Index of prolonged air leak score validation in case of video-assisted thoracoscopic surgery anatomical lung resection: results of a nationwide study based on the French national thoracic database, EPITHOR.

OBJECTIVES

The incidence rate of prolonged air leak (PAL) after lobectomy, defined as any air leak prolonged beyond 7 days, can be estimated to be in between 6 and 15%. In 2011, the Epithor group elaborated an accurate predictive score for PAL after open lung resections, so-called IPAL (index of prolonged air leak), from a nation-based surgical cohort constituted between 2004 and 2008. Since 2008, video-assisted thoracic surgery (VATS) has become popular in France among the thoracic surgical community, reaching almost 14% of lobectomies performed with this method in 2012. This minimally invasive approach was reported as a means to reduce the duration of chest tube drainage. The aim of our study was thus to validate the IPAL scoring system in patients having received VATS anatomical lung resections.

METHODS

We collected all anatomical VATS lung resections (lobectomy and segmentectomy) registered in the French national general thoracic surgery database (EPITHOR) between 2009 and 2012. The area under the receiver operating characteristic (ROC) curve estimated the discriminating value of the IPAL score. The slope value described the relation between the predicted and observed incidences of PALs. The Hosmer-Lemeshow test was also used to estimate the quality of adequacy between predicted and observed values.

RESULTS

A total of 1233 patients were included: 1037 (84%) lobectomies and 196 (16%) segmentectomies. In 1099 cases (89.1%), the resection was performed for a malignant disease. Ninety-six patients (7.7%) presented with a PAL. The IPAL score provided a satisfactory predictive value with an area under the ROC curve of 0.72 (0.67-0.77). The value of the slope, 1.25 (0.9-1.58), and the Hosmer-Lemeshow test (χ(2) = 11, P = 0.35) showed that predicted and observed values were adequate.

CONCLUSION

The IPAL score is valid for the estimation of the predictive risk of PAL after VATS lung resections. It may thus a priori be used to characterize any surgical population submitted to potential preventive measures.