This retrospective cohort study used de-identified electronic health record data from the TriNetX US Collaborative Network (2014–2026) to evaluate whether patients with treatment-defined moderate-to-severe chronic spontaneous urticaria (CSU) have a higher risk of developing immune-mediated comorbidities compared with matched patients with mild CSU. Data capture and analysis were performed on May 24, 2026. The study was exempt from institutional review board review due to use of de-identified data.
The analysis included adults (≥18 years) with at least two ICD-10 diagnoses of idiopathic urticaria (codes L50.1, L50.8, L50.9) recorded at least six weeks apart. The moderate-to-severe CSU cohort comprised patients treated with omalizumab, dupilumab, or cyclosporine; these therapies were selected as commonly used options for patients who fail second-generation antihistamines. The mild cohort comprised patients with the same diagnostic criteria who were never treated with biologics or cyclosporine. Patients with a diagnosis of atopic dermatitis were excluded to reduce potential confounding related to dupilumab use for that indication. ICD-10 and RxNorm codes used for cohort creation and matching were reported in the source article.
Patients with moderate-to-severe CSU were propensity score matched 1:1 to mild CSU patients using logistic regression and greedy nearest-neighbor matching with a caliper of 0.1 pooled SD. Matching variables included age, sex, race, ethnicity, baseline comorbid asthma and obesity, tobacco use, family history of selected autoimmune conditions, and socioeconomic factors related to insurance and income. After matching, each cohort contained 16,413 patients. The matched cohort had a mean age at diagnosis of 45.7 years and was predominantly female (77.6%). Racial and ethnic composition and other baseline characteristics after matching were reported in the source article.
The primary analysis calculated odds ratios (ORs) for new-onset immune-mediated comorbidities occurring at least one day after the index date (the date of the second idiopathic urticaria diagnosis). Outcomes present prior to the index date were excluded. Odds ratios were adjusted for demographic and clinical covariates and interpreted with Bonferroni-adjusted significance (alpha ≤ 0.00333).
After adjustment and multiplicity correction, the study found that treatment-defined moderate-to-severe CSU was associated with a statistically significant increased odds of developing rheumatoid arthritis (OR = 1.34; 95% CI: 1.12–1.60; p = 0.001) and Sjögren’s syndrome (source-reported OR and confidence interval reported in the original text; p < 0.0001). For the other immune-mediated conditions assessed—including autoimmune thyroiditis, vitiligo, pernicious anemia, systemic lupus erythematosus, celiac disease, Graves’ disease, Crohn’s disease, ulcerative colitis, scleroderma, fibromyalgia, type 1 diabetes mellitus, psoriatic arthritis, and psoriasis—no statistically significant differences in odds were observed between the moderate-to-severe and mild CSU cohorts.
Incidence rates and time-to-event analyses were reported in addition to ORs. The incidence rate for rheumatoid arthritis was 1.52 cases per 1000 person-years in the moderate-to-severe cohort versus 1.14 per 1000 person-years in the mild cohort (incidence rate ratio 1.33; 95% CI: 1.12–1.60; p = 0.0014). For Sjögren’s syndrome the incidence rate was 1.20 per 1000 person-years in the moderate-to-severe cohort versus 0.59 per 1000 person-years in the mild cohort (incidence rate ratio 2.03; 95% CI: 1.62–2.53; p < 0.0001).
Cumulative incidence over the study period showed 285 patients (1.8%) in the moderate-to-severe cohort developed rheumatoid arthritis versus 217 patients (1.4%) in the mild cohort (p = 0.0013). For Sjögren’s syndrome, 226 patients (1.4%) in the moderate-to-severe cohort developed the condition versus 115 patients (0.7%) in the mild cohort (p < 0.0001). The absolute risk differences reported were a 0.5% increased absolute risk for rheumatoid arthritis and a 0.7% increased absolute risk for Sjögren’s syndrome in the moderate-to-severe group compared with mild CSU.
The authors note that prior literature supports possible shared autoimmune and genetic pathways between CSU and rheumatoid arthritis, including increased prevalence of rheumatoid factor and antinuclear antibodies in CSU patients and genetic associations involving loci such as PTPN22 and polygenic risk score analyses. The findings in this study are presented as concordant with those prior genetic and serologic observations and suggest that more severe, treatment-defined CSU correlates with higher subsequent risk of rheumatoid arthritis.
For Sjögren’s syndrome, the authors report more limited prior evidence and hypothesize shared immune-mediated mechanisms could underlie the observed association; they emphasize that further research is needed to clarify biological pathways.
The absence of a significant association between moderate-to-severe CSU and several other autoimmune disorders led the authors to suggest that not all immune-mediated conditions share the same pathophysiology with CSU and that specific subtypes of CSU (for example type IIb) may be linked to particular autoimmune diseases such as autoimmune thyroid disease in prior studies.
The study authors acknowledge several limitations inherent to use of large EHR datasets and to the treatment-defined severity approach used here:
Treatment type was used as a proxy for disease severity because validated CSU activity measures were not available in TriNetX; this can create misclassification because treatment decisions may be influenced by access, insurance, prescribing patterns, and off-label use. The database does not provide medication indication, which can further confound classification.
Detection bias is possible: patients with more severe CSU may have greater healthcare utilization and therefore increased opportunity for diagnosis of comorbidities.
Use of ICD-10 coding in EHR data carries inherent risks of misclassification of diagnoses.
The authors emphasize that findings should be interpreted as associations with treatment-defined moderate-to-severe CSU rather than direct measures of clinical severity.
Within the constraints of the dataset and methods, the analysis suggests that patients with treatment-defined moderate-to-severe CSU may warrant closer interdisciplinary surveillance for the development of rheumatoid arthritis and Sjögren’s syndrome. The authors recommend clinicians consider longitudinal monitoring for these comorbidities among patients with more severe treatment-defined CSU and call for further research to clarify mechanisms and to validate findings with additional data sources and measures of CSU activity.