Nine studies were included in this review (n=366): two RCTs (n=46) and seven controlled studies (n=320). Only one RCT described the process of randomisation. In the second study there were important differences in the length of patient follow-up. Sample sizes ranged from 13 to 157 patients. Eight of the studies were included in the meta-analysis.
Ovarian preservation (two RCTs, six controlled studies): GnRHa therapy conveyed significant ovarian preservation according to a random-effects model (RR 1.68, 95% CI 1.34 to 2.1). Adding in one controlled study with no events produced a similar result (RR 1.70, 95% CI 1.36 to 2.13), but with significant statistical heterogeneity (as did use of a fixed effects analysis).
Pregnancy (two RCTs, three controlled studies): Women who received GnRHa had significantly more pregnancies than those who did not (RR 1.65, 95% CI 1.03 to 2.6). It appeared that this was based on a fixed-effects model. Inclusion of studies with no events did not significantly alter these results (RR 1.64, 95% CI 1.01 to 2.65). Results were similar using a random effects model (RR 1.63, 95% CI 1.004 to 2.6).