Table 3.
Odds ratios (OR) and 95% confidence intervals (CIs) of breast cancer histological subtypes by adherence to the Western diet score (tertiles).
| Tumour subtype | p for heterogeneity2 lobular vs. ductal | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| IDC | ILC | ||||||||
| Tertile 1 OR (95% CI) |
Tertile 2 OR (95% CI) |
Tertile 3 OR (95% CI) |
p trend | Tertile 1 OR (95% CI) |
Tertile 2 OR (95% CI) |
Tertile 3 OR (95% CI) |
p trend | ||
| All participants | – | ||||||||
| Crude1 | 1 (reference) | 0.24 (0.02, 0.45) | 0.79 (0.56, 1.03) | 0.01 | 1 (reference) | 0.65 (0.21, 1.08) | 1.22 (0.77, 1.66) | 0.001 | 0.07 |
| Model 1a | 1 (reference) | 1.34 (1.06, 1.70) | 2.44 (1.88, 3.16) | < 0.001 | 1 (reference) | 0.84 (0.16, 2.93) | 1.70 (0.67, 3.35) | 0. 01 | 0.04 |
| Model 2b | 1 (reference) | 1.33 (1.05, 1.69) | 2.45 (1.88, 3.17) | < 0.001 | 1 (reference) | 1.06 (0.23, 3.13) | 1.63 (0.63, 3.25) | 0.06 | 0.03 |
| Pre-menopause | – | ||||||||
| Crude | 1 (reference) | 0.18 (0.06, 0.43) | 0.91 (0.52, 1.30) | 0.06 | 1 (reference) | 0.88 (0.34, 1.42) | 2.03 (1.39, 2.68) | 0.001 | 0.12 |
| Model 1a | 1 (reference) | 1.26 (0.96, 1.66) | 2.95 (1.91, 4.55) | 0.001 | 1 (reference) | 2.01 (1.10, 3.62) | 5.38 (2.62, 11.03) | < 0.001 | 0.65 |
| Model 2b | 1 (reference) | 1.25 (0.94, 1.64) | 2.95 (1.91, 4.56) | < 0.001 | 1 (reference) | 2.36 (1.28, 4.33) | 5.25 (0.54, 10.82) | 0.03 | 0.02 |
| Post-menopause | |||||||||
| Crude | 1 (reference) | 0.46 (0.01, 0.92) | 0.99 (0.59, 1.40) | 0.17 | 1 (reference) | 0.15 (0.01, 0.91) | 0.43 (0.02, 1.10) | 0.05 | 0.44 |
| Model 1a | 1 (reference) | 1.20 (0.72, 1.99) | 2.14 (1.38, 3.34) | 0.001 | 1 (reference) | 1.00 (0.43, 2.32) | 1.37 (0.65, 2.86) | 0.35 | 0.02 |
| Model 2b | 1 (reference) | 1.20 (0.79, 1.99) | 2.16 (1.39, 3.37) | < 0.001 | 1 (reference) | 1.05 (0.45, 2.45) | 1.35 (0.64, 2.85) | 0.38 | 0.04 |
IDC invasive ductal carcinoma, ILC invasive lobular carcinoma, OR odds ratio, CI confidence interval.
Bolded p-values represent statistically significant associations (Pp < 0.05).
1Multinomial logistic regression.
2Using Wald-test of the hypothesis that both subtypes of breast cancer share the same odds ratio for each exposure under study.
aAdjusted for energy intake, family history of BC, smoking status, OCP, chest X-ray, history of benign breast disease, BMI, physical activity, age at first delivery (year), breastfeeding (month), history of miscarriage, menarche age (year) and menopausal status.
bAdjusted for model 1+ fruit and vegetable intakes.