“A Message in the Bottle”

We read with interest the work by Dres et al.¹  We would like to highlight some aspects that deserve particular consideration. First of all, the authors scanned the basal pleural space, estimating the effusion volume according to the British Thoracic Society classification as small, moderate, or large.²  They also used the equation proposed by Balik et al.³  to estimate the effusion volume at the maximal end-expiratory pleural distance between the parietal and visceral pleura. Specifically, the authors stated that they performed the ultrasound exam in the semirecumbent position, with the patient’s torso reclined at about 45°, as free fluids accumulate at the lung bases due to gravity. We would tend to disagree with the authors at this point considering that the patients of Balik et al.³  were investigated supine with a mild torso elevation of 15°; furthermore, the mean prediction error of this equation is quite high (158 ± 160 ml). The patient’s position has a high impact on the estimation of pleural effusion volume. Furthermore, a consistent evaluation of the effusion volume is very challenging for many reasons (e.g., tall people have a larger thoracic cavity area, diaphragm position, phrenic nerve palsy, diaphragmatic hernia), which was not considered or commented on by the authors. The authors also failed to report the laterality of the pleural effusions: It is well known that the ultrasound assessment of pleural effusions is overestimated on the left side, because the heart increases the fluid level (like a stone in a water receptacle). To overcome these limitations, other methods of pleural effusion estimation have been proposed using a transthoracic lung ultrasound approach.^3–6  However, as correctly stated in the study limitations acknowledged by the authors, the biggest problem lies in the numbers. The overwhelming majority of patients had a small pleural effusion, which has barely any impact on patient ventilation and hardly affects complex outcomes such as duration of mechanical ventilation, weaning success, and intensive care unit length of stay. The very low number of patients with moderate to large pleural effusions is thus an important limitation to the study. Consequently, a much larger study is needed. The authors calculated the sample size for their study, starting from the random assumption that a proportion of patients with a pleural effusion of 25% would be found in the group of patients with successful weaning. From these assumptions, the calculation of 136 patients for the sample size is correct. However, the authors should have recruited 68 patients per group, and not 51 patients with pleural effusion versus 85 patients without. To conclude, we believe the study is up to date and interesting; however, keeping the aforementioned limitations in mind, we are still far from reaching a definite conclusion on the real impact of pleural effusion on weaning.

The authors declare no competing interests.