Three RCTs with 751 participants and 6 cohort studies with 1,316 participants were included in the review. A further 12 studies of animals were included in the discussion, but these were not addressed in this abstract.
None of the included studies measured fully up to current standards for study design, conduct of trial, analysis and presentation. Many did not include a justification for sample size, were unclear about their method of randomisation, and did not include a discussion of clinical versus statistical significance.
For primary outcomes, the studies were discussed individually but not combined for review purposes. Only one study showed improvement in steroid group.
The weight of evidence provided by these studies, with regard to mortality and morbidity, is weak. They cannot reliably provide reassurance of safety, and they provide weak evidence of some increase in complications, particularly infection-related ones.