Systematic review with meta-analyses of studies on the association between cirrhosis and liver metastases.

The hemodynamic, cellular and metabolic changes seen in patients with cirrhosis may reduce the risk of liver metastases. The aim of this case-control is to compare the risk of liver metastases from extrahepatic malignant diseases among patients with or without cirrhosis. Electronic searches (Medline, Embase, and Web of Science) and manual searches were combined (October 2010) to identify observational studies on patients with malignant disease reporting the risk of liver metastases among cases (with cirrhosis) and controls (without liver disease). Meta-analysis was performed using random effects models due to an expected clinical heterogeneity. Sixteen studies were included. Evidence of liver metastases was diagnosed in 22% of cases and 38% of controls based on autopsies (n = 14 studies) or laparoscopy (n = 2 studies). Random effects meta-analysis suggested that patients with cirrhosis had a lower risk of liver metastases (relative risk = 0.53; 95% confidence interval = 0.42-0.66). The conclusions were confirmed in sensitivity and subgroup analyses accounting for the year of publication, matching for age, sex and location of tumors (within the portal vein). No statistical evidence of bias was identified and the analyses were confirmed when adjusting for multiple testing. The present review suggests that cirrhosis reduces the risk of liver metastases. However, additional evidence from prospective studies adjusting for confounding factors is still needed.