Objective: To assess the effectiveness of treatments for bullous pemphigoid. Methods: The Cochrane Library search strategy was used to identify randomized controlled trials from MEDLINE and EMBASE, from their inception to September 30, 2001. All randomized controlled trials on interventions for bullous pemphigoid, confirmed by immunofluorescence studies, were included. Results: We found 6 randomized controlled trials with a total of 293 patients. Two trials, one comparing prednisolone, 0.75 mg/kg per day, with prednisolone, 1.25 mg/kg per day, and the other comparing methylprednisolone with prednisolone, did not find any significant difference in effectiveness. The higher dose of prednisolone, however, was associated with more severe adverse effects. Combination treatments of prednisone with azathioprine in one trial and of prednisolone with plasma exchange in another were useful in halving the corticosteroid dose required (mean±SD, 0.52±0.28 mg/kg in the plasma exchange-treated group vs 0.97±0.33 mg/kg in the prednisolone only-treated group). However, a fifth trial, including all 3 treatment groups (prednisolone alone, prednisolone and azathioprine, and prednisolone and plasma exchange), failed to confirm the benefit of combination treatment over prednisolone alone. A trial of 20 patients, comparing prednisone with tetracycline and niacinamide, found no statistically significant difference in response between the 2 groups, but the prednisone-treated group had more serious adverse effects. Conclusions: There is inadequate evidence for a recommendation of a specific treatment for bullous pemphigoid, and there is a need for larger randomized controlled trials with adequate power. Starting doses of prednisolone greater than 0.75 mg/kg per day do not seem to give additional benefit, and it seems that lower doses may be adequate for disease control. The effectiveness of the addition of plasma exchange or azathioprine to corticosteroids has not been established. Combination treatment with tetracycline and niacinamide seems useful, although this needs further validation.
