Impact of time since diagnosis and mortality rate on cancer-associated venous thromboembolism - the Scandinavian Thrombosis and Cancer (STAC) cohort.

Background: Venous thromboembolism (VTE) is a common complication in cancer, and studies have suggested that aggressive cancers create the highest risk of VTE. However, competing risk by death may result in overestimation of VTE risk in patients with cancers associated with high mortality. Therefore, we estimated the risk of VTE by cancer site, accounting for the differential mortality between cancers. Methods: The Scandinavian Thrombosis and Cancer cohort included 144 952 participants followed from 1993-1997 to 2008-2012. Incidence rates, cause-specific hazard ratios (HRs) and subdistribution HRs (SHRs) were assessed for overall cancer and by cancer site according to time intervals since cancer diagnosis. Results: During follow-up, 14 272 subjects developed cancer, and 567 had cancer-related VTE. In cause-specific analyses, the VTE risk was highest in the first 6 months after cancer diagnosis (HR 17.5, 95% confidence interval [CI] 15.1-20.3), and declined rapidly thereafter. However, when mortality was taken into account, the risk was similar in the periods 6 months before (SHR 4.8, 95% CI 3.6-6.4) and 6 months after (SHR 4.6, 95% CI 3.9-5.4) cancer diagnosis. The range of the 2-year cumulative VTE incidence rates was substantially narrowed for all cancer sites after competing risk by death was taken into account (from 1-10% to 1-4%). Conclusion: VTE risk by cancer site was influenced by the mortality rate and the time since cancer diagnosis. Our findings suggest that the cancer itself is a major contributor to VTE risk, and that competing risk by death should be taken into account when VTE risk in cancer is explored.