To search for laboratory variables having independent prognostic signification in patients with unresected colorectal liver metastases.

We have systematically reviewed the biomedical literature using the methodology recommended by the Committee on Evidence-Based Laboratory Medicine of the International Federation of Clinical Chemistry and Laboratory Medicine, and taking into account the Consolidated Standards of Reporting Trials Statement.

Of 644 publications retrieved, the application of strict exclusion and inclusion criteria allowed us to include only eight studies in our systematic review. The main laboratory variables evaluated in these eight studies were serum carcino-embryonic antigen, alkaline phosphatase, albumin, bilirubin, and plasma prothrombin time. None of these variables were unanimously found to have an independent prognostic significance. A meta-analysis was not possible, mainly because of heterogeneity within the primary studies and these contradictory results.

Current evidence would not support the routine use of laboratory variables as independent prognostic variables in patients with unresected colorectal liver metastases. Taking into account the inadequate quality of the published studies, this negative conclusion might be provisory only. Until better designed studies are published, a number of arguments would support to recommend pre-treatment measurement of serum carcino-embryonic antigen and alkaline phosphatase in patients participating in clinical trials.