Diet-wide analyses for risk of colorectal cancer: prospective study of 12,251 incident cases among 542,778 women in the UK

Abstract

Uncertainty remains regarding the role of diet in colorectal cancer development. We examined associations of 97 dietary factors with colorectal cancer risk in 542,778 Million Women Study participants (12,251 incident cases over 16.6 years), and conducted a targeted genetic analysis in the ColoRectal Transdisciplinary Study, Colon Cancer Family Registry, and Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO). Alcohol (relative risk per 20 g/day=1.15, 95% confidence interval 1.09-1.20) and calcium (per 300 mg/day=0.83, 0.77-0.89) intakes had the strongest associations, followed by six dairy-related factors associated with calcium. We showed a positive association with red and processed meat intake and weaker inverse associations with breakfast cereal, fruit, wholegrains, carbohydrates, fibre, total sugars, folate, and vitamin C. Genetically predicted milk consumption was inversely associated with risk of colorectal, colon, and rectal cancers. We conclude that dairy products help protect against colorectal cancer, and that this is driven largely or wholly by calcium.

Fig. 1. Volcano plot showing results from diet-wide study method evaluating associations between 97 dietary risk factors and colorectal cancer risk.

The Y axis shows p values (two-sided) for the associations between each of the 97 dietary factors and colorectal cancer incidence calculated separately using Cox proportional hazards regression models stratified by year of birth, date of completion of the dietary survey (which is the baseline for this study), and region of residence (10 geographical regions: 9 in England and 1 in Scotland), and adjusted for area-based deprivation (fifths, based on the Townsend deprivation score, unknown), highest educational qualification (none, technical, secondary, tertiary, unknown), body mass index ( < 20, 20-22.49, 22.5-24.9, 25.0-27.49, 27.5-29.9, 30-32.49, 32.5–34.9, 35+ kg/m², unknown), height ( < 160, 160–164.9, ≥ 165 cm, unknown), strenuous exercise (none, ≤ once per week, > once per week, unknown), dietary energy intake (except for the analysis of energy and risk; fifths, unknown), alcohol (none, 1–5, 6–10, ≥ 11 drinks per week, unknown), smoking (never, past, current 1–4, current 5–9, current <10, current 10–14, current 15–19, current 20–24, current 25–29, current ≥ 30 cigarettes per day, unknown), current use of hormonal therapy for menopause (no, yes, unknown), and family history of bowel cancer (no, yes). The X axis shows relative risks (see Supplementary Data 1 for increments). Dietary factors associated with risk of colorectal cancer with a False Discovery Rate (FDR) p value < 0.05 are shaded in pink and those with a p > 0.05 are shaded in grey. For each of the 62 quantitatively measured dietary factors, we created a continuous variable using the re-measured mean intakes for each baseline category. Log-linear trends in risk across categories of baseline intakes were then calculated using the listed increments.

Fig. 2. Associations of the top eight FDR-significant dietary factors (p < 0.001) and colorectal cancer by intake categories.

Mean daily intakes taken from the mean of the 24-hour dietary assessments. Wholegrain intake represents actual grams of wholegrains. Associations between each of the foods or nutrients and colorectal cancer incidence calculated separately using Cox proportional hazards regression models stratified by year of birth, date of completion of the dietary survey (which is the baseline for this study), and region of residence (10 geographical regions: 9 in England and 1 in Scotland), and adjusted for area-based deprivation (fifths, based on the Townsend deprivation score, unknown), highest educational qualification (none, technical, secondary, tertiary, unknown), body mass index ( < 20, 20–22.49, 22.5–24.9, 25.0–27.49, 27.5–29.9, 30–32.49, 32.5–34.9, 35+ kg/m², unknown), height ( < 160, 160–164.9, ≥ 165 cm, unknown), strenuous exercise (none, ≤ once per week, > once per week, unknown), dietary energy intake (except for the analysis of energy and risk; fifths, unknown), alcohol (none, 1–5, 6–10, ≥ 11 drinks per week, unknown), smoking (never, past, current 1–4, current 5–9, current <10, current 10–14, current 15–19, current 20–24, current 25–29, current ≥30 cigarettes per day, unknown), current use of hormonal therapy for menopause (no, yes, unknown), and family history of bowel cancer (no, yes).

Fig. 3. Associations of the nine less FDR-significant dietary factors and colorectal cancer (p < 0.01) by intake categories.

Mean daily intakes taken from the mean of the 24 h dietary assessments. Associations between each of the foods or nutrients and colorectal cancer incidence calculated separately using Cox proportional hazards regression models stratified by year of birth, date of completion of the dietary survey (which is the baseline for this study), and region of residence (10 geographical regions: 9 in England and 1 in Scotland), and adjusted for area-based deprivation (fifths, based on the Townsend deprivation score, unknown), highest educational qualification (none, technical, secondary, tertiary, unknown), body mass index ( < 20, 20–22.49, 22.5–24.9, 25.0–27.49, 27.5–29.9, 30–32.49, 32.5–34.9, 35+ kg/m², unknown), height ( < 160, 160–164.9, ≥165 cm, unknown), strenuous exercise (none, ≤ once per week, > once per week, unknown), dietary energy intake (except for the analysis of energy and risk; fifths, unknown), alcohol (none, 1–5, 6–10, ≥ 11 drinks per week, unknown), smoking (never, past, current 1–4, current 5–9, current <10, current 10–14, current 15–19, current 20–24, current 25–29, current ≥30 cigarettes per day, unknown), current use of hormonal therapy for menopause (no, yes, unknown), and family history of bowel cancer (no, yes).