Dietary nitrate and nitrite intake and risk of colorectal cancer in the Shanghai Women’s Health Study

Abstract

Nitrate and nitrite are precursors of endogenously formed N-nitroso compounds (NOC), known animal carcinogens. Nitrosation reactions forming NOCs can be inhibited by vitamin C and other antioxidants. We prospectively investigated the association between dietary nitrate and nitrite intake and risk of colorectal cancer in the Shanghai Women’s Health Study, a cohort of 73,118 women ages 40 to 70 residing in Shanghai. We evaluated effect modification by factors that affect endogenous formation of NOCs: vitamin C (at or above/below median) and red meat intake (at or above/below median). Nitrate, nitrite and other dietary intakes were estimated from a 77-item food frequency questionnaire administered at baseline. Over a mean of 11 years of follow-up, we identified 619 colorectal cancer cases (n=383, colon; n=236, rectum). Hazard ratios (HR) and 95% confidence intervals (CI) were estimated using Cox proportional hazard regression. Overall, nitrate intake was not associated with colorectal cancer risk (HR = 1.08; 95% CI: 0.73–1.59). However, among women with vitamin C intake below the median (83.9 mg/day) and hence higher potential exposure to NOCs, risk of colorectal cancer increased with increasing quintiles of nitrate intake (highest vs. lowest quintile HR = 2.45; 95% CI: 1.15–5.18; p-trend = 0.02). There was no association among women with higher vitamin C intake. We found no association between nitrite intake and risk of colorectal cancer overall or by intake level of vitamin C. Our findings suggest that high dietary nitrate intake among subgroups expected to have higher exposure to endogenously-formed NOCs increases risk of colorectal cancer.

Introduction

Nitrate and nitrite are found in high concentrations in certain foods, especially green, leafy and root vegetables¹ and processed meats², respectively. Ingested nitrate is reduced to nitrite by the bacterial flora in the mouth and digestive tract. In turn, nitrite reacts with amines, amides and other nitrosation precursors in the gastrointestinal tract to form N-nitroso compounds (NOCs). Most NOCs are potent animal carcinogens.^3–4 There is also limited evidence of an association between NOC exposure and increased risk of certain cancers in humans, including colon and rectum cancers.^5–7 Specifically, N-nitrosodimethylamine (NDMA) from food sources has been associated with increased risks of colon and rectum cancers in European populations.^5–6 Because endogenous nitrosation is estimated to account for 45–75% of total NOC exposure,⁸ dietary intake of nitrate and nitrite, precursors for endogenous nitrosation, may be important colorectal cancer risk factors.

Evidence supporting an association between nitrate and nitrite intake and colorectal cancer risk to date has been limited. A case-control study in Iowa observed a positive association between dietary nitrite intake and risk of colon and rectum cancers.⁹ While no association was found with dietary nitrate, risk of colon cancer was increased among individuals with low vitamin C intake (<131.8 mg/day) and with exposure to higher levels of nitrate in drinking water (>5mg/L as nitrate-nitrogen [N]; one-half of the U.S. maximum contaminant level [MCL]) for over 10 years. Vitamin C and other antioxidants such as vitamin E and some carotenoids, inhibit endogenous NOC formation.^10–12 European⁶ and U.S.¹³ cohorts found no association between colorectal cancer incidence and dietary nitrate intake; whereas, the U.S. NIH-AARP Diet and Health Study cohort found an inverse association.¹⁴ Nitrite was not associated with colorectal cancer risk in two European cohorts^5–6; In the NIH-AARP study, nitrite intake overall was not associated with colorectal cancer risk; however, nitrate and nitrite from processed meat sources were positively associated with risk.¹⁴ In all four cohort studies, the interactive effect of vitamin C on dietary nitrate and nitrite intake and risk of colorectal cancer was not investigated.

Although there have been few studies comprehensively investigating dietary nitrate and nitrite and risk of colorectal cancers, other dietary factors have received greater attention. There is strong evidence of an association between red and processed meat consumption and risk of colorectal cancer.^{14–16,17–18} There are several potential mechanisms for an increased risk of colorectal cancer associated with red and processed meats, including catalysis of endogenous NOC formation by heme in red meats and exogenous NOC exposure from processed meats.¹⁸ There is also some evidence of a protective effect associated with vegetable¹⁹ and/or fruit intake.^20–21 However, almost all previous studies of dietary factors, including nitrate and nitrite intake and risk of colorectal cancer, have been based on populations with Western style diets.

The Shanghai Women’s Health Study (SWHS) is a large prospective cohort study of women residing in Shanghai, China. Unlike Western diets, where dietary nitrite is largely derived from processed meats, fruits and vegetables are the greatest contributors to both nitrate and nitrite intake in the SWHS.²² This is largely due to lower intake of processed meats and different patterns of vegetable consumption compared to Western diets. To further investigate these exposures in an Asian population, we evaluated the associations between dietary intake of nitrate and nitrite overall and among subgroups expected to have higher endogenous nitrosation and risk of colorectal cancer in the SWHS.

Materials and Methods

Study population

The SWHS is a prospective cohort of women residing in Shanghai, China. Complete details of the cohort have been described previously.²³ In brief, a roster of all (N=81,170) women aged 40 to 70 years residing in seven urban communities in Shanghai, China was obtained. A total of 75,221 (92.7%) women participated in the study and completed a baseline interview during 1996 to 2000. Trained interviewers conducted in-person interviews using a detailed questionnaire to collect information on demographic characteristics, dietary habits, physical activity, family history of cancer, smoking, alcohol consumption, medical conditions, body measurements, and surgery history. In-person follow-up interviews were conducted with 99.8% (2000–2002), 98.7% (2002–2004) and 96.7% (2004–2007) of cohort members or their next-of-kin.

Of the 75,221 study participants, we excluded 279 who were found to be outside the 40–70 year age range, 1,578 women who had prior malignancies reported at baseline, 128 women who reported extreme caloric intake (>3,500 kcal/day or <500 kcal/day), 47 women who had incomplete dietary nitrate or nitrite information and 1 record that was found to be a duplicate. A total of 73,188 women were included in final analyses.

Incident colon and rectum cancer cases were identified through biennial in-person follow-up interviews and by linkage to the Shanghai Cancer Registry. Cancer diagnoses were verified by review of medical charts and pathology slides at diagnostic hospitals. Causes of death were identified from death certificate data collected from the Shanghai Vital Statistics Unit. A total of 619 colorectal cancer cases (383 colon, 236 rectum) were identified over 793,043 person-years of follow-up through December 2009.

Assessment of dietary nitrate and nitrite

A 77-item food frequency questionnaire (FFQ) was administered at baseline. Intake of each food line item was assessed by asking each participant how frequently (daily, weekly, monthly, annually, never) they consumed the food during the past year and, except in the case of preserved food items, followed by questions on amount of intake per unit of time. Nutrient intakes, except for nitrate and nitrite, were calculated by multiplying daily intake by the nutrient content obtained from the Chinese Food Composition Tables and summed across all foods.²⁴

The FFQ was validated in the SWHS cohort²⁵, although nitrate and nitrite intakes were not assessed at that time. Details of the validation study can be found elsewhere.²⁵ Briefly, in a subset of 200 participants, FFQ intakes at baseline were compared to multiple 24-hr recalls conducted twice a month over a period of 12 months. A second FFQ was administered at the end of the 12 month period. Usual intakes of most major food groups and nutrients were found to be accurately and reliably measured with the exception of seasonal fruits and vegetables, which were overestimated. Therefore, seasonal fruit and vegetable intakes were adjusted based on a subset of randomly selected women (~1,000/season) who reported the number of months out of the year the food was eaten.²⁵

Calculation of nitrate and nitrite intake has been described in detail previously.²² Briefly, nitrate and nitrite contents for 29 vegetables, 19 meats, 6 processed foods, rice and noodles, 13 dessert and bean items, and 8 fruits were obtained from published literature focusing on Chinese foods. Nitrate and nitrite concentrations for vegetables were available from Shanghai.²⁶ Values for other foods were mostly from surveys of foods in China or other Asian countries.²⁷ Nitrate and nitrite content for fresh meat,²⁸ breads,²⁹ apples, pears, grapes, peaches and strawberries³⁰ were assigned from Western countries because no measurements from Asia were available. The nitrate and nitrite content for each FFQ line item was calculated as the mean of published values for that food, weighted by number of samples analyzed. Dietary nitrate and nitrite intake was calculated for all foods and for plant, animal and processed meat sources separately. Details on the contribution of specific food items to nitrate and nitrite intake have been presented elsewhere.²² In brief, the major contributors to nitrate intake were greens/Chinese greens (42.7%) and watermelon (12.5%). Watermelon was also the major contributor to nitrite intake (15.2%) followed by fresh bean curd (9.7%) and salted/preserved vegetables (9.1%).

Although nitrate from drinking water can be a substantial source of nitrate intake when levels in drinking water are high (>10mg/L nitrate-N),^31–32 nitrate intake from drinking water sources was not considered in our assessment of exposure because we determined that exposures from tap water in Shanghai were low. A total of 70,586 women (96%) lived in urban Shanghai for at least 10 years prior to enrollment. Of the 54,189 women (74%) who responded to a question about their drinking water source, 52,229 (96%) indicated they drank tap water in the previous 10 years. Average nitrate levels in the drinking water supplies in Shanghai measured in 2004 were low (approximately 2 mg/L as nitrate-N).^33–34 Thus, nitrate from drinking water would not be expected to contribute a substantial proportion of total nitrate intake.

Statistical analysis

Cox proportional hazards regression models with person-years as the underlying time metric were used to estimate hazard ratios (HR) and 95% confidence intervals (CIs) for the association between colon, rectum and colorectal cancers combined and quintiles of nitrate and nitrite intake. The proportional hazards assumption was tested and upheld in all analyses. All analyses were adjusted for total energy intake using the nutrient density method³⁵ and nitrate and nitrite intakes are expressed as intake per 1,000 kilocalories (kcal). All models were adjusted for age, total energy intake, education (college/high school/middle school/≤elementary school), physical activity (metabolic equivalent tasks-hours/day, dietary vitamin C intake (mg/1,000 kcal), carotene and folate. We also considered other potential risk factors for colorectal cancer including; marital status, body mass index (<18, 18 to <25, 25 to <30, ≥30 kg/m²), smoking status (ever/never), alcohol consumption (ever/never), history of familial polyp (yes/no), family history of cancer (yes/no), use of aspirin within the 12 months prior to enrollment (ever/never) and dietary intakes of vitamin E, calcium, fiber and fat. However, they are not presented in final models because their inclusion did not substantially change HR estimates (≥10%).

We examined risk associated with nitrate and nitrite intake overall, and from animal and plant sources separately. We also evaluated nitrate and nitrite intake from specific food sources including preserved foods overall (salted vegetables, salted egg, salted fish, salted meat, sausage, smoked meat), preserved meats (salted meat, salted fish, sausage, smoked meat), and preserved vegetables. Additionally, we investigated the risk of colorectal cancer associated with daily intake of red meat, greens/Chinese greens (greatest contributor of nitrate), watermelon (greatest contributor of nitrite), and individual preserved foods.

We conducted analyses of nitrate and nitrite intake stratified by the median intakes (at or above/below median) of vitamin C, vitamin E, red meat, and tea to evaluate effect modification by factors potentially affecting endogenous formation of NOCs. To evaluate the consistency of the associations, we stratified by age (at or above/below median [50 years]) and education (high school or greater/less than high school). Results of stratified analyses were confirmed by examining the interaction cross-product terms in multivariable models.

We conducted two sensitivity analyses. The first excluded colon (n=37) and rectum (n=27) cancer cases diagnosed within 2 years of enrollment. Second, we excluded women who had lived in urban Shanghai less than 10 years prior to enrollment (n=2,602) because they may have had higher exposures to nitrate from drinking water sources at previous residences and women who did not report drinking tap water within 10 years of enrollment (n=1,920).

Tests for linear trend were conducted using the median value of each exposure category as a continuous variable in the model. All p-values were 2-sided; α = 0.05 indicates statistical significance, and HRs and 95% CIs were calculated using SAS statistical software version 9.2 (SAS Institute, Inc., Cary, North Carolina).

Results

The median daily intakes of dietary nitrate and nitrite were 300.7 mg/day (IQR: 214.5–412.5 mg/day) and 1.4 mg/day (IQR: 1.1–1.8 mg/day), respectively. The majority of dietary nitrate intake (median= 298.6 mg/day) and nitrite intake (median= 1.2 mg/day) was from plant sources. Characteristics of the cohort by quintiles of nitrate and nitrite intake are shown in Table 1. Women in the highest quintiles of nitrate and nitrite intake reported lower rates of ever smoking and were more physically active than women in the lowest quintiles. A greater percentage of women in the highest quintile of nitrite, but not nitrate intake had at least a high school education compared to women in the lowest quintile. Higher nitrate and nitrite intake was also associated with greater consumption of vegetables, fruits, preserved foods, fiber, protein, vitamins C and E, carotene, folic acid and calcium.

Table 1.

Characteristics of the Shanghai Women’s Health Study cohort, by daily nitrate and nitrite intake

	Nitrate			Nitrite
	Q1	Q3	Q5	Q1	Q3	Q5
Energy adjusted median intake (mg/1000kcal)	98.7	182.4	313.2	0.56	0.87	1.23
Age at baseline (years)	52.7	51.6	52.4	53.8	51.5	51.5
Family history of any cancer (%)	24.6	25.1	25.1	23.1	25.6	26.5
Currently married (%)	88.2	90.3	86.4	85.1	90.3	89.5
Body mass index (kg/m2)	23.8	23.9	24.4	24.2	23.9	24.2
Ever smoker (%)	3.7	2.3	2.9	4.3	2.4	2.4
Ever regular drinker of alcohol (%)	2.3	2.3	2.5	2.1	2.1	2.8
Education, high school or more (%)	40.6	43.2	37.3	29.5	44.9	46.2
Physical activity (met /day/yr)a	0.54	0.67	0.9	0.67	0.73	0.89
Daily dietary intake (mean)
Energy (kcal)	1697.1	1698.9	1610.3	1637.6	1706.8	1653.5
Vegetables (g/1000kcal)	88.1	164.7	296.2	117.5	169.9	253.7
Cruciferous vegetables (g/1000kcal)	21.4	50.2	109.8	45.2	54.6	74.7
Fruits (g/1000kcal)	103.5	164.8	211.7	73.2	157.9	258.6
Red meat (g/1000kcal)	29.5	30.3	27.3	28.9	30.5	27.5
Preserved food (g/1000kcal)	24.4	29.4	31.6	15.0	27.3	45.1
Salted preserved meat (g/1000kcal)	0.78	0.87	0.87	0.59	0.80	1.25
Smoked meat/bacon (g/1000kcal)	0.34	0.41	0.40	0.17	0.37	0.73
Preserved vegetables (g/1000kcal)	17.4	21.2	23.0	8.9	19.6	34.8
Fat (g/1000kcal)	16.3	17.5	17.7	14.8	17.7	18.8
Fiber (g/1000kcal)	4.8	6.2	8.1	4.8	6.2	8.0
Protein (g/1000kcal)	36.7	39.7	42.5	35.5	40.0	43.1
Iron (mg/1000kcal)	9.6	10.9	12.8	9.8	11.0	12.4
Vitamin C (mg/1000kcal)	29.8	53.1	87.9	34.5	53.8	81.2
Vitamin E (mg/1000kcal)	5.4	7.0	8.8	4.8	7.0	9.5
Carotene (mg/1000kcal)	780.4	1502.8	2759.4	1121.6	1566.2	2260.3
Folic acid (μg/1000kcal)	139.2	170.2	215.9	143.5	171.9	207.7
Calcium (mg/1000kcal)	209.1	268.5	355.7	200.3	273.5	352.3

^ametabolic equivalent tasks-hours/day/year; 1 met = 1 kcal/kg/hr

Dietary nitrate and nitrite intakes overall, and from animal and plant sources separately, were not significantly associated with risk of colorectal cancer (Table 2). Compared to women in the lowest quintile of nitrate intake from preserved food sources, risk of colorectal cancer was significantly elevated in the fourth quintile of intake (HR=1.31; 95% CI: 1.02–1.69), but not in the highest quintile (p-trend = 0.75). Examining colon and rectum cancer risk separately, we found no association with intakes of nitrate and nitrite overall or from animal sources (Table 2). Compared with the lowest quintile of nitrite intake from plant sources, we observed a significantly increased risk of colon cancer among women in the third quintile (HR = 1.43; 95% CI: 1.02–2.02), but not in higher quintiles (p-trend=0.23). We also observed significantly increased risks of colon cancer associated with higher intakes of nitrate and nitrite from preserved foods, although the trends were not monotonic (p-trend=0.61 and 0.44, respectively). There was a suggestion of an inverse relationship between rectum cancer risk and intake of nitrite from plant sources, which did not achieve statistical significance (p-trend=0.08). In analyses of food groups and specific foods, we found no significant dose-response relationships between intakes of preserved foods overall, preserved meats, red meat, preserved vegetables, greens/Chinese greens, and watermelon, and risks of colon, rectum and colorectal cancers combined (results not shown).

Table 2.

Hazard ratios (HR) and 95% confidence intervals (CI) of colorectal, colon and rectum cancer for quintiles of total nitrate and nitrite, and nitrate and nitrite intakes from animal, plant, and preserved food sources

Quintile of exposure	Median intake (mg/day)	Colorectal Cancer		Colon Cancer		Rectum Cancer
		Cases	HRa (95% CI)	Cases	HRa (95% CI)	Cases	HRa (95% CI)
TOTAL
Nitrate
Q1	98.7	129	1.00	83	1.00	46	1.00
Q2	144.1	109	0.90 (0.69, 1.17)	70	0.90 (0.65, 1.25)	39	0.90 (0.58, 1.40)
Q3	182.4	106	0.88 (0.66, 1.16)	65	0.84 (0.59, 1.20)	41	0.95 (0.60, 1.50)
Q4	229	138	1.14 (0.85, 1.54)	87	1.13 (0.77, 1.66)	51	1.17 (0.72, 1.90)
Q5	313.2	137	1.08 (0.73, 1.59)	78	0.98 (0.59, 1.63)	59	1.26 (0.69, 2.32)
P trend			0.39		0.75		0.29
Nitrite
Q1	0.56	129	1.00	72	1.00	57	1.00
Q2	0.74	126	1.09 (0.85, 1.40)	81	1.27 (0.92, 1.76)	45	0.87 (0.58, 1.29)
Q3	0.87	123	1.10 (0.85, 1.43)	75	1.23 (0.88, 1.73)	48	0.94 (0.63, 1.42)
Q4	1.01	122	1.10 (0.84, 1.44)	80	1.34 (0.94, 1.90)	42	0.81 (0.52, 1.25)
Q5	1.23	119	1.05 (0.77, 1.42)	75	1.26 (0.85, 1.86)	44	0.80 (0.49, 1.29)
P trend			0.78		0.27		0.35
ANIMAL SOURCE
Nitrate
Q1	0.51	130	1.00	77	1.00	53	1.00
Q2	0.85	142	1.31 (1.03, 1.66)	86	1.34 (0.98, 1.82)	56	1.26 (0.86, 1.85)
Q3	1.13	113	1.12 (0.87, 1.45)	72	1.21 (0.87, 1.68)	41	1.00 (0.66, 1.52)
Q4	1.45	124	1.28 (0.99, 1.66)	79	1.38 (1.00, 1.91)	45	1.14 (0.76, 1.73)
Q5	2.06	110	1.15 (0.88, 1.50)	69	1.22 (0.87, 1.72)	41	1.05 (0.68, 1.61)
P trend			0.46		0.33		0.97
Nitrite
Q1	0.05	151	1.00	90	1.00	61	1.00
Q2	0.08	121	0.96 (0.75, 1.23)	74	0.99 (0.72, 1.35)	47	0.92 (0.63, 1.36)
Q3	0.11	121	1.06 (0.82, 1.35)	78	1.14 (0.83, 1.56)	43	0.93 (0.62, 1.39)
Q4	0.14	107	0.97 (0.75, 1.26)	72	1.09 (0.79, 1.51)	35	0.79 (0.51, 1.23)
Q5	0.19	119	1.17 (0.90, 1.51)	69	1.14 (0.81, 1.59)	50	1.21 (0.81, 1.82)
P trend			0.27		0.38		0.49
PLANT SOURCE
Nitrate
Q1	97.6	128	1.00	83	1.00	45	1.00
Q2	142.8	110	0.92 (0.70, 1.19)	70	0.90 (0.65, 1.26)	40	0.94 (0.61, 1.46)
Q3	181.1	106	0.88 (0.67, 1.17)	64	0.83 (0.58, 1.19)	42	0.99 (0.63, 1.56)
Q4	227.8	137	1.15 (0.85, 1.55)	87	1.15 (0.78, 1.68)	50	1.17 (0.72, 1.90)
Q5	311.9	138	1.10 (0.75, 1.63)	79	1.01 (0.61, 1.68)	59	1.28 (0.70, 2.36)
P trend			0.35		0.66		0.31
Nitrite
Q1	0.47	120	1.00	65	1.00	55	1.00
Q2	0.63	130	1.17 (0.91, 1.50)	82	1.38 (0.99, 1.93)	48	0.91 (0.62, 1.35)
Q3	0.75	133	1.22 (0.94, 1.58)	82	1.43 (1.02, 2.02)	51	0.97 (0.65, 1.45)
Q4	0.89	117	1.06 (0.80, 1.40)	76	1.34 (0.93, 1.91)	41	0.75 (0.48, 1.16)
Q5	1.11	119	1.03 (0.76, 1.39)	78	1.36 (0.92, 2.01)	41	0.67 (0.41, 1.09)
P trend			0.88		0.23		0.08
Preserved Foods
Nitrate
Q1	0.14	110	1.00	59	1.00	51	1.00
Q2	0.62	117	1.16 (0.90, 1.51)	72	1.35 (0.96, 1.91)	45	0.94 (0.63, 1.41)
Q3	1.65	124	1.20 (0.93, 1.55)	79	1.42 (1.01, 1.99)	45	0.95 (0.64, 1.42)
Q4	2.99	137	1.31 (1.02, 1.69)	92	1.64 (1.18, 2.28)	45	0.93 (0.62, 1.39)
Q5	8.24	131	1.13 (0.87, 1.45)	81	1.28 (0.92, 1.80)	50	0.95 (0.64, 1.40)
P trend			0.75		0.61		0.87
Nitrite
Q1	0.01	113	1.00	60	1.00	53	1.00
Q2	0.02	126	1.25 (0.97, 1.62)	78	1.48 (1.06, 2.08)	48	1.00 (0.67, 1.48)
Q3	0.06	123	1.22 (0.94, 1.57)	76	1.42 (1.01, 1.99)	47	0.99 (0.67, 1.47)
Q4	0.11	124	1.21 (0.94, 1.56)	85	1.56 (1.12, 2.18)	39	0.81 (0.53, 1.22)
Q5	0.29	133	1.16 (0.90, 1.49)	84	1.36 (0.98, 1.90)	49	0.92 (0.62, 1.36)
P trend			0.78		0.44		0.59

^aAdjusted for age, energy intake, education, physical activity, dietary vitamin C intake, carotene and folate

Analyses of nitrate intake stratified by dietary vitamin C (≥83.9/<83.9 mg/day) and the risk of colorectal cancer (Table 3) showed evidence of higher risk among women with vitamin C intake below the median, although the test for interaction did not reach statistical significance (p-interaction = 0.19). Among women with vitamin C intake below the median, colorectal cancer risk increased with increasing nitrate intake (p-trend=0.02), with a significantly increased risk among those in the highest quintile of intake compared to the lowest quintile (HR = 2.45; 95% CI: 1.15–5.18). No association between nitrate intake and risk of colorectal cancer was observed among women with vitamin C intake at or above the median. Nitrite intake was not associated with colorectal cancer risk among women with high or low vitamin C intakes (Table 3). When we evaluated colon and rectum cancers separately by vitamin C intake, we observed a risk pattern for rectum cancer that was similar to colorectal cancer overall; whereas, colon cancer risk, although elevated, did not vary by vitamin C intake (Supplemental Table 1).

Table 3.

Hazard ratios (HR) and 95% confidence intervals (CI) of colorectal cancer risk for quintiles of total nitrate and nitrite, by vitamin C intake

Quintile of exposure	Median intake (mg/day)	Cases	HRa (95% CI)
Low vitamin C intake (<83.9mg/day)
Nitrate
Q1	97.1	119	1.00
Q2	142.5	83	1.19 (0.84, 1.67)
Q3	180.4	55	1.27 (0.81, 1.97)
Q4	224.8	36	1.63 (0.93, 2.86)
Q5	291.5	20	2.45 (1.15, 5.18)
P trend			0.02
Nitrite
Q1	0.55	107	1.00
Q2	0.74	85	1.15 (0.86, 1.55)
Q3	0.87	59	1.11 (0.79, 1.57)
Q4	1.0	38	1.03 (0.69, 1.54)
Q5	1.2	24	1.10 (0.68, 1.79)
P trend			0.74
High vitamin C intake (≥83.9mg/day)
Nitrate
Q1	111	10	1.00
Q2	147.2	26	0.76 (0.36, 1.57)
Q3	184.4	51	0.84 (0.42, 1.68)
Q4	230.8	102	1.14 (0.57, 2.26)
Q5	317.5	117	0.93 (0.44, 1.96)
P trend			0.69
Nitrite
Q1	0.61	22	1.00
Q2	0.75	41	0.87 (0.52, 1.47)
Q3	0.88	64	0.96 (0.59, 1.58)
Q4	1.01	84	1.00 (0.61, 1.62)
Q5	1.24	95	0.90 (0.55, 1.49)
P trend			0.85

^aAdjusted for age, energy intake, education, physical activity, dietary vitamin C intake, carotene and folate

Among women with red meat intake at or above the median (≥43.5 mg/day) (Table 4), risk of colorectal cancer was significantly increased in the fourth quintile of nitrate intake (HR = 1.66; 95% CI: 1.02–2.68), but not in the highest quintile (p-trend=0.14) compared to the lowest quintile. There was no association with nitrate intake among women with red meat intake below the median. Nitrite intake was not associated with colorectal cancer risk in either the high or low red meat intake groups.

Table 4.

Hazard ratios (HR) and 95% confidence intervals (CI) of colorectal cancer risk for quintiles of total nitrate and nitrite, by red meat intake

Quintile of exposure	Median intake (mg/day)	Cases	HRa (95% CI)
Low red meat intake (<43.5mg/day)
Nitrate
Q1	97.0	79	1.00
Q2	144.3	60	0.85 (0.61, 1.21)
Q3	182.5	47	0.65 (0.44, 0.95)
Q4	229.3	65	0.86 (0.59, 1.27)
Q5	319.3	90	0.94 (0.58, 1.52)
P trend			0.97
Nitrite
Q1	0.55	79	1.00
Q2	0.74	64	1.04 (0.74, 1.46)
Q3	0.87	64	1.07 (0.76, 1.51)
Q4	1.01	64	1.06 (0.74, 1.52)
Q5	1.25	70	0.96 (0.65, 1.43)
P trend			0.88
High red meat intake (≥43.5mg/day)
Nitrate
Q1	100.5	50	1.00
Q2	143.9	49	1.01 (0.67, 1.52)
Q3	182.3	59	1.30 (0.85, 2.01)
Q4	228.6	73	1.72 (1.07, 2.79)
Q5	305.7	47	1.40 (0.73, 2.70)
P trend			0.14
Nitrite
Q1	0.58	50	1.00
Q2	0.74	62	1.16 (0.79, 1.70)
Q3	0.87	59	1.16 (0.78, 1.73)
Q4	1.01	58	1.17 (0.77, 1.78)
Q5	1.22	49	1.17 (0.73, 1.89)
P trend			0.56

^aAdjusted for age, energy intake, education, physical activity, dietary vitamin C intake, carotene and folate

We did not observe any significant effect modification between nitrate and nitrite intake and risk of colon, rectum and colorectal cancers by vitamin E intake, tea consumption, age or education.

Risk estimates from two separate sensitivity analyses, one excluding cases diagnosed within 2 years of enrollment and the other excluding women living in urban Shanghai for less than continuous 10 years prior to enrollment and women who did not report drinking tap water, were similar to HRs estimated from the full study population (results not shown).

Discussion

Overall, we did not observe associations between dietary nitrate and nitrite intakes and risk of colorectal cancer in this prospective cohort of women in Shanghai. However, risk was significantly increased among subgroups of the cohort who would be expected to have higher endogenous formation of NOCs, specifically those with low vitamin C intake. Vitamin C is an antioxidant known to inhibit endogenous formation of NOCs, which are compounds that have demonstrated carcinogenicity in animal studies.^{3, 36} Endogenous formation of NOCs in the stomach and intestinal tract, in the absence of high vitamin C, is a plausible mechanism for colorectal cancer risk. We also observed increased risk of colon cancer associated with nitrate and nitrite intake from preserved foods, although there was not a monotonic exposure-response.

To date there has been limited investigation of dietary nitrate and nitrite intake and risk of colorectal cancer and to our knowledge no study has examined the potential interaction of dietary intake and vitamin C. A Finnish cohort study found no association between nitrate or nitrite intake and risk of colorectal cancer; however, colorectal cancer was increased with higher intake of NDMA from dietary sources.⁶ NDMA intake was also associated with increased risk of rectum cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC) study.⁵ Nitrite intake was not associated with risk of gastrointestinal cancers in the EPIC cohort and nitrate intake was not evaluated.⁵ There was also no association between dietary nitrate and risk of colon and rectum cancers in a cohort study of older women in Iowa; dietary nitrite and NDMA intakes were not evaluated.¹³

A case-control study in Iowa found an increased risk of colon and rectum cancers associated with dietary nitrite intake.⁹ In contrast, dietary nitrate intake was inversely associated with colon cancer risk and not associated with risk of rectum cancer.⁹ Although potential interactive effects between dietary nitrate and vitamin C were not examined, colon cancer risk was significantly increased among those with higher nitrate intake from drinking water and low vitamin C intake (<131.8 mg/day) or high meat intake (>1.5 servings/day).⁹ Studies have found that drinking water can be a major source of nitrate intake and may exceed intake from diet when levels are above the MCL established by the U.S. Environmental Protection Agency (≥10 mg/L nitrate-N) ^31–32 or the World Health Organization limit of 11.3 mg/L as nitrate-N. Unlike plant sources of nitrate, which were the primary source of nitrate intake in the Iowa study, nitrate from water is ingested without antioxidants that can inhibit endogenous NOC formation.

Nitrate, after it is reduced to nitrite by bacteria in the oral cavity and digestive tract, and ingested nitrite react with amines, amides and other nitrosatable compounds to form NOCs.³⁷ Red and processed meats are sources of amine, amides, and heme, which catalyze endogenous nitrosation, leading to increased fecal NOC formation.^38–39 The increase in apparent total NOCs measured in feces after ingestion of nitrate⁴⁰ or meat³⁸ provides evidence of colonic NOC exposure, likely via endogenous NOC formation. Furthermore, there is evidence that colonic NOC exposure is associated with gene expression changes that may play a role in colorectal cancer development.^41–42 Consistent with this hypothesis, nitrate intake from processed meat has been associated with risk of colorectal cancer in the U.S. based NIH-AARP Diet and Health Study.¹⁴ While we did find a suggestion of increased risk of colorectal cancer among individuals with higher nitrate intake and high red meat intake, a strict dose-response was not observed. And we did not observe an association between nitrate or nitrite from preserved meats (results not shown) and risk of colorectal cancer overall in the SWHS. However, intake of nitrate and nitrite from animal and processed meat sources is lower among Shanghai women than in U.S. populations,^43–45 For example, nitrite intake from salted/preserved meats, which can be a substantial source of nitrite in U.S. populations,^43–45 represented a small fraction (1.2%) of nitrite intake in the SWHS population.

Vegetables and fruits are the main contributors of nitrate and nitrite in the SWHS. They contain antioxidants, such as vitamin C, and polyphenols, which, if present together with ingested nitrate and nitrite inhibit endogenous formation of NOCs. Although few studies have evaluated dietary nitrate or nitrite directly, a strong association between intake of nitrate and nitrite from plant sources and colorectal cancer would not be expected if risk is mediated through endogenous formation of NOC. Indeed, most studies have shown no association^{19, 46} or even decreased²⁰ risks of colorectal cancers in women associated with fruit and vegetable intake. The Netherlands Cohort Study on Diet and Cancer observed a significant positive trend between brassica vegetable (cruciferous vegetables, including cabbages) intake and risk of rectum cancer among women.²⁰ However, no association was observed between total vegetable consumption and risk of rectum cancer and an inverse association was observed between total vegetable and fruit intake and risk of colon cancer.²⁰

As with other studies, we did not observe an association between nitrate and nitrite from plant sources and colorectal cancer. Our findings do suggest, however; nitrate and nitrite intake from preserved foods, of which the majority of intake came from preserved vegetables, may increase the risk of colon cancer. Chinese preserved vegetables contain measureable amounts of NOCs^47–48 and may contain lower levels of beneficial nutrients than their fresh counterparts.⁴⁹

It is important to note the higher dietary nitrate intake in the Shanghai women’s cohort compared with the U.S. and other Western cohorts. For example, median daily dietary nitrate intake was 3 times higher among Shanghai women (301 mg/day) than in U.S. populations (~100 mg/day).^43–45 The high level of dietary nitrate intake among Shanghai women may explain the increased colorectal cancer risk among those with lower vitamin C intake. Such low intakes have not been reported in Western populations who have lower dietary nitrate exposures than those reported here. ^{6, 9} In addition, the interactive effect of high nitrate intake coupled with lower vitamin C intake, which is expected to result in high endogenous NOC exposure, has not been investigated in previous cohort studies of Western populations.

Strengths of our study include the prospective study design, large study population, and the completeness of follow-up. Further, we used a comprehensive FFQ to assess consumption of a wide variety of food sources. Nitrate levels in most vegetables were estimated from literature values specific for Shanghai. There can be substantial regional variation in nitrate levels in fruits and vegetables throughout China,²⁶ underscoring the need to use Shanghai-specific estimates.

A limitation of our exposure assessment was that dietary nitrate and nitrite intakes were not specifically evaluated in the validation study. However, the questionnaire was previously validated for macro- and micronutrient intakes.²⁵ In terms of assessment of dietary intakes, the FFQ was self-reported and limited to the year prior to baseline. Any changes in diet over time were therefore not captured. Although estimates of nitrate levels for most vegetables and fruits were specific for Shanghai, we were not able to capture annual variations in nitrate levels that may occur due to changes in fertilizer use, irrigation or other growing conditions.⁵⁰ Additionally, drinking water nitrate was not included in our assessment of exposure. However, most women lived in Shanghai for more than 10 years prior to enrollment and drank tap water, which contained low levels of nitrate. Therefore, we would not expect drinking water to contribute a substantial portion of total nitrate intake. Misclassification due to our exposure assessment methodology may act to attenuate the true associations between nitrate and nitrite intakes and colorectal cancer risks. We were also limited in our ability to assess colon and rectum cancer risk by vitamin C intake or in other stratified analyses due to small numbers of cases in specific strata. Finally, we did not estimate intake of preformed NOCs, which have been measured in many preserved vegetables and foods in China.^47–48

The dietary pattern of Shanghai women allowed us to compare colorectal cancer risk over a range of nitrate intakes that included much higher intake levels than those observed in Western populations. Our findings suggest that high nitrate intake among subgroups expected to have high endogenous NOC exposure, e.g. women with low vitamin C intake, are associated with an increased risk of colorectal cancer. In addition, our findings of increased risk of colorectal cancer among women with low vitamin C intake supports endogenous formation of NOCs as a mechanism for carcinogenesis related to nitrate intake and colorectal cancer risk. Given the substantially higher dietary nitrate intakes among Shanghai women compared to Western populations, further investigation addressing the effects of nitrate and nitrite on colorectal cancer at high exposures is warranted.

Novelty and impact statement.

Nitrate and nitrite are precursors in the endogenous formation of potentially carcinogenic N-nitroso compounds (NOC). These nitrosation reactions can be inhibited by vitamin C and other antioxidants. We investigated the potential interaction of dietary nitrate and nitrite and vitamin C on the risk of colorectal cancer, which has not been studied previously in an Asian population. Our findings suggest that high nitrate intake among individuals expected to have high endogenous NOC exposure, e.g. women with low vitamin C intake, is associated with an increased risk of colorectal cancer.

List of abbreviations

CI

confidence interval

FFQ

food frequency questionnaire

HR

hazard ratio

IQR

interquartile range

kcal

kilocalorie

MCL

maximum contaminant level

NDMA

-nitrosodimethylamine

NOC

N-nitroso compounds

SWHS

Shanghai Women’s Health Study