Video-assisted thoracoscopic lobectomy: intention-to-treat analysis to convey realistic outcomes.

Abstract

The paper from Laursen et al. [1] is a well-written retrospective analysis comparing the early outcomes of video-assisted thoracoscopic surgery (VATS) versus open lobectomies performed in the eastern region of Denmark. The authors found that VATS patients had a 46% lower incidence of major cardiopulmonary complications compared with the open ones (20 vs 37%). After adjusting for several confounders by using multivariable analysis, they found that the odds ratio of having major complications after thoracotomy compared with VATS was 1.9. The message of this work is not new. Several other retrospective single and multicentre studies have found similar results [2–4]. What is new is the inclusion of patients who started in VATS and were converted to open surgery for major intraoperative complications. The study is therefore an intention-to-treat analysis and showed the real impact of VATS lobectomy. The authors reported a 4.5% incidence of conversion rate in their series. In this regard, this analysis adds to the existing literature mostly based on organizational databases and in which an intention-to-treat analysis is mostly unfeasible. It would be desirable that future comparative studies regarding VATS lobectomy would include patients converted to thoracotomy for major intraoperative complications to provide more realistic findings. A recent paper from the European Society of Thoracic Surgeons (ESTS) database and endorsed by the ESTS Minimally Invasive Interest Group [5] showed similar results on a case-matched comparison between 2721 pairs of patients submitted to open versus VATS lobectomies. The authors found a 19% reduction in major cardiopulmonary morbidity (15.9 vs 19.6%) and a 47% reduction in in-hospital mortality (1 vs 1.9%) after VATS compared with thoracotomy. The two major differences between these two papers are that the ESTS analysis [5] is not an intention to treat one (there is no way to understand whether converted patients have been included among the VATS group as no specific variable is present at the moment in the ESTS database) and that the Laursen et al. [1] analysis is not matched. Although patients in the VATS group are slightly older, they had less comorbidities and their tumours were smaller compared with the open cases. This may have introduced an inherent selection bias. Matching, on the other hand, would have inevitably reduced the sample size and would have reduced the representativeness of the actual population, who were subject to this operation during the study period. This paper is certainly another confirmation of the positive impact of the minimally invasive approach on the early outcome of lung resection. Unfortunately, the authors did not analyse longterm survival in their cohort of patients. This would have been of utmost interest considering the highest number of lymph nodes harvested during VATS compared with open surgery in their series and the inclusion of patients converted to thoracotomy. The authors should be commended for a nicely written paper and for their effort to communicate a real clinical picture after this type of operation.