This study was performed to determine the prognostic role of preoperative serum carbohydrate antigen (CA) 19-9 levels in the survival of patients with cholangiocarcinoma. Articles published up to June 1, 2010 that evaluated preoperative CA19-9 levels and the prognosis of cholangiocarcinoma were collected for meta-analysis. The required information for calculating individual relative risk (RR) was extracted from the studies, and a combined overall RR was estimated. Nine eligible studies were included. One study dealt with extra-hepatic cholangiocarcinoma, while the other eight studies analyzed intra-hepatic cholangiocarcinoma. The mean methodological quality score was 74.1%, ranging from 65.5% to 82.5%. The overall RR for the nine studies was 1.28 (95% confidence interval= 1.10–1.46), and the Z-score for overall effect was 13.83 (<0.001). The association between serum CA19-9 level and lymph node involvement was also assessed. The combined RR was 1.471 (95% confidence interval= 0.411–5.264) and Z-score for overall effect was 0.59 ( = 0.553). CA19-9 levels were associated significantly with the prognosis of patients with cholangiocarcinoma. This meta-analysis shows that elevation of preoperative CA19-9 levels is correlated with a poor prognosis of patients with cholangiocarcinoma. However, larger scale and randomized studies are needed to draw a more substantive conclusion.