Background Secukinumab (SEC) is approved for the treatment of active PsA in adults who have responded inadequately to DMARD therapy. The FUTURE 1 and 2 RCTs which included biologic-naïve and biologic-experienced patients, demonstrated superiority of SEC 150 mg and 300 mg over placebo. As no direct trial comparisons are yet available there is a need for comparative effectiveness assessment.
Objectives Comparative assessment of SEC via network meta-analysis (NMA) vs. licensed biologics and apremilast in adults with active PsA who had failed conventional DMARD therapy.
Methods A systematic review of the literature (November 2015) identified 20 relevant RCTs. NMAs were conducted following health technology assessment guidance using Bayesian fixed-effects models. PASI and ACR responses and their credible intervals (CrIs), which can be interpreted similarly to CIs, were modelled using a conditional binomial likelihood NMA with probit link function. The results of pairwise comparisons are presented for SEC using relative risks (RRs) [95%CrI]. Non-responder imputation data were used for all comparators except golimumab QM (GOL) and infliximab 5 mg/kg Q2M (INF) for which only last observation carried forward data were available. Analyses were performed in biologic-naïve, biologic-experienced, and mixed populations at week 16 for ACR and at weeks 12–16 for PASI.
Results At week 16 in the biologic-naïve population, ACR20 responses were statistically significantly higher for SEC 150 mg (61%) and 300 mg (58%) than for ustekinumab QM (UST) 45 mg (RR 1.95 [1.29–2.86] and RR 1.84 [1.21–2.74], respectively), and for SEC 150 mg vs UST 90 mg (RR 1.52 [1.03–2.16]). ACR20 responses were statistically significantly higher for SEC 150 and 300 mg vs. apremilast BID (APR) 20 mg (RR 2.51 [1.52–4.29] and RR 2.37 [1.39–4.12] respectively), and for SEC 150 mg vs. APR 30 mg (RR 1.72 [1.09–2.71]). Comparisons between SEC and anti-TNFs did not show statisticall significance. For ACR50 and 70, both SEC doses showed the same pattern of significance as for ACR20. Sensitivity analyses using the mixed population were in general agreement with these observations.
For PASI50, 75, and 90 at weeks 12–16 in the biologic-naïve population, only a small evidence network with SEC, GOL 50/100mg, adalimumab 40 mg (ADA) and INF was available and there were no statistically significant differences between SEC and these comparators. In the mixed population, SEC was statistically significantly superior to ADA, APR 20/30 mg, certolizumab pegol (CZP) 200/400 mg, etanercept (ETN) 25/50 mg BIW and 50 mg QW, and (SEC 300 mg only) GOL 50 mg.
Conclusions SEC shows statistically significantly higher results vs. UST and APR in ACR20, 50 and 70, with at least comparable response rates across all ACR and PASI endpoints vs. all therapies included in this NMA. For PASI50, 75, and 90 responses, SEC had high response rates with statistically significant results vs. ADA, APR, CZP, ETN, and (SEC 300 mg only) GOL 50 mg.
Disclosure of Interest I. McInnes Consultant for: Novartis, Janssen, Abbvie, Pfizer and UCB, P. Nash Grant/research support from: Novartis, Consultant for: Novartis, C. Ritchlin Grant/research support from: Abbvie, Amgen, UCB, Consultant for: Amgen, Abbvie, Jannsen, Novartis, Sun Pharma, Sanofi, UCB, Boeringer Ingelheim, H. Thom Consultant for: Novartis Pharma AG and for a short time in 2015 for Eli Lilly, S. Kanters Consultant for: Novartis, E. Palaka Employee of: Novartis employee, K. Gandhi Shareholder of: Novartis, Employee of: Novartis, S. Mpofu Shareholder of: Novartis, Employee of: Novartis, S. Jugl Shareholder of: Novartis, Employee of: Novartis