Consumption of whole grains and cereal fiber in relation to cancer risk: a systematic review of longitudinal studies

Abstract

Context: Evidence from previous reviews is supportive of the hypothesis that whole grains may protect against various cancers. However, the reviews did not report risk estimates for both whole grains and cereal fiber and only case–control studies were evaluated. It is unclear whether longitudinal studies support this conclusion. Objective: To evaluate associations between whole grains and cereal fiber in relation to risk of lifestyle-related cancers data from longitudinal studies was evaluated. Data Sources: The following 3 databases were systematically searched: PubMed, EMBASE, and Cochrane CENTRAL. Study Selection: A total of 43 longitudinal studies conducted in Europe and North America that reported multivariable-adjusted risk estimates for whole grains (n = 14), cereal fiber (n = 23), or both (n = 6) in relation to lifestyle-related cancers were included. Data Extraction: Information on study location, cohort name, follow-up duration, sample characteristics, dietary assessment method, risk estimates, and confounders was extracted. Data Synthesis: Of 20 studies examining whole grains and cancer, 6 studies reported a statistically significant 6%–47% reduction in risk, but 14 studies showed no association. Of 29 studies examining cereal fiber intake in relation to cancer, 8 showed a statistically significant 6%–49% reduction in risk, whereas 21 studies reported no association. Conclusions: This systematic review concludes that most studies were suggestive of a null association. Whole grains and cereal fiber may protect against gastrointestinal cancers, but these findings require confirmation in additional studies.

INTRODUCTION

Whole grains are widely recommended as an integral part of a healthy diet,¹ so their significance in chronic disease prevention is of strong interest. Whole grains represent the intact, ground, cracked, or flaked kernel of grains for which the starchy endosperm, germ, and bran are found in the same relative portions as they exist in the intact grain even after separation and reconstitution of the grain.¹ Whole grains provide both nutrients and non-nutrients, which confer numerous health benefits.^2,3 Whole grains can be consumed as a single food, such as ready-to-eat breakfast cereals, brown rice, popcorn, buckwheat, bulgur, millet, oatmeal, quinoa, grain barley, whole rye, and whole wheat.¹ Alternatively, whole grains can be consumed as an ingredient in foods, such as whole-wheat flour incorporated into crackers or breads. Refined grains have been processed to remove the outer bran and inner germ, retaining only the endosperm. This results in a substantial reduction in content of dietary fiber, B vitamins, iron, magnesium, vitamin E, and other components; many refined grain products may be subsequently enriched with these vitamins and minerals, but not with fiber.¹

The 2015–2020 Dietary Guidelines for Americans recommend consuming at least half of the daily grain servings as whole grains to reduce chronic disease risk.¹ Similarly, the World Cancer Research Fund and American Institute for Cancer Research cancer prevention guidelines recommend eating relatively unprocessed cereals with every meal and limiting the intake of refined starchy foods.⁴ Nevertheless, whole grain intakes fall short of dietary guidance. Data from a nationally representative sample suggest that less than 10% of American adults consume the recommended 3 or more whole-grain servings per day,⁵ with no significant increases in whole-grain intakes over the past 10 years.⁶

A review of case–control evidence suggests that whole grains protect against various cancers.^7,8 Cancer is a major public health burden in the United States, with high prevalence and incidence rates.⁹ The World Cancer Research Fund and American Institute for Cancer Research claim that altering diet may play a pivotal role in reducing cancer incidence.⁴ A protective impact of whole grains on cancer risk is biologically plausible.^10–12 Whole grains are a rich source of antioxidants, vitamins, trace minerals, phytate, phenolic acids, lignans, and phytoestrogens with anticarcinogenic properties. Whole grains are also a major source of fiber. Due to the hypothesized association between dietary fiber and cancer,^13,14 it is important to consider whether cereal fiber alone or whole grains, as a collective package of measured and unmeasured dietary constituents, are associated with cancer risk.

Evidence on whole grains and cereal fiber has been previously reviewed in relation to obesity, type 2 diabetes, and cardiovascular disease.² A recent review showed that high intakes of cereal fiber and whole grains had a protective impact on the risk of these chronic diseases.² However, for cancer, previous reviews on the association between whole grains and risk of first incident cancer have been limited to evidence from case–control studies and do not report risk estimates for cereal fiber.^7,8 Moreover, only one review, which focused on colorectal cancer, summarized the evidence from prospective cohorts and nested case–controls on the impact of whole grains and total dietary fiber but not fiber from grains.¹⁴ Therefore, the purpose of the present systematic review is to summarize the evidence from longitudinal epidemiological studies evaluating the impact of both whole grain and cereal fiber consumption on the risk of any cancer. The insights from this literature review provide important information that helps identify gaps in the current literature and the methodological inconsistencies in assessing dietary consumption of whole grains; the findings may also be useful for guiding clinical practice and policy initiatives for preventing cancer.

POTENTIAL MECHANISMS LINKING WHOLE GRAINS TO CANCER RISK

Whole grains may impact cancer risk through a variety of mechanisms. First, whole grains promote increased chewing and higher satiation, by virtue of their fiber content, structural integrity, and particle size, leading to lower body adiposity.¹⁵ Whole-grain foods are also characterized by lower energy density, resulting in reduced caloric intake without a concomitant increase in hunger.¹⁵ Whole grains have also been shown to influence body composition and measures of central adiposity.^16,17 In particular, higher consumption of whole grains is inversely associated with percentage body fat, trunk fat mass, waist circumference, and waist-to-hip ratio.^16,17 In addition to their impact on weight gain, whole grains are associated with lower risk of type 2 diabetes.² High body mass index (BMI) and diabetes are, in turn, associated with higher cancer risk due to metabolic dysregulation observed among obese and diabetic individuals.^18,19 Therefore, improved glycemic response and reduced insulin resistance may be a potential mechanism by which they are protective against cancer.^20,21

Whole grains may also mediate cancer risk through their impact on sex hormones.¹⁰ Whole-grain foods are a source of phytoestrogens, such as lignans and isoflavones, which influence endogenous sex hormone production, metabolism, and biological activity. These substances reduce circulating estrogen levels, inhibit tumor initiation and growth, and decrease early markers of risk for mammary and colon carcinogenesis (reviewed in Slavin¹⁰). Based on these proposed mechanisms, whole grains may be particularly beneficial in reducing the risk of hormone-related cancers.¹⁰ Whole-grain foods are also a rich source of antioxidants, including vitamins (e.g., vitamin E), trace minerals (e.g., selenium), phenolic acids, lignans, and phytoestrogens, which collectively reduce oxidative damage and stress, thereby reducing cancer risk.²² Whole grains are an important source of magnesium, which has been associated with reduced risk of colorectal cancer.²³

The protective role of whole grains in cancer risk, specifically colorectal cancer, may be attributed to the health benefits of their dietary fiber content.²² Dietary fiber, particularly from cereal sources, increases fecal bulk and reduces intestinal transit time, thereby diluting carcinogens and reducing their absorption by the intestinal epithelium. Cereal fiber, in addition to the resistant starch and oligosaccharides in whole grains, is fermented in the colon into short-chain fatty acids including butyrate.¹⁵ Butyrate is the preferred fuel for mucosal cells and has proapoptotic and antineoplastic potential, thus decreasing tumor growth. They also lower intestinal pH, thereby reducing the solubility of free bile acids and diminishing their availability for carcinogenic activity.¹⁵ Finally, cereal fiber has been associated with lower body adiposity² and lower serum estrogen levels (reviewed in Aune et al.¹³), thereby potentially reducing risk for adiposity-related cancers.

METHODS

Approach and methodology

For this systematic review, a comprehensive search of 3 databases—PubMed, EMBASE, and Cochrane CENTRAL—for articles in the English language published from January 1, 1990, through October 17, 2014, was conducted. The Population, Intervention, Comparison, Outcomes, and Study Design criteria²⁴ (Table 1) were used to formulate and narrow the focus of the following research question: Among adults, is consumption of whole grains and cereal fiber in the highest vs lowest category of intake associated with reduced risk of first-incident cancer in longitudinal studies? The Preferred Reporting Items for Systematic Reviews and Meta-Analyses method²⁴ was used, as applicable for the topic of this review, to track and report progress on the systematic search for cohort and experimental studies examining the association between whole grains and cereal fiber in relation to cancer risk among adults.

Table 1.

PICOS criteria for inclusion and exclusion of studies

Parameter	Inclusion criteria	Exclusion criteria
Population	Adults aged >18 y who are free of cancer at baseline Sample size ≥200	Participants aged <18 y Participants include cancer survivors Sample size <200
Intervention/Exposure	Highest category of whole grain and cereal fiber intake (quintiles, quartiles, servings/week, servings/day, grams/day)
Comparison	Lowest category of whole grain and cereal fiber intake (quintiles, quartiles, servings/wk, servings/d, g/d)
Outcomes	Incidence of first primary lifestyle-related cancer: Female cancers Genitourinary cancers Gastrointestinal cancers Head and neck cancers Hematologic cancers	Cancer recurrence Cancer mortality Metastases
Study design	Longitudinal studies A. Observational studies: Prospective cohort studies Retrospective cohort studies B. Intervention studies: Randomized controlled trials Nonrandomized controlled trials	Editorials Case reports Cross-sectional studies Case–control studies Reviews Meta-analyses

The following search terms were used, as appropriate, for each of the 3 databases to capture articles that report risk estimates on whole grain and cereal fiber consumption in relation to cancer risk: (“whole-grain” OR “whole grains” OR “wholegrains” OR “whole meal” OR “wholemeal” OR “cereal” OR “cereals” OR “cereal fiber” OR “cereal fibers” OR “grain fiber” OR “grain fibers” OR “grain fibre” OR “bran” OR “brans”) AND (“cancer” OR “cancers” OR “neoplasm” OR “neoplasms”).

The literature search process is outlined in Figure 1. A total of 1443 research articles resulted from our search: 479 from PubMed, 854 from EMBASE, and 110 from Cochrane CENTRAL. After discarding 327 duplicates (95 from Cochrane CENTRAL and 232 from EMBASE), a total of 1116 research articles were identified for screening by 2 researchers who independently reviewed all manuscripts. Additionally, bibliographies were manually searched to supplement the online search process and ensure that all relevant studies were captured. After removing 1053 articles that were not relevant to the research topic based on abstract screening and identifying 6 additional studies manually, 69 full-text manuscripts were reviewed. There were no relevant randomized controlled trials.

Figure 1.

Flow diagram of the literature search process. A total of 43 studies were included in the final analysis after applying inclusion and exclusion criteria. Fourteen studies reported risk estimates on whole grains, 23 studies reported risk estimates on cereal fiber, and 6 studies reported risk estimates on both whole grains and cereal fiber in relation to cancer, respectively.

A total of 43 longitudinal studies were selected for inclusion in this systematic review. Fourteen studies reported risk estimates for whole grains, 23 reported risk estimates for cereal fiber, and 6 studies reported risk estimates for both, resulting in a total of 20 studies on whole grains and 29 on cereal fiber in relation to cancer, respectively. These studies evaluated whole grains and cereal fiber in relation to certain lifestyle-related cancers, which include female cancers (breast, ovarian, and endometrial), genitourinary cancers (renal, bladder, and prostate cancers), gastroinstestinal cancers (esophageal, gastric, liver, pancreatic, small intestinal, and colorectal cancers), head and neck cancers (oral cavity, pharyngeal, laryngeal, and overlapping regions cancers), and cancers of the reticuloendothelial system (lymphoma, myeloma, and leukemia). There was 1 study on lymphoma, 15 studies on female cancers, 20 studies on gastrointestinal cancers, 6 studies on genitourinary cancers, and 1 study on head and neck cancers.

Inclusion and exclusion criteria

Original research manuscripts were included if they 1) were intervention studies and observational studies, 2) reported estimates for risk of any cancer, 3) had a total sample size of at least 200 subjects, 4) presented hazard ratios or rate ratios, 5) conducted follow-up in cancer cases, and 6) presented multivariate analyses (not univariate analyses). Editorials, reviews, meta-analyses, cross-sectional studies, and case–control studies were excluded. Studies published in a language other than English were also excluded. Research articles that reported risk estimates for whole-grain food sources and total, soluble, and insoluble fiber with no data on total intake of whole grains and cereal fiber were also excluded from this review.

Data extraction

From each original research article, the following data were extracted: the lead author’s last name, year of publication, study location, cohort name, mean or median duration of follow-up, sample size, sex, age, number of cancer cases, method of dietary assessment (food frequency questionnaires [FFQs], diet records, or 24-h recalls), exposure (whole grains and/or cereal fiber), contrast (highest vs lowest category of intake), risk estimates (relative risks, hazard ratios, odds ratios), 95% confidence intervals (CIs), and confounders adjusted for in the final reported models.

Two investigators conducted the search. Both investigators reviewed all abstracts and, subsequently, full-text articles to identify relevant manuscripts for this review based on the inclusion and exclusion criteria. Both authors extracted the data from the relevant research articles and subsequently checked the extracted data to include study setting, year, design, sample, risk estimates, CIs, and covariates for accuracy. Both persons were required to independently assess the articles and then come to an agreement about any articles in question based on meeting the inclusion criteria. The objective parameters, including the sample size, study design, reporting of multivariable-adjusted risk estimates for whole grains and cereal fiber, and appropriate study outcome (first incidence of cancer), were considered to determine if the study’s quality merited its inclusion in this review. Investigators agreed that all included studies were deemed of good quality; however, an individual quality rating was not provided for each study.

RESULTS

Whole grains and cancer risk

A total of 20 longitudinal studies^25–44 evaluated the associations between whole-grain intake and cancer risk, reporting risk estimates from tertile, quartile, and quintile analysis and/or regression analysis (Table 2; Figure 2). Fourteen studies^25–38 reported a null association between whole-grain intake and cancer risk. Four studies reported a borderline significant reduction in cancer risk,^31,36–38 whereas 1 study observed a nonsignificant increase in risk with higher levels of whole-grain intake.²⁷ Six studies reported a 6%–47% statistically significant reduction in cancer risk among participants in the highest vs lowest category of whole-grain intake.^39–44 These studies focused on colorectal,^42–44 upper digestive tract,⁴¹ renal,⁴⁰ and head and neck cancers³⁹ as outcomes. The results from all studies are summarized by cancer site in the sections below, as the impact of whole grains on cancer may vary by cancer site.

Table 2.

Cohort studies evaluating whole-grain consumption in relation to cancer risk^a

Reference	Location	Sample (n)	Outcome	Contrast	RR/HR (95% CI)	Covariates
Female cancers
Kasum et al. (2001)25	USA; Iowa Women’s Health Study	23,014 postmenopausal women Age: 55–69 y Follow-up: 13 y	Endometrial cancer 382 cases	Whole-grain intake: Quintile 5 vs 1 (18.5–108 vs 0–3.5 servings/wk)	0.89 (0.61–1.29); Ptrend = .24	Age, energy, education, BMI, smoking, vitamin use, fruits and vegetables, red meat, refined/whole grain, total fat, saturated fat, age at menarche, age at menopause, number of live births, and HRT
Nicodemus et al. (2001)27	USA; Iowa Women’s Health Study	29,119 menopausal women Age: 55–69 y Follow-up: 9 y	Breast cancer 977 cases	Whole-grain intake: Quintile 5 vs 1 (19–108.5 vs 0–3.5 servings/wk)	1.21 (0.96–1.5); Ptrend = .02	Age, energy, HRT, history of benign breast disease, family history of breast cancer, mammography status, age at first live birth, number of live births, weight, waist-to-hip ratio, vitamin use, education, vitamin A, and refined/whole grain
Fung et al. (2005)28	USA; Nurses Health Study	71,058 postmenopausal women Age: 30–55 y Follow-up: 16 y	Breast cancer 512 ER− cases	Whole-grain servings: Frequency of consumption (2–3.9/d vs <1/wk) Per serving increase	1.04 (0.63–1.69); Ptrend = .77 0.99 (0.91–1.07)	Age, smoking status, BMI, multivitamin use, energy, physical activity, family history of breast cancer, history of benign breast disease, duration of menopause, age at menopause, HRT, age at menarche, parity, age at first birth, BMI at age 18, weight change since age 18, adult height, and alcohol
Egeberg et al. (2009)30	Denmark; Diet, Cancer and Health Cohort Study	25,278 postmenopausal women Age: 50–64 y Follow up (mean): 9.6 y	Breast cancer 978 cases Follow-up (mean): 9.6 y	Whole-grain product intake: Quartile 4 vs 1 (>163 vs <72 g/d) Per each 50 g/d increment	1.03 (0.85–1.24); Ptrend = N/A 1.01 (0.96–1.07)	Parity, age at first birth, education, HRT, duration of HRT use, alcohol, and BMI
Aarestrup et al. (2012)26	Denmark; Diet, Cancer and Health Cohort	24,418 postmenopausal women Age: 50–64 y	Endometrial cancer 217 cases	Whole-grain intake: Quartile 4 vs 1 (>52 vs <25 g/d) Per 25 g increment Whole-grain product intake: Quartile 4 vs 1 (>52 vs <25 g/d) Per 50 g increment	1.08 (0.71–1.65); Ptrend = N/A 1.00 (0.99–1.01) 1.08 (0.71–1.65); Ptrend = .83 1.01 (0.89–1.15)	Menopausal status, HRT, HRT type, parity, age at first pregnancy, BMI, smoking, and energy
Genitourinary cancers
Egeberg et al. (2011)29	Denmark; Diet, Cancer and Health Cohort Study	26,691 men Age: 50–64 y Follow-up (median): 12.4 y	Prostate cancer1081 cases	Whole-grain product intake: Quartile 4 vs 1 (>200 vs <103 g/d) Per 50 g increment	1.06 (0.89–1.26) 1.00 (0.96–1.05)	Height, weight, education, red and processed meat, dairy products, and smoking
Nimptsch et al. (2011)33	USA; Health Professionals Follow-up Study	49,934 men Age: 40–75 y Follow up: 1986–2007	Prostate cancer 5112 cases	Whole-grain intake: Quintile 5 vs 1 (>34.3 vs <6.5 g/d)	1.13 (1.03–1.24); Ptrend = .001	BMI, height, history of diabetes, family history of prostate cancer, race/ethnicity, smoking, vigorous physical activity, energy, alcohol, calcium, alpha-linolenic acid, and tomato sauce
Drake et al. (2012)34	Sweden; Malmo Diet and Cancer Cohort	8128 men Age: 45–73 y Follow-up: 15 y (median)	Prostate cancer 817 cases	Whole-grain intake: Quintile 5 vs 1 (median intake: 51.5 vs 0 g/d)	1.00 (0.78–1.28); Ptrend = .803	Age, year of study entry, season of data collection, energy, height, waist, physical activity, smoking, educational, birth in Sweden, alcohol, calcium, selenium, competing risk by death from all causes except prostate cancer
Daniel et al. (2013)40	USA; NIH–AARP Diet and Health Study	491,841 men and women Age: 50–71 y Follow-up (mean): 9 y	Renal cell carcinoma 1816 cases	Whole-grain intake: Quintile 5 vs 1 (median intake: 1.1 vs 0.13 servings/1000 kcal)	0.84 (0.73–0.98); Ptrend = .05	Age, sex, education, race, marital status, family history of any cancer, BMI, smoking, hypertension, diabetes, alcohol, red meat, energy, fruits, vegetables, legumes
Gastrointestinal cancers
Kasum et al. (2002)41	USA; Iowa Women’s Health Study	34,651 postmenopausal women Age: 55–69 y Follow-up: 14 y	Upper aerodigestive tract cancer 169 cases	Whole-grain intake: Tertile 3 vs 1 (13–108.5 vs 0–6.5 servings/wk)	Upper aerodigestive tract cancers: 0.53 (0.34–0.81); Ptrend = .005 Esophageal cancera: 0.47 Gastric cancera: 0.61	Age, smoking, alcohol, and energy
McCullough et al. (2003)35	USA; Cancer Prevention Study II Nutrition Cohort	62,609 men 70,554 women Age: 50–74 y	Colorectal cancer 508 cases	Whole-grain intake: Quintile 5 vs 1 (>11 vs <2 servings/wk)	Men: 0.95 (0.64–1.42); Ptrend = .78 Women: 1.17 (0.73–1.87); Ptrend = .52	Age, physical activity, aspirin, smoking, family history of colorectal cancer, BMI, education, energy, multivitamin use, total calcium, red meat, and HRT
Wu et al. (2004)36	USA; Health Professionals Follow-up Study	47,311 men Age: 45–75 y Follow-up: 14 y	Colorectal cancer 561 cases	Whole-grain intake: Quintile 5 vs 1	0.75 (0.57–1.00); Ptrend = N/A	Age, family history of colorectal cancer, history of endoscopy, physical activity, smoking before age 30, race, aspirin use, and energy
Larsson et al. (2005)37	Sweden; Swedish Mammography Cohort	61,433 women Mean age: 53–55 y Mean follow-up: 14.8 y	Colorectal cancer 805 cases	Whole-grain intake: ≥4.5 vs <1.5 servings/d	Colorectal cancer: 0.80 (0.60–1.06); Ptrend = .16 Colon cancer: 0.67 (0.47–0.96); Ptrend = .06 Rectal cancer: 1.11 (0.67–1.83); Ptrend = .85	Age, BMI, education, energy, saturated fat, calcium, red meat, fruits, and vegetables
Schatzkin et al. (2007)42	USA; NIH–AARP Diet and Health Study	291,988 men and 197,623 women Age: 50–71 y Follow-up: 1995–2000	Colorectal cancer 2974 cases	Whole-grain intake: Quintiles 5 vs 1 (median intake: 5.7 vs 1.7 g/1000 kcal/d)	0.79 (0.70–0.89); Ptrend < .001	Sex, physical activity, smoking, menopausal status, HRT, red meat, dietary calcium, dietary folate, and energy
Schatzkin et al. (2008)38	USA; NIH–AARP Diet and Health Study	291,988 men and 197,623 women Age: 50–71 y Follow-up: 1995–2003	Small intestinal cancer 165 cases	Whole-grain intake: Quintiles 5 vs 1 (1.4 vs 0.2 servings/1000 kcal/d)	0.59 (0.33–1.05); Ptrend = .06	Age, sex, education, family history of cancer, smoking, BMI, physical activity, HRT, red meat, total fat, energy, and fiber
Egeberg et al. (2010)44	Denmark; Diet, Cancer and Health Cohort Study	26,630 men and 29,189 women Age: 50–64 y Follow up (median): 10.6 y	Colorectal cancer 461 colon cancer cases 283 rectal cancer cases Follow-up (median): 10.6 y	Whole-grain product intake: Quartile 4 vs 1 (>160 vs <75 g/d) Per 50 g increment	Colon cancer: Men: 0.61 (0.43–0.86) Women: 0.92 (0.63–1.35); Ptrend = N/A Rectal Cancer: Men: 0.88 (0.57–1.36) Women: 0.81 (0.50–1.30); Ptrend = N/A Colon cancer: Men: 0.85 (0.77–0.94) Women: 0.98 (0.88–1.10) Rectal cancer: Men: 0.90 (0.80–1.01) Women: 1.02 (0.88–1.19)	BMI, alcohol, education, red and processed meat, HRT, and leisure time physical activity
Fung et al. (2010)43	USA; Health Professionals Follow-up Study Nurses Health Study	45,490 men Age: 40–75 y Follow-up: 20 y 87,256 women Age: 34–59 y Follow-up: 26 y	Colorectal cancer 1032 cases among men 1432 cases among women	Whole-grain intake: Per additional serving/d	Men: 0.94 (0.88–0.99) Women: 0.95 (0.89–1.02)	Age, BMI, alcohol, family history of colorectal cancer, physical activity, aspirin, colonoscopy, history of polyps, smoking, energy, and multivitamin use
Kyro et al. (2013)31	Scandinavia (Denmark, Sweden, Norway)	108,000 men and women Age: 40–55 y Follow-up: 11 y (median)	Colorectal cancer 1123 cases	Total whole-grain intake: Quartile 4 vs 1 (>71 vs <31 g/d in males and >68 vs <30 g/d in females) Per 25 g increase	0.86 (0.69–1.06); Ptrend = .11 0.94 (0.88–1.01)	Alcohol. smoking, education, HRT, red and processed meat, BMI, energy, and other grain products
Head and neck cancers
Kasum et al. (2002)41	USA; Iowa Women’s Health Study	34,651 postmenopausal women Age: 55–69 y Follow-up: 14 y	Upper aerodigestive tract cancer 169 cases	Whole-grains intake (servings/wk): Tertile 3 vs 1 (13–108.5 vs 0–6.5 servings/wk)	Upper aerodigestive tract cancers: 0.53 (0.34–0.81); Ptrend = .005 Oral/pharangeala: 0.47 Nasopharangeal/salivarya: 0.67 Laryngeala: 0.44	Age, smoking, alcohol, and energy
Lam et al. (2011)39	USA; NIH–AARP Diet and Health Study	494,991 men and women Age: 50–71 Follow-up:11 y	Head and neck cancers 1867 cases	Whole-grain intake: Quintile 5 vs 1 (median intake: 1.1 vs 0.15 servings/1000 kcal) Quintile 5 vs 1 (median intake: 1.1 vs 0.12 servings/1000 kcal)	Women: 0.71 (0.51–0.97); Ptrend = .005 Men: 0.85 (0.72–1.01); Ptrend = .139	Energy, education, BMI, physical activity, alcohol, smoking dose, red meat, and fiber
Hematologic cancers
Thompson et al. (2010)32	USA; Iowa Women’s Health Study	35,159 postmenopausal women Age: 55–69 y Follow-up: 19 y	Non-Hodgkin’s lymphoma 415 cases	Whole-grain intake: >17 vs <4.6 servings/wk	0.88 (0.66–1.17)	Age and energy

^a Dietary intakes of whole grains were measured using a food frequency questionnaire in all studies except the study by Drake et al. (2012)³⁴, where a 7-day menu book and 1-hour dietary interview were also used.

^b These hazard ratios do not have an associated confidence interval due to the small number of site-specific cancer cases, which limited statistical testing.

Abbreviations: BMI, body mass index; CI, confidence interval; ER, Estrogen Receptor; HR, hazard ratio; HRT, hormone replacement therapy; N/A, Not Applicable; NIH–AARP, National Institutes of Health–American Association of Retired Persons; PR, Progesterone Receptor; RR, relative risk.

Figure 2.

Hazard ratios/relative risk ratios (95% confidence intervals) for associations between whole-grain intake and risk of (A) female cancers, (B) genitourinary cancers, and (C) gastrointestinal cancers. For studies of gastrointestinal cancers, in which risk estimates were not available for the sample as a whole, risk estimates for men and women are presented. For studies on colorectal cancers that provided risk estimates for colorectal, colon, and rectal cancers, results for colorectal cancer are presented, with the exception of the study by Egeberg et al. (2010),⁴⁴ for which risk estimates for colon cancer are shown.

Female cancers.

A total of 5 prospective cohort studies (Figure 2A) evaluated whole-grain intake in relation to risk of breast^27,28,30 and endometrial cancer.^25,26 There was no significant association between whole-grain intakes and risk of these female cancers in any study. However, 1 analysis using the Iowa Women’s Health Study cohort was suggestive of a nonstatistically significant 21% higher risk of postmenopausal breast cancer among participants in the highest vs lowest quintile of whole-grain intake, and a statistically significant linear trend across the categories of intake was observed (relative risk, 1.21; 95% CI, 0.96–1.50; P_trend = .02).²⁷

Genitourinary cancers.

Four prospective cohort studies (2 conducted in Scandinavia^29,34 and 2 in the United States^33,40; Figure 2B) evaluated whole grains in relation to cancers of the genitourinary organs—namely, prostate cancer^29,33,34 and renal cell carcinoma.⁴⁰ Of 3 studies examining whole grains and prostate cancer risk,^29,33,34 2 reported a null association,^29,34 whereas 1 study was suggestive of a 13% increased risk of prostate cancer with higher consumption of whole grains (P = .001).³³ Contrary to these findings, the 1 US prospective study on renal cell carcinoma found a statistically significant 16% reduction in the risk of this cancer among participants in the highest vs lowest quintile of whole grain consumption (hazard ratio [HR], 0.84; 95% CI, 0.73–0.98; P_trend = .05).⁴⁰

Gastrointestinal cancers.

Whole grain consumption and risk of gastrointestinal cancers has been evaluated in 9 studies (3 conducted in Europe^31,37,44 and 6 in the United States^{35,36,38,41–43} (Figure 2C), primarily in relation to colorectal cancer^{31,35–37,42–44} but also in relation to small intestinal³⁸ and upper aerodigestive tract cancers, which included esophageal and gastric cancers.⁴¹ Findings from approximately half of these studies were suggestive of whole grains protecting against gastrointestinal cancers.^41–44 Four studies reported a statistically significant 6%–53% reduction in cancer risk with higher whole grain consumption.^41–44 Among these studies, 1 study of a US cohort⁴³ and 1 study of a Danish cohort⁴⁴ found a protective impact among men only; however, in the Danish cohort, this was only observed in relation to colon but not rectal cancer.⁴⁴ In contrast, there was no statistically significant association between whole grains and gastrointestinal cancers in 5 studies.^31,35–38 Nevertheless, in 4 of these studies, evidence was suggestive of a borderline significant 14%–41% reduction in risk when comparing participants in the highest vs lowest category of whole-grain intake.^31,36–38

Head and neck cancers.

The impact of whole grain consumption on risk of head and neck cancers was evaluated in 2 prospective cohort studies,^39,41 which were suggestive of a protective impact of whole grains. In the National Institutes of Health–American Association of Retired Persons (NIH-AARP) cohort, higher whole-grain intakes were associated with a statistically significant 29% reduction in the risk of these cancers among women (HR, 0.71; 95% CI, 0.51–0.97), and a statistically significant linear trend was observed across the increasing categories of whole-grain intake (P_trend = .005). However, only a non-significant 15% reduction in risk was observed among men (HR, 0.85; 95% CI, 0.72–1.01; P_trend = .139).³⁹ Similarly, in the Iowa Women’s Health Study, higher consumption of whole grains was associated with a 39%–66% lower risk of oral/pharyngeal, nasopharyngeal/salivary, and laryngeal cancers.⁴¹

Hematologic cancers.

One prospective cohort study evaluated whole grains in relation to hematologic cancers.³² This study examined whole-grain intake in relation to non-Hodgkin’s lymphoma using data from the Iowa Women’s Health Study cohort. The results were indicative of a null association. There were no studies on whole grains and lymphoma in men.

Cereal fiber and cancer risk

A total of 29 prospective cohort studies^{26,38–42,45–67} in US, Canadian, and European populations evaluated cereal fiber consumption in relation to cancer risk (Table 3; Figure 3). Eight studies reported a statistically significant 6%–49% reduction in risk of gastrointestinal and head and neck cancers.^{38,39,41,42,45–48} In contrast, 21 studies^{26,40,49–61} reported a null association, primarily in relation to female and genitourinary cancers. However, 2 studies were suggestive of reduced risk of breast⁵² and colorectal cancer⁶¹ with higher cereal fiber intake, but these associations were not statistically significant. These results are discussed separately in the sections below for female, genitourinary, gastrointestinal, and head and neck cancers.

Table 3.

Cohort studies evaluating cereal fiber consumption in relation to cancer risk^a

Reference	Location	Sample (n)	Outcome	Contrast	RR/HR (95% CI)	Covariates
Female cancers
Terry et al. (2002)49	Canada; Canadian National Breast Screening Study	89,835 women Age: 40–59 y Follow-up (mean): 16.2 y	Breast cancer 2536 cases	Cereal fiber intake: Quintile 5 vs 1 (>5.6 vs <2.6 g/d)	0.90 (0.78–1.04); Ptrend = .13	Study center, treatment allocations, age, BMI, smoking, education, vigorous physical activity, oral contraceptives, HRT, parity, history of benign breast disease, history of breast self-exam, family history of breast cancer, menopausal status, energy, alcohol, dietary calcium, vitamin C, vitamin E, folic acid, and saturated fat
Cho et al. (2003)50	USA; Nurses Health Study II	90,655 women Age: 26–46 y Follow-up: 8 y	Breast cancer 714 cases	Cereal fiber intake: Quintile 5 vs 1 (median intake: 8.8 vs 4.2 g/d)	0.91 (0.69–1.21); Ptrend = .21	Smoking, height, parity, age at first birth, BMI, age at menarche, family history of breast cancer, history of benign breast disease, oral contraceptives, menopausal status, alcohol, energy, and animal fat
Giles et al. (2006)51	Australia; Melbourne Collaborative Cohort Study	12,273 post-menopausal women Age: 40–69 y Follow-up (mean): 9.1 y	Breast cancer 324 cases	Cereal fiber intake (g/d): Per increment of 1 serving/wk	1.08 (0.95–1.23); Ptrend = .33	Age at attendance, country of birth, energy, and HRT
Cade et al. (2007)52	UK; UK Women’s Cohort Study	35,792 women Age: 35–69 y Follow up (mean): 7.5 y	Breast cancer 807 cases	Cereal fiber intake: Quintile 5 vs 1 (>13 vs <4 g/d)	Premenopausal: 0.59 (0.32–1.10); Ptrend = .05 Postmenopausal: 1.15 (0.68–1.94); Ptrend = .89	Age, BMI, physical activity, smoking, oral contraceptives, HRT, number of children, alcohol, and energy
Silvera et al. (2007)57	Canada; Canadian National Breast-Feeding Study	49,613 women Age: 40–59 y Follow-up (mean): 16.4 y	Ovarian cancer 264 cases	Cereal fiber intake: Quartile 4 vs 1 (>5.2 vs <2.6 g/d)	Total sample: 1.09 (0.71–1.70); Ptrend = .75 Premenopausal: 1.30 (0.72–2.36); Ptrend = .67 Postmenopausal: 0.90 (0.47–1.73); Ptrend = .63	Age, BMI, alcohol, HRT, oral contraceptives, parity, age at menarche, menopausal status, energy, vigorous physical activity, study center, and treatment allocation
Suzuki et al. (2008)53	Sweden; Swedish Mammography Cohort	51,823 postmenopausal women Age (mean): 57.9 y Follow-up (mean): 8.3 y	Breast cancer 1188 cases	Cereal fiber intake: Quintile 5 vs 1 (>19 vs <12 g/d)	0.91 (0.75–1.11); Ptrend = .098 ER+/PR+: 0.99 (0.77–1.29); Ptrend = .46 ER+/PR−: 0.86 (0.56–1.32); Ptrend = .26 ER−/PR−: 0.69 (0.39–1.24); Ptrend = .21	Age, height, BMI education, parity, age at first birth, age at menarche, age at menopause, type of menopause, oral contraceptives, HRT, family history of breast cancer among first-degree relatives, history of benign breast disease, energy, energy-adjusted total fat, fruits and vegetables, alcohol, fruit fiber, vegetable fiber, and other fibers
Park et al. (2009)54	USA NIH–AARP diet	185,598 postmenopausal women Age (mean): 62 y Follow-up (mean): 7 y	Breast cancer 5461 cases	Cereal fiber intake: Quintile 5 vs 1 (median intake: 8.9 vs 2.5 g/d)	0.93 (0.85–1.02); Ptrend = .27	Race, education, BMI, age at first birth, parity, family history of breast cancer, age at menopause, physical activity, smoking, HRT, breast biopsy, gynecologic surgery, alcohol, fruits and vegetables, total fat, and energy
Hedelin et al. (2011)58	Sweden	47,140 women Age: 30–49 y Follow-up (median): 16 y	Ovarian cancer 163 cases	Cereal fiber intake: Quartile 4 vs 1 (mean intake: 1.8 vs 0.8 g/d.MJ)	1.17 (0.74–1.87); Ptrend = N/A	Age, oral contraceptives, age at menarche, parity, HRT, energy, alcohol, saturated fat, meat, and fish
Aarestrup et al. (2012)26	Denmark; Diet, Cancer and Health Cohort	24,418 postmenopausal women Age: 50–64 y	Endometrial cancer 217 cases	Cereal fiber intake (g/d): Quartile 4 vs 1 (>11 vs <7 g/d) Per 5 g increment	1.17 (0.74–1.84); Ptrend = .51 1.08 (0.86–1.36)	Menopausal status, HRT, HRT type, parity, age at first pregnancy, BMI, smoking, and energy
Deschasaux et al. (2013)55	France; Supplementation en Vitamines et Mineraux Antioxydants (SU.VLMAX) study	4684 women Age (mean): 55.8 y Follow-up (mean): 12.6 y	Breast cancer 167 cases	Cereal fiber intake: Quartile 4 vs 1 (mean intake: 8.7 vs 4.1 g/d)	1.43 (0.81–2.53); Ptrend = .1	Age, intervention group, smoking, education, physical activity, height, BMI, number of dietary records, without-alcohol energy, alcohol, total fat, overall healthy dietary pattern, family history of breast cancer, menopausal status, HRT, and number of children
Ferrari et al. (2013)56	Europe; European Prospective Investigation into Cancer and Nutrition cohort	334,849 women Age: 35–70 y Follow-up (median): 11.5 y	Breast cancer 11,576 cases	Cereal fiber intake: Quintile 5 vs 1 (>10.9 vs <4.7 g/d)	1.01 (0.95–1.08); Ptrend = .85	Menopausal status, weight, interaction between weight and menopausal status, height, smoking, education, physical activity, age at menarche, age at first full-term birth, contraceptives, HRT, age at menopause, energy, and alcohol
Genitourinary cancers
Suzuki et al. (2009)59	Europe; European Prospective Investigation into Cancer and Nutrition cohort	142,590 men Age (mean): 52 y Follow-up (mean): 8.7 y	Prostate cancer 2747 cases	Cereal Fiber intake: Quintile 5 vs 1 (mean intake: 15.1 vs 6.3 g/d)	1.03 (0.89–1.18); Ptrend = .85	Age, energy, height, weight smoking, education, and marital status
Daniel et al. (213)40	USA; NIH–AARP Diet and Health Study	491,841 men and women Age: 50–71 y Follow-up (mean): 9 y	Renal cell carcinoma 1816 cases	Cereal fiber intake: Quintile 5 vs 1 (median intake 5.7 vs 1.7 g/1000 kcal)	0.99 (0.84–1.15); Ptrend = .59	Age, sex, education, race, marital status, family history of any cancer, BMI, smoking, hypertension, diabetes, alcohol, red meat, energy, fruits, vegetables, and legumes
Deschasaux et al. (2014)60	France; Supplementation en Vitamines et Mineraux Antioxydants (SU.VLMAX) study	3313 men Follow-up (median): 12.6 y	Prostate cancer 139 cases	Cereal fiber intake (g/d): Quartile 4 vs 1 (>9.6 vs <6.6 g/d)	1.05 (0.83–1.75); Ptrend = .7	Age, energy without alcohol, intervention group, number of 24-h dietary records, smoking, education, physical activity, height, BMI, alcohol, family history of prostate cancer, prostate-specific antigen, dietary calcium, processed meat, tomato product, vitamin E, and blood selenium
Gastrointestinal cancers
Giovannucci et al. (1994)62	USA; Health Professionals Follow-up Study	47,949 men Age: 40–75 y Follow-up: 6 y	Colon cancer 205 cases	Cereal fiber intake: Quintile 5 vs 1 (median intake: 15.3 vs 2.3 g/d)	1.28 (0.78–2.09); Ptrend = .16	Age, energy, previous polyps, previous endoscopic screening, parental history of colorectal cancer, smoking, aspirin use, red meat, methionine, and alcohol
Terry et al. (2001)63	Sweden; Swedish Mammography Cohort Study	61,463 women Age: 40–74 y Mean Follow-up: 9.6 y	Colorectal cancer 460 cases	Cereal fiber intake: Quartile 4 vs1 (median intake: 13.6 vs 5.7 g/1000 kcal/d)	0.91 (0.69–1.20); Ptrend = .82	Age, red meat, dairy products, and energy
Kasum et al. (2002)41	USA; Iowa Women’s Health Study	34,651 postmenopausal women Age: 55–69 y Follow-up: 14 y	Upper aerodigestive tract cancer 169 cases	Whole-grain fiber intake: Tertile 3 vs 1 (3.99–35.75 vs 0–1.49 g/d)	Upper aerodigestive tract cancers: 0.56 (0.37–0.84); Ptrend = .005 Esophageal cancera: 0.35 Gastric cancera: 0.53	Age, smoking, alcohol, and energy
Bingham et al. (2003)45	Europe; European Prospective Investigation into Cancer and Nutrition cohort	519978 men and women Age: 25–70 y Follow-up (mean): 4.5 y	Colorectal cancer 1065 cases	Cereal fiber intake: Quintile 5 vs 1 (mean intake 12.05 vs 4.72 g/d)	0·78 (0.62–0.98); Ptrend = .06	Age, weight, height, sex, nonfat energy, and energy from fat
Mai et al. (2003)65	USA; Breast Cancer Detection Demonstration Project	45491 women Mean age: 62 y Mean follow-up: 8.5 y	Colorectal cancer 487 cases	Cereal fiber intake: Quintile 5 vs1 (>4.75 vs <1.80 (g/1000 kcal/d)	1.02 (0.76–1.37); Ptrend = N/A	Age, nonsteroidal anti-inflammatory drugs, smoking, alcohol, calcium, vitamin D, red meat, height, BMI, and education
Bingham et al. (2005)64	Europe; European Prospective Investigation into Cancer and Nutrition cohort	519978 men and women Age: 25–70 y Mean follow-up: 6.2 y	Colorectal cancer 1721 cases	Cereal fiber intake: Quintile 5 vs 1 (>13.5 vs <4.6 g/d for women and >10.6 vs <3.9 for men)	0.93 (0.76–1.15); Ptrend = .44	Age, sex, energy from nonfat sources, energy from fat sources, height, weight, folate, red and processed meat, physical activity, alcohol, smoking, and education
Lin et al. (2005)66	USA; Women’s Health Study	39,976 women Age: ≥45 y Mean follow-up: 10 y	Colorectal cancer 223 cases	Cereal fiber intake: Quintile 5 vs 1 (median intake: 6.1 vs 3.1 g/d)	0.97 (0.66–1.42); Ptrend = .69	Age, BMI, treatment assignment, family history of colorectal cancer, colon polyps, physical activity, smoking, aspirin, red meat, alcohol, energy, menopausal status, and HRT
Michels et al. (2005)61	USA; Nurses Health Study and Health Professionals Follow-up Study	47,279 men Age: 40–75 y Follow-up: 14 y 76,947 women Age: 38–63 y Follow-up: 16 y	Colorectal cancer 593 cases in men 919 cases in women	Cereal fiber intake: Quintile 5 vs 1 (mean intake 8 vs 2.3 and 11.45 vs 2.8 g/1000 kcal/d for women and men, respectively)	Women: 0.89 (0.71–1.12); Ptrend = .63 Men: 0.79 (0.60–1.05); Ptrend = .19 Pooled: 0.85 (0.71–1.02); Ptrend = .21	Age, time period, family history of colorectal cancer, sigmoidoscopy or colonoscopy, height, BMI, physical activity, aspirin, smoking, multivitamins, energy, alcohol, dietary folate, calcium, red and processed meat, glycemic load, methionine, HRT, and menopausal status
Mendez et al. (2007)48	Europe; European Prospective Investigation into Cancer and Nutrition cohort	435000 men and women Age: 25–70 y Follow-up (median): 6.7 y	Gastric adenocarcinoma 312 cases	Cereal fiber intake (g/100 kcal): Quartile 4 vs 1	0.69 (0.48–0.99); Ptrend = .01	Sex, height, weight, education, smoking, other types of fiber, and stratification by age and center
Nomura et al. (2007)67	USA; Multiethnic Cohort Study	85903 men and 105 108 women Age: 45–75 y Mean follow-up: 7.3 y	Colorectal cancer 1138 cases in men 972 cases in women	Cereal fiber intake: Quintile 5 vs 1 (median intake: 15.6 vs 2.8 and 14 vs 2.4 g/d for men and women, respectively)	Men: 0.86 (0.70–1.07); Ptrend = .479 Women: 1.00 (0.78–1.27); Ptrend = .675	Age, ethnicity, time since cohort entry, family history of colorectal cancer, history of colorectal polyps, smoking, BMI, physical activity, aspirin use, multivitamin use, HRT, alcohol, red meat, folate, vitamin D, calcium, and energy
Schatzkin et al. (2007)42	USA; NIH–AARP Diet and Health Study	291,988 men and 197,623 women Age: 50–71 y Follow-up: 1995–2000	Colorectal cancer 2974 cases	Cereal fiber intake: Quintiles 5 vs 1 (median intake: 5.7 vs 1.7 g/1000 kcal/d)	0.86 (0.76–0.98); Ptrend = .01	Sex, physical activity, smoking, menopausal status, HRT, red meat, dietary calcium, dietary folate, and energy
Schatzkin et al. (2008)38	USA; NIH–AARP Diet and Health Study	291,988 men and 197,623 women Age: 50–71 y Follow-up: 1995–2003	Small intestinal cancer 165 cases	Cereal fiber intake: Quintiles 5 vs 1 (>8.4 vs <3.7 g/d)	0.51 (0.29–0.89); Ptrend = .01	Age, sex, education, family history of cancer, smoking, BMI, physical activity, HRT, red meat, total fat, energy, and whole grain
Hansen et al. (2012)46	Scandinavia; HELGA cohort	108,081 men and women Age: 30–64 y Follow-up (median): 11.3 y	Colorectal cancer 1168 cases	Cereal fiber intake: Quartile 4 vs 1 (>17.8 vs <9.5 g/d) Per 2g/d increment: Quartile 4 vs 1 (>13.7 vs <7.5 g/d)Per 2 g/d increment	Men: 0.66 (0.46–0.93); Ptrend = N/A; 0.94 (0.91–0.98) Women: 0.75 (0.56–1.02); Ptrend = N/A; 0.97 (0.93–1.00)	BMI, smoking, education, HRT, alcohol, red, and processed meat
Murphy et al. (2012)47	Europe; European Prospective Investigation into Cancer and Nutrition cohort	477,312 men and women Age: >35 y Follow-up (mean): 11 y	Colorectal cancer 4517 cases	Cereal fiber intake: per 10 g/d: Quintile 5 vs 1 (>12.3 vs <4.6 g/d) per 10 g/d: Quintile 5 vs 1 (>12.3 vs <4.6 g/d) per 10 g/d: Quintile 5 vs 1 (>12.3 vs <4.6 g/d)	Colorectal: 0.89 (0.82–0.97); 0.87 (0.77–0.99); Ptrend = .003 Colon: 0.89 (0.80–0.99); 0.88 (0.76–1.03); Ptrend = .032 Rectum: 0.89 (0.78–1.01); 0.86 (0.70–1.06); Ptrend = .031	Energy, BMI, physical activity, smoking, smoking intensity, education, contraceptives, HRT, menopausal status, alcohol, folate, red and processed meat, calcium, fiber from other sources and stratified by age, sex, and center
Head and neck cancers
Kasum et al. (2002)41	USA; Iowa Women’s Health Study	34,651 postmenopausal women Age: 55–69y Follow-up: 14 y	Upper aerodigestive tract cancer 169 cases	Whole-grain fiber intake: Tertile 3 vs 1 (3.99–35.75 vs 0–1.49 g/d)	Upper aerodigestive tract cancers combined: 0.56 (0.37–0.84); Ptrend = .005 Oral/pharangeala: 0.52 Nasopharangeal/salivarya: 0.57 Laryngeala: 1.12	Age, smoking, alcohol, and energy
Lam et al. (2011)39	USA; NIH–AARP Diet and Health Study	494,991 men and women Age: 50–71 y Follow-up: 11 y	Head and neck cancers 1867 cases	Cereal fiber intake: Quintile 5 vs 1 (1.1 vs 0.15 servings/1000 kcal) Quintile 5 vs 1 (1.1 vs 0.12 servings/1000 kcal)	Women: 0.66 (0.47–0.91); Ptrend = .003 Men: 0.95 (0.81–1.12); Ptrend = .915	Energy, education, BMI, physical activity, alcohol, smoking dose, red meat, and grains

^a Cereal fiber intake was measured using a food frequency questionnaire in all studies, except the study by Drake et al. (2012),³⁴ for which a 7-day menu book and a 1-hour dietary interview were also used, and the studies by Deschasaux et al. (2013),⁵⁵ Suzuki et al. (2009)⁵⁹ and Deschasaux et al. (2014),⁶⁰ for which 24-hour recalls were used.

^b These hazard ratios do not have an associated confidence interval due to the small number of site-specific cancer cases, which limited statistical testing.

Abbreviations: BMI, body mass index; CI, confidence interval; ER, Estrogen Receptor; HR, hazard ratio; HRT, hormone replacement therapy; N/A, Not Applicable; NIH–AARP, National Institutes of Health–American Association of Retired Persons; PR, Progesterone Receptor; RR, relative risk.

Figure 3.

Hazard ratios/relative risk ratios (95% confidence intervals) are shown for associations between cereal fiber intake in relation to risk of (A) female cancers, (B) genitourinary cancers, and (C) gastrointestinal cancers. For female cancers, risk estimates are presented for the whole sample, with the exception of one study by Cade et al. (2007),⁵² for which risk estimates were only reported separately for pre- and postmenopausal women. For gastrointestinal cancers, risk estimates are reported for the whole sample unless risk estimates were only reported separately for men and women.

Female cancers.

Eleven prospective cohort studies conducted in Europe and North America (Figure 3A) examined cereal fiber consumption in relation to the risk of breast,^49–56 ovarian,^57,58 and endometrial cancers.²⁶ There was no statistically significant association in any study, although a borderline significant 41% reduction in breast cancer risk with higher cereal fiber consumption was observed among premenopausal women in 1 study performed in the United Kingdom (HR, 0.59; 95% CI, 0.32–1.10; P_trend = .05).⁵²

Genitourinary cancers.

Three prospective cohort studies (Figure 3B) evaluated cereal fiber in relation to genitourinary cancers.^40,59,60 A null association was found in both studies on prostate cancer^59,60 and in the study on renal cell carcinoma.⁴⁰

Gastrointestinal cancers.

The majority of studies on cereal fiber and cancer focused on gastrointestinal cancers. Of 14 prospective cohort studies (Figure 3C), 11 studies examined cereal fiber in relation to colorectal cancers.^{42,45–47,61–67} However, there was only 1 study per cancer type for small intestinal cancer,³⁸ upper aerodigestive tract cancers,⁴¹ and gastric adenocarcinoma.⁴⁸ Similar to the studies on whole grains, approximately half of the evaluated studies were indicative of a protective impact of cereal fiber on gastrointestinal cancer risk.^42,45–47 Seven studies were indicative of a null association between cereal fiber and colorectal cancer.^61–67 However, 1 analysis of the Nurses Health Study and Health Professional Follow-up Study showed borderline significant 15% and 21% reductions in colorectal cancer risk in the pooled sample (HR, 0.85; 95% CI, 0.71–1.02) and among men (HR, 0.79; 95% CI, 0.60–1.05), respectively.⁶¹ In contrast, 7 studies in American and European cohorts indicated a statistically significant 6%–65% reduction in risk of colorectal cancer,^42,45–47 small intestinal cancer,³⁸ upper aerodigestive tract cancers, including esophageal and gastric cancers,⁴¹ and gastric adenoma.⁴⁸ A statistically significant linear trend was observed across the categories of cereal fiber intake.^38,41,42,48

Head and neck cancers.

Two prospective cohort studies that evaluated cereal fiber consumption in relation to head and neck cancers were suggestive of a protective association.^39,41 In the NIH-AARP cohort, results indicated a statistically significant 44% reduction in head and neck cancer risk among women (HR, 0.66; 95% CI, 0.47–0.91), and a statistically significant linear trend was observed across the categories of cereal fiber intake (P_trend = .003). However, there was no association among men. Similarly, in the Iowa Women’s Heath Study, higher cereal fiber consumption was associated with approximately 48% lower risk of oral/pharyngeal and nasopharyngeal/salivary cancers.⁴¹

DISCUSSION

This systematic review integrates epidemiological evidence from longitudinal observational studies on whole grains and cereal fiber in relation to cancer risk. Taken together, the majority of these studies were indicative of a null association. However, most of the studies that reported lack of an association evaluated whole grains and cereal fiber in relation to lymphoma, female, and prostate cancers as an outcome. In contrast, approximately half of the studies evaluating gastrointestinal cancers and the 2 studies on renal cell carcinoma and head and neck cancers were suggestive of a protective impact of whole grains and cereal fiber on cancer risk, but evidence was limited to draw firm conclusions. It is notable that among studies reporting risk estimates for both whole grains and cereal fiber, results were consistent for both exposures with the exception of one study. In that study,⁴⁰ an inverse association was observed between whole-grain intake and renal cell carcinoma risk, but no association was observed for cereal fiber, suggesting that consuming whole grains may confer an additional protective impact independent of their cereal fiber content. However, this finding is unusual given the high correlation between whole-grain and cereal fiber intake, estimated at 0.7 in a recent report from the Iowa Women’s Health Study.⁶⁸

Among the reviewed studies, it is notable that whole grains and cereal fiber have been studied mostly in relation to colorectal cancer. Hence, this review highlights a knowledge gap that remains to be addressed to better understand these relationships for other types of cancer. Furthermore, there were no studies on whole grains in relation to ovarian cancer. A single study evaluated whole grains in relation to lymphoma and was indicative of a null result. Similarly, there was only 1 study per cancer site for renal cell adenoma, gastric adenoma, small intestinal cancers, and aerodigestive tract cancers, and evidence was limited for prostate and endometrial cancers. Therefore, additional research is needed to clarify associations for these cancer sites.

Despite the biological plausibility of the hypothesized associations between whole grains and cereal fiber in relation to cancer risk, several methodological issues should be considered when examining the epidemiological evidence. Importantly, the inability to accurately and consistently measure whole-grain intake represents a confounding impediment to the interpretation of the existing evidence. First, studies may have varied in the definition of whole grains. Initially, whole-grain foods were defined as foods with 25% or more whole grains and bran by weight, including high-fiber bran cereals.⁶⁹ Recently, high-fiber bran cereals have been excluded from this definition.⁷⁰ Moreover, this definition contradicts the Food and Drug Administration definition that a whole-grain food is one that contains 51% or more whole-grain ingredients by weight per reference amount that is typically consumed.⁷⁰ Most of the studies reviewed herein did not explicitly state their definition of whole-grain foods, which may have accounted for, at least in part, the differences in results.

Moreover, some studies in the literature report risk estimates for one or multiple whole-grain food sources in relation to cancer without reporting associations for total whole-grain intake. Although a food source may capture the majority of whole-grain intake in some countries (e.g., rye bread in Finland), this is not the case in other countries such as the United States. In contrast, other studies that evaluate total whole-grain intake do not report risk estimates for the various food sources, which is important due to a potential differential physiologic role for different types of grains. Hence, future studies should report risk estimates for both total whole-grain consumption and whole-grain foods in relation to cancer.

The reviewed studies were published over a time period of more than 2 decades during which the methodology used to measure cereal fiber—namely, the AOAC (Association of Official Analytical Chemists) method—has been refined.⁷¹ This may have resulted in varying estimates of the fiber content of cereal products over time, which may account, at least in part, for differences in study findings. Studies also varied in the measurement tools used to assess intakes of whole grains and cereal fiber. Although most studies used food frequency questionnaires, some studies also used diet records and 24-hour recalls, which may lead to various degrees of measurement error, making study results difficult to compare. Moreover, most food frequency questionnaires were not designed to measure consumption of whole grains and, therefore, include limited relevant food sources. This, coupled with the generally low consumption of whole-grain foods in the general population, may have made the range of intake too narrow to detect an association.

Another methodological issue and potential explanation for inconsistency in results across the studies is lack of adjustment for important cancer risk factors in some studies. Studies did not consistently adjust for BMI, family history of cancer, cancer screening habits, education or socioeconomic status, physical activity, red and processed meat intake, fat intake, antioxidant intake, and other factors. This is important because the impact of diet on cancer risk may vary by tumor pathology in addition to clinical, demographic, and other lifestyle characteristics. Moreover, not all studies evaluated these associations separately in men and women. In some of the studies in this review, a protective impact of whole grains on cancer risk was observed in men only,^43,44 indicating that interactions with sex should be evaluated because associations between diet and cancer may be different for men and women. Additionally, not all studies on whole grains in relation to cancer adjusted for fiber and vice versa; this may have led to significant associations due to unaddressed confounding and difficulties in separating their independent roles.

Another important consideration when interpreting these results is study design. Although some short-term randomized controlled trials have evaluated whole grains and cereal fiber in relation to other chronic diseases, no such studies were located for cancer. Randomized controlled trials with sufficient sample size and long duration of follow-up that comprehensively adjust for relevant dietary, medical, and lifestyle factors may help clarify these associations and help assess the impact of confounding as an explanation for findings thus far. However, intervention studies may not be feasible due to cost and duration considerations. Therefore, well-designed observational studies that use reliable dietary measures, consistent definitions of whole grains, and sufficient follow-up and that comprehensively control for all appropriate confounders may be able to clarify these associations, especially for gastrointestinal cancers, for which the presently available evidence is most promising.

Finally, the literature on racial differences in nutritional factors and cancer risk is generally limited. All observational studies in this systematic review were conducted in European and North American cohorts. The available evidence is largely for Caucasian populations, although some studies enrolled racially diverse cohorts.^{28,35,38–40,42,43} Those studies did not stratify by race, however, possibly due to limited sample sizes. Additional research is warranted among other racial and ethnic groups which may have a unique cancer risk profile due to health disparities and for whom associations between diet and cancer may vary. Moreover, these groups may have limited access to healthful foods such as whole-grain products and, consequently, different levels of whole-grain intake. Therefore, health initiatives targeting whole-grain consumption may need to be tailored to the needs of these specific population groups.

CONCLUSION

Currently, the American Institute for Cancer Research and the American Cancer Society cancer prevention guidelines recommend choosing whole grains instead of refined grains as part of a comprehensive lifestyle approach to reduce cancer incidence.^4,72 Based on the results of the present qualitative review, a potential protective impact of whole grains and cereal fiber is limited to certain types of cancer, including head and neck cancers, renal cell carcinoma, and gastrointestinal cancers, which collectively constitute approximately a quarter of incident cancers.⁹ Therefore, at an individual and population level, the consumption of whole-grain foods should be encouraged, primarily due to their protective impact on cardiovascular disease and diabetes, but also due to their potential role in reducing risk of these cancers, especially among high-risk individuals. However, government policy thus far has provided little regulation of whole-grain labeling⁷³ and has eased the standards that require more whole-grain foods as part of the healthier school meals initiative.⁷⁴ Moreover, whole-grain health claims sanctioned by the US Food and Drug Administration have been limited to their potential role in reducing the risk of heart disease.⁷³ Thus, future policy and public health initiatives should focus on providing more rigorous regulation for the labeling of whole-grain products and promoting sufficient whole-grain consumption in various population groups as an integral component of a healthy diet that reduces the risk of chronic disease, including cancer.