Eleven RCTs (n=1,161) were included.
Three of the included RCTs were deemed to be of high quality: one trial was double-blind with a sample size calculation and two trials used intention-to-treat analyses. According to the authors, the calculated fail-safe N of 104 indicated that no important publication bias was present.
The pooled analysis showed that there were fewer deaths with IPT than surgery alone (OR 0.51, 95% CI: 0.40, 0.65). No statistically significant heterogeneity was identified and the exclusion of trials with low Jadad scores did not affect the results of the meta-analysis. The subgroup analysis of chemotherapy group suggested that there were fewer deaths with IHCP (7 RCTs) and IPT with CH (2 RCTs) than with IPT (2 RCTs) compared with surgery; the ORs were 0.48 (95% CI: 0.35, 0.67), 0.52 (95% CI: 0.29, 0.94) and 0.57 (95% CI: 0.35, 0.92), respectively. The subgroup analysis of setting suggested that trials based in Asian countries (9 RCTs) had fewer deaths than those from non-Asian countries (2 RCTs), ORs of 0.49 (95% CI: 0.38, 0.64) and 0.67 (95% CI: 0.32, 1.41), respectively. The subgroup analysis of length of follow-up suggested that the benefit of IPT compared with surgery was more significant in trials with a follow-up of less than 60 months (6 RCTs) than in those with a follow-up of more than 60 months (5 RCTs), ORs 0.40 (95% CI: 0.27, 0.59) and 0.60 (95% CI: 0.44, 0.82), respectively.
Five of the 11 included RCTs reported mild complications, three reported no significant differences between the IPT group and surgery group, and two reported complications in the chemotherapy group. One RCT, which was conducted in Austria, was terminated early due to serious adverse events and death in the chemotherapy group.