Letters to the Editor

CONSORT

An adequate nutritional status is crucial for optimum function of cells and organs, and for wound healing. Options for artificial nutrition have greatly expanded in the past few decades, but have concomitantly shown limitations and potential side-effects. Few rigorous randomised controlled trials (RCTs) have investigated enteral or parenteral nutritional support, and evidence-based clinical guidance is largely restricted to the first week of critical illness. In the early stages of critical illness, whether artificial feeding is better than no feeding intervention has been given little attention in existing RCTs. Expected beneficial effects of various forms of early feeding interventions on rates of morbidity or mortality have generally not been supported by results of recent high-quality RCTs. Thus, whether nutritional interventions early in an intensive care unit (ICU) stay improve outcomes remains unclear. Trials assessing feeding interventions that continue after the first week of critical illness and into the post-ICU and post-hospital settings are clearly needed. Although acute morbidity and mortality will remain important safety parameters in such trials, primary outcomes should perhaps, in view of the adjunctive nature of nutritional intervention in critical illness, be focused on physical function and assessed months or even years after patients are discharged from the ICU. This Series paper is based on results of high-quality RCTs and provides new perspectives on nutritional support during critical illness and recovery.

The revised Consolidated Standards of Reporting Trials (CONSORT) statement recognizes the importance of including data so that readers are able to judge the reliability or relevance of randomized controlled trial (RCT) findings. However, it is unclear whether journals consistently report the pre-randomization information necessary to assess the general applicability of trial results. The objective of this study was to assess the extent of adherence to pre-randomization data reporting in 4 leading general medicine journals, as recommended by the revised CONSORT statement over a specific time frame. A total of 480 RCTs met the eligibility criteria for review. In 2004, 138 of 246 (56%) articles reported the full CONSORT pre-randomization information, of which 121 (88%) presented this information in diagram form. In 2006, 136 of 234 (58%) reported the full information, of which 127 (93%) presented the information in a diagram. These numbers indicate no improvement in the reporting of pre-randomization information. In conclusion, reporting of the full CONSORT data remains inconsistent. We recommend that journals continue to support the CONSORT statement and ensure that adequate information is provided upon submission before commencing the peer review process. The presentation of clear, precise data will ensure that readers are able to better interpret the applicability of findings.