of the patients in this study required mechanical ventilation, which in turn significantly influences changes in IVC diameter (IVC distensibility for ventilated patients) and its value as an indicator of RV preload. The investigators did not adjust for this variable in the binary logistic regression analyses, and this can introduce a degree of bias when considering IVC collapsibility as predictor of outcome. Therefore, before the data can be correctly interpreted, an adjustment of the analysis for such variables (thrombolysis and mechanical ventilations) is needed. Finally, it would be interesting if the investigators could look into the echocardiographic follow-up of survivors, evaluating both early (within weeks) and late (after a few months) improvements in RV function and looking for predictors of good functional RV recovery. By reporting these data, the investigators would greatly contribute in providing further valuable insight on this interesting topic.