Lung ultrasound is non-inferior to bronchoscopy for confirmation of double-lumen endotracheal tube positioning: a randomized controlled noninferiority study

Background: Appropriate placement of left-sided double-lumen endotracheal tubes (LDLTs) is paramount for optimal visualization of the operative field during thoracic surgeries that require single lung ventilation. Appropriate placement of LDLTs is therefore confirmed with fiberoptic bronchoscopy (FOB) rather than clinical assessment alone. Recent studies have demonstrated lung ultrasound (US) is superior to clinical assessment alone for confirming placement of LDLT, but no large trials have compared US to the gold standard of FOB. This noninferiority trial was devised to compare lung US with FOB for LDLT positioning and achievement of lung collapse for operative exposure. Methods: This randomized, controlled, double-blind, noninferiority trial was conducted at the Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand from October 2017 to July 2019. The study enrolled 200 ASA classification 1-3 patients that were scheduled for elective thoracic surgery requiring placement of LDLT. Study patients were randomized into either the FOB group or the lung US group after initial blind placement of LDLT. Five patients were excluded due to protocol deviation. In the FOB group (n = 98), fiberoptic bronchoscopy was used to confirm lung collapse due to proper positioning of the LDLT, and to adjust the tube if necessary. In the US group (n = 97), lung ultrasonography of four pre-specified zones (upper and lower posterior and mid-axillary) was used to assess lung collapse and guide adjustment of the tube if necessary. The primary outcome was presence of adequate lung collapse as determined by visual grading by the attending surgeon on scale from 1 to 4. Secondary outcomes included the time needed to adjust and confirm lung collapse, the time from finishing LDLT positioning to the grading of lung collapse, and intraoperative parameters such has hypotension or hypertension, hypoxia, and hypercarbia. The patient, attending anesthesiologist, and attending thoracic surgeon were all blinded to the intervention arm. Results: The primary outcome of lung collapse by visual grading was similar between the intervention and the control groups, with 89 patients (91.8%) in the US group compared to 83 patients (84.1%) in the FOB group (p = 0.18) experiencing adequate collapse. This met criteria for noninferiority per protocol analysis. The median time needed to confirm and adjust LDLT position in the US group was 3 min (IQR 2-5), which was significantly shorter than the median time needed to perform the task in the FOB group (6 min, IQR 4-10) (p = 0.002). Conclusions: In selected patients undergoing thoracic surgery requiring LDLT, lung ultrasonography was noninferior to fiberoptic bronchoscopy in achieving adequate lung collapse and reaches the desired outcome in less time.