Multistate modelling of probability of on-treatment clinical response and time remaining in response in patients with moderate-to-severe psoriasis treated with brodalumab or ustekinumab in the AMAGINE-2 and -3 studies

BACKGROUND: Relative changes in Psoriasis Area and Severity Index (PASI) are used as outcomes in psoriasis clinical trials but are limited when analysing long-term data and in routine practice. Absolute PASI may be more clinically useful.

OBJECTIVES: To develop and implement a methodology for assessing the probability of achieving and maintaining a “response” in patients with psoriasis, defined using absolute PASI.

MATERIALS & METHODS: This analysis included pooled data from the Phase III AMAGINE-2 and -3 trials. Absolute PASI was described using all available data. Multistate modelling was used to compare the probabilities of achieving (absolute PASI = 0) and maintaining (absolute PASI ≤2) a response, and the time in the response state, in patients receiving brodalumab vs ustekinumab.

RESULTS: Higher proportions of patients achieved lower absolute PASI over 52 weeks with brodalumab vs ustekinumab. The probability of achieving the response state was greater with brodalumab vs ustekinumab over 52 weeks (hazard ratio: 1.96; 95% confidence interval [CI]: 1.66–2.31, p<0.001). At Week 52, there was a higher probability of being in response with brodalumab vs ustekinumab (81% [95% CI: 74–89%] vs 60% [95% CI: 54–67%], respectively). Mean time in response was longer with brodalumab (215 days; 95% CI: 197-233) vs ustekinumab (145 days; 95% CI: 130–160); a difference of 70 days (95% CI: 46–94; p<0.001).

CONCLUSION: Using a novel multistate modelling approach based on absolute PASI, we found that patients had a greater probability of achieving and maintaining a response with brodalumab vs ustekinumab.