Routine monitoring with pleural manometry during therapeutic large-volume thoracentesis to prevent pleural-pressure-related complications: a multicentre, single-blind randomised controlled trial

Background: In patients with non-expandable lung, removal of pleural fluid can result in excessively negative pleural pressure, which is associated with chest discomfort, pneumothorax, and re-expansion pulmonary oedema. Pleural manometry is widely used to safeguard against pressure-related complications during thoracentesis despite little evidence to support the approach. We investigated whether monitoring of pleural pressure with manometry during thoracentesis could protect against complications compared with assessment of symptoms alone. Methods: We did a prospective randomised single-blind trial involving patients with large pleural effusions at two academic medical centres in, Nashville, TN, and Baltimore, MD, USA. Eligible patients were adults with free-flowing effusions estimated to be at least 0·5 L who could remain seated throughout the procedure. Patients were randomly assigned 1:1 to receive thoracentesis guided by symptoms only (control)or by symptoms plus manometry at timepoints based on volume drained. The randomisation schedule was computer generated, used permuted blocks of four and six, and was stratified by participating institution. Patients, who were masked to study-group assignment, were asked to rate chest discomfort on 100 mm visual analogue scales before, during, and after drainage. In both groups drainage was discontinued before complete evacuation of pleural fluid if patients developed persistent chest discomfort, intractable cough, or other complications. In the manometry group, an additional criterion for stopping was if end-expiratory pleural pressure was lower than −20 cm H ₂ O or declined by more than 10 cm H ₂ O between two measurements to a value less than or equal to −10 cm H ₂ O. The primary outcome was overall chest discomfort from before the start to after the procedure measured by patients 5 min after the end of drainage. Analysis was by modified intention to treat (ie, included all patients with any procedure or outcome data). This trial is registered with ClinicalTrials.gov, number NCT02677883. Findings: Between March 4, 2016, and Sept 8, 2017, 191 patients were screened, of whom 128 were randomly assigned treatment and 124 were included in the final analysis (62 in each group). Four patients were excluded because of manometer malfunction (n=2), inability to access effusion due to pleural tumour burden (n=1), and inability to remain seated (n=1). Groups did not differ for the primary outcome (mean difference in chest discomfort score 2·4 mm, 95% CI −5·7 to 10·5, p=0·56). Six (10%)of 62 patients in the control group had asymptomatic pneumothorax ex vacuo compared with none in the manometry group (p=0·01). No serious complications occurred in either group. Interpretation: Measurement of pleural pressure by manometry during large-volume thoracentesis does not alter procedure-related chest discomfort. Our findings do not support the routine use of this approach. Funding: Centurion Medical Products.