Table 1. Sensitivity analyses following Higgins et al. (2008). IMORE and IMORC were calculated based on expert opinion regarding the missing data from the included studies. Effect estimates were calculated in “R” using the “metasens” package. There were no missing data for the outcomes “Miscarriage” and “Live birth”. OR: odds ratio, ICA: imputed case analysis, IMOR: informative missingness odds ratio for experimental group (E) or control group (C).
		Yilmaz et al., 2006	Lewis et al., 2006	George et al., 2007	Thomas et al., 2019	Meta-analysis
		OR	OR	OR	OR	Pooled OR (95% CI)	I2
Clinical pregnancy
Reference analysis	ICA-r	1.31	1.07	1.73	1.12	1.23 (0.91, 1.66)	0%
Sensitivity analysis	IMORE = 2, IMORC = 2	1.38	0.93	1.71	1.18	1.22 (0.83, 1.77)	0%
	IMORE = 0.5, IMORC = 0.5	1.25	1.22	1.74	1.11	1.26 (0.97, 1.63)	0%
	IMORE = 0.5, IMORC = 2	1.21	0.81	1.68	0.91	1.04 (0.65, 1.66)	0%
	IMORE = 2, IMORC = 0.5	1.42	1.41	1.77	1.44	1.47 (1.27, 1.69)	0%
	Gamble-Hollis	1.31	1.08	1.73	1.14	1.35 (0.98, 1.86)	0%
Ovulation
Reference analysis	ICA-r	1.63		1.84		1.70 (0.84, 3.43)	0%
Sensitivity analysis	IMORE = ∞, IMORC = ∞	1.79		1.81		1.80 (1.67, 1.94)	0%
	IMORE = 0.5, IMORC = 0.5	1.53		1.85		1.63 (0.52, 5.09)	0%
	IMORE = 0.5, IMORC = ∞	1.46		1.73		1.54 (0.55, 4.30)	0%
	IMORE = ∞, IMORC = 0.5	1.88		1.94		1.90 (1.57, 2.30)	0%
	Gamble-Hollis	1.63		1.84		1.70 (0.83, 3.52)	0%
Multiple pregnancy
Reference analysis	ICA-r	2.21			2.92	2.56 (0.43, 15.11)	0%
Sensitivity analysis	IMORE = 2, IMORC = 2	2.36			3.30	2.82 (0.34, 23.38)	0%
	IMORE = 0, IMORC = 0	2.03			3.03	2.51 (0.20, 31.82)	0%
	IMORE = 0, IMORC = 2	1.94			2.27	2.11 (0.78, 5.74)	0%
	IMORE = 2, IMORC = 0	2.47			4.40	3.36 (0.09, 129.73)	0%
	Gamble-Hollis	2.21			3.19	2.48 (0.28, 22.00)	0%