A review of the true methodological quality of nutritional support trials conducted in the critically ill: time for improvement.

In this review we sought to appraise the true methodological quality of nutritional support studies conducted in critically ill patients and to compare these findings to the methodological quality of sepsis trials. An extensive literature search revealed 111 randomized controlled trials conducted in critically ill patients evaluating the impact of nutritional support interventions on clinically meaningful outcomes. Compared with sepsis trials, nutritional support studies were significantly less likely to use blinding (32 of 40 versus 35 of 111, P < 0.001) or present an intention-to-treat analysis (37 of 40 versus 64 of 111, P < 0.001). There was a trend toward the less frequent use of randomization methods that are known to maintain allocation concealment (12 of 40 versus 19 of 111, P = 0.10). Although nutritional support studies demonstrated a significant increase in the use of blinding after the publication of the CONSORT statement in 1996 (9 of 47 versus 26 of 64 post-CONSORT, P = 0.023), there were no improvements in other key areas. Previous publications have described the overall methodological quality of sepsis trials as "poor." Nutritional support studies were significantly worse than sepsis trials in all aspects of methodological quality, and there were few improvements noted over time. To detect important differences in clinically meaningful outcomes in critical care, the methodological quality of future studies must be improved.