Evidence-based nutrition support in the intensive care unit: an update on reported trial quality

Meta-epidemiological reviews report that trials of nutritional support in critical illness rarely fulfil basic quality requirements, with overall quality rated as 'worse than poor'. This update reviews recently published trials to determine whether current evidence meets or exceeds basic quality requirements.

Although recent trials were significantly more likely to report blinding, there is a concerning trend towards a decrease in overall trial quality. Many recent trials fail to report the use of 'any' of three key validity criteria: use of blinding, presentation of intention-to-treat analysis and the maintenance allocation concealment.

Future researchers must improve the quality with which trials are conducted and reported. Submitting a clinical trial to an approved registry prior to enrolling patients provides transparency of conduct. Investigators must ensure that an intention-to-treat analysis is reported, especially when a subset efficacy analysis is presented, even if the intention-to-treat analysis requires imputing missing data values. Investigators also need to improve reporting details concerning allocation concealment and blinding. Finally, until investigators, editors and reviewers embrace these measures, we strongly recommend that readers should become familiar with the appropriate evidence-based medicine users' guides so that they can base clinical decisions on valid studies.