To the Editor In the report of the ELAIN trial,1 we think some issues need clarification. First, it is not clear whether the measure of effect for the primary outcome (hazard ratio for 90-day mortality from the Cox model) was presented as crude or adjusted. If it was adjusted, it is not clear what covariates were included. Following the Consolidated Standards of Reporting Trials (CONSORT) recommendations and the published ELAIN protocol,2 it is important to first present a crude measure of effect for the primary outcome and afterwards the adjusted result. Covariate adjustment in trials can provide additional useful information, such as an effect measure clinically closer to the patient level.3 Furthermore, the covariates to be adjusted for in the main analysis should be specified a priori. Nevertheless, the authors used a method (backward elimination) subject to important biases when trying to answer causal questions.4 Additionally, the reported adjustment was not congruent with what was specified in the published protocol,2 wherein the crude measure of effect would be estimated first and then the interaction between the intervention and the baseline risk of death (estimated with specific covariates).
Bruno Adler Maccagnan Pinheiro Besen, Otavio T. Ranzani, Marcelo Park. Timing of Initiation of Renal Replacement Therapy in Critically Ill Patients With Acute Kidney Injury. JAMA. 2016;316(11):1213. doi:10.1001/jama.2016.11332