Acute kidney injury (AKI) is common after major abdominal surgery. Selection of candidate kidney protective strategies for testing in large trials should be based on robust preliminary evidence.


We conducted a secondary analysis of the Restrictive versus Liberal Fluid Therapy in Major Abdominal Surgery (RELIEF) trial, in adult patients undergoing major abdominal surgery and randomly assigned to a restrictive or liberal perioperative fluid regimen. The primary outcome was maximum AKI stage prior to hospital discharge. We developed two multivariable ordinal regression models and tested the primary hypothesis that modifiable risk factors associated with increased maximum stage of postoperative AKI could be identified. Each model used a separate approach to variable selection to assess the sensitivity of our findings to modelling approach. For model 1 variable selection was informed by investigator opinion; for model 2 the Least Absolute Shrinkage and Selection Operator (LASSO) technique was used to develop a data-driven model from available variables.


Of 2444 patients analyzed, stages 1, 2 and 3 AKI occurred in 223 (9.1%), 59 (2.4%), and 36 (1.5%) patients, respectively. In multivariable modelling by model 1 administration of a nonsteroidal anti-inflammatory drug (NSAID) or cyclooxygenase-2 (Cox-2) inhibitor, intraoperatively only (OR 1.77 [99% CI 1.11-2.82]), and preoperative day-of-surgery administration of an angiotensin converting enzyme inhibitor or angiotensin receptor blocker compared to no regular use (OR 1.84 [99% CI 1.15-2.94]) were associated with increased odds for greater maximum stage AKI. These results were unchanged in model 2, with the additional finding of an inverse association between nadir hemoglobin concentration on POD-1 and greater maximum stage AKI.


Avoiding intraoperative NSAIDs or Cox-2 inhibitors is a potential strategy to mitigate the risk for postoperative AKI. Our findings strengthen the rationale for a clinical trial comprehensively testing the risk-benefit ratio of these drugs in the perioperative period.