Table 1. GnRH antagonist versus GnRH-a (long protocol and fl are-up) : cycle cancellation rate due to poor response
Trail	GnRHantagonist	GnRHagonist	%Weights	Odds ratio	95% CI
Long protocol
Cheunget al., 2005	10/31	11/32	21.8	0.90	(0.27, 2.94)
Marciet al., 2005	1/30	4/30	11.5	0.22	(0.004, 2.51)
D’Amato et al., 2004	4/85	20/60	66.6	0.098	(0.023, 0.32)
Pooled odds ratio = 0.29 (0.1479 to 0.569)Fixed effects (Mantel-Haenszel, Robins-Breslow-Greenland) Chi2 (test odds ratio differs from 1) = 13.5 p = 0.0002Non-combinability of studiesBreslow-Day = 8.47 (df=2) p = 0.0145 Cochran Q = 8.05 (df=2) p = 0.0179
Flare-up
Akmanet al., 2001	2/24	2/24	14.3	1.00	(0.07,14.94)
Malmusi et al.,2005	7/25	6/30	30.7	1.55	(0.37, 6.63)
Schmidt et al.,2005	10/24	12/24	54.9	0.71	(0.19, 2.57)
Pooled odds ratio = 1.014 (0.468 to 2.19)Fixed effects (Mantel-Haenszel, Robins-Breslow-Greenland) Chi2 (test odds ratio differs from 1) = 0.025 p = 0.87Non-combinability of studiesBreslow-Day = 0.8175 (df=2) p = 0.6645 Cochran Q = 0.8132 (df=2) p = 0.6658