Aims To investigate the association between incident retinal vein occlusion (RVO) and the subsequent development of cancer.

Methods In this nationwide population-based retrospective study using 2002–2013 National Health Insurance Service database which covers the entire South Korean population, 186 701 incident RVO patients and their 1:1 propensity-score matched controls were included. We defined the fixed cohort from January 1st, 2004 to December 31st, 2013; the cohort included patients who suffered incident RVO after entering the cohort and their matched controls, and excluded patients having any cancer history before entering the cohort. The association of RVO and cancer was assessed by time-varying covariate Cox regression models; Model 1 included RVO as a time-varying covariate, Model 2 included Model 1 plus demographic information and Model 3 included Model 2 and comorbidities.

Results RVO was associated with an increased risk of subsequent cancer (HR=1.29; 95% CI, 1.26–1.31 in Model 1), which was consistent in Models 2 and 3. The incidence rate of overall cancer during the study period was 25.55 (95% CI, 25.19–25.91) per 1000 person-years in the RVO group and 18.62 (95% CI, 18.46–18.79) per 1000 person-years in the control group. In the subgroup analysis, haematological malignancies showed the highest association with RVO (HR=1.65; 95% CI, 1.49–1.83).

Conclusion Patients with RVO have an increased risk of subsequent cancer development even after adjusting for demographic factors and comorbidities. Further study is warranted to elucidate these associations to provide proper recommendations for RVO patients regarding the cancer screening.