OBJECTIVES: One of the reported advantages of digital pleural drainage system is the possibility of predicting the occurrence of prolonged air leak (PAL) based on the recorded pleural pressures and/or air flow through chest tubes. Nevertheless, this fact has never been well supported. The objective of this investigation is to evaluate if the occurrence of PAL can accurately be predicted using clinical data and air leak measurements 24 h after lung resection on conventional pleural drainage system (CPDS). METHODS: Prospective observational study on 100 consecutive non-complicated patients undergoing anatomical lung resection (segmentectomy, lobectomy or bilobectomy). Prior to the operation, the risk of PAL was evaluated according to the score previously published. Twenty-four hours after surgery, two independent observers measured the air flow at forced deep expiration on a CPDS with graduated analogical leak monitor. The agreement between both observers was determined and in case of discrepancy, the mean of both observations was calculated. After discharge, the occurrence of PAL (defined as persistent air leak 5 or more days after the operation) was recorded. A logistic regression model was constructed including two independent categorical variables (PAL score and air flow) and the performance of the model was assessed by non-parametric receiver operating characteristic curves. RESULTS: The series includes 81 lobectomies, 8 bilobectomies and 11 anatomical segmentectomies. Median preoperative PAL score was 1 (range 0-3.5). Any postoperative air flow was observed in 30 cases with a median value of 0 (0-3.5). The prevalence of PAL in the series was 10% (10 of 100 cases). Both independent variables entered in the multivariate model (PAL score P = 0.050, air flow: 0.016) and C-index was 0.83. CONCLUSION: The performance of this simple predictive model, without any electronic recording, warrants a larger multi-institutional study to validate its usefulness in clinical decision-making regarding the management of patients with air leak after lung resection.
OBJECTIVES: One of the reported advantages of digital pleural drainage system is the possibility of predicting the occurrence of prolonged air leak (PAL) based on the recorded pleural pressures and/or air flow through chest tubes. Nevertheless, this fact has never been well supported. The objective of this investigation is to evaluate if the occurrence of PAL can accurately be predicted using clinical data and air leak measurements 24 h after lung resection on conventional pleural drainage system (CPDS). METHODS: Prospective observational study on 100 consecutive non-complicated patients undergoing anatomical lung resection (segmentectomy, lobectomy or bilobectomy). Prior to the operation, the risk of PAL was evaluated according to the score previously published. Twenty-four hours after surgery, two independent observers measured the air flow at forced deep expiration on a CPDS with graduated analogical leak monitor. The agreement between both observers was determined and in case of discrepancy, the mean of both observations was calculated. After discharge, the occurrence of PAL (defined as persistent air leak 5 or more days after the operation) was recorded. A logistic regression model was constructed including two independent categorical variables (PAL score and air flow) and the performance of the model was assessed by non-parametric receiver operating characteristic curves. RESULTS: The series includes 81 lobectomies, 8 bilobectomies and 11 anatomical segmentectomies. Median preoperative PAL score was 1 (range 0-3.5). Any postoperative air flow was observed in 30 cases with a median value of 0 (0-3.5). The prevalence of PAL in the series was 10% (10 of 100 cases). Both independent variables entered in the multivariate model (PAL score P = 0.050, air flow: 0.016) and C-index was 0.83. CONCLUSION: The performance of this simple predictive model, without any electronic recording, warrants a larger multi-institutional study to validate its usefulness in clinical decision-making regarding the management of patients with air leak after lung resection.