An increased risk of prostate cancer associated with a family history of prostate cancer has been documented in multiple published reports. Risk has been shown to vary by degree of relationship and age of onset of disease in the affected relative. Several studies, using various designs, have estimated the relative risk (RR) for these associations. The purpose of our study was to identify and summarize published reports on the relationship between risk of prostate cancer and family history, which is defined as having a father, brother, any first- or second-degree relative or other relative affected with prostate cancer. A Medline and manual search from 1982 to 2000 identified 24 studies that reported RR and confidence intervals (CI) and satisfied inclusion criteria. Pooled RR estimates based upon a weighted average model were as follows: any affected family member RR = 1.93, CI 1.65-2.26; affected first-degree relative RR = 2.22, CI 2.06-2.40; affected second-degree relative RR = 1.88, CI 1.54-2.30; father with prostate cancer RR = 2.12, CI 1.82-2.51; and brother with prostate cancer RR = 2.87, CI 2.21-3.73). Statistical comparison of pooled data demonstrated that the RR is significantly higher for affected brother than for affected father (p < 0.03). A sensitivity analysis demonstrated that these results are robust with respect to population bias. This meta-analysis confirms that risk of prostate cancer is associated with family history of disease and improves the quantification of this risk.