To the Editor:

We read with interest the study by Minami et al. which recently appeared in the Journal of Anesthesia [1]. We would like to discuss several methodological issues associated with this study.

Firstly, variables with a significance of P < 0.05 were selectively entered into the multivariable Cox regression model. To adequately adjust for as much confounding as possible, the criteria for including variables in the multivariable model may be loosened, for example, by using screening criteria of P < 0.30. In addition, propensity score analysis may further reduce any potential confounding. Secondly, the authors do not provide any information on suprarenal or infrarenal fixation of the endograft. This information is important because suprarenal fixation may be a risk factor for acute kidney injury (AKI) [2]. More information on the aneurysm, including morphology class, and on the procedure, including distal anastomosis type, would also be useful. Thirdly, an analysis of hydroxyethyl starch according to estimated glomerular filtration rate in the subgroup analysis would be interesting, although the sample size may not be sufficient. Fourthly, the severity of AKI should be considered. Stage 1 AKI may be only functional and transient [3]. Using more severe stages, such as AKI stage 2 or 3, as the primary outcome would suggest that cases of AKI are more prone to represent parenchymal injury.