Twenty-seven studies (n=979) were included in the review. The study designs included were double-blind (n=560), evaluator blind (n=10) or open parallel (n=121) randomised controlled trials (RCTs), and open uncontrolled studies (n=288).
Of the 27 studies included, nine had follow-up of less than 80%, with three of these not using an intention-to-treat analysis, 11 RCTs did not report on randomisation concealment, and blinding was deemed inadequate in 9 studies, of which three were RCTs.
Effectiveness.
Cyclosporin A: 11 studies reported a decrease in disease activity with cyclosporine, with superiority over placebo in all 6 RCTs.
Systemic glucocorticosteroids: 2 RCTs in children reported short-term decreases in severity of 22% and 39% after treatment with beclomethasone diproprionate and flunisolide, respectively.
Interferon gamma: 2 poor-quality RCTs reported interferon as superior to placebo in children and adults. One uncontrolled study reported a 30% decrease in the intensity of lesions, while another uncontrolled study reported low response rates.
Intravenous immunoglobulin: 3 small studies reported limited responses to immunoglobulin.
Mycophenolate mofetile: 2 small uncontrolled studies reported short-term decreases in severity of 55% and 68% after 8 and 12 weeks’ treatment, respectively.
Azathioprine: one double-blind RCT reported a 27% decrease in disease activity after 12 weeks.
Infliximab: one small uncontrolled study reported a benefit of 50% or more in 2 out of 9 patients.
Chinese herbal medicines: 3 double-blind RCTs reported conflicting results, with 2 poor-quality studies reporting beneficial results and the third study reporting no difference between treatment and placebo.
Adverse events.
Most studies either did not report on serious adverse events, or reported that no serious adverse events occurred. Reported adverse events included abdominal pain, herpes simplex infection, acute cholecystitis, hypertension, haematurea, increased serum creatinine, serum sickle-like illness and serious infusion reaction. The proportion of patients withdrawing due to adverse events ranged from 0 to 10%.