Diet and vitamin or mineral supplementation and risk of rectal cancer in Canada

Abstract

The study examines the relation of diet and vitamin or mineral supplementation with risk of rectal cancer. Mailed questionnaires were completed by 1 380 newly diagnosed patients with histologically confirmed rectal cancer and 3 097 population controls between 1994 and 1997 in seven Canadian provinces. Measurement included information on socio-economic status, lifestyle, diet and vitamin or mineral supplementation. We derived odds ratios and 95% confidence intervals through unconditional logistic regression. Total of consumption of vegetables, fruit and whole-grain products did not reduce the risk of rectal cancer. Consumption of cruciferous vegetables was inversely associated with risk of rectal cancer among women only, as did chicken intake among men. The strongest dietary association with increased rectal cancer risk appeared in males with increasing total fat intake and in females with bacon intake. Vitamin and mineral supplementation showed significant inverse associations with rectal cancer in women only. These findings suggest that dietary risk factors for rectal cancer in women may differ from those in men.

Colorectal cancer is the third most common cancer in Canada, and the second and third leading cause of death from cancer in men and women respectively [1]. Epidemiologic studies suggest that diet contributes to the aetiology of colorectal cancer as an important environmental factor [2].

Diet and colorectal cancer have been studied widely. A number of case control studies showed an inverse association between the consumption of vegetables, fruit and fiber and colorectal cancer [2–4]. A dietary pattern with greater consumption of fruit and vegetables may reduce the risk of adenoma recurrence in women [5] and men [6]. However, recently prospective cohort studies except two [7], [8] have found no inverse association between fruit, vegetable and fiber intake and the risk of colorectal cancer [9–11]. However some food items, legume fiber and/or other related sources might reduce the risk of colorectal cancer [11]. On the other hand, meat intake, specifically red meat and processed meat has been consistently reported as a risk factor in colorectal cancer [12–14]. Recently cohort studies further confirmed that colorectal cancer risk was positively associated with intake of red and processed meat [15], [16]. In addition, several large prospective studies reported that calcium and vitamin D may reduce the risk of colorectal cancer [17–19]. Results from randomized clinical trials also reported that calcium might play a role in the prevention of adenoma recurrence [20], [21]. However, a meta-analysis did not support the role of calcium in reducing the risk of colorectal cancer [22].

Although numerous studies have been extended to explore the role of diet in the aetiology of colorectal cancer, only limited studies have reported on rectal cancer individually [23–28]. Therefore, little recent data exist on the association between diet and rectal cancer, specifically any nation-wide results for Canada. In the present study, we used data from the National Enhanced Cancer Surveillance System (NECSS) to assess the relation of diet and vitamin or mineral supplementation to the risk of rectal cancer in Canada by sex.

Materials and methods

Between 1994 and 1997, the NECSS collected individual data from a population-based sample that covered 19 types of cancer (including a total of 20 819 cases) and 5 039 population controls in the Canadian provinces of British Columbia, Alberta, Saskatchewan, Manitoba, Ontario, Prince Edward Island, Nova Scotia and Newfoundland. The present study did not include subjects from Ontario.

Cases

According to pathology reports, participating provincial cancer registries ascertained a total of 2 247 (931 female and 1 316 male) histologically confirmed incident cases of rectal cancer between 1994 and 1997. Of these, 98 patients (4.4%; 22 female and 76 male) had died by the time of physician contact, and 151 (6.7%; 80 female and 71 male) were not contacted because the attending physician refused consent (generally because the patient was too ill). The patients for inclusion are (a) with a first primary neoplasm; (b) reported to cancer registries between 1994 and 1997; (c) aged 20–74 years at the time of diagnosis; (d) resident in a participating province at the time of diagnosis, and those cases for whom the attending physician's approval could be obtained. Of 1 998 questionnaires sent, 1 447 questionnaires were completed, yielding a response rate of 64.4% of cases ascertained and 72.4% of patients contacted. This study involved 1 380 (830 male and 550 female) histologically confirmed cases of rectal cancer as defined by the second edition of the International Classification of Diseases for Oncology (ICDO-2) [29]. A total of 67 cases were excluded: 66 cases of anal cancer (39 female and 27 male) and one male case with an unclear ICDO-2 code.

Controls

In the NECSS, population controls were frequency matched to the overall collection of cases for 19 types of cases. Individuals without cancer were selected from a random sample of individuals within a province, with an age/sex distribution similar to that of all cancer cases in the NECSS (i.e., 19 cancer types: liver, testis, pancreas, brain, stomach, bladder, kidney, colon, rectum, prostate, breast, ovary, lung, bone, salivary, leukaemia, multiple myeloma, non-Hodgkin's lymphoma and mesothelioma). Provincial cancer registries collected information from controls using the same protocol as for the cases. The strategies for selecting population controls varied by province; depending on data availability and accessibility. In Prince Edward Island, Nova Scotia, Manitoba, Saskatchewan and British Columbia, age group-and sex-stratified random samples of each province's population were obtained through the provincial health insurance plans. In Ontario, Ministry of Finance data were used to obtain a stratified random sample. Newfoundland and Alberta used random digit dialling to obtain a population sample.

Of 8 060 questionnaires sent to potential controls, 573 questionnaires were returned because of a wrong address; of the remainder, 5 039 were completed, representing 62.3% of controls ascertained and 67.3% (5 039/7 487) of controls contacted. A total of 3 097 controls (1 635 male and 1 462 female) were involved in the study. Data from Ontario were not included.

Data collection

The cancer registries identified most cases within one to three months of diagnosis through pathology reports. After obtaining physician consent, questionnaires were mailed to cancer cases and controls by the cancer registries. If the questionnaire was not completed and returned, a reminder postcard was sent out after 14 days and a second copy of the questionnaire at four weeks. After six weeks, telephone follow-up was used, if required, to complete the questionnaire. Information was collected on socio-economic status, employment history, residential history, height, weight, smoking history, physical activity, alcohol use, dietary history and use of vitamin or mineral supplements.

For weight, the questionnaire collected information on how much subjects weighed “about two years ago” (in pounds or kilograms) and the most subjects had ever weighed. Body mass index (BMI), calculated as weight (kg)/height squared (m²), was used to assess overweight and obesity. BMI was classified according to the World Health Organization's standards for adults [30], as follows: underweight (BMI < 18.50), normal weight (BMI = 18.50–24.99) and overweight (BMI ≥ 25.00), which includes pre-obesity (BMI 25.00–29.99) and obesity (BMI ≥ 30.00).

Physical activity two years before the survey was assessed based on session frequency, seasons of participation and average time per session for each of 12 categories of the most common types of moderate and strenuous leisure-time physical activity in Canada: walking for exercise, jogging or running, gardening or yard work, home exercise or exercise class, golf, racquet sports, bowling or curling, swimming or water exercise, skiing or skating, bicycling, social dancing and other strenuous exercise. We asked which season (spring, summer, fall, winter), how often (less than 1 per month, 1–3 per month, 1–2 per week, 3–6 per week, every day and time per session) and time per session (less than 15 minutes, 15–30 minutes, 31–60 minutes, 60+ minutes). We calculated total hours of moderate exercise per year, per month (including walking, gardening or yard work, home exercise or exercise class, golf, bowling or curling and dancing) and of strenuous exercise (including jogging, swimming or water exercise, skiing, bicycling or other strenuous exercise).

The diet portion of the questionnaire, which examined eating habits two years previous to the survey, was based on the short version of the Block Questionnaire [31]. The 70 food items were adjusted slightly to reflect Canadian dietary patterns by collaboration with the Bureau of Biostatistics and Computer Applications at Health Canada, as well as general changes in the individual's diet compared with 20 years ago. For each food item, cases and controls were asked to describe how often (per day, per week, per month) on average they ate the serving size specified of the item. Information on supplementation with vitamins (multiple, A, beta-carotene, B-complex, C and E) and minerals (calcium, iron, zinc and selenium) was also collected. Participants were also asked how often (no, yes, regularly, yes, fairly regularly) and for how many years in total (<1, 1–2, 3–5, 6–9, 10–19, 20+ years) they took the vitamins and mineral supplements mentioned above. In addition, information on alcohol consumption, specifically of beer, wine and liquor was collected. The food items and food groups were categorized based on quartile cut-off points defined by the consumption reported by controls. Estimates of total weekly caloric intake and dietary fat intake were calculated for each individual by applying the number of kilojoules (kJ) and grams of fat for each of the items in the diet questionnaire using the Canadian Nutrient Guide [32].

Statistical analysis

We computed odds ratios (ORs) and 95% confidence intervals (CIs) as a measure of the relative risk of rectal cancer. Unconditional logistic regression was used for multivariate analyses with SAS software. We selected the potential confounding variables: education, BMI, total alcohol use and total energy intake. We adjusted the dietary analyses for these confounding variables: education, BMI (≤24.99, 25.00–29.99, ≥30) and total energy intake for both sexes, as well as alcohol use and smoking status (never, ever) for male rectal cancer cases only. We also included age group (≤49, 50–59, 60–69, ≥70) and province (as described above) as variables. Tests for trend were assessed for each study variable by substituting the variable in the model in continuous form.

Results

The distribution of rectal cancer cases and controls by sex and age is shown in Table I. Among the cases, 88.8% of males and 84.9% of females were 50 years old or more; while the controls were 74.8% and 70.9% of males and females respectively; the percentages of aged 50 or over in controls were lower than in cases.

Table I.  Distribution of rectal cancer by age and sex, NECSS Study, Canada, 1994–1997.

	Males		Females
Age group	Controls n (%)	Cases n (%)	Controls n (%)	Cases n (%)
20–29	99 (6.1)	1 (0.1)	42 (2.9)	1 (0.2)
30–39	176 (10.7)	16 (1.9)	97 (6.6)	15 (2.7)
40–49	137 (8.4)	76 (9.2)	287 (19.6)	67 (12.2)
50–59	239 (14.6)	172 (20.7)	366 (25.0)	140 (25.5)
60–69	581 (35.5)	369 (44.5)	462 (31.6)	205 (37.2)
≥70	403 (24.7)	196 (23.6)	208 (14.2)	122 (22.2)
Total	1635	830	1462	550

Odds ratios and mean values (standard deviation) for selected variables are presented in Table II. For both sexes, we observed significantly elevated ORs for rectal cancer with increased BMI, along with a dose-response relation (test for trend p < 0.0001 for males, and p = 0.001 for females). We also saw an inverse association between high education level and risk of rectal cancer in both men and women. However, an increased risk associated with alcohol use appeared among men only. Compared with men who never drank alcohol, those who consumed 18 or more servings per week had an adjusted OR of 1.7 (95% CI = 1.3–2.2; p for trend 0.0001). Income level was not reported for 20.8% of male cases and 30.2% of female cases, nor for 20.9% of male controls and 25.2% of female controls.

Table II.  Odds ratios and mean values (standard deviation) for rectal cancer by sex, NECSS Study, Canada, 1994–1997.

	Males				Females
	Controls n	Cases n	Age and province adjusted OR (95% CI)	p-value for trend	Controls n	Cases n	Age and province adjusted OR (95% CI)	p-value for trend
Family income
Low	284	135	1.0 ref.	0.43	300	118	1.0 ref.	0.84
Lower middle	313	159	0.9 (0.7–1.2)		272	85	1.1 (0.7–1.6)
Upper middle	424	206	1.1 (0.8–1.5)		355	123	0.9 (0.6–1.3)
High	273	157	1.1 (0.8–1.4)		167	58	1.1 (0.7–1.6)
Not reported	341	173			368	166
Education
Mean (SD)	12.0 (3.9)	11.4 (3.5)			12.0 (3.1)	11.5 (3.3)
1–8 yrs	299	163	1.0 ref.	0.003	172	85	1.0 ref.	0.06
9–13 yrs	803	447	1.0 (0.8–1.2)		886	342	0.9 (0.6–1.2)
≥14 yrs	511	203	0.7 (0.5–0.9)		389	114	0.7 (0.5–1.0)
Not reported	22	17			15	9
Body mass index (kg/m^2)*
<25	639	260	1.0 ref.	<0.0001	822	254	1.0 ref.	0.001
25–29	753	408	1.4 (1.1–1.7)		423	185	1.3 (1.1–1.7)
≥30.00	235	160	1.7 (1.3–2.2)		212	105	1.5 (1.2–2.1)
Not reported	8	2			5	6
Moderate activity (hr/month)**
Mean (SD)	16.1 (15.0)	18.5 (16.7)			15.3 (13.2)	16.9 (15.0)
≤4.2	572	269	1.0 ref.	0.25	321	112	1.0 ref.	0.72
4.3–11.6	354	165	1.0 (0.8–1.3)		321	114	0.9 (0.7–1.2)
11.7–24.4	355	176	1.0 (0.8–1.2)		322	104	0.8 (0.6–1.2)
≥24.5	354	220	1.2 (0.9–1.5)		320	140	1.1 (0.8–1.5)
Not reported	0	0			178	80
Strenuous activity (hr/month)**
Mean (SD)	5.7 (11.1)	5.5 (11.1)			2.9 (6.8)	2.9 (7.0)
None	575	316	1.0 ref.	0.75	571	252	1.0 ref.	0.63
≤1.9	277	149	1.0 (0.8–1.3)		217	68	0.9 (0.6–1.2)
2.0–9	249	108	0.9 (0.7–1.2)		208	63	0.8 (0.6–1.1)
≥9.1	265	126	1.0 (0.8–1.4)		237	92	1.0 (0.7–1.4)
Not reported	269	131			229	75
Smoking status**
Never	414	164	1.0 ref.		709	256	1.0 ref.
Ever	1182	629	1.2 (1.0–1.5)		736	267	1.0 (0.8–1.3)
Not reported	39	37			17	27
Total alcohol (servings/week)**
Mean (SD)	6.4 (11.1)	8.4 (12.0)			2.4 (5.4)	2.7 (5.7)
Never	502	220	1.0 ref.	< 0.0001	706	281	1.0 ref.	0.57
1–6	625	273	1.0 (0.8–1.3)		582	194	0.8 (0.6–1.0)
7–17	288	164	1.4 (1.0–1.8)		108	45	0.9 (0.4–1.4)
≥18	218	173	1.7 (1.3–2.2)		64	30	1.0 (0.6–1.7)
Not reported	2	0			2	0

*Adjusted for 10-year age group, province and education.

**Adjusted for 10-year age group, province, education, body mass index (<25, 25–29.9, ≥30); physical activity in females: ≤5.2, 5.3–12, 12.1–22.8, ≥22.9 for moderate activity (hr/month), none, ≤0.8, 0.9–3.7, ≥3.8 for strenuous activity.

Tables III and IV show the associations between intake frequencies of selected foods or food groups and risk of rectal cancer among males and females, respectively. Among males, total consumption of vegetables, fruit and whole-grain products had no effect on risk of rectal cancer. However, high consumption of juices was related to increased rectal cancer risk. An increased risk was also associated with consumption of total meat, hamburger, processed meat and eggs. On the other hand, consumption of white meat, chicken in particular, reduced the risk of rectal cancer in males; the adjusted OR for high versus low chicken intake was 0.4 (95% CI = 0.2–0.8; p for trend 0.01). We also observed a significantly increased risk of rectal cancer in males with increasing total fat intake; the adjusted OR for high versus low intake was 1.7 (95% CI = 1.1–2.6; p for trend 0.02). The remaining foods and food groups were not associated with any effect for male rectal cancer cases.

Table III.  Odds ratios* for rectal cancer in males by consumption of selected foods and food groups**, NECSS Study, Canada, 1994–1997.

	Quartiles n for cases/controls					Mean & (Standard Deviation)
Foods or food groups** (servings/per week)	I (low)	II	III	IV (high)	p-value for trend	Cases	Controls
Total fruit	189/411	199/412	216/402	226/408
OR (95% CI)	1.0	1.1 (0.9–1.5)	1.2 (0.9–1.6)	1.3 (1.0–1.7)	0.45	10.6 (8.9)	10.2 (9.4)
Total vegetables	187/408	210/409	226/410	207/406
OR (95% CI)	1.0	1.0 (0.7–1.3)	0.9 (0.7–1.2)	0.8 (0.6–1.1)	0.19	13.2 (11.5)	13.4 (9.9)
Total juices	188/440	217/385	204/407	221/401
OR (95% CI)	1.0	1.4 (1.1–1.9)	1.3 (1.0–1.7)	1.6 (1.2–2.2)	0.002	9.4 (12.6)	8.6 (10.5)
Yellow-orange vegetables	347/656	285/529	187/427
OR (95% CI)	1.0	1.0 (0.8–1.3)	0.8 (0.6–1.1)		0.15	2.8 (3.3)	2.9 (3.1)
Dark green vegetables	142/302	326/624	163/305	188/367
OR (95% CI)	1.0	1.0 (0.8–1.4)	0.9 (0.7–1.3)	0.8 (0.6–1.1)	0.13	2.4 (3.5)	2.4 (3.1)
Cruciferous vegetables (0.5 cup/week)	102/243	349/618	186/390	193/382
OR (95% CI)	1.0	1.3 (0.9–1.7)	1.0 (0.7–1.4)	0.9 (0.5–1.3)	0.10	2.5 (3.8)	2.5 (3.1)
Tomatoes (Ave. No. of 0.5 cup/week)	77/209	366/707	274/467	87/200
OR (95% CI)	1.0	1.3 (0.9–1.8)	1.2 (0.9–1.7)	0.9 (0.6–1.4)	0.34	2.2 (2.5)	2.2 (2.8)
Total meat	169/423	207/406	239/395	215/409
OR (95% CI)	1.0	1.4 (1.0–1.8)	1.5 (1.2–2.1)	1.3 (0.9–1.8)	0.22	10.5 (9.3)	9.8 (8.6)
Total fish (4 oz/week)	107/208	476/898	247/527
OR (95% CI)	1.0	0.9 (0.7–1.3)	0.8 (0.6–1.2)		0.29	1.4 (2.1)	1.5 (2.0)
Chicken (Ave. No. of 4 oz/week)	565.0	1054	243/511
OR (95% CI)	1.0	0.8 (0.7–1.0)	0.4 (0.2–0.8)		0.01	1.6 (1.6)	1.7 (1.9)
Beef, pork or lamb as a main dish (Ave. No. of 4 oz/week)	331/729	358/688	112/155
OR (95% CI)	1.0	1.0 (0.8–1.3)	1.1 (0.8–1.5)		0.80	2.7 (2.9)	2.4 (2.5)
Hamburger (Ave. No./week)	124/308	540/979	149/312
OR (95% CI)	1.0	1.6 (1.2–2.1)	1.4 (1.0–2.0)		0.06	1.2 (1.9)	1.1 (1.3)
Bacon (Ave. No. of slice/week)	205/472	392/757	206/354
OR (95% CI)	1.0	1.1 (0.9–1.4)	1.2 (0.9–1.6)		0.16	1.4 (2.4)	1.4 (2.9)
Sausage (Ave. No./week)	225/568	478/851	100/160
OR (95% CI)	1.0	1.3 (1.0–1.6)	1.3 (0.9–1.8)		0.05	0.9 (1.8)	0.8 (1.4)
Processed meat products	157/420	216/402	250/408	207/403
OR (95% CI)	1.0	1.3 (1.0–1.8)	1.6 (1.2–2.2)	1.2 (0.9–1.7)	0.10	5.3 (6.1)	5.1 (6.8)
Total milk (8 oz/week)	215/418	371/700	244/515
OR (95% CI)	1.0	1.1 (0.9–1.4)	0.9 (0.7–1.2)		0.65	8.3 (9.0)	8.7 (9.8)
Dairy products	181/412	213/403	246/422	190/396
OR (95% CI)	1.0	1.2 (0.9–1.6)	1.3 (1.0–1.7)	1.0 (0.7–1.4)	0.72	13.0 (10.3)	13.2 (11.1)
Eggs (Ave. No. /week)	286/688	389/646	139/254
OR (95% CI)	1.0	1.3 (1.1–1.6)	1.3 (1.0–1.7)		0.04	3.1 (3.2)	2.8 (3.2)
Cheese (No. of 1 oz/week)	326/676	301/543	180/373
OR (95% CI)	1.0	1.0 (0.8–1.3)	0.9 (0.7–1.2)		0.64	3.0 (3.4)	3.1(3.8)
Butter on bread or vegetables (Ave. No. of tsp/week)	351/714	244/536	209/321
OR (95% CI)	1.0	0.9 (0.7–1.1)	1.1 (0.9–1.4)		0.56	4.6 (8.0)	3.9 (7.2)
Margerine on bread or vegetables (Ave. No. of tsp/week)	163/312	262/561	319/571	64/145	0.72	9.2 (10.0)	9.2 (9.9)
OR (95% CI)	1.0	1.0 (0.8–1.4)	1.2 (0.9–1.5)	0.9 (0.6–1.3)
Used fat /cooking oil to cook (times/week)	367/780	264/489	180/313
OR (95% CI)	1.0	1.0 (0.8–1.2)	0.9 (0.7–1.2)		0.54	3.8 (4.1)	3.6 (3.9)
Whole-grain products	170/416	221/405	256/434	183/378
OR (95% CI)	1.0	1.3 (1.0–1.8)	1.1 (0.9–1.6)	0.9 (0.6–1.2)	0.19	18.3 (19.6)	18.4 (20.3)
Total fat intake (g/week)	141/364	208/364	186/365	207/367
OR (95% CI)	1.0	1.5 (1.1–2.0)	1.4 (1.0–2.0)	1.7 (1.1–2.6)	0.02	395.3 (196.2)	379.2 (203.0)
Total energy intake (kJ/week)	152/362	169/366	222/367	199/365		55500.4	53328.0
OR (95% CI)	1.0	1.0 (0.8–1.4)	1.4 (1.0–1.8)	1.2 (0.9–1.6)	0.05	(22475.2)	(21356.2)

*Adjusted for 10-year age group, province, education, body mass index (<25, 25–29.9, ≥30), smoking status (never, ever), alcohol use and total energy intake.

**Total fruit: apples, pears, oranges, bananas, cantaloupe or other fruit, fresh or canned.

Total vegetables: tomatoes, carrots, broccoli, cabbage, cauliflower, brussel sprouts, spinach or other greens, yellow squash, green beans, corn, peas or any other vegetable.

Total juice: fresh or frozen fruit end vegetable juices, and drinks from powdered drink crystals.

Total meat: beef, pork, lamb, hamburger, hotdogs, bacon, sausage, smoked meat, luncheon meat and liver.

Processed meat products: hotdogs, lunch meat, smoked meat, bacon and sausage.

Total milk: 1% milk, 2% milk and skim milk.

Dairy products: milk, cheese and ice cream.

Whole-grain products: dark or whole grain bread, granola and cooked cereals.

Note: Totals may vary due to missing values.

Table IV.  Odds ratios* for rectal cancer in females by consumption of selected foods and food groups**, NECSS Study, Canada, 1994–1997.

	Quartiles n for cases/controls					Mean & (Standard Deviation) (servings/per week)
Foods or food groups** (servings/per week)	I (low)	II	III	IV (high)	p-value for trend	Cases	Controls
Total fruit	183/552	141/440	129/278	97/190
OR (95% CI)	1.0	0.8 (0.6–1.1)	1.1 (0.8–1.4)	1.2 (0.9–1.8)	0.16	13.7 (10.4)	11.9 (8.8)
Total vegetables	145/376	110/356	148/364	147/364
OR (95% CI)	1.0	0.7 (0.5–1.0)	0.8 (0.6–1.1)	0.8 (0.5–1.1)	0.27	16.7 (13.9)	15.8 (13.5)
Total juices	129/365	111/365	155/367	155/363
OR (95% CI)	1.0	0.8 (0.6–1.1)	1.0 (0.8–1.4)	1.1 (0.8–1.5)	0.26	10.6 (12.3)	9.1 (10.8)
Yellow-orange vegetables	141/376	194/517	117/264	93/295
OR (95% CI)	1.0	0.9 (0.7–1.2)	1.0 (0.8–1.5)	0.7 (0.5–1.0)	0.16	3.7 (4.5)	3.7 (3.9)
Dark green vegetables	119/301	180/503	107/321	136/321
OR (95% CI)	1.0	0.8 (0.6–1.1)	0.8 (0.5–1.1)	0.9 (0.6–1.3)	0.60	3.2 (3.3)	3.1 (3.9)
Cruciferous vegetables (0.5 cup/week)	100/201	227/659	95/267	128/333
OR (95% CI)	1.0	0.6 (0.4–0.8)	0.6 (0.4–0.8)	0.6 (0.4–0.8)	0.03	3.1(3.5)	3.1 (3.6)
Tomatoes (Ave. No. of 0.5 cup/week)	48/141	213/617	185/468	91/205
OR (95% CI)	1.0	0.9 (0.6–1.3)	0.9 (0.6–1.4)	1.0 (0.6–1.5)	0.83	2.6 (3.3)	2.4 (2.9)
Total meat	108/367	143/357	129/371	170/365
OR (95% CI)	1.0	1.4 (1.0–2.0)	1.2 (0.9–1.7)	1.6 (1.1–2.2)	0.05	8.0 (6.8)	7.5 (10.1)
Total fish (4 oz/week)	65/187	307/858	178/415
OR (95% CI)	1.0	0.9 (0.6–1.3)	1.1 (0.7–1.6)		0.38	1.5 (1.8)	1.4 (2.9)
Chicken (Ave. No. of 4 oz/week)	333/852	191/555	23/47
OR (95% CI)	1.0	0.8 (0.7–1.0)	1.0 (0.6–1.8)		0.22	1.9 (2.3)	1.9 (2.3)
Beef, pork or lamb as a main dish	362/778	214/526	56/120
(Ave. No. of 4 oz/week)
OR (95% CI)	1.0	1.2 (0.9–1.5)	1.2 (0.8–1.7)		0.20	2.4 (2.7)	2.1 (2.3)
Hamburger (Ave. No./week)	103/247	363/998	74/201
OR (95% CI)	1.0	0.9 (0.7–1.2)	0.9 (0.6–1.3)		0.47	0.9 (1.0)	1.0 (1.5)
Bacon (Ave. No. of slice/week)	173/513	286/755	75/143
OR (95% CI)	1.0	1.0 (0.8–1.3)	1.6 (1.1–2.3)		0.04	0.8 (1.1)	0.7 (1.7)
Sausage (Ave. No. /week)	224/643	277/732	29/47
OR (95% CI)	1.0	1.0 (0.8–1.3)	1.4 (0.8–2.3)		0.40	0.5 (0.8)	0.5 (1.4)
Processed meat products	118/359	148/421	122/316	162/364
OR (95% CI)	1.0	1.1 (0.8–1.5)	1.3 (0.9–1.8)	1.4 (1.0–1.9)	0.05	3.3 (3.8)	3.1 (6.5)
Total milk (8 oz/week)	163/409	204/580	183/471
OR (95% CI)	1.0	0.8 (0.6–1.1)	0.9 (0.7–1.2)		0.53	8.8 (10.3)	8.4 (9.8)
Dairy products	134/370	154/383	125/348	137/359
OR (95% CI)	1.0	1.1 (0.8–1.5)	0.8 (0.5–1.0)	0.9 (0.6–1.3)	0.24	12.9 (11.1)	12.6 (10.9)
Eggs (Ave. No./week)	35/100	215/719	250/515	40/107
OR (95% CI)	1.0	0.7 (0.5–1.1)	1.1 (0.7–1.7)	0.9 (0.5–1.6)	0.10	2.2 (2.1)	2.0 (2.3)
Cheese (No. of 1 oz/week)	211/585	228/521	104/337
OR (95% CI)	1.0	1.1 (0.9–1.5)	0.7 (0.5–1.0)		0.11	2.9 (2.8)	3.1 (3.4)
Butter on bread or vegetables	351/714	244/536	209/321
(Ave. No. of tsp/week)
OR (95% CI)	1.0	0.9 (0.7–1.1)	1.1 (0.9–1.4)		0.56	4.4 (7.6)	4.5 (7.4)
Margerine on bread or vegetables (Ave. No. of tsp/week)	163/312	262/561	319/571	64/145	0.72	8.8 (9.7)	8.5 (9.1)
OR (95% CI)	1.0	1.0 (0.8–1.4)	1.2 (0.9–1.5)	0.9 (0.6–1.3)
Used fat /cooking oil to cook
(times/ week)	254/740	174/467	110/246
OR (95% CI)	1.0	1.1 (0.8–1.4)	1.2 (0.9–1.6)		0.29	3.6 (3.4)	3.3 (3.5)
Whole-grain products	109/355	133/377	175/391	133/337
OR (95% CI)	1.0	1.2 (0.9–1.7)	1.3 (0.9–1.8)	1.0 (0.7–1.5)	0.87	19.6 (18.1)	17.6 (17.3)
Total fat intake (g/week)	104/338	143/335	128/338	128/337
OR (95% CI)	1.0	1.2 (0.9–1.7)	1.0 (0.7–1.5)	0.8 (0.5–1.3)	0.29	339.0 (151.1)	334.5 (267.3)
Total energy intake (kJ/week)	104/338	125/336	111/336	163/338		50623.6	48386.6
OR (95% CI)	1.0	1.2 (0.9–1.7)	1.1 (0.8–1.5)	1.7 (1.2–2.2)	0.003	(18071.5)	(26778.3)

*Adjusted for 10-year age group, province, education, body mass index (<25, 25–29.9, ≥30) and total energy intake.

**Foods or food groups are the same as in Table III.

Note: Totals may vary due to missing values.

Among females (Table IV), consumption of vegetables, particularly cruciferous vegetables, was inversely associated with risk of rectal cancer; the adjusted OR for cruciferous vegetables was 0.6 (95% CI = 0.4–0.8; p for trend 0.03). We observed an elevated rectal cancer risk in women with increased consumption of total meat and processed meat; the effect of bacon intake was particularly pronounced. The adjusted OR for high versus low intake of bacon was 1.6 (95% CI = 1.1–2.3; p for trend 0.04). Total energy intake also increased the risk of rectal cancer in women; the adjusted OR for high versus low intake was 1.7 (95% CI = 1.2–2.2; p = 0.003). However, a similar relation did not appear for total fat intake.

Table V presents the risks of rectal cancer associated with vitamin or mineral supplementation among males and females. For females only, we observed significant inverse associations with rectal cancer among those taking B-complex vitamins, calcium and zinc for more than five years. The adjusted ORs were 0.6 (95% CI = 0.4–0.8; p = 0.001), 0.6 (95% CI = 0.4–0.7; p = 0.0001) and 0.5 (95% CI = 0.3–0.9; p = 0.004) respectively. In addition, consumption of vitamin A, vitamin E and selenium supplements was also associated with reduced risk of rectal cancer in females. We found no clear association between intake of vitamin or mineral supplements and risk of rectal cancer among males.

Table V.  Odds ratios* for rectal cancer by consumption (years) of selected vitamin and mineral spplements, NECSS Study, Canada, 1994–1997.

	Males n for cases/controls				Females n for cases/controls
Vitamin or mineral supplements	Never OR	1–4 OR (95% CI)	≥5 OR (95% CI)	p-value for trend	Never OR	1–4 OR (95% CI)	≥5 OR (95% CI)	p-value for trend
Multiple vitamins	449/890	212/416	120/247		208/506	193/549	111/334
	1.0	1.0 (0.8–1.2)	1.0 (0.6–1.5)	0.34	1.0	0.9 (0.7–1.2)	0.9 (0.6–1.4)	0.03
Vitamin A	624/1272	81/150	40/87		402/1054	49/170	25/89
	1.0	1.0 (0.7–1.3)	0.8 (0.5–1.2)	0.42	1.0	0.7 (0.5–1.0)	0.6 (0.4–1.0)	0.009
Beta-carotene	641/1326	81/113	26/58		392/1049	65/211	18/58
	1.0	1.3 (0.9–1.8)	0.8 (0.5–1.3)	0.95	1.0	0.7 (0.5–1.0)	0.8 (0.4–1.3)	0.06
B-complex vitamins	595/1219	100/180	52/105		335/851	100/313	45/167
	1.0	1.0 (0.8–1.4)	0.9 (0.6–1.3)	0.66	1.0	0.8 (0.6–1.0)	0.6 (0.4–0.8)	0.001
Vitamin C	423/829	217/427	133/271		285/797	110/266	108/305
	1.0	0.9 (0.7–1.1)	0.8 (0.7–1.1)	0.18	1.0	0.8 (0.6–1.0)	0.7 (0.5–1.0)	0.02
Vitamin E	494/1047	178/302	92/175		336/898	85/251	72/197
	1.0	1.0 (0.8–1.2)	0.9 (0.7–1.2)	0.51	1.0	0.7 (0.6–0.9)	0.7 (0.5–1.0)	0.009
Calcium	594/1208	103/201	54/100		256/632	168/524	74/229
	1.0	0.9 (0.7–1.2)	1.0 (0.7–1.4)	0.61	1.0	0.6 (0.5–0.8)	0.5 (0.4–0.7)	<0.0001
Iron	605/1314	113/143	20/50		272/742	166/456	47/131
	1.0	1.7 (1.3–2.3)	0.8 (0.5–1.4)	0.07	1.0	1.0 (0.8–1.3)	1.0 (0.7–1.4)	0.98
Zinc	690/1295	29/68	22/50		421/1092	43/156	15/66
	1.0	0.9 (0.6–1.3)	0.8 (0.5–1.4)	0.32	1.0	0.7 (0.5–1.0)	0.5 (0.3–0.9)	0.004
Selenium	682/1415	41/52	11/30		441/1177	25/89	7/37
	1.0	1.4 (0.9–2.2)	0.7 (0.4–1.5)	0.88	1.0	0.6 (0.4–1.0)	0.4 (0.2–1.0)	0.01

*Adjusted for 10-year age group, province, education and body mass index (<25, 25–29.9, ≥30), plus alcohol intake for males.

Note: Totals may vary due to missing values.

Discussion

High intake of vegetables, fruit and dietary fibre has been associated with a reduction in the risk of colorectal cancer [2], [3]. A cohort study, the European Prospective Investigation into Cancer and Nutrition (EPIC) conducted in ten European countries has shown a clear protective effect of fiber with a dose-response relationship [7]. An approximate doubling of total fiber intake from food could reduce the risk of colorectal cancer by 40%. Another study of the prospects for colorectal cancer in New Zealand to estimate the protection offered by significantly increasing the consumption of fruit and vegetables indicated that increasing the consumption of fruit and vegetables provided potential longer-term reduction in colorectal cancer incidence and mortality [33]. Results from The Self-Defense Forces Health Study, a cross-sectional survey, also showed that a dietary pattern including greater consumption of dairy products and fruit and vegetables may be associated with a decreased risk of colorectal adenoma in Japanese men [6]. However, a report from two prospective cohort studies, the Nurses’ Health Study (NHS) and the Health Professionals’ Follow-up Study (HPFS), indicated that high consumption of fruit and vegetables did not appear to be protective against colon and rectum cancers in the large US cohort [24]. Further repeated assessment from the same studies with a long follow-up did not indicate an important association between fiber intake and colorectal cancer [34]. Other cohort studies, the Women's Health Study in the US [11], and the Japan Public Health Center study (JPHC) [9] and The Miyagi Cohort Study [10] in Japan also did not support an association between fruit, vegetable and fibre intake and colorectal cancer. The JPHC study also observed no significant association between healthy dietary pattern and rectal cancer [35]. Also, a diet high in fiber, fruit and vegetables did not influence the risk of recurrence of colorectal adenomas in the White Bran Fiber trial [36] and The Polyps Prevention Trial [37]. Our study results indicated that total consumption of vegetables, fruit and whole-grain products was not related to reduced risk of rectal cancer in either sex. However, a recent study estimating the global burden of disease attributable to low consumption of fruit and vegetables by 14 geographical regions [38] showed that increasing fruit and vegetable intake reduced many non-communicable diseases. Increasing individual fruit and vegetable consumption to up to 600 g (the baseline of choice) could reduce the total worldwide burden of disease by 1.8%. For colorectal cancer, the potential reduction was 2%.

A case-control study reported that cooked vegetables and citrus fruit were inversely associated with the risk of rectal cancer in men and women combined [3]. Our results also showed that consumption of one vegetable group in our study—cruciferous vegetables—was inversely associated with risk of rectal cancer in females only. Results from The Swedish Mammography Screening Cohort also reported an inverse association between rectal cancer in women and total vegetable and fruit intake but not cereal fibre intake [8]. In contrast, a case-control study in Uruguay reported that plant-based foods, raw vegetables, cooked vegetables and citrus fruit were associated with reduced risk of rectal cancer in men only [25]. A recent European intervention trial, including 21 centres from ten countries indicated that a Mediterranean diet might reduce the risk of colorectal adenoma recurrence in women but not in men [5]. However, results from two other cohort studies, the large US cohort mentioned before and the Netherlands Cohort Study on Diet and Cancer, showed that consumption of vegetables, fruit and specific groups of vegetables and fruit was not related to reduced rectal cancer risk in either sex [24], [26]. Few foods or nutrients remain as candidates for cancer prediction or prevention [39]. The results from studies on the relation between plant-based foods and rectal cancer risk differed by sex [8],[25], including our study]. Consumption of specific vegetable groups might play a certain role in the differences in rectal cancer risk between the sexes. More studies are needed to clarify the relation of specific vegetable and fruit groups to the risk of rectal cancer by gender.

High dietary fat consumption is likely to increase the incidence of colorectal cancer [40]. A high-fat diet has generally been implicated in the etiology of colorectal cancer [41]. A previous case-control study reported the risk of rectal cancer increased with increasing intake of fat in males [23]. Our results further supported that total fat intake was related to increased risk of rectal cancer in males. Diets containing high amounts of animal fat also provide high amounts of cholesterol. A prospective study in Finland found that high cholesterol intake was associated with colon and rectal cancer, but fat and fatty acids were not associated [42].

High consumption of meat, especially red meat and processed meat elevated the risk of colorectal cancer in a case-control study [3], but not in another [43]. A meta-analysis of 12 prospective studies in seven countries showed a positive association between colorectal cancer and consumption of all meat, red meat and processed meat [12]. Another meta-analysis of case-control and prospective studies reported positive associations with intake of red meat, but the strongest for processed meat [14]. Recent results from EPIC data further confirmed that colorectal cancer risk was positively associated with the intake of red and processed meat [15]. Our results also showed that elevated risk of rectal cancer was associated with high consumption of total meat and processed meat, particularly of bacon by females. However, the research findings are not consistent. In contrast, total consumption of meat, red meat and processed meat showed a non-significant relation with rectal cancer in Finland [42] and in the US of NHS and HPFS [27]. A recent large, population-based cohort in the US, using data on meat information in 1982 and again in 1992/1993 further confirmed that high consumption of red meat and processed meat increased the risk of rectal cancer [16]. The study also emphasized the value of long follow-up and multiple measurement of diet. On the other hand, white meat consumption was not associated with either colorectal cancer [3], [43] or rectal cancer [42], [16]. This pattern is consistent with our findings. Furthermore, our results also showed a significant inverse association between chicken intake and rectal cancer in males only. Consumption of red meat (but not of white meat) is associated with a high risk of rectal cancer. Dietary heme induces faecal cytotoxicity and hyperproliferation of the colonic mucosa in rats [44]. The heme content of red meat is ten-fold higher than that of white meat [45]. Red meat consumption increases risk of colorectal cancer, which is also possibly due to the exposure to heterocyclic amines, mainly formed during the cooking of meat at high temperatures [46], [47]. A clinic-based case-control study reported that increased risk of colorectal adenomas was more strongly associated with benzo[a]pyrene (as a surrogate for total carcinogenic polycyclic aromatic hydrocarbons) intake estimated from all foods [48]. However, meat preparation habits did not increase colorectal adenoma risk in a Dutch population: the associations were not influenced by relevant genetic polymorphisms [49].

A pooled analysis of ten cohort studies in five countries reported that milk intake was related to the reduced risk of rectal cancer [22]. Although our results did not show the association between milk intake and rectal cancer, we found that egg intake was associated with rectal cancer in men only, unlike the results of other studies [42], [43].

A previous case-control study reported that calcium intake from diet was inversely associated with the risk of colorectal cancer [50], as did the above-mentioned pooled analysis of prospective studies [22]. According to the latter study, assuming that the association with colorectal cancer risk is causal, if individuals who consumed less than 1 000 mg/day of calcium increased their intake to 1 000 mg/day or more, 15% and 10% of colorectal cancer cases in the study population would have been avoided for women and men, respectively [22]. A report from the Cancer Prevention Study II Nutrition Cohort in the US also indicated that total calcium intake (from diet and supplements) was associated with marginally lower colorectal cancer in both genders. The association was strongest for calcium from supplements [17]. Our results showed that use of calcium supplement for more than five years reduced the risk of rectal cancer in females only. Results from The Calcium Polyp Prevention Study in the US showed that calcium supplementation was associated with a significant-through moderate-reduction in the risk of recurrent colorectal adenomas [21]. Also calcium supplementation was associated with a modest but not significant reduction in the risk of adenoma recurrence in the European Cancer Prevention Study [20]. A systematic review from two randomized controlled trials to assess the effect of supplementary dietary calcium on the incidence of colorectal cancer and the incidence or recurrence of adenomatous polyps suggested that calcium supplementation might contribute to a moderate degree to the prevention of colorectal adenomatous polyps [51]. However, the research findings have been somewhat inconsistent. The results from the prospective studies of The NHS and HPFS in the US or the two studies combined above mentioned showed that calcium intake from diet was not related to colon and rectal cancers [27]. Calcium prevents colorectal carcinogenesis [52] and reduces colonic epithelial cell proliferation [53]. In colorectal cancer, the molecular mechanisms of the calcium-sensing receptor (CASR) indicate that specific signalling pathways involved in cell growth and differentiation are activated by calcium through the CASR [54]. Recently, Peters and colleagues reported that variants in the CASR intracellular signalling region were significantly associated with risk of advanced adenoma, with supporting laboratory evidence on the major role of the CASR in the chemopreventive effects of calcium [55]. In addition, a statistically significant inverse association between calcium intake and colorectal cancer was found within the highest tertile of total vitamin D intake [22].

Epidemiological studies in adults suggested that antioxidants such as caroteroids, which are abundant in fruit and vegetables may protect against cancer [56], [57] and other chronic diseases [58]. Basic research supports the concept that reactive oxygen species precipitate changes that results in oxidative damage to lipid, protein, and DNA biomolecules [59]. A clinic trial on antioxidant status and risk of cancer in the SU.VI.MAX study in France reported that the effect of supplementation at baseline antioxidant status is related to risk of cancer in men, but not in women [60]. Oxidative stress increases frameshift mutations have been demonstrated in human colorectal cancer cells [61]. Results on the association between dietary carotenoids and antioxidants and risk of colorectal cancer have also been inconsistent. An inverse association between dietary intake of beta-carotene or carotene and rectal cancer was observed in some studies [50], [62]. The Polyp Prevention Trial on the relation of serum and dietary carotenoids and vitamin A to adenomatous polyp recurrence in a low fat, high fiber, high fruit and vegetable dietary intervention supported the protective association of serum or dietary alpha-carotene and vitamin A with adenoma recurrence, especially in non-smokers and non-drinkers [63]. A recent case-control study suggested that lycopene might modestly reduce the risk of rectal cancer in women [28]. However, a case-control analysis was taken in a large prospective study of Canadian women who were enrolled in the Canadian National Breast Screening Study found no association between dietary carotenoids and colorectal cancer [64]. Similarly, intake of carotene and vitamin A from food or from supplements was not associated with risk of colorectal cancer in another case-control study [65]. However, our results showed that use of vitamin A supplement, but not beta-carotene, for more than five years was associated with reduced risk of rectal cancer in females only. It is possible that the association of beta-carotene with rectal cancer is through the regulation of cell growth. The availability of beta-carotene as a precursor to vitamin A and the maintenance of cell integrity may be more important for estrogen-positive women, since estrogen may increase cell proliferation [28].

Recently, Murtaugh and colleagues suggested that high intake of dietary vitamin E was associated with a modest reduction of rectal cancer risk in women only, and the association was modified by estrogen status in a case-control study [28]. Our results also showed that use of vitamin E supplement for five or more years might reduce the risk of rectal cancer in females only. However, results from the American Cancer Society's Cancer Prevention Study II cohort during 14 years follow-up reported that regular use of vitamin E supplement among men and women combined was not associated with mortality in colorectal cancer or rectal cancer [66]. The association was also not found in other studies (either for colorectal adenomatous polyps [65] or for rectal cancer in men and women [23] or that did not examine sex-specific associations [50]). Further research is needed to clarify the association between vitamin E and rectal cancer, and assess whether estrogen status influences the relation of vitamin E to rectal cancer risk.

Vitamin C intake from food or from supplements was not associated with risk of either rectal cancer (in our study and another one) [28] or colorectal cancer [65], [66], but an inverse association between vitamin C intake and rectal cancer was found in another study [50]. A previous case-control study reported that iron intake was associated with a significant increase in risk of rectal cancer at low levels of vitamin C intake [25]. In addition, taking B-complex vitamin supplement for more than five years reduced the risk of rectal cancer among women only in our study but not in another [50].

Results from three randomized trials suggested that higher selenium status might be related to decreased risk of colorectal cancer [67]. Our findings also showed that taking selenium supplement for more than five years was associated with reduced rectal cancer in women only; whereas, no such association appeared in another study [28]. We also found that taking a zinc supplement reduced the risk of rectal cancer in women only, but another study did not find an association [50]. In addition, dietary iron intake was associated with significant increases in risk of rectal cancer among men and women in one study [25] but not in another case-control study [68] or in the present study.

A major strength of this study was the large population-based sample from seven of ten Canadian provinces. Another strength was the use of a well-known food frequency questionnaire to collect dietary information. Nevertheless, several limitations should be mentioned. The first is the potential misclassification of exposure, particularly from the quality of information concerning dietary habits two years ago. The possibility of differential misclassification (recall bias) cannot be excluded in retrospective case-control studies. Individuals with cancer recall their diet differently from the healthy controls. Participants may have under-or-over-estimated their actual food consumption. However, non-differential misclassification between cases and controls would bias the odds ratios toward unity in most instances [69]; consequently, actual risks may have been stronger than observed.

The second limitation is that the potential confounding effects of medical history could not be considered. In particular, we did not have data on family history of rectal cancer, for which associations have been reported in one study [70] but not in another [27]. Associations with family history are the strongest for colon cancer rather than for rectal cancer [27], [70]. The expression of familial susceptibility can be substantially modified by adult life risk factors [71].

The third limitation is that we could not analyse in detail nutrients such as protein and vitamins, and we categorized use of vitamin or mineral supplements only by duration of use. We thus present risks separately for foods and supplements.

A final limitation is that selection bias might have been introduced into our data. Although the NECSS was conducted on a large population-based sample from eight Canadian provinces and the cases were 100% coverage in participating cancer registries, 11.1% of the rectal cancer cases (who were too ill or had died) were not included in this study. The response rate was 64.4% of cases ascertained and 72.4 of patients contacted, which might lead to selection bias. However, the significant associations between consumption of plant-based foods and meat and risk of rectal cancer in our results are consistent with results from other studies [12], [24], [26]. Thus, we think that the results obtained were unlikely to be biased by missing data.

For the controls, the method of selection varied by province. Two provinces used random digit dialing and the rest of provinces used provincial health plans in the rectal cancer study to obtain a population sample. This might have potential selection bias. However, we did interaction analyses to examine the differences between the two methods to select controls in the study, and no difference was found in selecting the controls between the two methods (p-values >0.05).

In conclusion, our findings add to the evidence that diet may play an important role in the aetiology of rectal cancer. These results suggest that some of the dietary risk factors for rectal cancer differ by sex.

The Canadian Cancer Registries Epidemiology Research Group comprises a principal investigator from each of the provincial cancer registries involved in the National Enhanced Cancer Surveillance System: Bertha Paulse, Newfoundland Cancer Foundation; Ron Dewar, Nova Scotia Cancer Registry; Dagny Dryer, Prince Edward Island Cancer Registry; Nancy Kreiger, Cancer Care Ontario; Heather Whittaker, Manitoba Cancer Treatment and Research Foundation; Diane Robson, Saskatchewan Cancer Foundation; Shirley Fincham, Division of Epidemiology, Prevention and Screening, Alberta Cancer Board; and Nhu Le, British Columbia Cancer Agency.