Adv Ther (2023) 40:869–886 https://doi.org/10.1007/s12325-022-02379-9 ORIGINAL RESEARCH Comparative Effectiveness of Biologics Across Subgroups of Patients with Moderate-to-Severe Plaque Psoriasis: Results at Week 12 from the PSoHO Study in a Real-World Setting Charles Lynde . Elisabeth Riedl . Julia-Tatjana Maul . Tiago Torres . Andreas Pinter . Gabrielle Fabbrocini . Flavia Daniele . Alan Brnabic . Catherine Reed . Stefan Wilhelm . Thorsten Holzkämper . Christopher Schuster . Luis Puig Received: October 6, 2022 / Accepted: November 11, 2022 / Published online: December 14, 2022 Ó The Author(s) 2022, corrected publication 2023 ABSTRACT study aimed to provide direct comparative effectiveness data at week 12 between anti-in- Introduction: In routine clinical care, impor- terleukin (IL)-17A biologics relative to other tant treatment outcomes among patients with approved biologics for the treatment of PsO moderate-to-severe plaque psoriasis (PsO) have across seven clinically relevant patient sub- been shown to vary according to patient groups in the real-world setting. demographics and disease characteristics. This Methods: From the international, non-inter- ventional Psoriasis Study of Health Outcomes (PSoHO), 1981 patients with moderate-to-severe Supplementary Information The online version PsO were grouped a priori according to seven contains supplementary material available at https:// clinically relevant demographic and disease doi.org/10.1007/s12325-022-02379-9. variables with binary categories, which were sex C. Lynde (&) G. Fabbrocini Lynde Institute for Dermatology, Markham, ON, Section of Dermatology, Department of Clinical Canada Medicine and Surgery, University of Naples Federico e-mail: derma@lynderma.com II, Naples, Italy E. Riedl  C. Schuster F. Daniele Department of Dermatology, Medical University of Medica Dermatologa en Swiss Medical Group, Vienna, Vienna, Austria Buenos Aires, Argentina J.-T. Maul A. Brnabic  C. Reed  S. Wilhelm  T. Holzkämper  Faculty of Medicine, University of Zurich, Zurich, C. Schuster Switzerland Eli Lilly and Company, Indianapolis, IN, USA T. Torres L. Puig Department of Dermatology and Dermatology Department of Dermatology, Hospital de la Santa Research Unit, Centro Hospitalar Universitário do Creu i Sant Pau, Universitat Autònoma de Porto, University of Porto, Porto, Portugal Barcelona, Barcelona, Spain A. Pinter Department of Dermatology, Venereology and Allergology, University Hospital Frankfurt, Frankfurt am Main, Germany 870 Adv Ther (2023) 40:869–886 (male or female), age (\ 65 or C 65 years), body mass index (B 30 or[ 30 kg/m2), race (White or Key Summary Points Asian), PsO disease duration (\ 15 or C 15 years), psoriatic arthritis (PsA) comorbidity Why carry out this study? (present or absent), and prior biologic use (never or C 1). Across these subgroups, effec- Strict inclusion criteria for clinical trials tiveness was compared between the anti-IL-17A can exclude many patients and report an cohort (ixekizumab, secukinumab) versus all averaged treatment effect, and varying other approved biologics and ixekizumab versus demographics and disease characteristics five individual biologics. The proportion of of patient subgroups can potentially patients in each subgroup who achieved 90% modify the effect of a treatment on improvement in Psoriasis Area and Severity clinical outcomes. Index (PASI90) and/or static Physician Global Newer psoriasis therapies have shown Assessment (sPGA) 0/1, PASI100, or PASI90 at robust responses among subgroups of week 12 were assessed. Comparative analyses patients in randomized clinical trials, yet were conducted using frequentist model aver- little is known about the comparative aging (FMA). Missing data were imputed using effectiveness of biologic treatments in non-responder imputation. population subgroups receiving care in Results: Patients in each of the seven sub- the real world. groups achieved similar response rates to those of the overall treatment cohort, apart from What was learned from the study? patients with PsA treated with other biologics The Psoriasis Study of Health Outcomes who had 7–10% lower response rates. Conse- (PSoHO) provides direct comparative quently, patients with comorbid PsA had sig- effectiveness of different biologics across nificantly higher odds of achieving skin seven clinically relevant subgroups of clearance at week 12 with anti-IL-17A biologics patients with moderate-to-severe psoriasis compared to other biologics. Patients in all in a real-world setting. subgroups had significantly higher odds of achieving PASI90 and/or sPGA (0,1), PASI100, Patients in the anti-interleukin-17A cohort and PASI90 in the anti-IL-17A cohort relative to versus the other biologics cohort had the other biologics cohort, except for the Asian significantly higher odds of achieving all subgroup. No sex- or age-specific differences in outcomes across all subgroups, except the treatment effectiveness after 12 weeks were Asian subgroup. identified, neither between the treatment The presence of comorbid psoriatic cohorts nor between the individual treatment arthritis may affect a patient’s clinical comparisons. response to some treatments evidenced by Conclusions: Despite relative consistency of the significantly higher odds of achieving comparative treatment effectiveness across skin clearance at week 12 with anti- subgroups, the presence of comorbid PsA may interleukin-17A biologics compared to affect a patient’s clinical response to some other biologics. treatments. Keywords: Psoriasis; Biologic; Demographic; Subgroup; Comorbidity; Treatment; Effectiveness; Real-world INTRODUCTION Although randomized clinical trials (RCTs) are considered the gold standard for evaluating the efficacy and safety of potential therapies, they do Adv Ther (2023) 40:869–886 871 not necessarily reflect the heterogeneous patient clinical characteristics, as well as all prescribed population and complex conditions encoun- biologics have been previously published [17]. tered in daily clinical practice [1, 2]. This is of Briefly, the PSoHO study enrolled 1981 adult particular relevance in routine clinical care as patients from 23 participating countries and important outcomes among patients with mod- with a confirmed diagnosis (at least 6 months erate-to-severe plaque psoriasis (PsO) have been prior to baseline) of moderate-to-severe PsO shown to vary according to patient demograph- who initiated or switched biologic treatment ics and disease characteristics [3–5]. Older gen- during routine medical care. Prescribed biolog- eration biologics, in particular, are associated ics were grouped into the anti-IL-17A antibodies with less favorable clinical outcomes among cohort [IXE and secukinumab (SEC)] and a sec- patients with comorbid obesity [3, 6, 7], psoriatic ond cohort of other biologics targeting the IL- arthritis (PsA) [5], failure with a prior biologic 17 receptor A (brodalumab [BROD]), tumor treatment [8, 9], or a longer disease duration necrosis factor (TNF)-a (adalimumab [ADA], [10]. There are also insufficient, often conflict- certolizumab, etanercept, infliximab), IL-12/23 ing, real-world data on the effect on clinical p40 (ustekinumab [UST]), and IL-23 p19 effectiveness of other demographic characteris- (guselkumab [GUS], risankizumab [RIS], and tics, such as sex [11–13], age [13, 14], and race tildrakizumab). The primary endpoint was the [4, 13, 15]. Direct comparative effectiveness of proportion of patients achieving at least a 90% treatments in patient subgroups may therefore improvement in the psoriasis area and severity contribute to the personalization of treatment index score (PASI90) and/or a static Physician selection that accounts for both patient- and Global Assessment score of clear or almost clear disease-related factors and may help to reduce (sPGA 0/1 on a 6-point scale) at week 12. Sec- the number of switches between treatments [16]. ondary outcomes included in this analysis were The Psoriasis Study of Health Outcomes the proportion of patients who achieved (PSoHO) is an ongoing, international, prospec- PASI100 or PASI90 at week 12. tive, non-interventional observational study In this study, patient subgroups were pre- that has been designed to investigate the com- specified a priori in the protocol and defined parative effectiveness of biologic treatments for according to seven clinically relevant demo- patients with moderate-to-severe PsO within a graphic and disease variables with binary cate- real-world setting [17]. The primary PSoHO gories: (i) sex (male or female) [11, 18–20], (ii) week 12 results provided additional evidence age (\ 65 or C 65 years) [13, 14], (iii) body mass that the early onset skin clearance and the high index (BMI [B 30 or [ 30 kg/m2]) [19–21], (iv) efficacy of anti-interleukin (IL)-17A inhibitors race (White or Asian) [4, 22–25], (v) PsO disease observed in RCTs correlated to their effective- duration (\ 15 or C 15 years) [13, 20], (vi) PsA ness in the real-world setting compared to other comorbidity (present or absent) [13, 20, 26], and biologics [17]. We aim to evaluate the real- (vii) prior biologic use (never or C 1) world effectiveness of anti-IL-17A biologics [8, 20, 27, 28]. Across all subgroup categories, versus other biologics across seven clinically pairwise effectiveness comparisons were com- relevant subgroups. We also seek to inform on pleted for the anti-IL-17A cohort versus the the pairwise comparative effectiveness of ixek- other biologics cohort and for IXE versus the izumab (IXE) versus five other individual bio- other individual biologics. Since the statistical logics across these patient subgroups. models did not converge for any treatment groups with fewer than 100 patients, pairwise comparisons are only shown for IXE versus SEC, METHODS GUS, RIS, ADA, and UST. For the same rationale, other races were not evaluated in this study. Study Design and Assessments The protocol, amendments, and consent documentation were approved by local institu- Details of the PSoHO study including eligibility tional review boards. The study was registered at criteria, baseline patient demographics and the European Network of Centers for 872 Adv Ther (2023) 40:869–886 Pharmacoepidemiology and Pharmacovigilance RESULTS (ENCEPP24207) [29] and was conducted according to International Conference on Har- Baseline Demographics and Disease monization, Good Clinical Practice guidelines Characteristics and the Declaration of Helsinki. All patients were required to give informed consent for Baseline demographics and disease characteris- participation in the study. tics of the select patient subgroups are provided in Table 1. Of the 1981 patients enrolled, 9.0% Statistical Analysis of patients were 65 years or older, 47.1% of patients had endured PsO for 15 years or longer, Descriptive statistics are reported as mean and and 64.3% of patients were biologic-naı̈ve. At standard deviation or median and quartiles 1 baseline, 39.0% (n = 773) initiated an anti-IL- and 3 (Q1, Q3) for continuous variables, and 17A biologic and 61.0% (n = 1208) received proportions and percentages for categorical other biologics. Patient profiles were compara- variables. Pairwise comparisons of baseline ble between the anti-IL-17A and other biologics demographics between the anti-IL-17A versus cohorts, with exceptions including the propor- the other biologics cohort and IXE versus indi- tion of patients in the two age categories vidual biologics were performed using Fisher’s (C 65 years: 11.5% vs. 7.5%; p = 0.001), or with exact test or chi-square for categorical variables comorbid PsA (29.4% vs. 19.4%; p = 0.001), and analysis of variance (ANOVA), Mood’s respectively. Across individual treatments, the median test or exact P value from median test greatest variation in demographic distribution (Monte Carlo estimate) for continuous vari- was in the age, race, and PsA comorbidity sub- ables. A P value of less than 0.05 was considered groups (Table 1). In patients who received the statistically significant. European Medicines Agency-approved on-label The comparative effectiveness analyses were dosing (1767/1981; 89.2%), results in those performed using a machine learning, data-dri- outcomes studied were comparable (Supple- ven approach, known as frequentist model mentary Material). averaging (FMA), with this method and its application previously described [17, 30, 31]. Comparative Effectiveness of Anti-IL-17A FMA was used for pairwise comparisons Biologics Versus Other Biologics Across between cohorts or treatments within each Patient Subgroups subcategory of the subgroups defined. Com- parative adjusted results are presented as odds Overall, patients in the anti-IL-17A cohort had ratios (OR) with 95% confidence intervals (CI), significantly higher odds of achieving the pri- formed using the 2.5th and 97.5th percentiles mary endpoint (sPGA 0/1 and/or PASI90), derived from 100 bootstrap samples. Statistical PASI100, and PASI90 at week 12 compared to significance is indicated when the CIs do not patients in the other biologics cohort (Fig. 1) cross the null hypotheses (OR = 1). In a few [17]. Patients in each of the seven subgroups cases, no analysis strategies converged, because achieved similar response rates to those of the of zero counts for the outcome in one group, overall treatment cohort, except for patients resulting in no treatment effect estimate. Miss- with PsA treated with other biologics, who had ing data were imputed using non-responder 7–10% lower response rates. Accordingly, imputation. Further details are provided in patients in all subgroups had significantly the Supplementary Material. higher odds of achieving these outcomes in the anti-IL-17A cohort relative to the other biolog- ics cohort, except for the Asian subgroup ORs (primary and PASI90), which did not reach sta- tistical significance. The most pronounced contrast between the cohorts was for patients Table 1 Demographics and disease characteristics of patients with moderate-to-severe plaque psoriasis at baseline Overall Anti-IL-17A Other biologics IXE SEC GUS RIS ADA UST (n = 1981) (n = 773) (n = 1208) (n = 532) (n = 241) (n = 303) (n = 259) (n = 284) (n = 127) Age, years 45.3 (13.6) 46.8 (13.7)* 44.4 (13.5) 47.4 (14.1) 45.4 (12.8) 44.2 (13.2) 44.1 (13.7)à 45.1 (13.0)à 46.4 (14.5) Age, n (%) \ 65 years 1802 (91.0) 684 (88.5)* 1118 (92.5) 460 (86.5) 224 (92.9)à 284 (93.7) 243 (93.8)à 258 (90.8) 114 (89.8) Adv Ther (2023) 40:869–886 C 65 years 179 (9.0) 89 (11.5)* 90 (7.5) 72 (13.5) 17 (7.1)à 19 (6.3) 16 (6.2)à 26 (9.2) 13 (10.2) Sex, n (%) Male 1143 (57.7) 442 (57.2) 701 (58.0) 313 (58.8) 129 (53.5) 179 (59.1) 161 (62.2) 163 (57.4) 77 (60.6) Female 838 (42.3) 331 (42.8) 507 (42.0) 219 (41.2) 112 (46.5) 124 (40.9) 98 (37.8) 121 (42.6) 50 (39.4) Weight, kg 85.0 (21.1) 85.6 (20.8) 84.6 (21.2) 86.3 (20.4) 83.9 (21.6) 84.0 (21.2) 83.8 (22.6) 86.7 (21.3) 82.9 (17.1) a 2 BMI , kg/m 29.0 (6.7) 29.2 (6.6) 28.9 (6.7) 29.4 (6.6) 28.85 (6.5) 29 (6.7) 28.6 (6.9) 29.3 (6.6) 28 (5.6)à BMI, n (%) B 30 kg/m2 1233 (63.3) 468 (61.7) 765 (64.3) 316 (60.5) 152 (64.1) 188 (63.3) 167 (65.0) 179 (63.7) 92 (74.8) 2 [ 30 kg/m 716 (36.7) 291 (38.3) 425 (35.7) 206 (39.5) 85 (35.9) 109 (36.7) 90 (35.0) 102 (36.3) 31 (25.2) Raceb, n (%) White 1441 (72.7) 576 (74.5) 865 (71.6) 394 (74.1) 182 (75.5) 162 (53.5) 169 (65.3)à 248 (87.3) 99 (78.0) Asian 296 (14.9) 103 (13.3) 193 (16.0) 67 (12.6) 36 (14.9) 100 (33.0) 53 (20.5)à 7 (2.5) 8 (6.3)à 873 Table 1 continued 874 Overall Anti-IL-17A Other biologics IXE SEC GUS RIS ADA UST (n = 1981) (n = 773) (n = 1208) (n = 532) (n = 241) (n = 303) (n = 259) (n = 284) (n = 127) Disease duration, 14.0 14.3 13.8 13.9 14.9 14.9 13.7 14.2 12.1 median years (6.8, 23.8) (6.4, 24.2) (7.1, 23.6) (6.7, 25.3) (6.0, 21.8) (7.8, 24.4) (8.2, 23.5) (6.3, 25.0) (6.3, 23.7) (Q1, Q3) Disease duration, n (%) \ 15 years 1047 (52.9) 401 (51.9) 646 (53.5) 279 (52.4) 122 (50.6) 154 (50.8) 144 (55.6) 151 (53.2) 71 (55.9) C 15 years 934 (47.1) 372 (48.1) 562 (46.5) 253 (47.6) 119 (49.4) 149 (49.2) 115 (44.4) 133 (46.8) 56 (44.1) PASI 14.5 (8.6) 14.6 (8.5) 14.5 (8.6) 14.4 (8.5) 15.03 (8.7) 14.6 (9.3) 15.4 (9.8) 13.3 (7.1) 14.4 (7.9) Percentage of BSA 21.3 (17.7) 21.1 (17.5) 21.5 (17.9) 20.6 (17.2) 22.3 (18.1) 21.7 (18.5) 20.6 (18.9) 20.6 (16.6) 22.6 (17.7) sPGA, n (%) Moderate 988 (50.7) 387 (50.7) 601 (50.8) 267 (50.6) 120 (50.8) 143 (47.7) 102 (40.8)à 170 (60.5)à 68 (54.8) Severe 610 (31.3) 242 (31.7) 368 (31.1) 176 (33.3) 66 (28.0) 101 (33.7) 93 (37.2) 69 (24.6)à 37 (29.8) à Very severe 76 (3.9) 34 (4.5) 42 (3.5) 16 (3.0) 18 (7.6) 14 (4.7) 15 (6.0) 5 (1.8) 2 (1.6) DLQIc 12.6 (7.8) 12.9 (7.9) 12.4 (7.8) 12.6 (7.9) 13.5 (7.7) 12.3 (8.1) 11.8 (7.3) 12.9 (7.6) 12.3 (8.0) Current 1.5 (1.8) 1.6 (1.8) 1.4 (1.8) 1.6 (1.8) 1.6 (1.9) 1.5 (1.7) 1.4 (1.8) 1.4 (1.7) 1.7 (2.1) comorbidities reportedd Adv Ther (2023) 40:869–886 Table 1 continued Overall Anti-IL-17A Other biologics IXE SEC GUS RIS ADA UST (n = 1981) (n = 773) (n = 1208) (n = 532) (n = 241) (n = 303) (n = 259) (n = 284) (n = 127) Psoriatic 461 (23.3) 227 (29.4)* 234 (19.4) 161 (30.3) 66 (27.4) 71 (23.4)à 32 (12.4) 64 (22.5)à 19 (15.0) arthritis, n (%)e Any previous 1565 (79.0) 573 (74.2)* 992 (82.1) 393 (74.0) 180 (74.7) 225 (74.3) 199 (76.8) 265 (93.3) 106 (83.5) Adv Ther (2023) 40:869–886 conventional therapy, n (%) Prior treatment 706 (35.7) 291 (37.7) 415 (34.4) 204 (38.4) 87 (36.1) 178 (58.7) 111 (42.9) 25 (8.8)à 35 (27.6)à with biologics, n (%)f All results are expressed as mean (standard deviation) of all available data for that measure, unless otherwise indicated. Anti-IL-17A cohort includes IXE- and SEC- treated patients In cases of n (%) not matching total in group, remainder is attributable to missing data for patients Italic text signifies significant difference (P \ 0.05) between anti-IL-17A cohort vs. the other biologics cohort Bold italic text signifies significant difference (P \ 0.05) vs. IXE ADA adalimumab, BMI body mass index, BSA body surface area, DLQI Dermatology Life Quality Index, GUS guselkumab, IXE ixekizumab, PASI Psoriasis Area and Severity Index, sPGA Static Physician Global Assessment, PASI Psoriasis Area and Severity Index, Q quartile, RIS risankizumab, SEC secukinumab, UST ustekinumab * P \ 0.001 vs. the other biologics cohort ** P \ 0.05 vs. the other biologics cohort P B 0.001 vs. IXE à P \ 0.05 vs. IXE a Information about BMI missing for 32 patients b Some individuals selected more than one race c DLQI was measured on a 0–30 scale d Comorbidities were captured on the basis of a pre-defined list e PsA diagnosis was recorded by the dermatologists on the basis of the medical history and/or information provided by the patient f Information about prior biologic use missing for 1 patient 875 876 Adv Ther (2023) 40:869–886 Fig. 1 Comparative effectiveness in anti-IL-17A versus patient subgroups. Data are non-responder imputation. other biologics cohorts across subgroups of patients with Results are statistically significant if 1 is not covered by the moderate-to-severe plaque psoriasis. Comparative adjusted 95% CI for the odds ratios. For the Asian subgroup, the analysis of primary (sPGA (0/1) and/or PASI90) and lower CI is 0.990 (primary) and 1.048 (PASI100). BMI secondary outcomes, PASI100 and PASI90, actual body mass index, CI confidence interval, IL interleukin, responses rates, and adjusted odds ratios at week 12 for PASI Psoriasis Area and Severity Index, PsA psoriatic anti-IL-17A cohort and other biologics cohort across arthritis, sPGA Static Physician Global Assessment with PsA, whereby the anti-IL-17A cohort had cohort than the SEC cohort across all subgroup 4.0-fold and 3.3-fold greater odds of achieving categories (Fig. 2). Compared to SEC-treated PASI100/90, respectively, versus the other bio- patients, IXE-treated patients had higher odds logics cohort. Moreover, the CIs associated with of achieving the three outcomes, with statistical the OR for PASI100 were non-overlapping significance reached for half of the subgroup between patients with PsA (OR 4.0; 95% CIs 2.5, categories. For the primary endpoint, statisti- 6.5) and patients without PsA (OR 1.8; 95% CIs cally significant differences were not reached in 1.4, 2.2; Fig. 1). Further stratification showed the following subcategories: male, that patients with PsA in either treatment age C 65 years, BMI [ 30, Asian, disease dura- cohort (but not in all individual biologics) had tion C 15 years, PsA comorbidity, or biologic- lower unadjusted response rates when biologic- experienced (Fig. 2). experienced compared to biologic-naı̈ve (Sup- Across all subgroups, GUS-treated patients plementary Material). had lower response rates for the three outcomes versus IXE-treated patients, with significantly Pairwise Comparisons of Ixekizumab higher ORs for all subgroups in the IXE cohort, Versus Five Individual Biologics except for the 65 years or older (PASI100/90), BMI [ 30 (PASI100), or Asian (primary and The unadjusted response rates for the three PASI90; Fig. 3) subgroups. Versus RIS-treated outcomes at week 12 were higher in the IXE patients, IXE-treated patients had higher Adv Ther (2023) 40:869–886 877 Fig. 2 Comparative effectiveness in ixekizumab versus the odds ratios. For the male subgroup, the lower CI is secukinumab cohorts across subgroups of patients with 0.973 (primary) and 1.040 (PASI100); for the female moderate-to-severe plaque psoriasis. Comparative adjusted subgroup, the lower CI is 1.049 (primary); for analysis of primary (sPGA (0/1) and/or PASI90) and the C 15 years subgroup, the lower CI is 0.974 (PASI90). secondary outcomes, PASI100 and PASI90, actual BMI body mass index, CI confidence interval, IXE responses rates, and adjusted odds ratios at week 12 for ixekizumab, PASI Psoriasis Area and Severity Index, PsA ixekizumab cohort and secukinumab cohort across patient psoriatic arthritis, SEC secukinumab, sPGA Static Physi- subgroups. Data are non-responder imputation. Results are cian Global Assessment statistically significant if 1 is not covered by the 95% CI for response rates, except for C 65 years (primary) the overall patient group (OR 4.0 vs. 2.1) and Asian subgroups, and the comparative (Fig. 1). effectiveness reached statistical significance Relative to ADA, the comparative effective- across the subgroup categories, except for male ness of IXE was significantly higher across all (primary, PASI90), female (PASI90), subgroup categories, with the IXE-treated age C 65 years, BMI [ 30, Asian, disease dura- cohort consistently demonstrating at least tion \ 15 years, C 15 years (primary), biologic- threefold higher odds of achieving PASI100/90 naı̈ve (PASI100), and biologic-experienced (pri- (Fig. 5). Since none of the seven Asian patients mary and PASI90; Fig. 4). While the ORs for receiving ADA achieved any of the outcomes, each subgroup category were comparable to the the FMA analysis could not be conducted for OR for the overall patient group (Figs. 3 and 4), this subgroup category. Likewise, none of the 19 the exception was the subgroup of patients with patients with PsA who received UST achieved PsA, whose odds of achieving PASI100 were PASI100 (Fig. 6). Additionally, as a result of the considerably higher than for the overall group low number of patients who received UST in the for IXE versus GUS (OR 3.0 vs. 2.1) and RIS (OR age C 65 years, Asian, comorbid PsA, and bio- 2.8 vs. 1.6). This mirrors the PASI100 result of logic-experienced subgroup categories, pairwise the anti-IL-17A versus the other biologics comparisons were unviable for some outcomes. cohort for patients with PsA when compared to Except for patients with comorbid PsA or the 878 Adv Ther (2023) 40:869–886 Fig. 3 Comparative effectiveness in ixekizumab versus subgroups. Data are non-responder imputation. Results are guselkumab cohorts across subgroups of patients with statistically significant if 1 is not covered by the 95% CI for moderate-to-severe plaque psoriasis. Comparative adjusted the odds ratios. For the C 65 years subgroup, the lower CI analysis of primary (sPGA (0/1) and/or PASI90) and is 1.043 (primary). BMI body mass index, CI confidence secondary outcomes, PASI100 and PASI90, actual interval, GUS guselkumab, IXE ixekizumab, PASI Psoriasis responses rates, and adjusted odds ratios at week 12 for Area and Severity Index, PsA psoriatic arthritis, sPGA ixekizumab cohort and guselkumab cohort across patient Static Physician Global Assessment biologic-experienced, IXE-treated patients had DISCUSSION significantly higher odds (OR 2.1–4.0) of achieving the primary endpoint than UST-trea- In this subgroup analysis of week 12 data from ted patients (Fig. 6). For all subgroups with the prospective, non-interventional PSoHO viable pairwise comparisons, patients had sig- study, anti-IL-17A biologics showed consis- nificantly higher odds of achieving PASI100 (OR tently greater effectiveness compared to other 2.6–6.6) or PASI90 (OR 2.8–6.5) when treated biologics for patients with moderate-to-severe with IXE compared to UST. PsO, irrespective of most baseline demographics Overall, the comparative treatment effect and all disease characteristics evaluated. The estimates across subgroups largely reflected the most pronounced difference in effectiveness effect estimate for the overall patient group, between the two treatment cohorts was for with few exceptions (Figs. 2, 3, 4, 5, 6). Results patients with comorbid PsA, whereby the anti- thereby showed that patients treated with IXE IL-17A cohort had 19.1% (70.0/50.9), 23.4% had higher response rates and odds of achieving (37.9/14.5), and 26.0% (55.5/29.5) higher pri- the primary endpoint, PASI100, and PASI90 at mary, PASI100, and PASI90 response rates, week 12 across most subgroups compared to respectively, as well as significantly higher odds patients treated with SEC, GUS, RIS, ADA, or of high-level skin clearance compared to the UST (Figs. 2, 3, 4, 5, 6) [17]. other biologics cohort. Moreover, while patients with comorbid PsA receiving other biologics had 15–40% proportionately lower response Adv Ther (2023) 40:869–886 879 Fig. 4 Comparative effectiveness in ixekizumab versus (primary), 1.000 (PASI100), and 0.977 (PASI90); for risankizumab cohorts across subgroups of patients with \ 15 years subgroup, lower CI is 0.981 (PASI100); for moderate-to-severe plaque psoriasis. Comparative adjusted C 15 years subgroup, lower CI is 1.046 (PASI100) and analysis of primary (sPGA (0/1) and/or PASI90) and 1.000 (PASI90); for biologic naı̈ve, lower CI is 0.992 secondary outcomes, PASI100 and PASI90, actual (PASI100). Where statistical models did not converge, responses rates, and adjusted odds ratios at week 12 for results are marked as non-convergence (NC). BMI body ixekizumab cohort and risankizumab cohort across patient mass index, CI confidence interval, IXE ixekizumab, NC subgroups. Data are non-responder imputation. Results are non-convergence, PASI Psoriasis Area and Severity Index, statistically significant if 1 is not covered by the 95% CI for PsA psoriatic arthritis, RIS risankizumab, sPGA Static the odds ratios. For male subgroup, lower CI is 0.984 Physician Global Assessment (PASI90); for female subgroup, lower CI is 1.037 rates across the studied outcomes than patients discontinuation [20, 26, 32, 33], there are few without joint disease, the anti-IL-17A-treated studies that have investigated whether the patients had similar response rates, regardless of presence of PsA can affect treatment response the presence of comorbid PsA. Proportionately rates and skin improvement [13, 21]. In PSoHO, lower response rates were also shown for the IL- no meaningful differences in skin outcomes 23 treatments, GUS (23–33%) and RIS were observed between patients with or without (16–50%), for patients with PsA compared to comorbid PsA, who were prescribed IXE, SEC, or those without PsA. This indicates that patients ADA, whereas a proportionately higher treat- with PsO and comorbid PsA may be more ment effectiveness across each study outcome challenging to treat (and in general, even more was observed for patients unaffected by PsA so if biologic-experienced), and that anti-IL-17A who were prescribed the IL-23 biologics, RIS or biologics may preferentially benefit these GUS. This variable treatment effectiveness for patients compared to treatment with other patients with PsA highlights the importance of biologics. addressing the paucity of real-world studies While recent real-world studies have shown directly comparing treatments across different that comorbid PsA is not associated with sig- patient subgroups [13]. Hence, direct compara- nificantly increased risk for drug tive effectiveness data, such as those presented 880 Adv Ther (2023) 40:869–886 Fig. 5 Comparative effectiveness in ixekizumab versus the odds ratios. For prior biologics C 1 subcategory, lower adalimumab cohorts across subgroups of patients with CI is 1.040 (primary). Where statistical models were moderate-to-severe plaque psoriasis. Comparative adjusted unviable because of low patient numbers, results are analysis of primary (sPGA (0/1) and/or PASI90) and marked as not applicable (NA). ADA adalimumab, BMI secondary outcomes, PASI100 and PASI90, actual body mass index, CI confidence interval, IXE ixekizumab, responses rates, and adjusted odds ratios at week 12 for NA not applicable, PASI Psoriasis Area and Severity Index, ixekizumab cohort and adalimumab cohort across patient PsA psoriatic arthritis, sPGA Static Physician Global subgroups. Data are non-responder imputation. Results are Assessment statistically significant if 1 is not covered by the 95% CI for here, can assist clinicians to optimize treatment results align with the findings of a Spanish selection for individual patients with varying registry study that found that the treatment demographics and disease characteristics. effectiveness of biologics was similar between This study evaluated comparative biologic male and female patients [11]. Also, in PSoHO, effectiveness across four clinically relevant treatment effectiveness was comparable patient demographic characteristics, namely between patients in the \ 65 and C 65 years age sex, age, BMI, and race subgroups, and three categories, although the small sample size in the disease- and treatment-related factors. Potential latter age category led to less stability of models sex-specific differences have been reported in and broader CIs. Prior studies investigating the other studies, such as in the prescription pattern comparative efficacy of biologics for patients in of systemic drugs [11, 18], disease burden [32], different age groups have shown conflicting treatment outcomes [12], and a higher likeli- results; one clinical trial showed that age did hood of biologic discontinuation in female not significantly influence the clinical outcome patients [19, 20]. In this study, however, no sex- at week 52 following treatment with RIS or SEC specific differences in treatment effectiveness [13], while another study indicated dispropor- after 12 weeks were identified, neither between tionately better clinical outcomes at week 48 for the treatment cohorts nor between the indi- patients aged 65 or older treated with GUS vidual treatment comparisons. These 12-week compared to SEC [34]. Adv Ther (2023) 40:869–886 881 Fig. 6 Comparative effectiveness in ixekizumab versus 1.033 (primary) and 1.024 (PASI100). Where statistical ustekinumab cohorts across subgroups of patients with models were unviable because of low patient numbers, moderate-to-severe plaque psoriasis. Comparative adjusted results are marked as not applicable (NA). Where analysis of primary (sPGA (0/1) and/or PASI90) and statistical models did not converge, results are marked as secondary outcomes, PASI100 and PASI90, actual non-convergence (NC). BMI body mass index, CI confi- responses rates, and adjusted odds ratios at week 12 for dence interval, IXE ixekizumab, NA not applicable, NC ixekizumab cohort and ustekinumab cohort across patient non-convergence, PASI Psoriasis Area and Severity Index, subgroups. Data are non-responder imputation. Results are PsA psoriatic arthritis, sPGA Static Physician Global statistically significant if 1 is not covered by the 95% CI for Assessment, UST ustekinumab the odds ratios. For BMI [ 30 subcategory, lower CI is The question of whether obesity There is a paucity of both RCT and real-world (BMI [ 30 kg/m2) affects the efficacy of biolog- data evaluating psoriasis treatments across dif- ical therapies also remains controversial. In ferent races and ethnicities, and particularly in PSoHO, treatment with SEC or RIS was similarly non-White populations [4, 22, 35]. In PSoHO, effective as treatment with IXE in patients with the largest proportions of patients selected a BMI [ 30. Other studies have indicated that White (72.7%) and Asian (14.9%) race, irre- obesity is associated with lower efficacy of bio- spective of their geographical location. As a logic treatment, with important consequences result of the small numbers of patients, sub- on drug survival [19–21]. For instance, one real- group analyses of other race or ethnicity sub- world evidence study has indicated that a groups were unviable. High-level skin response higher BMI was associated with an increased rates were generally lower in Asian patients risk of drug discontinuation for SEC and BROD, compared to White patients across all treat- but not IXE, GUS, and RIS [20]. Other studies ments. An explanation is not readily evident, have shown that obesity has little or no impact but it should be noted that for most treatment on the efficacy of some, generally newer-gen- cohorts, the Asian subgroup comprised low eration biologics [20, 21]. numbers of patients that rendered effectiveness analyses unviable for some treatment 882 Adv Ther (2023) 40:869–886 comparisons, while others did not reach signif- patients treated with any other biologic, the icance. Differences in PsO disease characteristics significance of the comparative effectiveness for between these two subgroups have been repor- biologic-experienced patients was more vari- ted previously [23, 24], but were not found to able, thus underscoring the need to tailor initial impact treatment efficacy [24]. A recent sys- treatment selection. tematic review of RCT data showed that IXE The PSoHO results presented here should be provided the highest skin clearance efficacy for interpreted in the context of the study design Asian patients compared with seven other drugs and its recently published primary data [17]. [4], and another study found no significant First, observational studies have inherent limi- difference in the short-term efficacy of anti-IL- tations, including measured and unmeasured 17 treatment for Caucasians [sic] and Asians confounding and other bias (including selec- [25]. Real-world data from Taiwan indicated tion bias) compared with RCTs. The application disparate effectiveness between treatments with of FMA can accommodate for some of these highest response rates with SEC, followed by uncertainties in model choice through the UST, ADA, and etanercept, although this study machine learning framework [31]. Statistical did not include more recently approved bio- analyses and strategies were pre-specified in the logics, such as RIS and IXE, for which data are protocol and included the grouping of non- eagerly awaited [22]. anti-IL-17A biologics into a single category. This Longer PsO disease duration has been shown study also presented an opportunity to develop to be a negative predictor of clinical outcomes and extend the initial application of FMA to the with IL-12/23 biologics in some studies [15], evaluation of treatment effects across patient although other studies concluded that disease subgroups. The execution and statistical preci- duration is not a significant determining factor sion of these comparative analyses were con- in treatment effectiveness for patients [13, 20]. strained by the number of representative In PSoHO, whether a patient had PsO for less patients in each treatment cohort, category, and than 15 years or for 15 years or longer had little the respective covariates used. There was no effect on response rates and comparative treat- enrolment restriction per treatment group and ment effectiveness. Patients with a longer dis- although the highest proportion of patients ease duration are usually more likely to have were prescribed IXE, other key biologics are prior exposure to biologic agents and other represented in sufficiently high numbers to systemic therapies [8], which may negatively facilitate reliable comparisons. These 12-week affect drug efficacy and consequentially result data provide insights into the speed of action of in lower drug survival of subsequently admin- different biologics and will complement longer- istered biologics [8, 20, 27, 28]. Hence, treat- term effectiveness data that will follow. ment of biologic-experienced patients represents a significant challenge to dermatol- ogists [27]. A recently published paper reported CONCLUSIONS that patients who were biologic-naı̈ve presented a lower hazard of discontinuation [20]. This Despite evidence that clinical outcomes for reinforces the importance of choosing the most patients with moderate-to-severe PsO may vary suitable agent ad initium. A lack of effectiveness on the basis of patient demographics and dis- is a key reason for drug switching or discontin- ease characteristics, this study showed relative uation and 12-week PSoHO data might provide consistency between comparative treatment a rationale for this finding, as biologic-experi- effectiveness, indicating that both the older and enced patients, including those with PsA newer biologics work consistently across patient comorbidity, and across cohorts and treat- subgroups in this routine clinical care setting. ments, largely had lower response rates relative Of note, our results indicate that the presence of to biologic-naı̈ve patients. While the ORs for comorbid PsA may affect a patient’s clinical IXE-treated patients who were biologic-naı̈ve all response to some treatments and that these reached statistical significance relative to patients have significantly higher odds of Adv Ther (2023) 40:869–886 883 achieving skin clearance at week 12 with anti- BMS, Celgene, Eli Lilly and Company, LEO IL-17A biologics compared to other biologics. Pharma, Janssen-Cilag, MSD, Novartis, Pfizer, Pierre Fabre, Roche, Sanofi, UCB. Tiago Torres has received consulting fees from AbbVie, ACKNOWLEDGEMENTS Almirall, Amgen, Arena Pharmaceuticals, Bio- cad, Boehringer Ingelheim, Bristol-Myers Eli Lilly and Company would like to thank the Squibb, Celgene, Eli Lilly and Company, Jans- participants, caregivers and investigators, with- sen, LEO Pharma, MSD, Novartis, Pfizer, Sam- out whom this work would not be possible. The sung, Bioepis, Sandoz, Sanofi, and Viatris; authors would also like to thank Dr. Alyssa payment or honoraria for lectures or presenta- Garrelts for review and valuable comments and tions from AbbVie, Almirall, Amgen, Arena Dr. Natalie Haustrup for project management Pharmaceuticals, Biocad, Boehringer Ingelheim, support and scientific communication Bristol-Myers Squibb, Celgene, Eli Lilly and expertise. Company, Janssen, LEO Pharma, MSD, Novar- tis, Pfizer, Samsung- Bioepis, Sandoz, Sanofi, and Viatris. Andreas Pinter has served as an Funding. Sponsorship for this study, as well investigator, speaker, and/or advisor for AbbVie, as Open Access and Rapid Service Fees, were Almirall-Hermal, Amgen, Biogen Idec, BioN- funded by Eli Lilly and Company. Tech, Boehringer Ingelheim, Celgene, GSK, Eli Lilly and Company, Galderma, Hexal, Janssen, Author Contributions. CS, AB, SW, CR and LEO Pharma, MC2, Medac, Merck Serono, Mit- ER contributed to the conceptualization of the subishi, MSD, Novartis, Pascoe, Pfizer, Tigercat ideas presented. CL, JTM, TT, GF, AP, LP con- Pharma, Regeneron, Roche, Sandoz Biophar- tributed to data acquisition. AB performed the maceuticals, Sanofi Genzyme, Schering-Plough statistical analysis. CS, AB, SW interpretation of und UCB Pharma. Gabrielle Fabbrocini reports the data. CL, ER, JTM, TT, AP, GF, FD, AB, CR, speaking or consultant fees from Abbvie, Eli SW, TH, CS and LP contributed to the inter- Lilly and Company, Novartis, Almirall, Janssen, pretation of the data. CS and AB wrote the Leo Pharma, Amgen and UCB. Flavia Daniele manuscript. All authors contributed to the reports speaking fees from Raffo, Eli Lilly and critical revision of the manuscript. All authors Company and ICB Montpellier, support for read and approved the final manuscript. travel and/or attending meetings from Sanofi, Disclosures. Charles Lynde has been a Montppellier Laboratorios, Raffo, Novartis, speaker, consultant and/or private investigator Vichy laboratories, Eli Lilly and Company, par- for: AbbVie, Altius, Amgen, Aralez, Arcutis, ticipation on an advisory board with Eli Lilly Bausch Health, Bayer, Boehringer Ingelheim, and Company and receipt of equipment or Bristol Myers Squibb, Celgene, Cipher, Concert, materials from Raffo, Andromaco, Vichy, La Dermavant, Eli Lilly and Company, Fresnius Roche Possay, Abbie, Merz and Lazzar labora- Kabi, GSK, Innovaderm, Intega Skin, Janssen, tories. Alan Brnabic, Catherine Reed, Christo- Kyowa, La Roche Posay, LEO Pharma, L’Oreal, pher Schuster, Stefan Wilhelm and Thorsten Medexus, Merck, P&G, Pediapharm, Pfizer, Holzkämper are employees and minor share- Regeneron, Roche, Sanofi Genzyme, Sentrex, holders of Eli Lilly and Company. Luis Puig SunPharma, TEVA, Tribute, UCB, Valeant, Via- reports grants to institution from AbbVie, tris, Volo Health. Elisabeth Riedl is a former Almirall, Amgen, Boehringer Ingelheim, Cel- employee of Eli Lilly and Company and a minor gene, Janssen, Leo Pharma, Eli Lilly and Com- stockholder and has been a speaker and con- pany, Novartis, Pfizer, Regeneron, Roche, sultant for Eli Lilly and Company. Julia-Tatjana Sanofi, UCB; consulting fees or honoraria from Maul has served as advisor and/or received AbbVie, Almirall, Amgen, Baxalta, Biogen, speaking fees and/or participated in clinical Boehringer Ingelheim, Celgene, Gebro, Janssen, trials sponsored by AbbVie, Almirall, Amgen, Leo Pharma, Eli Lilly and Company, Merck Serono, MSD, Mylan, Novartis, Pfizer, 884 Adv Ther (2023) 40:869–886 Regeneron, Roche, Sandoz, Samsung-Bioepis, in this article are included in the article’s Sanofi, UCB; participation in company spon- Creative Commons licence, unless indicated sored speaker’s bureau for Celgene, Janssen, Eli otherwise in a credit line to the material. If Lilly and Company, MSD, Novartis, Pfizer. material is not included in the article’s Creative Commons licence and your intended use is not Compliance with Ethics Guidelines. The permitted by statutory regulation or exceeds the protocol, amendments and consent documen- permitted use, you will need to obtain permis- tation were approved by local institutional sion directly from the copyright holder. To view review boards. The study was registered at the a copy of this licence, visit http:// European Network of Centers for Pharma- creativecommons.org/licenses/by-nc/4.0/. coepidemiology and Pharmacovigilance (ENCEPP24207) [29] and was conducted according to International Conference on Har- monization, Good Clinical Practice guidelines REFERENCES and the Declaration of Helsinki. All patients were required to give informed consent for 1. Malatestinic W, Nordstrom B, Wu JJ, et al. Charac- participation in the study. We confirm that the teristics and medication use of psoriasis patients necessary central or local IRB and/or ethics who may or may not qualify for randomized con- trolled trials. J Manag Care Spec Pharm. 2017;23: committee approvals have been obtained for 370–81. this multi-site, international study by United BioSource LLC (UBC). 2. Masson Regnault M, Castañeda-Sanabria J, Diep Tran MHT, et al. Users of biologics in clinical practice: would they be eligible for phase III clinical Data Availability. Data are available on studies? Cohort Study in the French Psoriasis Reg- reasonable request. Lilly provides access to all istry PSOBIOTEQ. J Eur Acad Dermatol Venereol. individual participant data collected during the 2020;34:293–300. study, after anonymization. Data are available 3. Armstrong AW, Harskamp CT, Armstrong EJ. The to request after primary publication acceptance. association between psoriasis and obesity: a sys- No expiration date of data requests is currently tematic review and meta-analysis of observational set once data are made available. Access is pro- studies. Nutr Diabetes. 2012;2: e54. vided after a proposal has been approved by an 4. Ferguson JE, Seger EW, White J, et al. Racial/ethnic independent review committee identified for differences in treatment efficacy and safety for this purpose and after receipt of a signed data moderate-to-severe plaque psoriasis: a systematic sharing agreement. Data and documents, review. Arch Dermatol Res. 2022. https://doi.org/ including the study protocol, statistical analysis 10.1007/s00403-022-02324-4. plan, clinical study report, blank or annotated 5. Kamata M, Tada Y. Efficacy and safety of biologics case report forms, will be provided in a secure for psoriasis and psoriatic arthritis and their impact data sharing environment. For details on sub- on comorbidities: a literature review. Int J Mol Sci. mitting a request, see the instructions provided 2020;21:1690. at www.vivli.org. 6. Chiricozzi A, Gisondi P, Girolomoni G. The phar- macological management of patients with comor- Open Access. This article is licensed under a bid psoriasis and obesity. Expert Opin Creative Commons Attribution-Non- Pharmacother. 2019;20:863–72. Commercial 4.0 International License, which 7. Lebwohl M, Yeilding N, Szapary P, et al. Impact of permits any non-commercial use, sharing, weight on the efficacy and safety of ustekinumab in adaptation, distribution and reproduction in patients with moderate to severe psoriasis: rationale any medium or format, as long as you give for dosing recommendations. J Am Acad Dermatol. appropriate credit to the original author(s) and 2010;63:571–9. the source, provide a link to the Creative 8. Papp KA, Gordon KB, Langley RG, et al. Impact of Commons licence, and indicate if changes were previous biologic use on the efficacy and safety of made. The images or other third party material brodalumab and ustekinumab in patients with Adv Ther (2023) 40:869–886 885 moderate-to-severe plaque psoriasis: integrated (PSoHO). J Eur Acad Dermatol Venereol. analysis of the randomized controlled trials AMA- 2022;36(11):2087–2100. GINE-2 and AMAGINE-3. Br J Dermatol. 2018;179: 320–8. 18. Guillet C, Seeli C, Nina M, et al. The impact of gender and sex in psoriasis: what to be aware of 9. Menter A, Papp KA, Gooderham M, et al. Drug when treating women with psoriasis. Int J Womens survival of biologic therapy in a large, disease-based Dermatol. 2022;8(2):e010. registry of patients with psoriasis: results from the Psoriasis Longitudinal Assessment and Registry 19. Mourad A, Straube S, Armijo-Olivo S, et al. Factors (PSOLAR). J Eur Acad Dermatol Venereol. 2016;30: predicting persistence of biologic drugs in psoriasis: 1148–58. a systematic review and meta-analysis. Br J Derma- tol. 2019;181:450–8. 10. Leonardi CL, Kimball AB, Papp KA, et al. Efficacy and safety of ustekinumab, a human interleukin- 20. Torres T, Puig L, Vender R, et al. Drug survival of IL- 12/23 monoclonal antibody, in patients with pso- 12/23, IL-17 and IL-23 inhibitors for psoriasis riasis: 76-week results from a randomised, double- treatment: a retrospective multi-country, multi- blind, placebo-controlled trial (PHOENIX 1). Lan- centric cohort study. Am J Clin Dermatol. 2021;22: cet. 2008;371:1665–74. 567–79. 11. Hernández-Fernández CP, Carretero G, Rivera R, 21. Pirro F, Caldarola G, Chiricozzi A, et al. Impact of et al. Effect of sex in systemic psoriasis therapy: body mass index on the efficacy of biological ther- differences in prescription, effectiveness and safety apies in patients with psoriasis: a real-world study. in the BIOBADADERM prospective cohort. Acta Clin Drug Investig. 2021;41:917–25. Derm Venereol. 2021;101: adv00354. 22. Chen Y-C, Huang Y-T, Yang C-C, et al. Real-world 12. Maul JT, Augustin M, Sorbe C, et al. Association of efficacy of biological agents in moderate-to-severe sex and systemic therapy treatment outcomes in plaque psoriasis: an analysis of 75 patients in Tai- psoriasis: a two-country, multicentre, prospective, wan. PLoS One. 2021;15: e0244620. noninterventional registry study. Br J Dermatol. 2021;185:1160–8. 23. Kim J, Oh C-H, Jeon J, et al. Molecular phenotyping small (Asian) versus large (Western) plaque psoriasis 13. Crowley JJ, Langley RG, Gordon KB, et al. Efficacy shows common activation of IL-17 pathway genes of risankizumab versus secukinumab in patients but different regulatory gene sets. J Investig Der- with moderate-to-severe psoriasis: subgroup analy- matol. 2016;136:161–72. sis from the IMMerge study. Dermatol Ther. 2022;12:561–75. 24. Shah SK, Arthur A, Yang YC, et al. A retrospective study to investigate racial and ethnic variations in 14. Blauvelt A, Papp KA, Griffiths CEM, et al. Efficacy the treatment of psoriasis with etanercept. J Drugs and safety of guselkumab, an anti-interleukin-23 Dermatol. 2011;10:866–72. monoclonal antibody, compared with adalimumab for the continuous treatment of patients with 25. Zhang D, Qiu J, Liao X, et al. Comparison of efficacy moderate to severe psoriasis: results from the phase of anti-interleukin-17 in the treatment of psoriasis III, double-blinded, placebo- and active compara- between caucasians and Asians: a systematic review tor–controlled VOYAGE 1 trial. J Am Acad Derma- and meta-analysis. Front Med (Lausanne). 2022;8: tol. 2017;76:405–17. 814938. 15. Papp KA, Langley RG, Lebwohl M, et al. Efficacy 26. Lunder T, Zorko MS, Kolar NK, et al. Drug survival and safety of ustekinumab, a human interleukin- of biological therapy is showing class effect: upda- 12/23 monoclonal antibody, in patients with pso- ted results from Slovenian National Registry of riasis: 52-week results from a randomised, double- psoriasis. Int J Dermatol. 2019;58:631–41. blind, placebo-controlled trial (PHOENIX 2). Lan- cet. 2008;371:1675–84. 27. Ger T-Y, Huang Y-H, Hui RC, et al. Effectiveness and safety of secukinumab for psoriasis in real-world 16. Amin M, No DJ, Egeberg A, et al. Choosing first-line practice: analysis of subgroups stratified by prior biologic treatment for moderate-to-severe psoriasis: biologic failure or reimbursement. Therap Adv what does the evidence say? Am J Clin Dermatol. Chronic Dis. 2019;10:2040622319843756. 2018;19:1–13. 28. Gniadecki R, Bang B, Bryld LE, et al. Comparison of 17. Pinter A, Puig L, Schäkel K, et al. Comparative long-term drug survival and safety of biologic effectiveness of biologics in clinical practice: agents in patients with psoriasis vulgaris. Br J Der- week 12 primary outcomes from an international matol. 2015;172:244–52. observational psoriasis study of health outcomes 886 Adv Ther (2023) 40:869–886 29. European Medicines Agency. Psoriasis Study of antibodies: a time-period-adjusted registry analysis. Health Outcomes—an international observational Br J Dermatol. 2021;184:1094–105. study of 3 year health outcomes in the biologic treatment of moderate to severe plaque psoriasis. 33. Kishimoto M, Komine M, Kamiya K, et al. Drug In: (Pharmacovigilance ENoCfPa, ed). https://www. survival of apremilast in a real-world setting. J Der- encepp.eu/encepp/viewResource.htm?id=25115: matol. 2019;46:615–7. EUPAS24207. 2018. Accessed 31 Aug 2022. 34. Blauvelt A, Armstrong AW, Langley RG, et al. Effi- 30. Zagar AJ, Kadziola Z, Lipkovich I, et al. Evaluating cacy of guselkumab versus secukinumab in sub- different strategies for estimating treatment effects populations of patients with moderate-to-severe in observational studies. J Biopharm Stat. 2017;27: plaque psoriasis: results from the ECLIPSE study. 535–53. J Dermatol Treat. 2021;33(4):2317–24.. 31. Zagar A, Kadziola Z, Lipkovich I, et al. Evaluating 35. Kaufman BP, Alexis AF. Psoriasis in skin of color: bias control strategies in observational studies using insights into the epidemiology, clinical presenta- frequentist model averaging. J Biopharmaceutic tion, genetics, quality-of-life impact, and treatment Stat. 2022:32(2):247–76. of psoriasis in non-white racial/ethnic groups. Am J Clin Dermatol. 2018;19:405–23. 32. Graier T, Salmhofer W, Jonak C, et al. Biologic drug survival rates in the era of anti-interleukin-17