(i) The exclusion of patients with mild (grade 1) intraoperative air leak. We included all patients with air leaks in our study as we did not want to make any assumptions about the length of time it would take for air leaks to resolve, whatever the grade of air leak. We allocated our patients into the two groups by minimization which ensured balance between the groups with respect to air leak grades among other factors [4]. Although it may appear that we had a smaller proportion of patients with moderate/severe air leaks (Grades 2/3), further inspection of the definition of air leak grades in fact indicate a likely cross-over in our grading system. D’Andrilli et al. graded mild as ‘countable bubbles’ and moderate as ‘a stream of bubbles’. In our grading system, Grade 1 was defined as ‘a single stream of bubbles’ and Grade 2 ‘multiple (2–5) but discrete bubble streams’. We considered it possible to count the number of streams which equates to the number of sources. It is likely that a substantial proportion of patients graded as 1 in our study would have been included in the Italian study as ‘moderate’ according to their grading criteria. However, the grading of air leak is extremely subjective and inter-observer variation is inevitable. It would be impossible to conclude with certainty if this has in fact contributed to the difference in results in the two trials. (ii) Dose of product applied in each patient. We would like to confirm as stated in our study, that patients in the CoSeal group did in fact receive an additional application of CoSeal if there was a persistent air leak after the first application, and often meant using a double volume of sealant similar to the Italian study. The air tightness test was performed for a third time after the second application, but no further CoSeal was applied if an air leak persisted at this stage.
Read full abstract