Re: Postoperative Complications do not Affect Long‐term Outcome in Esophageal Cancer Patients

Abstract

It is with great interest that we read the recent report by Lindner and colleagues in this Journal [1]. In their report, the authors describe the results of a retrospective review of outcomes for transthoracic esophagectomy for esophageal cancer (including both squamous cell carcinoma and adenocarcinoma), with stapled anastomosis, at a single centre. Using multivariate Cox regression analysis, they conclude that postoperative complications do not play a significant role in determining long-term survival.While the authors' contribution to this discussion is welcomed, we are concerned about the validity of their findings. The authors' results contradict a growing body of evidence suggesting that, in fact, patients who experience postoperative complications suffer reduced long-term survival as a result. This has been examined in a number of recent publications, many involving large-scale database analysis, across numerous specialties including operations for colorectal [2], lung [3], esophageal [4], gastric [5], and hepatic [6] malignancies. Though not yet fully elucidated, molecular and immune mechanisms underlying this relationship are being actively investigated; pathways including neutrophil-cancer cell interactions [7, 8] and bacterial antigen receptor cell activation [9, 10] have been previously implicated. A recent meta-analysis of long-term survival following complications after surgery published by our group [11], incorporating results for over 20,000 patients, has reported similar conclusions, with a significant reduction in long-term survival for patients with postoperative complications (hazard ratio and 95 % confidence interval: 1.28 [1.21, 1.34]). For infectious complications, this relationship was stronger still (1.92 [1.50, 2.35]).Considering this discrepancy, we note that in Lindner and colleagues' examined cohort, adenocarcinoma patients who suffered major complications had a reduced survival (median 29.9 months) compared to those with no complications (33.2 months), with a similar trend for patients who suffered anastomotic leak (33.7 months) compared to those who did not (42.8 months). Though this did not reach statistical significance, this trend would appear to potentially contradict the authors' conclusions.Lindner and colleagues acknowledge that one limitation of their study is the moderate sample size, involving only 134 patients in total. However, we are concerned that this acknowledgement understates the problem present in this study. The authors have conducted a subset analysis for both adenocarcinoma and squamous cell carcinoma, and further subdivided these groups for major, minor, and no complications. At this level of subdivision, the compared groups' (the individual size of which have not been reported) risk is being reduced to numbers so small that they are not suitable for meaningful analysis. Only a total of 21 patients experienced a ''major'' complication, and how many of these were in each diagnostic subcategory is unclear. If we presume an equal risk between the groups, this wou...