Six RCTs were included in the review (n=1,476, range 30 to 327). Two RCTs (n=137) were Phase II (small preliminary studies). Quality scores were 2 (four RCTs), 3 (one RCT) and 4 (one RCT) from a possible maximum of 5. Only one RCT was double-blinded.
IP was associated with significantly higher rates of overall response to treatment (RR 1.10, 95% CI 1.00 to 1.21; five RCTs) and overall survival (HR 0.81, 95% CI 0.66 to 0.99; four RCTs) than EP. There was no significant difference between the groups in progression-free survival (four RCTs).
IP was associated with significantly lower rates of anaemia (OR 0.51, 95% CI 0.36 to 0.72; six RCTs), neutropenia (OR 0.26, 95% CI 0.12 to 0.54; six RCTs) and thrombocytopenia (OR 0.29, 95% CI 0.20 to 0.41; six RCTs). IP was associated with significantly higher rates of vomiting (OR 1.51, 95% CI 1.01 to 2.25; five RCTs) and diarrhoea (OR 10.52, 95% CI 5.94 to 18.65; six RCTs). There was no significant difference between the groups in treatment-related mortality (four RCTs).
There was significant heterogeneity in the analyses of overall survival (I2=67%), progression-free survival (I2=79%) and neutropaenia (I2=84%). No evidence of significant publication bias was found. Subgroup analyses restricted to studies that used cisplatin found no significant difference between the groups for efficacy outcomes.