Table 2.

The effect of mediators on endometrioid ovarian cancer estimated by Mendelian randomization and follow-up sensitivity analyses^a

exposure	outcome	method	OR	95% CI	MR P-value	Q	Q P-value	Egger intercept	value intercept
Body mass index	ENOC	Inverse variance weighted	1.22	1.02–1.47	0.03	434.523	0.194
		MR Egger	1.87	1.14–3.05	0.01	431.027	0.218	–0.008	0.069
Hip circumference	ENOC	Inverse variance weighted	1.36	1.12–1.65	<0.01	421.000	0.067
		MR Egger	1.80	1.05–3.07	0.03	419.650	0.069	–0.006	0.271
Waist circumference	ENOC	Inverse variance weighted	1.44	1.14–1.82	<0.01	361.423	0.237
		MR Egger	2.52	1.31–4.84	0.01	358.075	0.264	–0.009	0.075
Arm fat mass	ENOC	Inverse variance weighted	1.36	1.12–1.65	<0.01	433.916	0.081
		MR Egger	1.68	0.99–2.87	0.06	433.143	0.079	–0.004	0.403
Arm predicted mass	ENOC	Inverse variance weighted	1.50	1.16–1.94	<0.01	509.396	0.114
		MR Egger	1.61	0.85–3.05	0.14	509.334	0.108	–0.001	0.810
18:2 linoleic acid	ENOC	Inverse variance weighted	1.23	1.06–1.42	0.01	14.950	0.455
		MR Egger	0.97	0.68–1.37	0.86	12.756	0.546	0.035	0.161

^aOnly qualified mediators are shown in Table 2. See Supplementary Data 13 for information on all 24 candidate mediators. Heterogeneity testing was conducted using Cochran’s Q statistics (Q) and Cochran’s Q P value (Q P-value). Horizontal pleiotropy testing was conducted using the intercept term (Egger intercept) and intercept P value (P-value intercept) in MR Egger regression.

OR, odd ratio; CI, confidence interval.