To the Editor:

With interest, we read the study by Oh et al. [1], which retrospectively analyzed a large cohort of patients admitted to the surgical intensive care unit. Although results are encouraging in the quest to find evidence that supports the relevance of intensivist in the management of post-surgical patients, two questions have come up in the interpretations of the results.

Firstly, parameters of intraoperative management which may affect the postoperative incidence of acute kidney injury (AKI) were not sufficiently included in the contributors of propensity score. Just like the patient management variables during postoperative days, intraoperative parameters including hemodynamics, vasopressor infusion, hydroxyethyl starch use, nephrotoxic agent use such as diuretics, and transfusion are equally or more important for the development of AKI especially during immediate postoperative period [2]. Secondly, in the sensitivity analysis of AKI according to admission time, there were only two time-points where the associations were significant and there seems no dose–response relationship between the duration of non-intensivist coverage and the incidence of AKI. To suggest causality in the retrospective analysis, dose–response relationship should be demonstrated in addition to consistency, biological plausibility and temporal relationship [3]. Further evidence are required to substantiate their hypothesis.