2045: Malignancy in Behçet’s Syndrome: A 24-Year Cohort Study In a Non-Endemicand Multiethnic Country

Background/Purpose: Behçet’s syndrome (BS) has been associated with an increased risk of malignancy, particularly in studies from endemic regions such as East Asia and the Middle East. Nonetheless, little is known about cancer risk in non-endemic, multiethnic populations. This study aimed to evaluate the frequency and profile of malignancies in a long-term BS cohort from South America.

Methods: We retrospectively conducted a data analysis of a single-center, multiethnic, prospective cohort of patients with BS followed between January 2000 and December 2024. Patients were diagnosed based on the International Study Group and/or the International Criteria for Behçet's Disease. Cancer occurrence was analyzed using descriptive statistics and compared across groups with chi-square, Fisher’s exact, and unpaired Student’s t-tests. Unadjusted odds ratios (OR) (95% CI) were calculated through bivariate logistic regression.

Results: A total of 302 BS patients were included, of whom 14 (4.6%; 95% CI: 2.3%–7.0%) were diagnosed with cancer. The cohort was predominantly female (n=186; 61.6%), with 8 women (57.1%) in the cancer group. The mean follow-up time was 160.3 months (±92.6) in the cancer group and 122.8 months (±80.3) in the non-cancer group. Patients who developed cancer had a mean age of 33.4 years (±8.6) at the time of BS diagnosis, compared to 31.4 years (±10.4) among those without cancer. Cancer was diagnosed 18(±13.1) years after BS diagnosis, at a mean age of 49,4(±17) years. The most common malignancies were endometrial and bladder cancer (2 cases each; 14.3%). Single cases (7.1% each) included cervical intraepithelial neoplasia, breast cancer, undifferentiated soft tissue neoplasm, adenocarcinomas of the lung, colon, prostate, and a metastatic tumor with unknown primary site; chronic myeloid leukemia, large granular lymphocyte leukemia, and central nervous system lymphoma. In univariate analysis, type 2 diabetes mellitus (DM2) was the only variable significantly associated with cancer occurrence (p = 0.004), with an OR of 5.8 (95% CI: 1.89–17.74). No statistically significant associations were found with sex, smoking, alcohol use, clinical manifestations, or immunosuppressive therapies, including cyclophosphamide. Due to the presence of only one significant predictor, a multivariate regression model could not be created.

Conclusion: In the present long-term cohort, overall malignancy rate was 4.6%, with hematologic malignancies being the most frequent type, which is consistent with findings from other BS cohorts and raises questions about underlying immunologic mechanisms or exposure to immunosuppressive agents. The average interval between BS diagnosis and cancer onset was 18 years, underscoring the need for long-term vigilance. Notably, DM2 was the only variable significantly associated with cancer, with a nearly sixfold increased risk. This finding reinforces the established link between metabolic comorbidities and carcinogenesis. No associations were found between malignancy and other variables. However, given the relatively small number of cancer cases, the study was underpowered to detect weaker associations.