Background <p>Case reports suggest an increased risk of both venous thromboembolism and arterial thromboembolism in patients with anorexia nervosa (AN), even in the absence of known risk factors. Population-based studies on thromboembolism in patients with AN are missing. This study aims to establish the association in a registry-based nation-wide data set from Denmark.</p> Methods <p>This nationwide, observational registry-based cohort study used mandatory health care registries to identify the risk of thromboembolism in a cohort of patients with AN and in population controls matched on date of birth, area of residence, and sex. Thromboembolic outcomes were validated using specific criteria. The two groups were compared using Fine-Gray regression models.</p> Results <p>In total, 12,676 patients (mean age 18.3 and 93.5% female) were matched with 63,049 controls. Median follow-up time was 9.5&#xa0;years. For venous thromboembolism, the cumulative risk during the full follow-up period was 0.88% (<i>n</i> = 112) in the AN group and 0.12% (<i>n</i> = 79) in the control group with a subhazard ratio of 7.1 [5.3–9.5]. For arterial thromboembolism, the cumulative risk during the full follow-up period was 0.54% (<i>n</i> = 68) in the AN group and 0.05% (<i>n</i> = 36) in the control group (subhazard ratio 9.5 [6.4–14.1]).</p> Conclusions <p>Patients with AN have a substantially increased risk of both arterial and venous thromboembolism, compared to the background population.</p>