Whether the high incidence of venous thromboembolism (VTE) in the elderly can be attributed to cancer is not well studied. We assessed the impact of cancer on risk of VTE in young, middle-aged and elderly. 26,094 subjects without a history of cancer or VTE were recruited from the Tromsø study. Incident cancer (n = 2,290) and VTE (n = 531) were recorded from baseline (1994-1995) through December 31st, 2009. Cox regression with cancer as time-varying exposure was used to calculate hazard ratios with 95 % confidence intervals (CI). Overt cancer was associated with a fivefold (95 %CI 4.3, 6.7) increased risk of VTE, with an age-dependent gradient from 26-fold (95 %CI 12.1, 56.5) increased in the young, ninefold (95 % CI 6.6, 12.7) increased in the middle-aged, and threefold (95 % CI 2.5, 4.5) increased risk in the elderly. The population attributable risks were 14, 27 and 18 %, respectively.
Conclusion: The relative risk of VTE by cancer were higher in young compared to elderly subjects, but the proportion of VTEs in the population due to cancer did not differ much across age groups. Our findings indicate that the increased risk of VTE by advancing age cannot be attributed to higher incidence of cancer in the elderly.