Fourteen trials (6,717 participants, range 50 to 1,345) were included in the review. Ten trials reported allocation concealment and 12 studies used intention-to-treat analyses. Most studies were reported to be appropriately blinded. Baseline characteristics for the treatment groups were described as similar across the studies.
Results of the mixed treatment comparison found that fingolimod administered at 0.5mg conferred significantly greater benefits than all comparators, with lower annualised relapse rates (ARR) compared to interferon beta 1a 30mcg (6 MIU) (ARR 1.67, 95% CI 1.32 to 2.10), interferon beta 1a 22mcg (ARR 1.55, 95% CI 1.26 to 1.90), interferon beta 1a 44mcg (ARR 1.93, 95% CI 1.59 to 2.34), interferon 1b 250mcg (ARR 1.51, 95% CI 1.22 to 1.860) and glatiramer acetate (ARR 2.32, 95% CI 1.95 to 2.77).
Covariate analyses showed significant effects of baseline Expanded Disability Scale Scores (ARR 1.63, 95% CI 1.04 to 2.56; 13 trials) and year of publication (ARR 0.96, 95% CI 0.94 to 0.98). Repeat mixed treatment analyses with adjustments for each covariate found no differences between actual relative treatment effects for the unadjusted models.
There were no significant differences in results from the mixed-treatment comparison model and those from head-to-head comparisons. Estimates provided by the reduced Χ² statistic indicated a reasonably good fit for both the adjusted and unadjusted models. Informal assessment of the model fit showed that the mixed-treatment comparison-derived relative rates closely matched to rates from head-to-head trials.