Twenty-six RCTs were included in the review (total number of participants not reported).
5-FU versus 5-FU combination chemotherapy (five RCTs): There was a statistically significant benefit for combination therapy in progression-free survival (HR 0.67, 95% CI 0.46 to 0.98; two RCTs, n=416, I2=72%) and in overall response rate (RR 0.43, 95% CI 0.25 to 0.74; five RCTs, n=574), but not for time to progression. There was statistically significantly more grade 3 or 4 vomiting (RR 3.76, 95% CI 1.67 to 8.44; two RCTs, n=320), nausea (RR 4.70, 95% CI 1.41 to 15.68; one RCT, n=123) and leucopenia (RR 1.68, 95% CI 1.09 to 2.60; one RCT, n=123) in the combination therapy groups. There were non-significant increases in diarrhoea, stomatitis, thrombocytopenia and neutropenia.
Gemcitabine versus 5-FU (two RCTs): There was a statistically significant benefit for gemcitabine on time to progression (HR 0.46, 95% CI 0.31 to 0.70), but not for progression-free survival or response rates. There was a higher level of grade 3 and 4 neutropenia in the gemcitabine group (25% versus 4.9%, p<0.001, one RCT).
Gemcitabine versus gemcitabine-based combination chemotherapy (19 RCTs): There were statistically significant benefits to combination chemotherapy on progression-free survival (HR 0.78, 95% CI 0.70 to 0.88; four RCTs, n=864), time to progression (HR 0.85, 95% CI 0.72 to 0.99; three RCTs, n=559) and overall response rate (RR 0.56, 95% CI 0.46 to 0.68; 17 RCTs, n=3,581). There were statistically significantly higher levels of the following toxicities in the combination chemotherapy groups: thrombocytopenia, leucopenia, neutropenia, anaemia, nausea, vomiting and diarrhoea (full details reported in the paper).
There was evidence of possible publication bias, although interpretation of funnel plots was difficult due to small numbers of included studies.