Sixteen randomised controlled trials were included in the review (6,443 participants, range 61 to 1,135). All of the included studies were at high or unclear risk of bias for multiple quality domains.
No statistically significant difference in overall survival time was observed with use of non-platinum doublets versus platinum-based doublets (HR 1.03, 95% CI 0.98 to 1.08; 16 studies; Ι²=5%). Similar results were found in a sensitivity analysis that excluded low quality studies and in subgroup analyses according to the type of non-platinum doublet regimen and the type of platinum-based doublet regimen.
Compared with non-platinum doublet regimens, platinum-based regimens demonstrated a slightly longer progression-free survival time (HR 1.06, 95% CI 1.01 to 1.12; 11 studies; Ι²=26%); this result was of borderline statistical significance. Subgroup analyses suggested that the advantage was observed with the use of cisplatin-based platinum doublets rather than carboplatin based platinum doublets. No statistically significant differences were found between the regimens in overall response.
No evidence of publication bias was found with the Egger's test (p=0.99); some asymmetry was evident from the funnel plot. Further details of the results were reported in the review (including data on Grade 3-4 toxicities).