As expected, in both treatment groups, the majority of patients had colon rather than rectal cancer. The proportions of inhibitor purchase patients who had previously received adjuvant chemotherapy were similar in the two groups and there was no significant difference between the groups with respect to extent or sites of metastatic disease. In general, there were no differences in the distribution of prognostic factors at baseline; although significant differences in serum alkaline phosphatase concentrations favouring the 5-FU/LV group were observed in one study (Hoff et al, 2001), no significant differences were apparent in the integrated data (Table 2 ). Table 1 Demographics and baseline characteristics Table 2 Distribution of prognostic factors in the two treatment groups (mean values��s.d.

) The median duration of treatment was 4.5 months in the capecitabine group and 4.6 months in the 5-FU/LV group. Safety data, including data on treatment interruption and dose modification, are described in detail elsewhere (Cassidy et al, 2002). Tumour responses Results from the integrated analysis demonstrate a significantly superior overall response rate with capecitabine compared with 5-FU/LV (26 vs 17%, P<0.0002; Table 3 ). Notably, the superior efficacy of capecitabine was confirmed by the IRC-assessed response rate (22 vs 13%, P<0.0001). Furthermore, analysis of the data according to subpopulations defined by baseline characteristics consistently demonstrated superior response rates for capecitabine compared with 5-FU/LV (P<0.05; Figure 1).

In both treatment arms, response rates decreased with increasing numbers of metastatic sites, as would be expected. Of note, in patients who had previously received adjuvant chemotherapy, the response rate with capecitabine was 21% (31 out of 147 patients) compared with only 9% (14 out of 155) in patients treated with 5-FU/LV. Table 3 Investigator-assessed tumour response rates Figure 1 Response rates by subpopulation. Response to treatment occurred at least as rapidly in patients treated with capecitabine compared with patients who received 5-FU/LV (median time to response: 1.7 vs 2.4 months, respectively). The median duration of response was also similar in the two treatment groups (8.1 and 9.4 months in the capecitabine and 5-FU/LV groups, respectively). Time to disease progression TTP was equivalent in the two treatment groups (hazard ratio (HR) 0.

997, 95% confidence interval (CI) 0.885?1.123, log-rank P=0.95). The median TTP was 4.6 months (95% CI 4.3?5.3) with capecitabine and 4.7 months (95% CI 4.3?5.4) with 5-FU/LV (Figure 2). Subgroup analysis according to baseline characteristics, including cancer type (rectal vs colon cancer), history of adjuvant chemotherapy and gender, demonstrated no significant differences between GSK-3 the two treatment groups for TTP. Figure 2 Time to disease progression.