Pleural effusion and thoracentesis in ICU patients: A longitudinal observational cross-sectional study

Background: Pleural effusion is common among patients in the intensive care unit (ICU) but reported prevalence varies. Thoracentesis may improve respiratory status, however, indications for this are unclear. We aimed to explore prevalence, development, and progression of pleural effusion, and the incidence and effects of thoracentesis in adult ICU patients. Methods: This is a prospective observational study utilizing repeated daily ultrasonographic assessments of pleurae bilaterally, conducted in all adult patients admitted to the four ICUs of a Danish university hospital throughout a 14-day period. The primary outcome was the proportion of patients with ultrasonographically significant pleural effusion (separation between parietal and visceral pleurae >20 mm) in either pleural cavity on any ICU day. Secondary outcomes included the proportion of patients with ultrasonographically significant pleural effusion receiving thoracentesis in ICU, and the progression of pleural effusion without drainage, among others. The protocol was published before study initiation. Results: In total, 81 patients were included of which 25 (31%) had or developed ultrasonographically significant pleural effusion. Thoracentesis was performed in 10 of these 25 patients (40%). Patients with ultrasonographically significant pleural effusion, which was not drained, had an overall decrease in estimated pleural effusion volume on subsequent days. Conclusion: Pleural effusion was common in the ICU, but less than half of all patients with ultrasonographically significant pleural effusion underwent thoracentesis. Progression of pleural effusion without thoracentesis showed reduced volumes on subsequent days.


Editorial Comment
This study confirms that pleural effusions are common in the intensive care unit (ICU). The effect of pleural effusions and of thoracocentesis in patients admitted to the ICU are still uncertain. This emphasizes the need for interventional clinical trials in this patient group to assess risk and benefit.

| INTRODUCTION
Pleural effusion is common in critical illness with a reported prevalence in patients admitted to the intensive care unit (ICU) between 8% and 62% for radiographically confirmed diagnosis, 1,2 and 37% and 81% for ultrasonographically confirmed diagnosis. 1,[3][4][5] Pleural effusion has been associated with restrictive ventilatory impairment, hypoxaemia, 6 and failure in weaning from mechanical ventilation, 7 but the full clinical implications have not yet been determined. In diagnosing pleural effusion, ultrasonography has the advantages of repeatability and bedside availability, and the sensitivity and specificity approaches that of computed tomography. 8,9 An earlier study utilizing ultrasonography in the ICU setting considered pleural effusion clinically relevant when the separation between the visceral and parietal pleurae was above 45 mm, 10 while a more recent study considered pleural effusion clinically relevant when pleural separation was equal to or greater than 20 mm together with a potential adverse effect on patient progress. 11 Criteria that clearly define clinical significance of pleural effusion in ICU patients, however, are lacking. 12 Pleural drainage of pleural effusion through thoracentesis is frequently conducted in the ICU 13 ; the complication rate is low with the most common complication being pneumothorax having an incidence of 0.8%, or even lower with procedural ultrasonographical guiding. 14,15 Observational studies indicate that thoracentesis in patients with pleural effusion improves oxygenation and ventilation with increased post-procedural partial pressure of arterial oxygen to fraction of inspired oxygen (PaO 2 /FiO 2 ) ratio and end-expiratory lung volumes, 15 although paradoxically, pleural drainage has been associated with increased mortality in ICU patients with pleural effusion. 16 The progression of pleural effusion without thoracentesis in ICU patients has never been quantified, and no randomized clinical trials of thoracentesis in ICU patients with pleural effusion have been conducted. 15 Consequently, indication and timing for when to drain pleural fluid is not yet standardized, nor are the clinical consequences of pleural drainage fully enlightened. 17 The aim of this study was to investigate the prevalence of ultrasonographically significant pleural effusion, and secondarily, to quantify this over time and to assess the incidence of thoracentesis, in adult patients admitted to the ICU.

| METHODS
This is a prospective longitudinal observational quality control study utilizing cross-sectional daily ultrasonographic measurements. The protocol was published on zenodo.org prior to study initiation. 18  Approval to obtain data from patient medical journals was obtained from the local head of department as required. The project was registered in the North Denmark Regional research registry (ID: 2021-136) as according to the Danish Data Protection agency.

| Study procedures and population
Bilateral bedside pleural ultrasonography was performed daily on all adult patients (≥18 years) admitted to the four ICUs at Aalborg University Hospital, Denmark (a cardiothoracic ICU, a neuro and trauma ICU, and two multidisciplinary ICUs, for a total of 26 patient beds) in a 14-day period from September 27 to October 10, 2021.
Ultrasonography was performed by two medical students and two physicians who had all previously completed an ultrasonography course including 6 h of e-learning and 6 h of workshop followed by a practical exam. To ensure adequate and equal proficiency with the procedure, a 4-h structured observation and a repeated exam, supervised by a cardiothoracic anaesthesiologist (A.M.) experienced in point-of-care ultrasonography, was conducted prior to study initiation.
Bilateral examinations could be conducted in 15-20 min and were conducted from 08:00 to 15:00 each day. To evaluate protocol adhesion, we report the proportion of patient days, defined as the patient being in the ICU at any time from 08:00 to 15:00 within the inclusion period, where ultrasonographic examinations were conducted.

| Ultrasonography
The ultrasonographic assessments of pleurae were performed with the patients in supine position and a 15 elevation of the torso. 10 Bilateral pleural views were obtained according to the Consensus Document ESC/EACVI for Focus Cardiac Ultrasound and Lung Ultrasound. 20 A transverse section perpendicular to the body axis was obtained with the intrapleural fluid visible as an anechoic or hypoechoic layer between the parietal and visceral pleurae. Fluid measurement was performed and recorded at the lung basis along a perpendicular line between the parietal and visceral pleurae at the largest pleural separation (Figure 1). Measurements were conducted at end-expiration for patients on mechanical ventilation, and endinspiration for non-ventilated spontaneously breathing patients. 10 To quantify the estimated volume of pleural effusion, the simplified Balik formula was used: where Vol = estimated volume (mL) of pleural effusion, and Sep = maximal separation between parietal and visceral pleurae (mm). 10 Ultrasonographically significant pleural effusion was defined as pleural separation >20 mm, equal to an estimated volume >400 mL.
Any pleural effusion was defined as pleural separation >5 mm.
All ultrasonographic measurements were conducted using Vivid S5 or S6, GE or SonoSite X-porte ultrasonography units with cardiac transducers (3ScRS).

| Outcome measures
The primary outcome was the proportion of patients with ultrasono-

| Data
Ultrasonographically collected data were compared to data from the

| Blinding
The examinations performed were not part of the daily clinical practice in the ICUs. Treating physicians were kept unaware of results of the ultrasonographic examinations. However, for safety reasons, if ultrasonographically significant pleural effusion was identified in patients with severe or progressing respiratory failure, as evaluated by one of the physicians responsible for the study conduct, the treating physicians were informed. Study examiners were aware of any thoracentesis performed.

| Statistical analyses
No power estimation was conducted, the inclusion period was defined by feasibility and therefore the number of patients should be considered a convenience sample.
Categorical data are presented as numbers and percentages, continuous data as means with standard deviations or 95% confidence intervals, or medians with interquartile ranges (IQR) as appropriate.
Normality of data was assessed by histograms and quantile-quantile

| Baseline parameters
The types of ICU admittance were evenly distributed between medical and surgical admissions. The most common reasons for ICU admission were pneumonia and sepsis. See Table 1 for baseline characteristics. sepsis (see Table 1). Thoracentesis was performed during ICU stay in 10 of the 25 patients (40%) with pleural separation >20 mm, in 3 of 13 patients (23%) with pleural separation >5 mm and ≤20 mm, and additionally, in three patients without any pleural fluid (pleural separation = 0 mm) in the last pre-procedural study ultrasonography conducted. In total, 15 patients (19%) underwent thoracentesis in the ICU regardless of ultrasonographically identified pleural effusion. See Table 2 for pleural effusion and thoracentesis-related outcomes.  Table 3 for clinical outcomes.

| Outcomes
We found no correlation between estimated volume of pleural effusion prior to thoracentesis and cumulated drained volume in 24 h after thoracentesis, as illustrated by Figure 3.
No significant differences were found in neither the 24-h PaO 2 / FiO 2 ratios of patients with ultrasonographically significant pleural effusion prior to and after pleural drainage, nor in the 24-h PaO 2 /FiO 2 ratios prior to and after verified ultrasonographically significant pleural effusion in patients who were never drained. See Table 4 for an overview of oxygenation-related parameters, including mechanical ventilation parameters, prior to and after thoracentesis or identified ultrasonographically significant pleural effusion.  drainage as compared to 40% in the current study. 11 Since the cut-off for ultrasonographical significant pleural effusion has been proposed to be the level which is clinically relevant when combined with potential adverse effect on patient progress, 11 the 40% drained may indeed be considered low, as large proportions of our patients were mechanically ventilated, or had pneumonia or other cardiopulmonary pathology, that is, conditions in which patient progress would likely be negatively affected by pleural effusion.

| DISCUSSION
We identified no significant effect of thoracentesis on the 24-h PaO 2 /FiO 2 ratio, which stands in contrast to previous studies pooling larger patient numbers, 15   Defined as maximal separation between parietal and visceral pleurae >20 mm at any side on any day in the ICU in the inclusion period. b Comparison between patients with and without ultrasonographically significant pleural effusion, Wilcoxon rank-sum tests for continuous data and Chi-squared tests for mortalities.   substitution, diuretics and other anticongestive drugs. 17,24 Nevertheless, this fluctuating nature may also question the efficacy of thoracocentesis. Unfortunately, current evidence is insufficient to evaluate effects of thoracentesis on clinical patient-important outcomes such as ICU length-of stay or duration of mechanical ventilation. 15

| Strengths and limitations
The current study has several strengths. These include a prepublished protocol, 18

FUNDING INFORMATION
Departmental funding only. This research did not receive any specific grant from funding agencies in the public, commercial, or notfor-profit sectors.