Venous thromboembolic events and organ-specific occult cancers: a review and meta-analysis

Background: Despite a recognized association between venous thromboembolic events (VTE) and cancer, little is known about the strength and the features of this association. We performed a meta-analysis in order to clarify this issue. Methods: We retrieved data from 40 reports published between 1982 and 2007: 12 contained cancer risk estimates for patients with either idiopathic or secondary VTE vs. subjects without VTE and 17 for patients with idiopathic vs. secondary VTE. We also pooled risk estimates from four cohort studies to assess the association between VTE and specific forms of cancer and conducted a proportional incidence study, based on the remaining 28 reports, which did not provide risk estimates. Results: The pooled relative risk (RR) of cancer was 3.2 [95% confidence interval (95% CI) 2.4-4.5] for patients with any form of VTE vs. no VTE, 2.7 (95% CI 1.9-3.9) for patients with idiopathic vs. no VTE and 3.8 (95% CI 2.6-5.4) for patients with idiopathic vs. secondary VTE. In the pooled cohort studies, RRs for VTE vs. no VTE were significantly elevated for cancers of the ovary (RR 7.0), pancreas (RR 6.1), liver (RR 5.6), blood (4.2), brain (RR 3.8), kidney (RR 3.4), lung (3.1), colon (2.9), and esophagus (2.1). In the proportional incidence study, cancers of the pancreas, colon, and blood were significantly more frequently observed than in the general population. Conclusions: Overall we found a 3-fold excess risk of occult cancer in patients with VTE. The risk varies according to tumor site and is highest for cancers of the ovary, pancreas, and liver.