Meta-analysis results do not reflect the real safety of biologics in psoriasis

摘要

Background In reported systematic reviews and meta-analyses of randomized controlled trials (RCTs) assessing treatments for psoriasis, the proportion of serious adverse events (SAEs) did not differ between treatments and placebo. Including cases of psoriasis worsening asSAEs may explain the lack of difference. Objectives This systematic review and meta-analysis aimed to explore this possibility. Methods Among the 140RCTs included in the Living Network Cochrane Review (last search on 8 May 2019), we selected those comparing a biologic treatment against placebo. The primary outcome was the numbers ofSAEs in the treatment and placebo arms after excluding cases of psoriasis worsening. Secondary outcomes were the number of adverse events (AEs) of special interest. The trial was registered onPROSPERO(CRD42019124495). Results We analysed 51RCTs. Of these, 21 included at least one anti-tumour necrosis factor (TNF)-alpha arm, 15 one anti-interleukin (IL)-17 arm, 11 one anti-IL-23 arm and nine one anti-IL-12/23 arm. With cases of psoriasis worsening included, the risk of occurrence ofSAEs between biologic treatments and placebo did not differ: risk ratio (RR) 1 center dot 09, 95% confidence interval (CI) 0 center dot 88-1 center dot 36. After excluding cases of psoriasis worsening, theRRbecame significant (RR1 center dot 30, 95%CI1 center dot 02-1 center dot 65). By drug class, theRRs were for anti-TNF-alpha, 1 center dot 68 (95%CI1 center dot 11-2 center dot 54; no missing data); anti-IL-17, 1 center dot 28 (95%CI0 center dot 88-1 center dot 85; no missing data); anti-IL-23, 0 center dot 95 (95%CI0 center dot 59-1 center dot 52; no missing data) and anti-IL-12/23, 1 center dot 18 (95%CI0 center dot 72-1 center dot 94; no missing data). We were unable to examine potential differences inAEs of special interest between biologic treatments and placebo arms because of the small number of events. Conclusions On excluding cases of worsening psoriasis, the risk of occurrence ofSAEs is higher in the biologic than in the placebo arm. Given the rare events, we could not highlight whether this higher risk ofSAEs was related toAEs of special interest. Reporting ofSAEs in clinical trials has to be changed to provide more transparency through the separate reporting of disease flares leading to hospital admission and otherSAEs.