Retrospective survey of unselected hospital patients with and without cancer comparing outcomes following venous thromboembolism

Background: Cancer patients are at risk of venous thromboembolism (VTE). Currently, management of these patients is based on studies performed in the general population. Aims: To determine whether the natural history of VTE in cancer patients differs sufficiently from patients in general such that a specific management approach is required. Methods: A retrospective survey was conducted using case records from three tertiary referral hospitals (1993?2001). One hundred and forty‐seven VTE patients with cancer and 82 control patients matched for age and gender were followed for a median of 24 months (0?104 months). Results: The overall prevalence of cancer among patients with VTE was 19.3% (vs 4% in patients with arterial thromboembolic events). In the follow‐up groups, cancer conferred a significantly worse functional status, although other risk factors for VTE were similar. Proximal sites of VTE were twice as common in cancer as in non‐cancer patients. The majority of patients in both groups were treated with oral anti­coagulants. Complications were more common in cancer patients (40%vs 23%) and occurred more frequently despite apparently adequate anticoagulation. Following VTE, cancer patient survival (median 4.1 months, 127 deaths) was markedly worse than non‐cancer patients (median not reached, eight deaths). Thirty‐five cancer deaths occurred within 5 weeks of VTE. Eleven deaths in cancer patients were directly related to VTE compared to three in patients without cancer. Conclusions: Cancer patients with VTE fare extremely poorly compared to the general hospital population. Prospective treatment studies, specifically in cancer patients with VTE, are needed. (Intern Med J 2002; 32: 437?444)