Surrogate markers may help predict the effects of first-line treatment on survival. This metaregression analysis examines the relationship between several surrogate markers and survival in women with advanced breast cancer after receiving first-line combination anthracycline chemotherapy 5-fluorouracil, adriamycin and cyclophosphamide (FAC) or 5-fluorouracil, epirubicin and cyclophosphamide (FEC). From a systematic literature review, we identified 42 randomised trials. The surrogate markers were complete or partial tumour response, progressive disease and time to progression. The treatment effect on survival was quantified by the hazard ratio. The treatment effect on each surrogate marker was quantified by the odds ratio (or ratio of median time to progression). The relationship between survival and each surrogate marker was assessed by a weighted linear regression of the hazard ratio against the odds ratio. There was a significant linear association between survival and complete or partial tumour response (P<0.001, R(2)=34%), complete tumour response (P=0.02, R(2)=12%), progressive disease (P<0.001, R(2)=38%) and time to progression (P<0.0001, R(2)=56%); R(2) is the proportion of the variability in the treatment effect on survival that is explained by the treatment effect on the surrogate marker. Time to progression may be a useful surrogate marker for predicting survival in women receiving first-line anthracycline chemotherapy and could be used to estimate the survival benefit in future trials of first-line chemotherapy compared to FAC or FEC. The other markers, tumour response and progressive disease, were less good.