Impact of biologic class and treatment line on psoriatic arthritis risk in psoriasis: a population-based cohort study Authors: Abdulla Watad, Yoav Elizur, Or Hen, Paula David, Yonatan Shneor Patt, Omer Gendelman, Amit Nachman, Alen Zabotti, Merav Lidar, Dennis McGonagle, Howard Amital Published: 2026 ABSTRACT AbstractBackgroundPatients with psoriasis (PsO) are at increased risk of developing psoriatic arthritis (PsA). Emerging evidence suggests that biologic therapies may differentially influence the development of PsA; however, there are no data on the relative ability of second-line biological therapy to prevent PsA.AimTo evaluate the impact of biologic therapy on the risk of incident PsA in patients with PsO with a focus on second-line biologic therapy.MethodsWe identified 2819 patients who initiated biologic therapy for PsO between 2002 and 2022, including Tumor Necrosis Factor (TNF) inhibitors, interleukin (IL)-17 inhibitors, IL-23p19 inhibitors and IL-12/23 inhibitors. The primary outcome was incident PsA. Kaplan-Meier survival analysis and Cox proportional hazards models were used to evaluate arthritis-free survival and HRs for the development of PsA. Risk was assessed separately for first-line and second-line biologic therapy.ResultsAmong 2819 patients with PsO initiating first-line biologic therapy, 400 (14.2%) developed PsA. In multivariable Cox regression, first-line anti-IL-23p19 agents were significantly associated with a lower risk of developing PsA compared with TNF inhibitors (HR 0.13; 95% CI 0.02 to 0.96; p=0.045). Among 1246 patients receiving second-line biologics, 125 (10.0%) developed PsA. Notably, in second-line therapy, only IL-17 inhibitors were independently associated with a statistically significant reduction in PsA risk versus TNF inhibitors (HR 0.37; 95% CI 0.16 to 0.85; p=0.019), with trends observed for anti-IL-12/23 (HR 0.14; 95% CI 0.02 to 1.05; p=0.056) and anti-IL-23p19 (HR 0.46; 95% CI 0.20 to 1.07; p=0.070).ConclusionsThis is the first study to evaluate second-line biological therapy in relation to PsA risk. IL-17 inhibitor therapy was significantly associated with a reduced risk of incident PsA compared with TNF inhibitors. WHAT IS ALREADY KNOWN ON THIS TOPIC Patients with psoriasis are at increased risk of developing psoriatic arthritis (PsA), and emerging real-world evidence suggests that some biologic therapies (notably first-line interleukin (IL)-23p19 inhibitors) may reduce this risk, but the preventive impact of second-line biologic sequencing is largely unknown. Patients with psoriasis are at increased risk of developing psoriatic arthritis (PsA), and emerging real-world evidence suggests that some biologic therapies (notably first-line interleukin (IL)-23p19 inhibitors) may reduce this risk, but the preventive impact of second-line biologic sequencing is largely unknown. WHAT THIS STUDY ADDS In a population-based cohort, second-line IL-17 inhibitors were associated with a significantly lower risk of incident PsA compared with second-line Tumor Necrosis Factor (TNF) inhibitors. In a population-based cohort, second-line IL-17 inhibitors were associated with a significantly lower risk of incident PsA compared with second-line Tumor Necrosis Factor (TNF) inhibitors. The study also independently supports prior observations of better PsA-free survival with first-line IL-23p19 inhibitors. The study also independently supports prior observations of better PsA-free survival with first-line IL-23p19 inhibitors. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings suggest that biologic class and treatment sequencing may influence PsA prevention, supporting IL-17 inhibitors as a preferred second-line option after TNF inhibitor failure and motivating prospective studies to guide prevention-oriented sequencing recommendations. These findings suggest that biologic class and treatment sequencing may influence PsA prevention, supporting IL-17 inhibitors as a preferred second-line option after TNF inhibitor failure and motivating prospective studies to guide prevention-oriented sequencing recommendations. Introduction Psoriasis (PsO) is a chronic inflammatory skin condition that progresses to psoriatic arthritis (PsA) in a subset of patients, yet the impact of biological therapies on this transition remains complex and may vary depending on the specific agent used. The risk of progression from PsO to PsA is estimated to be approximately 20–30% over a patient’s lifetime, with an annual incidence of 1–3% among patients with PsO.13 Several factors contribute to this risk, including disease severity, nail involvement, obesity, genetic predisposition and systemic inflammation.13 Subclinical musculoskeletal inflammation, detectable via imaging, has been identified in patients with PsO even before the onset of clinical arthritis, suggesting that PsA may develop along a continuum rather than as a distinct event.46 Understanding how biological therapy influences this transition is critical, as early intervention could potentially alter the disease course and prevent long-term joint damage. Several studies have explored the association between PsO and PsA development; with the exception of one study,7 studies having shown that biological therapy is associated with a lower risk of PsA.811 Furthermore, numerous studies have shown that the interleukin (IL)-23 inhibitors appear to be superior for PsA prevention when used as first-line agents.12 13 IL-23p19 blockade has emerged as a potentially promising biological strategy associated with lower PsA risk, with multiple independent studies supporting this association. However, existing evidence has focused almost exclusively on first-line biologic therapy, and no studies to date have systematically evaluated the effect of second-line or subsequent biologic classes on PsA prevention among patients with PsO. In addition, regulatory restrictions in many countries often limit the use of IL-23 inhibitors as first-line agents, leading clinicians to initiate treatment with Tumor Necrosis Factor (TNF) inhibitors by default. Here, we focused on the impact of second-line biologic therapy in PsO on the incidence of PsA where there is currently a paucity of data to investigate whether the apparent benefit of anti-IL-23p19 blockade as a first-line therapy carries over to second-line therapy. Material and methods Study design and data source Given the retrospective nature of the analysis and the use of de-identified data, the requirement for informed consent was waived. This retrospective cohort study was conducted using electronic medical records from the Clalit Health Services database, the largest healthcare provider operating under the national health insurance system in Israel. The database provides comprehensive longitudinal data, integrating medical records from multiple healthcare providers, including demographic information, diagnostic codes, prescription records, procedures and laboratory test results. The study focused on individuals diagnosed with PsO, identified using the International Classification of Diseases, Ninth Revision (ICD-9) codes. The study period spanned from 1 January 2002 to 31 December 2022. Patient selection Eligible participants were adults (≥18 years) with a documented diagnosis of PsO (ICD-9 codes 696, 696.0, 696.1). Patients were excluded if they met any of the following criteria: A diagnosis of PsA (PsA or other PsA) preceding the PsO diagnosis or within 30 days following PsO diagnosis. A diagnosis of PsA (PsA or other PsA) preceding the PsO diagnosis or within 30 days following PsO diagnosis. Initiation of biologic therapy prior to the diagnosis of PsO. Initiation of biologic therapy prior to the diagnosis of PsO. Initiation of biological therapy after the first recorded diagnosis of PsA. Initiation of biological therapy after the first recorded diagnosis of PsA. Exposure and outcome measures Baseline demographic and clinical characteristics extracted from the database included age, sex, ethnicity, body mass index, smoking status and comorbidities such as diabetes, hypertension, obesity, coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease and hyperlipidaemia. Patients included in the study were required to have dispensed a biological therapy approved for PsO treatment, with the date of the first pharmacy dispensing considered as the treatment initiation date. The biological therapies were categorised into the following classes: TNF inhibitors: adalimumab, etanercept, infliximab, certolizumab, golimumab. TNF inhibitors: adalimumab, etanercept, infliximab, certolizumab, golimumab. IL-17 inhibitors: secukinumab, ixekizumab. IL-17 inhibitors: secukinumab, ixekizumab. IL-23p19 inhibitors: guselkumab, risankizumab. IL-23p19 inhibitors: guselkumab, risankizumab. IL-12/23 inhibitor: ustekinumab. IL-12/23 inhibitor: ustekinumab. The index date was defined as the initiation date of the first dispensed biological therapy, occurring on or after the first recorded diagnosis of PsO. The primary outcome was the onset of PsA, defined by newly recorded ICD-9 diagnostic codes entered by senior rheumatologists following the initiation of biological therapy. The risk of PsA was assessed separately for first-line and second-line biological therapy. Time to event was calculated from the index date until the first recorded diagnosis of PsA, switch to another biological class, treatment discontinuation, death or the study endpoint (1 January 2024), whichever occurred first. Statistical analysis Baseline characteristics were summarised using descriptive statistics for the entire cohort and stratified by biological class. Continuous variables were reported as means with SD for parametric data and compared using Student’s t-tests (two-group comparisons) or one-way analysis of variance (ANOVA) for multigroup comparisons. For non-parametric data, medians with IQRs were reported and compared using Mann-Whitney U (two-group comparisons) or Kruskal-Wallis tests (multigroup comparisons), followed by post hoc analyses where appropriate. Continuous variables were reported as means with SD for parametric data and compared using Student’s t-tests (two-group comparisons) or one-way analysis of variance (ANOVA) for multigroup comparisons. For non-parametric data, medians with IQRs were reported and compared using Mann-Whitney U (two-group comparisons) or Kruskal-Wallis tests (multigroup comparisons), followed by post hoc analyses where appropriate. Categorical variables were reported as frequencies and percentages and compared using χ2 tests; Fisher’s exact test was applied when expected cell counts were <5. Categorical variables were reported as frequencies and percentages and compared using χ2 tests; Fisher’s exact test was applied when expected cell counts were <5. Time-to-event analysis The cumulative incidence of PsA following biological therapy initiation was estimated using Kaplan-Meier survival curves, with incidence rates expressed as events per 100 person-years. Patients were censored at the earliest occurrence of PsA diagnosis, switch to a different biological class, discontinuation of biological therapy, loss to follow-up, death or end of the study period (1 January 2024). Multivariable Cox proportional hazards regression models were constructed to assess the risk of developing PsA. TNF inhibitors served as the reference biological class. Models were adjusted for age, sex, smoking status and obesity. HRs and 95% CIs were estimated to compare the relative risk of PsA across biological classes. All statistical analyses were performed using R statistical software (V.4.4.3). A two-tailed p value<0.05 was considered statistically significant. Sensitivity analyses We performed sensitivity analyses across all three models—first-line therapy, second-line therapy and time-dependent exposure—by using a natural cubic spline to model age, instead of treating it as a straight linear variable. We also performed a multinomial propensity score analysis for first-line therapy using an inverse probability of treatment weighting Cox regression model. Covariate balance was assessed with absolute mean differences, with values of 0.1 or less indicating adequate balance. Results Study population Out of 157 360 patients with PsO identified in the database, 7445 (4.7%) were excluded due to a diagnosis of PsA prior to PsO (n=4915), biologic treatment before PsO diagnosis (n=379) or biologic initiation after the onset of PsA (n=2151). The final cohort included 149 915 patients (figure 1). Among these, 2819 patients received biological agents and were included in the primary analysis. Baseline characteristics, stratified by first-line biological agent, are presented in table 1. Flow chart. IL, interleukin; PsO, psoriasis; TNF, Tumor Necrosis Factor. Table 1Characteristics by first biologic class (psoriasis treatment)Characteristics by first biologic line (psoriasis treatment)VariableFirst biologic classTotaln=2819*Anti-TNF n=2517*Anti-IL-17 n=124*Anti-IL-23 n=132*Anti-IL-12/23 n=46*P value†Age at psoriasis onset (years)36 (26, 51)36 (26, 51)38 (28, 52)34 (24, 49)49 (33, 61)0.003Sex0.24 Male1684 (60%)1493 (59%)84 (68%)77 (58%)30 (65%) Female1135 (40%)1024 (41%)40 (32%)55 (42%)16 (35%)Sector<0.001 Jewish2246 (80%)1971 (78%)108 (87%)125 (95%)42 (91%) Arab564 (20%)537 (21%)16 (13%)7 (5.3%)4 (8.7%) Time from psoriasis to first biologic (years)6.9 (3.2, 11.7)6.6 (3.1, 11.3)10.1 (4.2, 13.6)10.7 (5.4, 15.9)6.0 (1.7, 10.2)<0.001 Obesity976 (35%)882 (35%)38 (31%)38 (29%)18 (39%)0.33 Smoking1563 (55%)1393 (55%)70 (56%)73 (55%)27 (59%)0.97 Diabetes mellitus570 (20%)499 (20%)30 (24%)24 (18%)17 (37%)0.020 Hyperlipidaemia1396 (50%)1247 (50%)57 (46%)65 (49%)27 (59%)0.54 Hypertension730 (26%)645 (26%)37 (30%)26 (20%)22 (48%)0.002 Peripheral vascular disease66 (2.3%)57 (2.3%)5 (4.0%)1 (0.8%)3 (6.5%)0.087 Chronic obstructive pulmonary disease138 (4.9%)126 (5.0%)3 (2.4%)4 (3.0%)5 (11%)0.10 Congestive heart failure94 (3.3%)77 (3.1%)6 (4.8%)5 (3.8%)6 (13%)0.002 Chronic renal failure111 (3.9%)95 (3.8%)7 (5.6%)1 (0.8%)8 (17%)<0.001 Ischaemic heart disease322 (11%)282 (11%)17 (14%)12 (9.1%)11 (24%)0.036First biologic line prescribed Anti-TNF2517 (89.3%)2517 (100%)––– Anti-IL-17124 (4.4%)–124 (100%)–– Anti-IL-23132 (4.7%)––132 (100%)– Anti-IL-12/2346 (1.6%)–––46 (100%)Second biologic line prescribed Anti-TNF654/1472 (44.4%)607/1377 (44.1%)18/46 (39.1%)22/30 (73.3%)7/19 (36.8%) Anti-IL-17382/1472 (26%)363/1377 (26.4%)7/46 (15.2%)5/30 (16.7%)7/19 (36.8%) Anti-IL-23324/1472 (22%)297/1377 (21.6%)20/46 (43.5%)2/30 (6.7%)5/19 (26.3%) Anti-IL-12/23112/1472 (7.6%)110/1377 (8%)1/46 (2.2%)1/30 (3.3%)–*Median (Q1, Q3); n (%).†Kruskal-Wallis rank sum test; Pearson’s χ2 test.IL, interleukin; TNF, Tumor Necrosis Factor. Characteristics by first biologic class (psoriasis treatment) Median (Q1, Q3); n (%). Kruskal-Wallis rank sum test; Pearson’s χ2 test. IL, interleukin; TNF, Tumor Necrosis Factor. First-line biologic therapy and PsA prevention Of 2819 patients initiating first-line biological therapy, most were treated with anti-TNF agents (n=2517; 89.3%), followed by IL-23p19 (n=132; 4.7%), IL-17 (n=124; 4.4%) and IL-12/23 (n=46; 1.6%) inhibitors. The median time from PsO diagnosis to the initiation of first-line biologic therapy was 6.9 years (IQR: 3.2–11.7), with significant variation observed across the different biologic classes (p<0.001). During follow-up on first-line biological therapy, 400 patients (14.2%) developed PsA. Kaplan-Meier survival analysis demonstrated a statistically significant difference in PsA-free survival between biological classes (log-rank p=0.023; figure 2A). As previously reported in other studies, patients treated with anti-IL-23p19 inhibitor agents had the highest arthritis-free survival probability (0.95; 95% CI 0.88 to 1.00), whereas those treated with TNF inhibitors had the lowest (0.64; 95% CI 0.60 to 0.69). (A) Kaplan-Meier survival curve for first biologic line. (B) Adjusted HRs for PsA after first biologic line. IL, interleukin; PsA, psoriatic arthritis; TNF, Tumor Necrosis Factor.* P < 0.05, ** P < 0.01. In the multivariable Cox regression analysis (figure 2B), treatment with anti-IL-23p19 agents was significantly associated with a lower risk of developing PsA compared with TNF inhibitor therapy (HR 0.13; 95% CI 0.02 to 0.96; p=0.045). Yet, IL-17 and anti-IL-12/23 inhibitors did not show a statistically significant reduction in PsA incidence when compared with TNF inhibitors treatment (anti-IL-17: HR 0.39; 95% CI 0.13 to 1.23; p=0.109 and anti-IL-12/23 inhibitors: HR 1.07; 95% CI 0.34 to 3.38; p=0.911). In terms of patient characteristics, female sex was associated with a significantly higher risk of developing PsA (HR 1.61; 95% CI 1.16 to 2.24; p=0.005), as was younger age at biologic initiation (HR 0.99 per year increase; 95% CI 0.98 to 1.00; p=0.03). Obesity and smoking status were not significantly associated with the risk of developing PsA. Second-line biologic therapy and PsA prevention Among the 1246 patients who received a second-line biologic agent and were at risk for PsA development, the most common treatments were anti-TNF (n=534; 42.8%), anti-IL-17 (n=317; 25.4%), anti-IL-23p19 (n=299; 24.0%) and anti-IL-12/23 (n=96; 7.7%). The median time from PsO diagnosis to the initiation of second-line biologic therapy was 8.9 years (IQR: 4.9–13.3), which varied significantly by biologic class (p<0.001). During follow-up on the second biologic line, 125 patients (10.0%) developed PsA. Kaplan-Meier survival analysis revealed significant differences in arthritis-free survival across biologic classes (log-rank p=0.013; figure 3A), with patients treated with anti-IL-23p19 agents demonstrating the highest survival probability (0.93; 95% CI 0.89 to 0.97) compared with those treated with anti-TNF therapy. (A) Kaplan-Meier survival curve for second biologic line. (B) Adjusted HRs for PsA after second biologic line. IL, interleukin; PsA, psoriatic arthritis; TNF, Tumor Necrosis Factor. * P < 0.05. However, in multivariable Cox regression (figure 3B), treatment with anti-IL-17 inhibitors was significantly associated with reduced PsA risk compared with TNF inhibitors (HR 0.37; 95% CI 0.16 to 0.85; p=0.019). Similar but non-significant trends were observed for anti-IL-12/23 (HR 0.14; 95% CI 0.02 to 1.05; p=0.056) and anti-IL-23p19 inhibitors (HR 0.46; 95% CI 0.20 to 1.07; p=0.070) compared with anti-TNF inhibitors. Younger age at therapy initiation remained an independent predictor of reduced PsA incidence (HR 0.98; 95% CI 0.96 to 1.00; p=0.03), while sex, obesity and smoking status were not significantly associated. Time-dependent biologic exposure analysis To assess the impact of biologic therapy over time, regardless of whether it was first line or second line, we conducted a time-dependent Cox regression analysis that included all patients from the primary cohort. This model accounted for treatment switching between different biologic classes during follow-up. In this analysis, treatment with anti-IL-23p19 agents was associated with a significantly lower risk of developing PsA compared with TNF inhibitors (HR 0.47; 95% CI 0.25 to 0.86; p=0.014). Therapies targeting IL-17 and IL-12/23 inhibitors also showed a trend toward reduced arthritis risk compared with TNF inhibitors, but these associations did not reach statistical significance (IL-17: HR 0.62; 95% CI 0.37 to 1.04; p=0.073; IL-12/23: HR 0.66; 95% CI 0.36 to 1.24; p=0.196). Among patient characteristics, female sex (HR 1.39; 95% CI 1.06 to 1.83; p=0.019) and younger age (HR 0.99 per year; 95% CI 0.98 to 1.00; p=0.006) were significantly associated with a higher risk of the development of PsA. In contrast, obesity and smoking status were not significantly associated with PsA risk (see online supplemental data). Sensitivity analyses The results remained consistent in all models, showing similar effect sizes and directions for all variables supporting the robustness of our findings (see onlinesupplemental figures S1S5). Discussion In this large, real-world cohort study, we investigated the impact of biologic therapy on the risk of developing PsA with a specific focus on second-line biologic use since many patients with PsA switch biological therapy class. Our findings indicate that the class of biologic therapy significantly influences the risk of incident PsA, with notable differences between treatment lines. Specifically, we observed that first-line treatment with IL-23p19 inhibitors was associated with the highest arthritis-free survival and a significantly reduced risk of PsA compared with TNF inhibitors, which has been repeatedly shown in earlier studies.12 13 Yet, the key novel finding of our study is that second-line IL-17 inhibitors were the only class associated with a statistically significant reduction in PsA risk compared with TNF inhibitors after multivariable adjustment. These results highlight the potential of IL-17 blockade as a preferred second-line strategy associated with reduced PsA risk in patients with PsO, but further validation and replication of these findings are needed. Our findings are consistent with and reinforce previous reports showing that IL-23p19 inhibitor therapy, when used as first-line treatment, is associated with a reduced risk of developing PsA in patients with PsO. This supports the growing body of evidence that IL-23p19 blockade is increasingly recognised as being associated with lower PsA risk early in the disease course, possibly via targeting of upstream inflammatory pathways involved in both skin and joint pathogenesis.14 By replicating these findings in a large real-world cohort, our study adds further support to the potential of early IL-23p19 inhibition to modify disease progression.15 Conversely, the significant reduction in PsA risk observed with anti-IL-17 agents in the second-line setting may reflect their potent anti-inflammatory effect in patients with established subclinical musculoskeletal disease who had suboptimal response to initial therapy.16 IL-17 inhibitors, such as secukinumab and ixekizumab, directly target a key effector cytokine in SpA pathogenesis and are well established in both PsA and axial SpA management.17 The trend toward reduced PsA risk seen with IL-12/23 and anti-IL-23p19 inhibitors in second-line therapy, compared with TNF inhibitor therapy, although not statistically significant, supports a broader therapeutic benefit of IL-targeting agents beyond TNF inhibition. The time-dependent biologic exposure analysis in our study further supports these findings. Our study also confirms previously reported demographic associations, including the increased PsA risk among female patients and younger individuals, highlighting the importance of demographic stratification in risk prediction and treatment selection.18 Notably, obesity and smoking status were not independently associated with the development of PsA in our adjusted models, contrasting with some earlier studies, possibly reflecting differences in study design, population characteristics or confounder adjustment.19 This study’s strengths include its large, real-world cohort, long follow-up and use of multiple analytic models showing consistent results. Limitations include its retrospective design, reliance on ICD-9 codes and lack of data on disease severity, imaging or biomarkers. Importantly, the lack of detailed clinical information at the time of biologic initiation prevents the assessment of whether a proportion of patients may have already presented with arthralgia or other clinical manifestations suggestive of a prodromal phase of PsA, even if not yet fulfilling diagnostic criteria. Additionally, the relatively small sample sizes within specific biologic subgroups, notably IL-12/23 inhibitors in the first-line analyses, resulted in wide CIs that limit the precision of these specific estimates. Residual confounding and unmeasured factors, such as treatment preferences or adherence, may have influenced biologic selection. In addition, evaluating biologic therapy beyond the second line remains particularly challenging, due to the long latency between PsO onset and arthritis development, as well as the limited number of patients who progress through multiple biologic switches. Consequently, larger longitudinal datasets will be needed to address this gap. Nonetheless, our findings support that the risk of the development of PsA is modifiable and influenced by biologic class and timing. Given that IL-17 and IL-23 inhibitors are commonly used for severe PsO, these results are highly relevant to real-world practice. In this population-based cohort study, biologic class and treatment sequencing were associated with differences in the risk of incident PsA. First-line IL-23p19 inhibitors were associated with higher PsA-free survival, consistent with prior reports. Notably, in the second-line setting, IL-17 inhibitors were associated with a significantly lower risk of incident PsA compared with TNF inhibitors. These findings suggest that biologic class and sequencing are associated with differences in PsA risk; however, prospective studies are required to determine causality. At a conceptual level, these findings support the idea that PsA risk may be modifiable and influenced by immunological pathway targeting and treatment sequencing. Supplementary material