Table 3.

Mendelian randomization estimates for the association of bioavailable and total testosterone levels with risk of lung or colorectal cancer among women in HUNT.

	Cases	Bioavailable testosteronea				Total testosteroneb
		HRc (95% CI)	p-value	Q-statistic	p of Q-statistic	HRd (95% CI)	p-value	Q-statistic	p of Q-statistic
Lung cancer	468	0.64 (0.36–1.14)	0.13	61	0.98	0.72 (0.50–1.05)	0.09	103	0.94
Lung adenocarcinoma	174	1.07 (0.41–2.79)	0.88	62	0.98	0.96 (0.51–1.80)	0.91	106	0.91
Lung non-adenocarcinoma	294	0.47 (0.23–0.96)	0.04	85	0.52	0.60 (0.37–0.98)	0.04	136	0.26
Colorectal cancer	984	0.90 (0.60–1.34)	0.60	80	0.66	0.94 (0.72–1.23)	0.67	119	0.67
Colon cancer	733	0.97 (0.61–1.54)	0.88	66	0.95	0.95 (0.70–1.29)	0.73	96	0.98
Rectal cancer	251	0.68 (0.28–1.64)	0.39	105	0.08	0.89 (0.50–1.59)	0.69	158	0.03

CI confidence interval, HR hazard ratio, IVW inverse-variance weighted, MR Mendelian randomization, SD standard deviation.

^aTwo-sample MR was performed using summary statistics from Ruth²³ for bioavailable testosterone level and from HUNT for lung or colorectal cancer.

^bTwo-sample MR was performed using summary statistics from Ruth²³ for total testosterone level and from HUNT for lung or colorectal cancer.

^cPer one-SD increase in genetically predicted bioavailable testosterone level, based on the IVW method.

^dPer one-SD increase in genetically predicted total testosterone level, based on the IVW method.