Dermatol Ther (Heidelb) (2022) 12:2867–2870 https://doi.org/10.1007/s13555-022-00812-z LETTER A Response to: Letter to the Editor Regarding Long- Term Benefit–Risk Profiles of Treatments for Moderate- to-Severe Plaque Psoriasis: A Network Meta-analysis April W. Armstrong . Ahmed M. Soliman . Keith A. Betts . Yan Wang . Yawen Gao . Vassilis Stakias . Luis Puig Received: August 24, 2022 / Accepted: September 5, 2022 / Published online: October 21, 2022 Ó The Author(s) 2022 We thank the journal for the opportunity to term benefit–risk profile is well supported by the respond to the letter to the editor regarding analyses and the results. ‘‘Long-Term Benefit–Risk Profiles of Treatments for Pettitt et al. suggest that the analysis based Moderate-to-Severe Plaque Psoriasis: A Network on the Surface Under the Cumulative Ranking Meta-Analysis’’ [1]. The analyses reported in (SUCRA) score does not support the benefit–risk Armstrong et al. [2] and Shear et al. [3] were profile summarized in our conclusion. However, conducted following a rigorous and well-ac- this approach is well in line with prevailing cepted methodology. Therefore, the authors methodological conventions applied in other believe that the conclusion that risankizumab recently published network meta-analyses was associated with the most favorable long- (NMAs) that also described the benefit–risk profiles of psoriasis treatments [4, 5]. For This reply refers to the comment available online at example, the NMAs by the British Association of https://doi.org/10.1007/s13555-022-00809-8. Dermatologists used the bidimensional SUCRA plot to cluster treatments into high vs. low A. W. Armstrong (&) efficacy and high vs. low tolerability [4]; similar Department of Dermatology, Keck School of interpretations were given to the bidimensional Medicine, University of Southern California, Health Sciences Campus, HC4 2000 1450 San Pablo, Los SUCRA plots in the Cochrane systematic review Angeles, CA 90033, USA of systemic treatments for psoriasis [5]. These e-mail: armstrongpublication@gmail.com examples used SUCRA for efficacy and safety A. M. Soliman outcomes to characterize the benefit–risk pro- AbbVie, Inc., 1 N. Waukegan Road, North Chicago, files of psoriasis treatments. Additionally, the IL 60064, USA novel approach developed by Mavridis et al. (a method suggested by Pettitt et al.) is an exten- K. A. Betts  Y. Wang  Y. Gao Analysis Group, Inc., 333 South Hope Street 27th sion of the bidimensional SUCRA plots in the Floor, Los Angeles, CA 90071, USA present study [6]. Moreover, Salanti et al. [7] examined multiple ranking metrics for NMAs V. Stakias AbbVie, Inc., 26525 N. Riverwoods Blvd, Mettawa, and concluded that the SUCRA score is the IL 60045, USA proper metric to address the treatment hierar- chy question of ‘‘which treatment has the largest L. Puig Department of Dermatology, IIB SANTPAU, fraction of competitors that it beats’’. To summa- Hospital de la Santa Creu i Sant Pau, Universitat rize, our methodology to compare the bene- Autònoma de Barcelona, Carrer de Sant Quintı́, 89, fit–risk profile of treatments using SUCRA is well 08041 Barcelona, Spain 2868 Dermatol Ther (Heidelb) (2022) 12:2867–2870 established and is in line with prevailing also has a statistically significantly lower rate of methodological conventions [4, 5, 7]. any AE compared with secukinumab, ustek- Moreover, although SUCRA is agnostic to the inumab, and bimekizumab by week 48–56. magnitude of treatment effects, Armstrong et al. These results are consistent with and augment [2] also identified and reported many clinically the superiority of risankizumab over other meaningful, statistically significant contrasts treatments based on SUCRA. Additionally, based on the rates of Psoriasis Area and Severity while Pettitt et al. mention that the treatment- Index (PASI) response and any adverse event specific PASI response rates and safety event (AE) across treatments. Both outcomes are well rates have overlapped credible intervals (CrIs), established, commonly used for assessing the these estimates are correlated, and thus over- efficacy and safety of psoriasis treatments. lapped CrIs do not necessarily translate to high Detailed results have been reported in the uncertainty for their joint distribution or sta- original publication. Briefly, as summarized in tistically insignificant contrasts between treat- Table 1, risankizumab is significantly superior to ments. A formal assessment based on odds all other included treatments in either or both ratios had been reported in Armstrong et al. of long-term efficacy and long-term safety. (2022), which informs whether the contrast Specifically, compared with brodalumab, between a pair of treatments is statistically sig- guselkumab, ixekizumab, secukinumab, ustek- nificant [2]. inumab, adalimumab, and etanercept, risanki- Pettitt et al. further suggest that random-ef- zumab has statistically significantly higher PASI fects models or meta-regression models should response rates by week 48–56. Risankizumab be used instead. At the same time, they admit that such models are not possible for the sparse networks in this study. Because applying a ran- Table 1 Summary of statistical comparisons (pairwise odds ratios) of risankizumab versus other treatments in dom-effects model to sparse networks would terms of PASI response rates and any AE rates in the long generate unreasonably wide CrIs, using the term (Tables 2 and 3 and Supplementary Table 2 in random-effects model in this study would not Armstrong et al. [2]) be in line with the recommendation by the Grading of Recommendations Assessment, Treatments PASI response rates (week 48–56) Development and Evaluation (GRADE) Work- Risankizumab Risankizumab ing Group: ‘‘reviewers conducting NMAs who significantly statistically believe that the common between-study hetero- superior to comparable to geneity across comparisons is unrealistic, or that it cannot be estimated reliably in their sparse net- Any AE (week 48–56) works—and that it is causing some network esti- Risankizumab SEC, UST BKZ mates to have CIs much wider than appears significantly sensible—may reasonably assume that such safer than between-study heterogeneity across comparisons is zero by conducting the NMA using fixed rather than Risankizumab GUS, ADA, IXE – random effects models—if, that is, results make statistically more intuitive sense than those of random effect comparable to models’’ [8]. Additionally, for sparse networks, Other BRO, ETA – meta-regression models would produce statisti- cally insignificant coefficients for adjusted treatments covariates, which would then not support such without safety adjustment. Armstrong et al. (2022) had trans- data parently acknowledged that the sparsity of the ADA adalimumab, AE adverse event, BKZ bimekizumab, networks precludes the use of random-effects or BRO brodalumab, ETA etanercept, GUS guselkumab, IXE meta-regression models, making the adjust- ixekizumab, PASI Psoriasis Area and Severity Index, SEC ment for heterogeneities infeasible [2]. secukinumab, UST ustekinumab Dermatol Ther (Heidelb) (2022) 12:2867–2870 2869 We agree with Pettitt et al. that an open indicate if changes were made. The images or dialogue about the standards for conducting other third party material in this article are and interpreting NMA would be useful. Overall, included in the article’s Creative Commons the interpretations of meta-analyses by Arm- licence, unless indicated otherwise in a credit strong et al. [2] and Shear et al. [3] are based on line to the material. If material is not included rigorous and accepted methodology also used in the article’s Creative Commons licence and by professional dermatology societies such as your intended use is not permitted by statutory the British Association of Dermatologists and regulation or exceeds the permitted use, you the Cochrane group. will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by- ACKNOWLEDGEMENTS nc/4.0/. Funding. Sponsorship for this study was funded by AbbVie Inc. AbbVie participated in REFERENCES interpretation of data, review, and approval of the original manuscript, as well as this response 1. Pettitt D, Plotnick M, Gagne J. Letter to the editor concerning the article: ‘‘Long-term benefit–risk pro- letter. files of treatments for moderate-to-severe plaque psoriasis: a network meta-analysis’’. Dermatol Ther. Author contributions. All named authors 2022;85:305. contributed to this publication. 2. Armstrong AW, Soliman AM, Betts KA, Wang Y, Gao Y, Stakias V, et al. Long-term benefit–risk profiles of Disclosures. April W. Armstrong serves as treatments for moderate-to-severe plaque psoriasis: a investigator and/or scientific advisor to AbbVie, network meta-analysis. Dermatol Ther. 2022;12(1): BMS, Incyte, Leo, UCB, Janssen, Lilly, Novartis, 167–84. Ortho Dermatologics, Sun, Dermavant, Der- 3. Shear NH, Betts KA, Soliman AM, Joshi A, Wang Y, mira, Sanofi, Regeneron, Pfizer, and Modmed. Zhao J, et al. Comparative safety and benefit–risk Ahmed M. Soliman and Vassilis Stakias are profile of biologics and oral treatment for moderate- employees of AbbVie and own AbbVie stock. to-severe plaque psoriasis: a network meta-analysis of Keith A. Betts, Yan Wang, and Yawen Gao are clinical trial data. J Am Acad Dermatol. 2021;85(3): 572–81. employed by Analysis Group, Inc., which received payment from AbbVie Inc. for partici- 4. Mahil SK, Ezejimofor MC, Exton LS, Manounah L, pation in this research. Luis Puig has served as Burden AD, Coates LC, et al. Comparing the efficacy investigator and/or consultant or paid speaker and tolerability of biologic therapies in psoriasis: an updated network meta-analysis. Br J Dermatol. to AbbVie, Almirall, Amgen, Biogen, Boehringer 2020;183(4):638–49. Ingelheim, Celgene, Janssen, Leo, Lilly, Merck, Novartis, Pfizer, and UCB. 5. Sbidian E, Chaimani A, Garcia-Doval I, Doney L, Dressler C, Hua C, et al. Systemic pharmacological Compliance with ethics guidelines. This treatments for chronic plaque psoriasis: a network meta-analysis. Cochrane Database Systemat Rev. article is based on previously conducted studies. 2022;5:CD011535. Open Access. This article is licensed under a 6. Mavridis D, Porcher R, Nikolakopoulou A, Salanti G, Creative Commons Attribution-NonCommer- Ravaud P. Extensions of the probabilistic ranking metrics of competing treatments in network meta- cial 4.0 International License, which permits analysis to reflect clinically important relative differ- any non-commercial use, sharing, adaptation, ences on many outcomes. Biom J. 2020;62(2):375–85. distribution and reproduction in any medium or format, as long as you give appropriate credit 7. Salanti G, Nikolakopoulou A, Efthimiou O, Mavridis D, Egger M, White IR. Introducing the treatment to the original author(s) and the source, provide hierarchy question in network meta-analysis. Am J a link to the Creative Commons licence, and Epidemiol. 2022;191(5):930–8. 2870 Dermatol Ther (Heidelb) (2022) 12:2867–2870 8. Brignardello-Petersen R, Murad MH, Walter SD, meta-analysis: avoiding spurious judgments of impre- McLeod S, Carrasco-Labra A, Rochwerg B, et al. cision in sparse networks. J Clin Epidemiol. 2019;105: GRADE approach to rate the certainty from a network 60–7.