Authors’ response to commentaries on rosuvastatin for delirium and cognitive impairment in sepsis-associated acute respiratory distress syndrome

We are encouraged by the interest (1,2) in our publication that evaluated the effect of rosuvastatin vs. placebo on delirium in the intensive care unit (ICU) and subsequent cognitive impairment (3). We appreciate the critique of our publication by Pourafkari et al. (2) commented about our study being underpowered for secondary analyses of cognitive impairment at 6- and 12-month follow-up. We reported a 2% absolute reduction in the proportion of patients with cognitive impairment for rosuvastatin vs. placebo at 6-month follow-up (36% vs. 38%). This 2% reduction was not statistically significant (treatment effect 0.93; 95% confidence interval: 0.39–2.22; P=0.87), which may be due to the study being underpowered (i.e., too few patients to detect such a small effect). However, the subsequent comparison of cognitive impairment at 12-months (30% vs. 28% for rosuvastatin vs. placebo, respectively) and the vast majority of the individual standardized tests used to evaluate cognitive impairment favored placebo relative to rosuvastatin. Hence, to be convinced that the lack of statistical significance at 6 months was attributable to the trial’s sample size, we would have expected non-significant results consistently favoring rosuvastatin, which was not the case.