Background: Psoriasis is an immune-mediated disease for which some people have a genetic predisposition. The condition manifests in inflammatory effects on either the skin or joints, or both, and it has a major impact on quality of life. Although there is currently no cure for psoriasis, various treatment strategies allow sustained control of disease signs and symptoms. Several randomised controlled trials (RCTs) have compared the efficacy of the different systemic treatments in psoriasis against placebo. However, the relative benefit of these treatments remains unclear due to the limited number of trials comparing them directly head to head, which is why we chose to conduct a network meta-analysis.

Objectives: To compare the efficacy and safety of conventional systemic agents (acitretin, ciclosporin, fumaric acid esters, methotrexate), small molecules (apremilast, tofacitinib, ponesimod), anti-TNF alpha (etanercept, infliximab, adalimumab, certolizumab), anti-IL12/23 (ustekinumab), anti-IL17 (secukinumab, ixekizumab, brodalumab), anti-IL23 (guselkumab, tildrakizumab), and other biologics (alefacept, itolizumab) for patients with moderate to severe psoriasis and to provide a ranking of these treatments according to their efficacy and safety.

Search methods: We searched the following databases to December 2016: the Cochrane Skin Specialised Register, the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, and LILACS. We also searched five trials registers and the U.S. Food and Drug Administration (FDA) and European Medicines Agency (EMA) reports. We checked the reference lists of included and excluded studies for further references to relevant RCTs. We searched the trial results databases of a number of pharmaceutical companies and handsearched the conference proceedings of a number of dermatology meetings.

Selection criteria: Randomised controlled trials (RCTs) of systemic and biological treatments in adults (over 18 years of age) with moderate to severe plaque psoriasis or psoriatic arthritis whose skin had been clinically diagnosed with moderate to severe psoriasis, at any stage of treatment, in comparison to placebo or another active agent.

Data collection and analysis: Three groups of two review authors independently undertook study selection, data extraction, 'Risk of bias' assessment, and analyses. We synthesised the data using pair-wise and network meta-analysis (NMA) to compare the treatments of interest and rank them according to their effectiveness (as measured by the Psoriasis Area and Severity Index score (PASI) 90) and acceptability (the inverse of serious adverse effects). We assessed the certainty of the body of evidence from the NMA for the two primary outcomes, according to GRADE; we evaluated evidence as either very low, low, moderate, or high. We contacted study authors when data were unclear or missing.

Main results: We included 109 studies in our review (39,882 randomised participants, 68% men, all recruited from a hospital). The overall average age was 44 years; the overall mean PASI score at baseline was 20 (range: 9.5 to 39). Most of these studies were placebo controlled (67%), 23% were head-to-head studies, and 10% were multi-armed studies with both an active comparator and placebo. We have assessed all treatments listed in the objectives (19 in total). In all, 86 trials were multicentric trials (two to 231 centres). All of the trials included in this review were limited to the induction phase (assessment at less than 24 weeks after randomisation); in fact, all trials included in the network meta-analysis were measured between 12 and 16 weeks after randomisation. We assessed the majority of studies (48/109) as being at high risk of bias; 38 were assessed as at an unclear risk, and 23, low risk.Network meta-analysis at class level showed that all of the interventions (conventional systemic agents, small molecules, and biological treatments) were significantly more effective than placebo in terms of reaching PASI 90.In terms of reaching PASI 90, the biologic treatments anti-IL17, anti-IL12/23, anti-IL23, and anti-TNF alpha were significantly more effective than the small molecules and the conventional systemic agents. Small molecules were associated with a higher chance of reaching PASI 90 compared to conventional systemic agents.At drug level, in terms of reaching PASI 90, all of the anti-IL17 agents and guselkumab (an anti-IL23 drug) were significantly more effective than the anti-TNF alpha agents infliximab, adalimumab, and etanercept, but not certolizumab. Ustekinumab was superior to etanercept. No clear difference was shown between infliximab, adalimumab, and etanercept. Only one trial assessed the efficacy of infliximab in this network; thus, these results have to be interpreted with caution. Tofacitinib was significantly superior to methotrexate, and no clear difference was shown between any of the other small molecules versus conventional treatments.Network meta-analysis also showed that ixekizumab, secukinumab, brodalumab, guselkumab, certolizumab, and ustekinumab outperformed other drugs when compared to placebo in terms of reaching PASI 90: the most effective drug was ixekizumab (risk ratio (RR) 32.45, 95% confidence interval (CI) 23.61 to 44.60; Surface Under the Cumulative Ranking (SUCRA) = 94.3; high-certainty evidence), followed by secukinumab (RR 26.55, 95% CI 20.32 to 34.69; SUCRA = 86.5; high-certainty evidence), brodalumab (RR 25.45, 95% CI 18.74 to 34.57; SUCRA = 84.3; moderate-certainty evidence), guselkumab (RR 21.03, 95% CI 14.56 to 30.38; SUCRA = 77; moderate-certainty evidence), certolizumab (RR 24.58, 95% CI 3.46 to 174.73; SUCRA = 75.7; moderate-certainty evidence), and ustekinumab (RR 19.91, 95% CI 15.11 to 26.23; SUCRA = 72.6; high-certainty evidence).We found no significant difference between all of the interventions and the placebo regarding the risk of serious adverse effects (SAEs): the relative ranking strongly suggested that methotrexate was associated with the best safety profile regarding all of the SAEs (RR 0.23, 95% CI 0.05 to 0.99; SUCRA = 90.7; moderate-certainty evidence), followed by ciclosporin (RR 0.23, 95% CI 0.01 to 5.10; SUCRA = 78.2; very low-certainty evidence), certolizumab (RR 0.49, 95% CI 0.10 to 2.36; SUCRA = 70.9; moderate-certainty evidence), infliximab (RR 0.56, 95% CI 0.10 to 3.00; SUCRA = 64.4; very low-certainty evidence), alefacept (RR 0.72, 95% CI 0.34 to 1.55; SUCRA = 62.6; low-certainty evidence), and fumaric acid esters (RR 0.77, 95% CI 0.30 to 1.99; SUCRA = 57.7; very low-certainty evidence). Major adverse cardiac events, serious infections, or malignancies were reported in both the placebo and intervention groups. Nevertheless, the SAEs analyses were based on a very low number of events with low to very low certainty for just over half of the treatment estimates in total, moderate for the others. Thus, the results have to be considered with caution.Considering both efficacy (PASI 90 outcome) and acceptability (SAEs outcome), highly effective treatments also had more SAEs compared to the other treatments, and ustekinumab, infliximab, and certolizumab appeared to have the better trade-off between efficacy and acceptability.Regarding the other efficacy outcomes, PASI 75 and Physician Global Assessment (PGA) 0/1, the results were very similar to the results for PASI 90.Information on quality of life was often poorly reported and was absent for a third of the interventions.

Authors' conclusions: Our review shows that compared to placebo, the biologics ixekizumab, secukinumab, brodalumab, guselkumab, certolizumab, and ustekinumab are the best choices for achieving PASI 90 in people with moderate to severe psoriasis on the basis of moderate- to high-certainty evidence. At class level, the biologic treatments anti-IL17, anti-IL12/23, anti-IL23, and anti-TNF alpha were significantly more effective than the small molecules and the conventional systemic agents, too. This NMA evidence is limited to induction therapy (outcomes were measured between 12 to 16 weeks after randomisation) and is not sufficiently relevant for a chronic disease. Moreover, low numbers of studies were found for some of the interventions, and the young age (mean age of 44 years) and high level of disease severity (PASI 20 at baseline) may not be typical of patients seen in daily clinical practice.Another major concern is that short-term trials provide scanty and sometimes poorly reported safety data and thus do not provide useful evidence to create a reliable risk profile of treatments. Indeed, we found no significant difference in the assessed interventions and placebo in terms of SAEs. Methotrexate appeared to have the best safety profile, but as the evidence was of very low to moderate quality, we cannot be sure of the ranking. In order to provide long-term information on the safety of the treatments included in this review, it will be necessary to evaluate non-randomised studies and postmarketing reports released from regulatory agencies as well.In terms of future research, randomised trials comparing directly active agents are necessary once high-quality evidence of benefit against placebo is established, including head-to-head trials amongst and between conventional systemic and small molecules, and between biological agents (anti-IL17 versus anti-IL23, anti-IL23 versus anti-IL12/23, anti-TNF alpha versus anti-IL12/23). Future trials should also undertake systematic subgroup analyses (e.g. assessing biological-naïve patients, baseline psoriasis severity, presence of psoriatic arthritis, etc.). Finally, outcome measure harmonisation is needed in psoriasis trials, and researchers should look at the medium- and long-term benefit and safety of the interventions and the comparative safety of different agents.