Seasonal and geographic variation in cutaneous immune‐related adverse events after treatment with immune checkpoint inhibitors

Immune checkpoint inhibitors (ICIs) have revolutionized cancer therapy. Up to 40% of patients receiving ICI treatment develop cutaneous immune-related adverse events (cirAEs),1 which may server as prognostic indicators of therapeutic response.2, 3 Identifying cirAE risk factors can thus help stratify ICI candidates at the highest risk of toxicity and shed light on immune response mechanisms during ICI treatment. This study aims to investigate the seasonal and regional variability in cirAE development using a multi-centre cohort of ICI recipients. We used the TriNetX Dataworks Network, which provides deidentified data on over 90 million US patients. Cancer patients who received ICI therapy between January 2010 and December 2019 were selected. To identify cirAEs, we compiled a list of International Classification of Diseases 10th Revision (ICD-10) codes corresponding to potential ICI-induced cutaneous toxicities, informed by existing literature and expert consensus.4 New diagnoses such as psoriasis, rash, vitiligo, drug hypersensitivity, eczema, erythema multiforme, Stevens–Johnson syndrome, toxic epidermal necrolysis, lichen planus, bullous dermatosis and pruritus were considered cirAEs if they occurred more than once after ICI initiation. Diagnoses were excluded if recorded ≥2 times within 6 months before ICI initiation or within 3 months of a chemotherapy. We built competing risks Fine–Gray and Prais–Winsten regression models to test the association between cirAE risk and the season of ICI initiation, controlling for age, sex, race and ethnicity, ICI target, ICI year, cancer type, regional variation and autocorrelation expected in time series data. The ICI cohort comprised 15,253 patients between January 2010 and December 2019, of whom 2413 patients (15.8%) developed cirAEs. The rate of cirAEs was highest in the fall (43%), followed by the spring (21%), winter (19%) and summer (17%) seasons. Using competing risks and Prais–Winsten models, the risk of cirAE was significantly associated with starting ICI in the winter (β, 0.006; 95% CI, 0.001–0.012, p = 0.013) (Table 1). This seasonality was robust to sensitivity analysis after excluding all diagnoses of eczema which often flares in the winter (β, 0.005; 95% CI, 0.003–0.008, p < 0.001). Analysis of geographic variation in cirAE development in the United States revealed the highest rate in the West (35%), followed by the South (29%), Northeast (21%) and Midwest (16%). Multivariate competing risks analysis showed significantly increased risk of cirAE development in the west (HR, 2.42; 95% CI, 2.18–2.68; p < 0.001), while there was decreased cirAE risk in the northeast (HR, 0.87; 95% CI, 0.77–0.99; p = 0.029) (Figure 1). Overall, cirAE incidence varied by season, with peak incidence in fall and a heightened risk associated with patients initiating ICI treatment in winter. It is possible that these trends align with variations in pathogen exposure. For example, cold viruses such as respiratory syncytial virus and human metapneumovirus are more prevalent during these seasons, potentially contributing to immune dysregulation and cirAE susceptibility.5 Additionally, decreased vitamin D exposure during winter, and seasonal fluctuations in CD4 T-cell expansion and regulatory T-cell levels, could partially explain these patterns.6 Similarly, geographic variation in pathogen landscapes and vitamin D exposure may influence immune responses and contribute to cirAE pathogenesis. This study highlights seasonal and regional variability in cirAE development, suggesting the importance of incorporating temporal considerations into risk assessment, surveillance and management. Future work needs to similarly explore large-scale seasonal and geographic patterns in patient survival after ICI, as recent work demonstrates seasonality in outcomes after ICI treatment for BRAF wild-type melanoma living in Denmark.7 Limitations of this study include its retrospective design, the broad region definitions in TriNetX to preserve deidentification and the use of ICD-10 codes to identify cirAEs which limits the identification of some uncoded cutaneous eruptions, although this approach previously showed an accuracy of 87%.4 YRS is supported by the Department of Defense, under award number W81XWH2110819, by the Dermatology Foundation and the National Institute of Health K23AR080791-01A1. The authors have no conflicts of interest to declare. This study utilized de-identified data from a multi-institutional registry and was deemed exempt from Institutional Review Board review, as it did not involve human subjects research or the collection of identifiable patient information. The data that support the findings of this study are available from the corresponding author upon reasonable request.