A randomised controlled trial to compare the safety, effectiveness and cost-effectiveness of doxycycline (200 mg/day) with that of oral prednisolone (0.5 mg/kg/day) for initial treatment of bullous pemphigoid: The bullous pemphigoid steroids and tetracyclines (BLISTER) trial

Background: Bullous pemphigoid (BP) is an autoimmune blistering skin disorder with increased morbidity and mortality in the elderly. Objectives: To evaluate the effectiveness, safety and cost-effectiveness of a strategy of initiating BP treatment with oral doxycycline or oral prednisolone. We hypothesised that starting treatment with doxycycline gives acceptable short-term blister control while conferring long-term safety advantages over starting treatment with oral prednisolone. Design: Pragmatic multicentre two-armed parallel-group randomised controlled trial with an economic evaluation. Setting: A total of 54 dermatology secondary care centres in the UK and seven in Germany. Participants: Adults with BP [three or more blisters at two sites and positive direct and/or indirect immunofluorescence (immunoglobulin G and/or complement component 3 immunofluorescence at the dermal-epidermal junction)] and able to give informed consent. Interventions: Participants were allocated using online randomisation to initial doxycycline treatment (200 mg/day) or prednisolone (0.5 mg/kg/day). Up to 30 g/week of potent topical corticosteroids was permitted for weeks 1-3. After 6 weeks, clinicians could switch treatments or alter the prednisolone dose as per normal practice. Main outcome measures: Primary outcomes: (1) the proportion of participants with three or fewer blisters at 6 weeks (investigator blinded) and (2) the proportion with severe, life-threatening and fatal treatment-related events at 52 weeks. A regression model was used in the analysis adjusting for baseline disease severity, age and Karnofsky score, with missing data imputed. Secondary outcomes included the effectiveness of blister control after 6 weeks, relapses, related adverse events and quality of life. The economic evaluation involved bivariate regression of costs and quality-adjusted life-years (QALYs) from a NHS perspective. Results: In total, 132 patients were randomised to doxycycline and 121 to prednisolone. The mean patient age was 77.7 years and baseline severity was as follows: mild 31.6% (three to nine blisters), moderate 39.1% (10-30 blisters) and severe 29.3% (> 30 blisters). For those starting on doxycycline, 83 out of 112 (74.1%) had three or fewer blisters at 6 weeks, whereas for those starting on prednisolone 92 out of 101 (91.1%) had three or fewer blisters at 6 weeks, an adjusted difference of 18.6% in favour of prednisolone [90% confidence interval (CI) 11.1% to 26.1%], using a modified intention-to-treat (mITT) analysis. Per-protocol analysis showed similar results: 74.4% compared with 92.3%, an adjusted difference of 18.7% (90% CI 9.8% to 27.6%). The rate of related severe, life-threatening and fatal events at 52 weeks was 18.2% for those started on doxycycline and 36.6% for those started on prednisolone (mITT analysis), an adjusted difference of 19.0% (95% CI 7.9% to 30.1%; p = 0.001) in favour of doxycycline. Secondary outcomes showed consistent findings. There was no significant difference in costs or QALYs per patient at 1 year between doxycycline-initiated therapy and prednisolone-initiated therapy (incremental cost of doxycycline-initiated therapy £959, 95% CI -£24 to £1941; incremental QALYs of doxycycline-initiated therapy -0.024, 95% CI -0.088 to 0.041). Using a willingness-to-pay criterion of < £20,000 per QALY gained, the net monetary benefit associated with doxycycline-initiated therapy was negative but imprecise (-£1432, 95% CI -£3094 to £230). Conclusions: A strategy of starting BP patients on doxycycline is non-inferior to standard treatment with oral prednisolone for short-term blister control and considerably safer in the long term. The limitations of the trial were the wide non-inferiority margin, the moderate dropout rate and that serious adverse event collection was unblinded. Future work might include inducing remission with topical or oral corticosteroids and then randomising to doxycycline or prednisolone for maintenance.