To the Editor: We read with interest the article published by Maharaj et al.1 The authors’ aim was to investigate the correlation between echocardiographic parameters and survival in patients supported by venovenous extracorporeal membrane oxygenation (VV-ECMO). The authors retrospectively collected clinical and echocardiographic data from 17 patients over 1 year period. The echocardiograms were a mixture of transthoracic (TTE) and transesophageal (TEE) scans. Visual assessment and right ventricle (RV) fractional area change (FAC) were used as the parameters for RV systolic function as measured by two independent blinded experts, and an average was used for statistical analysis. In patients who did not survive, significant RV dysfunction by visual assessment was more prevalent in nonsurvivors. In addition, median value for FAC in nonsurvivors was significantly lower (11.3% [10.3–15] vs. 30.1% [27.6–31.4], respectively). RV FAC is performed by tracing the endocardial border area of the RV in end systole (ESA) and end diastole (EDA), and it is calculated as EDA − ESA/EDA.2 This method renders FAC open to intra- and interobserver variabilities. A previous study demonstrated FAC to exhibit limited and nonstatistically significant intraobserver (r, 0.56; p = 0.06) and interobserver (r, 0.46; p = 0.16) agreements.2 In addition, mean intra- and interobserver differences were high (mean, 1.86%; 95% CI, −9.25% to −12.96%; and mean, 4.00%; 95% CI, −7.72% to −15.72%, respectively). FAC variability was worst compared with other RV systolic function parameters (e.g., Tricuspid annular plane systolic excursion and velocity). Where FAC is used for RV systolic function assessment, intra- and interobserver variabilities should be assessed and cited providing a guide to results reliability. The authors used a mixture of TEE and TTE to obtain FAC. The two modalities were considered comparable, and thus, values from both modalities were incorporated in the FAC comparison. A recent study compared TTE and TEE in the assessment of RV function using the same parameters in 100 patients undergoing cardiac surgery.3 The TTE and TEE were performed in close temporal proximity under the same loading conditions. The study demonstrated that RV parameters had fair agreement between modalities but not identical. FAC comparison demonstrated correlation concordance coefficient between TTE and TEE of 0.31 (95% CI, 0.12–0.48), suggesting limited agreement and that FAC results are not interchangeable between the two modalities. This result casts a shadow on the feasibility of incorporating FAC from TTE and TEE in the same statistical comparison. Association between a continuous independent parameter (e.g., RV FAC) and categorical dependent parameter (e.g., mortality) requires in-depth statistical analysis. Logistic regression is the most commonly statistical method used to investigate such association.4 A multivariate model is constructed to incorporate all likely and significantly different independent parameters to investigate which parameters predict outcome. Furthermore, receiver operator curve (ROC) analysis is usually performed to establish sensitivity and specificity of the independent parameter in predicting the dependent parameter.5 Without such analysis, association between FAC and mortality could not be convincingly established. Fractional area change is a method of RV systolic function assessment with high observer variability. FAC is not interchangeable between TTE and TEE. Further statistical analysis is required to establish association between RV FAC and outcome.