Placebo Response in Moderate-to-Severe Psoriasis: Prevalence Meta-Analysis of Randomized Controlled Trials

Despite the increase in available treatments and their effectiveness, placebo use has not decreased over time in psoriasis randomized controlled trials (RCTs) ( Afach et al., 2021b Afach S. Evrenoglou T. Oubaya N. Le Cleach L. Sbidian E. Most randomized controlled trials for psoriasis used placebo comparators despite the availability of effective treatments. J Clin Epidemiol. 2021; 133: 72-79 Abstract Full Text Full Text PDF PubMed Scopus (2) Google Scholar ). Furthermore, more patients in the placebo group experienced a serious adverse event owing to psoriasis worsening (n = 23, 20%) than in the treatment group (n = 17, 5%) ( Afach et al., 2021a Afach S. Chaimani A. Evrenoglou T. Penso L. Brouste E. Sbidian E. et al. Meta-analysis results do not reflect the real safety of biologics in psoriasis. Br J Dermatol. 2021; 184: 415-424 Crossref PubMed Scopus (15) Google Scholar ). These results raise the ethical question of using a placebo group in future psoriasis trials. According to the European Medicines Agency, variability in placebo response is one reason for using a placebo as a comparator in RCTs evaluating new treatments for plaque psoriasis ( Committee for Medicinal Products for Human Use, 2004 Committee for Medicinal Products for Human UseEuropean Medicines Agency. Guideline on clinical investigation of medicinal products indicated for the treatment of psoriasis. https://www.ema.europa.eu/en/clinical-investigation-medicinal-products-indicated-treatment-psoriasisDate: 2004 Date accessed: July 15, 2020 Google Scholar ). The elements identified as responsible for the variability in response are the variability of the placebo effect and the natural disease course. However, the variability of lesions in moderate-to-severe psoriasis over time may be low. Therefore, using a placebo group in RCTs does not seem justified ( Papakostas and Daras, 2001 Papakostas Y.G. Daras M.D. Placebos, placebo effect, and the response to the healing situation: the evolution of a concept. Epilepsia. 2001; 42: 1614-1625 Crossref PubMed Scopus (61) Google Scholar ; Spuls et al., 2004 Spuls P.I. Witkamp L. Bossuyt P.M.M. Bos J.D. The course of chronic plaque-type psoriasis in placebo groups of randomized controlled studies. Arch Dermatol. 2004; 140 (discussion 344): 338-344 Crossref PubMed Scopus (14) Google Scholar ). To explore the reality of this statement, we aimed to (i) assess the therapeutic response in the placebo group from RCTs of an existing Cochrane review and network meta-analysis ( Sbidian et al., 2022 Sbidian E. Chaimani A. Garcia-Doval I. Doney G. Dressler C. Hua C. et al. Systemic pharmacological treatments for chronic plaque psoriasis: a network meta-analysis. Cochrane Database Syst Rev. 2022; 5 (CD011535) PubMed Google Scholar ) and (ii) identify factors associated with variability in response in the placebo group if any. We selected all RCTs with a placebo group from the latest version of the living Cochrane review (until October 2021) ( Sbidian et al., 2022 Sbidian E. Chaimani A. Garcia-Doval I. Doney G. Dressler C. Hua C. et al. Systemic pharmacological treatments for chronic plaque psoriasis: a network meta-analysis. Cochrane Database Syst Rev. 2022; 5 (CD011535) PubMed Google Scholar ), including adults (>18 years) with moderate-to-severe psoriasis and assessing the following systemic treatments: (i) nonbiologic agents (fumaric acid esters, acitretin, cyclosporine, methotrexate), (ii) small molecules (apremilast, tofacitinib, deucravacitinib), (iii) anti–TNF-α agents (infliximab, etanercept, adalimumab, certolizumab), (iv) anti–IL-12/23 agent (ustekinumab), (v) anti–IL-17 agents (secukinumab, brodalumab, ixekizumab, bimekizumab), and (vi) anti–IL-23 agents (tildrakizumab, guselkumab, risankizumab, mirikizumab). The primary outcomes were the proportion of randomized patients reaching PASI-50/75/90 responses in the placebo group during the induction phase (<24 weeks after randomization). We performed prevalence meta-analyses. We used a random-effects model with the DerSimonian and Laird approach to estimate pooled proportions of PASI-50/75/90 responses in the placebo group with their 95% confidence intervals ( Nyaga et al., 2014 Nyaga V.N. Arbyn M. Aerts M. Metaprop: a Stata command to perform meta-analysis of binomial data. Arch Public Health. 2014; 72: 39 Crossref PubMed Scopus (1254) Google Scholar ). Heterogeneity was assessed with I2 statistic, with 0–40% considered not important, 30–60% possibly moderate heterogeneity, 50–90% possibly substantial heterogeneity, and 75–100% considerable heterogeneity ( Higgins and Thompson, 2002 Higgins J.P.T. Thompson S.G. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002; 21: 1539-1558 Crossref PubMed Scopus (22314) Google Scholar ). In the presence of heterogeneity, according to the categories used for I2 values, we first conducted subgroup analyses, then univariable and multivariable random-effects meta-regressions ( Thompson and Higgins, 2002 Thompson S.G. Higgins J.P.T. How should meta-regression analyses be undertaken and interpreted?. Stat Med. 2002; 21: 1559-1573 Crossref PubMed Scopus (2144) Google Scholar ). The choice of variables used for subgroup analyses is listed in the tables. We performed sensitivity analyses using an alternative statistical model (i.e., binomial-normal) ( Stijnen et al., 2010 Stijnen T. Hamza T.H. Özdemir P. Random effects meta-analysis of event outcome in the framework of the generalized linear mixed model with applications in sparse data. Stat Med. 2010; 29: 3046-3067 Crossref PubMed Scopus (400) Google Scholar ).