Five trials (2,919 patients, range 20 to 1,068) were eligible for inclusion. There was moderate-to-high risk of bias as at least two domains of concern were unclear or unmet for all trials. Blinding was unclear in four trials and the fifth was unblinded.
There was no difference in overall tumour response rate (RR 0.89, 95% CI 0.67 to 1.19; four trials; no significant heterogeneity) or overall survival (HR 1.01, 95% CI 0.91 to 1.13; four trials; no significant heterogeneity) between the two treatments. However, there were significant differences in the frequency of occurrence of reported toxicities. Leucopenia (OR 0.089 95% CI 0.067 to 0.119) and febrile neutropenia (OR 0.02, 95% CI 0.004 to 0.102) were significantly less prominent in the oral uracil-tegafur regimens, as were severe stomatitis (OR 0.075, 95% CI 0.039 to 0.146), any infection (OR 0.672, 95% CI 0.547 to 0.826) and any diarrhoea (OR 0.743, 95% CI 0.626 to 0.881). Other results were presented for subgroups based on different grades of cancer. Funnel plot asymmetry was not statistically significant (four trials).