Background Hydroxychloroquine is a disease-modifying anti-rheumatic drug for patients with Sjögren's syndrome who have an increased risk of malignancies, particularly lymphoma. Inhibition of systemic inflammation and autophagy by hydroxychloroquine has been linked to antineoplastic effects but there is a lack of evidence on the effect of hydroxychloroquine on the risk of cancers in patients with Sjögren's syndrome.
Objectives This study aimed to explore the effect of hydroxychloroquine on the risk of cancers in patients with Sjögren's syndrome.
Methods We compared cancer incidence in incident patients with Sjögren's syndrome who received hydroxychloroquine for at least 6 months within one year or three years to those did not using the Taiwan National Health Insurance Research Database. Landmark analysis was used to account for immortal time bias and propensity score matching was used to control for potential effects of known confounders.
Results Of 4,262 Sjögren's syndrome patients was identified between 2000 to 2005, we included 4,175 patients who were alive and did not have cancer at one year from initial diagnosis of Sjögren's syndrome. The propensity score matched 1,141 patients who exposed to 6 months of hydroxychloroquine prescription on the date 1 year from the initial diagnosis and 1,141 patients who was not exposed to hydroxychloroquine. The median follow year was 6 years from one-year landmark. There were 71 hydroxychloroquine users and 56 non-hydroxychloroquine users had cancer during the follow-up. The hazard ratio for cancer was 1.31 (95% confidence interval [CI], 0.92–1.86) in the one year landmark analysis. We further conducted three-year landmark analysis and 6-month hydroxychloroquine exposure was associated with a hazard ratio of 1.33 (95% CI, 0.98–1.81) in the three-year landmark analysis between hydroxychloroquine users and non-hydroxychloroquine users, was 1.31 (0.92–1.86) in the one landmark analysis. There was no difference in overall survival based on Kaplan-Meier estimates between hydroxychloroquine user and non-users in both the one-year (log-rank test p=0.15) and three-year landmark analysis (log-rank test p=0.09).
Conclusions This propensity score matched landmark analysis in a population of incident patients with Sjögren's syndrome in Taiwan found a neutral effect of 6-month hydroxychloroquine treatment early in the clinical course on the risk of cancer, but whether a higher dose or an longer exposure to hydroxychloroquine could influence the risk of cancer remains to be determined.
Disclosure of Interest None declared