A total of 65 trials were included in the review (n=12,233 patients, of which 11,180 patients were considered in survival analyses, sample size range 4 to 926). Fifteen trials used neoadjuvant chemotherapy, 23 trials used concurrent chemotherapy and 27 trials used neither of the two regimens.
Chemotherapy plus radiotherapy versus radiotherapy alone: There were 22 randomised comparisons of chemotherapy plus radiotherapy versus the same radiotherapy alone in advanced cervical cancer where survival date was reported (n=3,837 patients). There was no statistically significant benefit from chemotherapy by either fixed-effect or random-effects models, though there was significant between-study heterogeneity (p=0.04), due mainly to contradictory results in earlier trials. When the analysis was restricted to trials undertaken between 1997 to 2006, the 11 comparisons showed no significant benefit, although there was no evidence of between-study heterogeneity.
Type of regimen: Analysis by regimen showed no survival benefit in trials of platinum monotherapy plus radiotherapy versus radiotherapy alone (three RCTs), nor for combinations of platinum plus non-platinum agents plus radiotherapy versus radiotherapy alone (12 RCTs).
Cycle length: The five trials using cisplatin-based regimens with short length cycles revealed a marginally significant survival benefit (summary relative hazard 0.80, 95% confidence interval (CI): 0.66 to 0.97), although the results were driven by one trial. The 11 trials using platinum-based regimens with longer cycles found a marginally significant deterioration of survival benefit (summary relative hazard 1.18, 95% CI: 1.02 to 1.38). There were low levels of heterogeneity in these analyses.
Timing of chemotherapy in relation to radiotherapy: Twelve trials used neoadjuvant chemotherapy followed by radiotherapy and 10 trials assessed the concurrent administration of chemotherapy with radiotherapy. There was no benefit reported for the neoadjuvant chemotherapy trials and a trend for benefit in the concurrent chemotherapy trials (summary relative hazard 0.85, 95% CI: 0.73 to 1.00) was due to the hydroxyurea trials.
Chemotherapy regimen direct comparison: Three trials (n=743 patients) compared non-platinum-based regimens against a combination of platinum and non-platinum agents (hydroxyurea) yielding superior results for the combination therapy (summary relative hazard 1.57, 95% CI: 1.13 to 2.19), but there was large between-study heterogeneity. Four trials (n=755 patients) compared platinum monotherapy against non-platinum regimens, with nominally better survival noted in the platinum monotherapy patients compared with the non-cisplatin regimens (summary relative hazard 0.74, 95% CI: 0.57 to 0.97), with modest between-study heterogeneity. There was no evidence for any survival differences for comparisons of platinum monotherapy versus platinum combinations with other regimens.