Trousseau's syndrome - what is the evidence? A population-based autopsy study.

Despite numerous studies documenting the association between cancer and venous thromboembolism (VTE), the reason for the excessive risk in certain cancers remains obscure. No large-scale studies have yet investigated the independent effects of cancer type, site and growth pattern. Between 1970 and 1982, 23,796 standardised autopsies were performed, representing 84% of all in-hospital deaths in an urban Swedish population. The relationship between cancer and PE was evaluated with logistic regression. The overall PE prevalence was 23%, and 10% of the population had a fatal PE. Forty-two per cent of pancreatic cancer patients had PE (OR 2.55; 95% CI 2.10-3.09) (p<0.001); gall bladder, gastric, colorectal and pulmonary adenocarcinomas were similarly independently associated with PE. In comparison with squamous cell lung cancer, patients with pulmonary adenocarcinoma had 1.65 times higher odds for PE (95% CI 1.20-2.29). Adenocarcinoma and metastatic cancer were independently associated with PE risk (OR 1.27; 95% CI 1.16-1.40; p<0.001, and OR 1.10;95% CI 1.01-1.20; p=0.024, respectively) but when controlling for cancer type and spread, pancreatic cancer was still associated with an OR of 2.10 (95% CI 1.71-2.58) of PE (p<0.001). We conclude that the risk of PE in cancer patients depends not only on the cancer site and spread but also on the histological type. The excess independent risk in pancreatic cancer is intriguing and should warrant further research.