Venous thromboembolism during active disease and remission in inflammatory bowel disease: a cohort study

Background Patients with inflammatory bowel disease who develop deep vein thrombosis or pulmonary embolism often have active disease at the time of thromboembolism. We therefore aimed to quantify the risk of venous thromboembolism prospectively during different activity phases of inflammatory bowel disease. Methods From the General Practice Research Database, we matched patients with prospectively recorded inflammatory bowel disease from November, 1987, until July, 2001 with up to five controls by age, sex, and general practice. A flare was defined as the period 120 days after a new corticosteroid prescription. We used Cox regression analysis with time-varying covariates to accommodate changes in the state of inflammatory bowel disease, and whether patients were at high risk of venous thromboembolism after hospitalisation. Findings 13 756 patients with inflammatory bowel disease and 71672 matched controls were included in the analysis, and of these 139 patients and 165 controls developed venous thromboembolism. Overall, patients with inflammatory bowel disease had a higher risk of venous thromboembolism than did controls (hazard ratio 3.4, 95% CI 2.7-4.3; p<0.0001; absolute risk 2.6 per 1000 per person-years). At the time of a flare, however, this increase in risk was much more prominent (8.4, 5.5-12.8; p<0.0001; 9.0 per 1000 person-years). This relative risk at the time of a flare was higher during non-hospitailised periods (15.8, 9.8-25.5; p<0.0001; 6.4 per 1000 person-years) than during hospitalised periods (3.2, 1.7-6.3; p=0.0006; 37.5 per 1000 person-years). Interpretation Trials of primary prophylaxis of venous thromboembolism are warranted to find out whether this important complication can be prevented.