Background: Although chemotherapeutic regimens containing a taxane or platinum agent have been widely recommended for unfavourable carcinoma of unknown primary (CUP), no evidence exists for the superiority of any administered regimens. To date, the efficacy has been mostly assessed in the limited setting of phase II trials, and few attempts have been made to synthesise all available data for survival outcomes.
Methods: Electronic databases were searched from 1980 to 2011. Survival results were combined for each pre-specified category of regimens using a random-effects model, and meta-regression models were used to adjust for heterogeneity in some known prognostic factors.
Results: A total of 32 studies were included for meta-analysis. Tendency towards better survival outcome by platinums or taxanes was indicated. After adjustment for important prognostic factors, however, the difference between the platinum-based and non-platinum regimens became no longer significant. Survival benefits by the taxane-based regimens remained significant, with a prolonged median survival time of 1.52 months (P=0.03) and a higher 1-year survival rate of 6.25% (P=0.05), but the benefit did not sustain for 2 years.
Conclusion: Although no effective therapies have been established, this meta-analysis helps to fill an important gap of evidence. However, caution should still be taken because of the potential unmeasured confounding.