Table 1.

Relative risk (RR) (95% CI) estimates of breast cancer risk in North American women comparing the median of extreme quintiles (lowest quintile as the reference) of dairy intake (kcals)

Dairy exposure	RR	Extreme valuesa	(95% CI)	P for trend	RR	(95% CI)	P for trend
	Model 1b				Model 2c
Total subjects
Total dairy kcal/d	1.17	18.9/349	1.03–1.33	0.017	1.22	1.05–1.40	0.008
Total dairy kcal/d	Regression calibration (extremes 12.2/265)d				1.35	1.09–1.71	0.003
Dairy milk kcal/d	1.42	3.2/152	1.18–1.72	<0.001	1.50	1.22–1.84	<0.001
Dairy milk_kcal/d	Regression calibration (extremes 1.2/105)d				2.37	1.55–3.88	<0.0001
Dairy cheese kcal/d	0.89	0/83.4	0.69–1.14	0.34	0.89	0.69–1.14	0.35
Yogurt kcal/d	0.92	0/68.5	0.78–1.10	0.36	0.98	0.82–1.17	0.80
Dairy fat g/d	1.13	1.9/20.5	0.95–1.34	0.15	1.13	0.93–1.36	0.22
Dairy protein g/d	1.17	1.0/20.4	0.99–1.40	0.067	1.23	1.01–1.51	0.042
Dairy CHO g/d	1.22	1.1/25.6	1.03–1.44	0.022	1.30	1.07–1.59	0.008
Post-menopausal
Total dairy kcal/d	1.17	14.9/346	1.02–1.35	0.032	1.23	1.05–1.45	0.011
Dairy milk kcal/d	1.44	2.7/153	1.16–1.77	0.001	1.54	1.22–1.93	<0.001
Dairy cheese kcal/d	0.90	0/79	0.69–1.17	0.43	0.90	0.69–1.17	0.43
Yogurt kcal/d	0.89	0/67.7	0.74–1.07	0.22	0.94	0.78–1.14	0.55
Dairy fat g/d	1.11	1.7/19.5	0.92–1.33	0.28	1.10	0.90–1.35	0.36
Dairy protein g/d	1.15	1.0/20.5	0.96–1.38	0.14	1.22	0.98–1.52	0.073
Dairy CHO g/d	1.15	1.0/20.5	0.96–1.38	0.14	1.30	1.05–1.61	0.016
Premenopausal
Total dairy kcal/d	1.23	26.1/357	0.85–1.78	0.27	1.20	0.80–1.80	0.37
Dairymilk kcal/d	1.45	4.4/146	0.85–2.47	0.18	1.38	0.77–2.46	0.28
Dairy cheese kcal/d	0.78	0/90.8	0.36–1.66	0.52	0.81	0.38–1.72	0.58
Yogurt kcal/d	1.24	0/70.0	0.75–2.04	0.40	1.28	0.77–2.14	0.33
Dairy fat g/d	1.37	2.3/21.9	0.83–2.27	0.22	1.30	0.75–2.25	0.36
Dairy protein g/d	1.45	1.2/20.1	0.86–2.42	0.16	1.37	0.75–2.51	0.31
Dairy CHO gm/d	1.44	1.2/24.9	0.88–2.35	0.15	1.38	0.78–2.47	0.27

CHO, carbohydrate; d, day.

^aExtremes are medians of extreme quintiles of FFQ data, except as indicated for calibrated models.

^bModel 1: Cox proportional hazards regression analyses with the listed exposure of interest, plus adjustment for the following categorical variables: race [reference (ref = White], family history of breast cancer (ref = none), time since mammography (ref = never, nested among those greater than 50 years of age), menopausal status (ref = no), oral contraceptive use (ref = never), nulliparous (ref = yes), smoking (ref = never), alcohol duration of use, min/week of vigorous physical activity; and the following continuous variables: age at menarche, age at menopause among post-menopausal women, cumulative months of breastfeeding among parous women, number of children among parous women, age at first childbirth among parous women, duration of estrogen replacement therapy among post-menopausal women, duration of progestogen therapy among post-menopausal women, BMI, BMI*Menopausal status.

^cModel 2: main exposure variable, plus covariates as in model 1, also total calcium, unprocessed and processed red meats, poultry, fish, seeds and nuts. Total dairy calorie models and dairy macronutrient models (full models 2) also have total isoflavones as a covariate. Dairy food models have other dairy and other soy foods as covariates (full models 2). Dairy macronutrient models contain only the one named of the dairy macronutrients, due to collinearity. All dietary variables are modelled as continuous variables.

^dAs is appropriate for the regression calibration analyses, the medians of extreme quintiles that are compared are those from the recalls.