Introduction
Inflammatory bowel disease (IBD), comprising Crohn’s disease and ulcerative colitis, is a chronic relapsing-remitting disorder that causes progressive tissue injury, complications, and impaired quality of life. Hidradenitis suppurativa (HS) is strongly associated with IBD: patients with HS have 2.1-2.5-fold higher odds of Crohn’s disease and 1.5-1.7-fold higher odds of ulcerative colitis than controls, with pooled HS prevalence of 17.3% in Crohn’s disease and 8.5% in ulcerative colitis.1–7 Concomitant HS is linked to more severe IBD activity, perianal disease, and increased body mass index.5–6 While extraintestinal manifestations (EIMs) affect up to one-quarter of IBD patients, HS functions as a dermatologic comorbidity that often parallels IBD activity.8 The bidirectional association between IBD and HS highlights the need for integrated care.5–6
Anti-tumor necrosis factor (TNF) agents, particularly adalimumab, remain first-line biologics for IBD with HS. Adalimumab is FDA-approved for both, while infliximab is also recommended for HS though dosing is less defined.9,10 Limitations include primary non-response, secondary loss of response, and paradoxical HS induction, especially with infliximab and etanercept.9–11 Interleukin-12/23 inhibitors, such as ustekinumab, target the Th1/Th17 pathway, while IL-23–selective inhibitors—including risankizumab, guselkumab, and mirikizumab—act specifically on the p19 subunit of IL-23 to modulate downstream Th17-mediated inflammation. Ustekinumab, approved for IBD, has shown modest HS benefits in small observational studies,9,12,13 whereas evidence for IL-23 inhibitors in HS remains limited; available data demonstrate inconsistent HiSCR responses and only occasional quality-of-life improvements, with insufficient support for reliable clinical efficacy of risankizumab in HS.12 Guidelines emphasize the need for larger studies.9 Janus kinase (JAK) inhibitors (tofacitinib, upadacitinib) are effective in IBD and show emerging HS efficacy, achieving Hidradenitis Suppurativa Clinical Response (HiSCR) and reducing abscesses and nodules in small cohorts,14,15 though safety concerns include infection, thromboembolism, and cardiovascular events.15,16
Direct comparative evidence between IL-12/23 or IL-23 inhibitor therapies and JAK inhibitors in patients with both IBD and HS is lacking. RCTs rarely include this comorbidity, and available subgroup data are underpowered. Real-world evidence on hospitalization, corticosteroid use, complications, and antibiotic exposure remains limited.
To address this gap, we compared IL-12/23 or IL-23 inhibitor therapies and JAK inhibitors in patients with IBD and HS using a large federated electronic health record (EHR) network. This analysis provides clinically actionable evidence to guide treatment selection in patients who have failed or cannot tolerate anti-TNF therapy.
Methods
Data source and study population
We conducted a retrospective cohort study using a federated research platform that aggregates de-identified EHR data from 155 healthcare organizations worldwide. The database includes diagnoses, procedures, medications, laboratory values, and demographic information from academic medical centers, specialty clinics, and community hospitals.
Eligible patients were adults (≥16 years) with a diagnosis of Crohn’s disease or ulcerative colitis, identified using ICD-10-CM codes, and a concurrent diagnosis of HS. Patients were included if they had documented exposure to either an IL-12/23 or IL-23 inhibitor (ustekinumab, risankizumab, guselkumab, mirikizumab) or a JAK inhibitor (tofacitinib, upadacitinib) during the study period. Exclusion criteria were applied to maintain mutually exclusive treatment groups: patients entering the IL-12/23 or IL-23 inhibitor cohort were excluded if they had any prior or concurrent exposure to a JAK inhibitor, and patients entering the JAK inhibitor cohort were excluded if they had any prior or concurrent exposure to an IL-12/23 or IL-23 inhibitor.
In addition, patients who’s qualifying IBD or HS diagnosis were first recorded more than 20 years before the index date were excluded to limit inclusion of remote or potentially miscoded diagnoses.
Detailed ICD-10-CM, RxNorm, and procedure codes used to define cohorts are provided in Table S1.
Exposure and comparator definitions
The exposure of interest was treatment with an interleukin (IL)-12/23 or IL-23 inhibitor, defined as receipt of ustekinumab (IL-12/23) or risankizumab, guselkumab, or mirikizumab (IL-23), without prior or concurrent use of JAK inhibitor. The comparator group consisted of patients treated with a JAK inhibitor, including tofacitinib or upadacitinib, without prior or concurrent use of any IL-12/23 or IL-23 inhibitors.
In accordance with the query criteria, all patients were required to have prior exposure to an anti-TNF agent, most commonly adalimumab or infliximab, before initiation of IL-12/23, IL-23, or JAK therapy. This ensured that both cohorts reflected real-world treatment sequencing after anti-TNF therapy.
The index date was defined as the first prescription or administration of the qualifying IL-12/23, IL-23, or JAK inhibitor.
Medication definitions were based on RxNorm and procedure codes, detailed in Table S1.
Outcomes of interest
The primary outcomes were:
Secondary outcomes included:
Detailed outcome definitions, including ICD-10-CM, CPT, and RxNorm codes, are provided in Table S2.
All outcomes were assessed during a follow-up window of 30-365 days after the index date. The first 30 days were excluded a priori to allow for a therapeutic onset (induction) period and to reduce protopathic bias, whereby events occurring immediately after treatment initiation may reflect preexisting disease activity or the clinical circumstances that prompted the treatment switch rather than the response to the newly initiated therapy. The 365-day upper limit was chosen to capture clinically meaningful medium-term outcomes while limiting loss to follow-up and subsequent treatment switching.
Patients who had a given outcome recorded prior to the index date were excluded from the corresponding outcome analysis to ensure accurate attribution to post-index treatment exposure.
Baseline covariates used for propensity score matching (PSM) comprised demographic characteristics—age at index; sex (male, female); race (White, Black or African American, Asian); and ethnicity (Hispanic or Latino, not Hispanic or Latino, unknown)—together with baseline systemic glucocorticoid exposure. Prior anti–TNF exposure was an inclusion requirement for both cohorts and therefore did not require matching.
To minimize confounding, we performed 1:1 PSM between the IL-12/23 or IL-23–treated cohort and the JAK inhibitor cohort. Matching was conducted using nearest-neighbor methodology without replacement, based on the logit of the propensity score. Covariate balance was assessed using standardized mean differences (SMD), with values <0.1 indicating adequate balance.
A complete list of covariates and baseline characteristics before and after matching is provided in Table S3.
Statistical analysis
Comparative analyses were performed on the propensity score–matched cohorts. The risk ratio (RR) was prespecified as the primary effect measure for all outcomes. Odds ratios (OR) and hazard ratios (HR) were estimated as supportive analyses to assess the consistency of findings across analytic approaches and are presented in Table S4.
Risk estimates were derived from matched risk comparisons, while ORs and HRs were derived from logistic regression and Cox proportional hazards models, respectively, using embedded analytic functions. All estimates are reported with 95% confidence intervals (CI), and statistical significance was defined as a two-tailed P-value <.05. Given the multiple outcomes examined and the observational design, results are framed primarily in terms of effect size and precision rather than statistical significance alone.
Patients with an outcome documented prior to the index date were removed from the corresponding analysis to ensure correct temporal attribution. All analyses were performed within the platform, which uses de-identified aggregate data and validated statistical procedures; patient-level data are not accessible.
Ethical considerations
This study was conducted using de-identified EHR data in accordance with applicable data protection regulations. All data were anonymized prior to analysis, and only aggregated results were available to investigators. Because no identifiable patient information was accessed, the study was exempt from institutional review board approval and informed consent requirements.
Additional regulatory details are withheld for blinded peer review.
Results
Cohort characteristics
A total of 300 patients met inclusion criteria after 1:1 PSM, with 150 patients in the IL-12/23 or IL-23 inhibitor cohort and 150 in the JAK inhibitor cohort. Baseline characteristics were well balanced between groups, with a mean age of 38 years and 75.3% female representation (Table S3).
Primary outcomes
Within the 30-365-day outcome window, rates of ED visits or hospitalization were similar between groups, occurring in 28.0% of IL-12/23 or IL-23 users and 23.3% of JAK inhibitor users (RR 1.20; 95% CI, 0.82-1.77; P = .355) (Table 1). Corticosteroid use was numerically higher among IL-12/23 or IL-23 users than JAK inhibitor users (61.3% vs 51.3%), but this difference did not reach statistical significance (RR 1.20; 95% CI, 0.98-1.46; P = .081).
Secondary outcomes
Rates of IBD-related complications were low and comparable between groups, occurring in 12.7% of IL-12/23 or IL-23 users and 12.0% of JAK users (RR 1.06; 95% CI, 0.58-1.93; P = .861). Overall systemic antibiotic use was also comparable, occurring in 60.0% of IL-12/23 or IL-23 users versus 52.7% of JAK inhibitor users (RR 1.14; 95% CI, 0.93-1.39; P = .200). In contrast, HS-specific antibiotic use was higher in the IL-12/23 or IL-23 cohort, occurring in 48.0% of patients compared with 32.7% in the JAK inhibitor cohort (RR 1.47; 95% CI, 1.11-1.95; P = .007).
This finding should be interpreted with caution and is discussed further below. Consistent OR and HRs for all outcomes are provided in Table S4.
Discussion
In this multicenter real-world analysis of patients with concomitant inflammatory bowel disease and HS, treatment with IL-12/23 or IL-23 inhibitors and treatment with JAK inhibitors resulted in comparable rates of hospitalization and IBD-related complications within the 30-365-day outcome window. Corticosteroid use was numerically higher among IL-12/23 or IL-23 users but did not reach statistical significance, whereas HS-specific antibiotic use was more frequent in this cohort. Although antibiotic utilization has been proposed as a more sensitive marker of HS disease activity and treatment burden than acute care utilization—particularly in patients with chronic, fluctuating disease managed largely in the outpatient setting—it remains an indirect proxy.
Antibiotic prescribing is also shaped by maintenance and prophylactic strategies, institutional protocols, and individual physician preference, and these data should therefore be regarded as hypothesis-generating.
Our findings—showing comparable risks of hospitalization and IBD-related complications between IL-12/23 or IL-23 inhibitors and JAK inhibitors, alongside higher corticosteroid and HS-directed antibiotic use in the IL-12/23 or IL-23 cohort—align with, and further contextualize, the existing literature across IBD and HS. RCTs have demonstrated that ustekinumab, risankizumab, guselkumab, and mirikizumab are effective for induction and maintenance therapy in both Crohn’s disease and ulcerative colitis. In Crohn’s disease, risankizumab has also outperformed ustekinumab in head-to-head endoscopic outcomes, supporting the use of IL-12/23 or IL-23 inhibitors as viable post–anti-TNF strategies in IBD, although without HS-specific enrollment.17–20 JAK inhibitors, tofacitinib approved for UC and upadacitinib approved for CD and UC, similarly show strong efficacy in moderate-to-severe IBD, though their use is shaped by safety concerns—including zoster, thromboembolism, and MACE—that require thoughtful patient selection and are emphasized in current clinical guidelines.21–24 In HS, adalimumab remains the guideline-endorsed first-line biologic with consistent RCT support, whereas early phase and systematic data indicate that JAK inhibitors may improve cutaneous disease activity.
In contrast, IL-23 inhibitors (guselkumab, risankizumab, mirikizumab) generally yield limited or inconsistent HiSCR responses, despite occasional quality-of-life benefits.25–33 Importantly, no head-to-head comparisons exist for patients with both IBD and HS, necessitating extrapolation from single-disease trials. By directly comparing IL-12/23 or IL-23 inhibitors with JAK inhibitors in a real-world cohort of patients with concurrent IBD and HS, our study addresses this gap and provides clinically applicable evidence supporting therapeutic sequencing after anti-TNF failure.17–33
Both IBD and HS share dysregulated immune pathways, particularly Th1/Th17 activation and JAK–signal transducer and activator of transcription (STAT)-mediated cytokine signaling, providing a mechanistic rationale for targeting these axes.34–37 IL-12/23 inhibitors modulate the Th1/Th17 pathway and are effective in IBD, but their slower onset of action and inconsistent cutaneous response often lead to interim use of corticosteroids or antibiotics during HS flares.38–39 In contrast, JAK inhibitors exert broader and more rapid immunomodulatory effects by blocking multiple cytokine signals, a mechanism that may account for their steroid-sparing potential despite overall complication rates similar to IL-12/23 therapy.37,40 These mechanistic differences carry important clinical implications for treatment sequencing after anti-TNF failure: IL-12/23 agents remain appropriate for patients prioritizing long-term safety, whereas JAK inhibitors may be preferred in individuals with severe or relapsing disease where rapid control and reduction of corticosteroid exposure are essential.38,40 HS-specific antibiotic use was selected a priori as a disease-relevant secondary outcome given its frequent use as a marker of HS flares and treatment escalation. In the absence of validated severity measures, however, antibiotic prescriptions remain an indirect surrogate, and the higher HS-specific antibiotic use observed in the IL-12/23 or IL-23 cohort should be interpreted as hypothesis-generating rather than as confirmatory evidence of poorer HS disease control.
From a clinical perspective, our findings demonstrate comparable rates of hospitalization and IBD-related complications between IL-12/23 or IL-23 inhibitors and JAK inhibitors in patients with concomitant IBD and HS. In contrast, HS-specific antibiotic use was more frequent among patients treated with IL-12/23 or IL-23 inhibitors, suggesting potential differences in HS disease control or treatment escalation between therapeutic classes; however, because antibiotic prescribing may also reflect institutional standard-operating procedures, maintenance or prophylactic regimens, and dermatologic practice patterns rather than true disease activity, this difference should be interpreted with caution and confirmed in studies incorporating validated HS severity measures. In the absence of clinically meaningful differences in acute care utilization or IBD-related outcomes, therapeutic decision-making in this population should be guided by individualized considerations, including comorbidities, safety profiles, route of administration, and patient preferences.
Collectively, these findings reinforce the importance of a tailored, multidisciplinary treatment strategy—integrating gastroenterology and dermatology—to optimize control of both intestinal and cutaneous disease manifestations.
This study has several notable strengths. It leverages a large, multicenter, real-world dataset, improving generalizability across diverse healthcare settings and patient populations, and applies to rigorous PSM that achieves well-balanced baseline characteristics between treatment groups. Several limitations should nonetheless be acknowledged.
First, the analysis relied on coded administrative outcomes—hospitalization, corticosteroid use, and antibiotic prescriptions—that serve only as indirect proxies for disease activity. The dataset does not contain validated severity or activity measures, such as Hurley stage or HiSCR response for HS or validated clinical and endoscopic activity indices for IBD, and lacks information on the clinical presentation and disease severity that preceded initiation of IL-12/23, IL-23, or JAK therapy. Consequently, the observed differences—particularly in antibiotic and corticosteroid use—may reflect physician practice patterns, institutional protocols, or management strategies rather than true differences in disease control.
Second, the exposure group combined agents with distinct mechanisms of action, onset of action, and levels of HS-specific evidence (ustekinumab, an IL-12/23 inhibitor, and the IL-23–selective inhibitors risankizumab, guselkumab, and mirikizumab); ustekinumab has some observational support in HS, whereas evidence for IL-23–selective agents is more limited and inconsistent. Pooling these therapies may obscure clinically meaningful between-agent differences, and the comparator group similarly combined the JAK inhibitors tofacitinib and upadacitinib. In addition, mutually exclusive per-agent counts could not be derived from the aggregate data platform after matching, precluding a quantitative breakdown by individual agent.
Third, although IBD comprises Crohn’s disease and ulcerative colitis, which differ in phenotype and treatment response, the primary analysis combined both subtypes; subtype-specific estimates were constrained by sample size and should be examined in larger cohorts. Fourth, despite PSM, residual and unmeasured confounding cannot be excluded. Matching in this study was limited to demographic characteristics and baseline systemic glucocorticoid exposure; clinically important factors—including IBD subtype, disease duration and severity, HS severity (eg, Hurley stage), perianal disease, body-mass index, smoking status, and other comorbidities—were not incorporated into the propensity model and therefore could not be balanced.
Because treatment selection in this population is also likely influenced by these and other unrecorded factors, as well as by safety considerations and multidisciplinary decision-making between gastroenterology and dermatology, the possibility of confounding by indication is substantial. The relatively small JAK inhibitor cohort available for matching may also have limited statistical power for some comparisons. Finally, reliance on ICD-10-CM and RxNorm coding introduces potential misclassification, and because the network provides only aggregate-level, de-identified data, patient-level analyses are not possible.
For these reasons, the findings should be regarded as associative and hypothesis-generating rather than causal.
In summary, IL-12/23 or IL-23 inhibitors and JAK inhibitors demonstrated comparable short-term safety and complication profiles in patients living with both IBD and HS, with meaningful differences in adjunctive treatment patterns—particularly HS-specific antibiotic use. This observation is hypothesis-generating and suggests a potential differential effect on HS disease control that merits further investigation. These findings offer pragmatic, real-world insights to guide therapeutic sequencing after anti-TNF failure in a population often underrepresented in clinical trials.
Future research should prioritize prospective studies and randomized controlled trials directly comparing IL-12/23, IL-23, and JAK inhibitors in patients with dual IBD–HS diagnoses, as this underrepresented population requires dedicated evaluation. Biomarker-driven approaches also warrant investigation to improve patient selection and refine treatment sequencing. Furthermore, integrated gastroenterology–dermatology care models will be essential to manage the complex and overlapping disease burden in these patients and to optimize both intestinal and cutaneous outcomes.
Together, these efforts will help advance personalized, multidisciplinary care for this challenging patient population.
Supplementary Material