Pharmacotherapy for Behcet's syndrome

To determine the effects of available pharmacological interventions in treating the different clinical features of Behcet's syndrome.

We searched the Cochrane Musculoskeletal Group's trials register, the Cochrane Controlled Trials Register, and Medline up to January 1998. The computer search was complemented by a hand search of all bibliographic references from the reference lists of included trials. Principal investigators were contacted to seek unpublished literature. All languages were included.

Studies were eligible if they fulfilled all of the four following criteria: 1. Randomized controlled trials, single or double-blind; 2. Participants were patients with Behcet's Syndrome as defined by the International Study Group, 1990 (Int Study Group, 1990); 3. Interventions included any pharmacological therapy compared to placebo or some other pharmacological intervention for the treatment of Behcet's syndrome. 4. Outcome measures included active ocular inflammatory processes, arthritis, mucocutaneous manifestations (oral ulcer, genital ulcer, erythema nodosum), laboratory changes and major events such as adverse effects and death.

The 32 potentially relevant references were assessed by two independent reviewers (MA, AS) according to the inclusion criteria. Ten trials fit the inclusion criteria and were included in this review. From the 10 included trials, data were independently extracted by the same two observers and crosschecked. The quality of the included trials was assessed independently by two observers (MA, AS) using a validated scale (Jadad 1996). For dichotomous measures, the treatment effect for each trial was calculated using a fixed effect model [Peto model (Petitti 1994)]. The weighted mean differences were based, if available, on end-of-trial results. The analysis was conducted separately for each different intervention. Since the trials could not be pooled it was not possible to carry out a sensitivity analysis by quality scores or a subgroup analysis by drug dosages. Because of this lack of comparability across trials and the small number of trials, we could not conduct a heterogeneity test or a funnel plot.

Ten trials and 679 patients were included. The main results were the lack of efficacy of some of the classic treatments for Behcet's syndrome, including colchicine, cyclophosphamide and steroids for eye involvement, azapropazone and colchicine for arthritis and acyclovir, colchicine and topical interpheron for aphthas. The results confirm the protective effects of cyclosporine and azathioprine for eye involvement and benzathine-penicillin for arthritis.

We conclude that further randomized, placebo-controlled, double-blind trials should be carried out to compare cyclosporine, azathioprine and benzathine-penicillin versus placebo in order to make the results generalizable and comparable.