Prediction Of The Course Of Air Leaks After Pulmonary Lobectomy Using Continuous Flow Data
Kwanyong Hyun1, Gongmin Rim1, Sook Whan Sung2, Deog Gon Cho3.
1Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea, Republic of, 2Ewha Womens University Seoul Hospital, Seoul, Korea, Republic of, 3St. Vincent's Hospital, The Catholic University of Korea, Suwon, Korea, Republic of.
BACKGROUND: The assessment for a degree of air leak has been conducted by the subjective analysis of air bubble and this leads to the difficulty for using air leak as an evaluation factor. The aim of this study was to extract predictive factors of prolonged air leak (PAL) and air leak cessation (ALC) including air flow data as objective parameter by digital drainage system.
METHODS: 352 patients who underwent lung lobectomy and had flow records were reviewed. The flow data were extracted at a designated interval (1, 2, and 3 postoperative hours (POH), then 3 times a day thereafter (06AM, 13PM, 19PM)). ALC was determined to be less than 20 mL/min for the past 12 hours and PAL was defined having ALC after 5 days. Using Kaplan-Meier estimate of time to ALC, cumulative incidence curves were obtained. Cox regression analysis were performed to determine the effect of variables on the rates of ALC.
RESULTS: Incidence of PAL was 18.2% (64/352). By ROC curve analysis, a cut-off value was set at 180 mL/min for 3 POH flow and 73.3 mL/min for 1 POD flow, having the sensitivity and specificity of 88.9% and 82.5%, respectively. The rates of ALC were 56.8% at 48 POH and 65.6% at 72 POH by Kaplan-Meier analysis. Multivariate Cox regression analysis revealed 3 POH flow (≤80 mL/min), operation time (≤220 min), and RML lobectomy independently predicted ALC.
CONCLUSIONS: The air flow measured by digital drainage system will be useful predictors of PAL and ALC, which may provide optimization of hospital course.
LEGEND: Cumulative incidence of ALC (air leak cessation) by postoperative hour (95% CI indicated by shaded area) 
Back to 2023 Abstracts