The coroner does not know!

Abstract

Dear Editor:We read with interest the study entitled BLong-term mortality and recurrence in patients treated for colonic diverticulitis with abscess formation: a nationwide register-based cohort study^by Gregersen et al. and would like to congratulate the authors [1].Higher recurrence rates were reported in patients following nonoperative management as compared to patients having undergone surgery. Such recurrence rates are in line with the current literature including retrospective data [2] as well as a recent randomized trial [3]. However, we are concerned that the breakdown of recurrence rates by drainage (23.6%) or antibiotics (15.5%) emphasized in the Abstract may be misleading to naïve readers. In fact, the authors did acknowledge at the end of the Discussion that abscess size and location were not known. Moreover, we hope that patients undergoing drainage have received antibiotics. Similarly, highlighting in the Abstract the breakdown of mortality rates by type of treatment for recurrence seems to insinuate a causality. On the other hand, the authors stated that Bthe cause of death was not extracted, and it was therefore not certain whether deaths were correlated wholly or partly to diverticulitis or other unrelated causes.^Furthermore, death rates related to drainage (2%) or antibiotics (1.1%) seem to be taken out of context and can be misleading as the reader is not provided with key details. It is also questionable whether the patients in the operative group were correctly identified as surgery is not the standard of care for diverticulitis complicated by abscess [4]. Many of the study's limitations stem from the inherent flaws of administrative databases with data collection carried out by clerical employees of a government. Compulsory databases are created for quality assurance purposes and lack details necessary for clinical research. Moreover, a large sample size may lead to statistical significance from small variations within the data that may confer no clinical significance [5]. These problems were not adequately addressed by the authors.In conclusion, although we are impressed with the exhaustive work that lies behind such a large database, we believe that attributing mortality to a disease may be misleading when the cause of death is unknown. Compliance with ethical standardsConflict of interest None.Previously copyrighted material None.Presentations None.