Table 3. Associations between tertiles of dietary iron intake and breast cancer risk from multivariable cox proportional hazards models, stratified by antioxidant/placebo group of the SU.VI.MAX trial, france, 1994–2007^a, ^b.

	Placebo group			Antioxidant supplementation group
	N for cases/non- cases	HR (95% CI)	P-trend	N for cases/non- cases	HR (95% CI)	P-trend	P-interactionc
All women			0.003			0.7	0.1
Tertile 1	25/772	1.00		28/723	1.00
Tertile 2	33/751	1.83 (1.03, 3.25)		24/741	0.70 (0.38, 1.29)
Tertile 3	42/719	2.80 (1.42, 5.54)		36/752	0.86 (0.43, 1.74)
Premenopausal women			0.02			0.2	0.04
Tertile 1	7/545	1.00		10/500	1.00
Tertile 2	10/549	1.83 (0.62, 5.39)		9/545	0.60 (0.21, 1.68)
Tertile 3	16/521	3.87 (1.16, 12.86)		7/530	0.39 (0.1, 1.56)
Postmenopausal women			0.03			0.6	0.6
Tertile 1	18/593	1.00		18/550	1.00
Tertile 2	23/544	1.90 (0.96, 3.76)		15/549	0.74 (0.35, 1.59)
Tertile 3	26/535	2.49 (1.08, 5.74)		29/572	1.18 (0.51, 2.73)

CI, confidence interval, HR, Hazard ratio, SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydants

^aMultivariable models were adjusted for age (timescale), energy intake without alcohol, intervention group of the initial SU.VI.MAX trial, number of 24-h dietary records, smoking status, educational level, physical activity, height, BMI, alcohol intake, family history of breast cancer, lipid intake, use of hormone replacement therapy, number of children and for premenopausal women: use of contraceptive pill, heavy period, and use of a hormonal intrauterine system.

^bCut-offs for tertiles of dietary iron intake were 9.3 and 11.9 mg/d.

^cBetween dietary iron intake and supplementation group