Twenty-four studies were included in the review (n=981): eight randomised controlled trials (RCTs, n=370); 11 controlled trials (n=489); and five observational studies (n=122). Quality scores ranged from 29 to 38 for RCTs, nine to 29 for controlled trials and 20 to 23 for observational studies.
Endoscopic remission: Mean rate of endoscopic remission (defined as no ulceration) with corticosteroids was 17% (95% CI 12% to 22%; n=212); which was significantly higher than placebo (p=0.02). The equivalent rate with enteric diet therapy was 43% (95% CI 33% to 52%; n=100) and was similar with infliximab, at 44% (95% CI 35% to 53%; n=107); both were significantly superior to corticosteroids (p=0.0001). Mean rate of endoscopic remission (variously defined) with azathioprine was 54% (95% CI 38% to 69%; n=137).
Endoscopic improvement: The mean reduction in CDEIS (Crohn's Disease Endoscopic Index of Severity) with corticosteroids was 46% (95% CI 39% to 53%; n=204) and with infliximab was 70% (95% CI 62% to 78%; n=121). The mean reduction on a four-point severity score with enteric diet therapy was 63% (95% CI 53% to 72%, n=91). When data were pooled across scoring tools, corticosteroids were significantly superior to placebo (45% versus 12%, p=0.0008). Enteric diets and infliximab did not differ significantly, but both were significantly superior to corticosteroids (61% p=0.01 for enteric diets and 70%, p<0.0001 for infliximab).
No statistically significant relationship was found between endoscopic and clinical remission (only one study reported this outcome). There was some evidence of a correlation between endoscopic improvement and clinical remission
Subgroup analyses had inconclusive findings. Other data were reported in the review, including 95% CIs for all outcomes.