Network meta-analyses in psoriasis: overview and critical discussion

Network meta-analyses (NMAs) increasingly assist in treatment decisions in disease areas such as psoriasis, where data from multiple clinical trials (CTs) on a growing number of different drugs become available. This study aimed to characterize NMAs published in psoriasis. A systematic literature search in PubMed was conducted using a structured search protocol based on the PRISMA criteria. Twenty-seven NMAs were identified, including an average of 43 CTs per NMA. Only eight of 27 NMAs (29.6%) were documented in the PROSPERO registry and only 17 (63%) reported following the PRISMA criteria. The mean number of patients per NMA was 19 624 (range: 6113-51 749). Across all NMAs, the drugs most frequently included were ustekinumab (n = 27 NMAs), followed by adalimumab (n = 25), infliximab and etanercept (n = 24 each). In all n = 27 NMAs, placebo comparisons and in n = 25, comparisons with active controls were used for bridging. Effect estimates were performed in all cases, SUCRA in 14. Most frequently used outcomes were Psoriasis Area and Severity Index (PASI) 75 (n = 25) and PASI 90 (n = 24), and Dermatology Life Quality Index (n = 10). NMAs mostly measured induction efficacy (weeks 10-16, n = 25) but rarely long-term outcomes (weeks 48-56, n = 4). Sensitivity analyses were performed in n = 17 (63%) of the studies. Main results varied considerably between studies and depended on the year of publication and thus the number of available drugs and studies. However, the concordance between NMA efficacy rankings based on PASI 75 was high. Although a large number of NMAs have been published on psoriasis showing highly comparable results on efficacy, no sufficient information on the quality criteria was reported, and PROSPERO registry criteria were not followed. This argues in favour of greater standardization of NMA methodology and reporting.