Dietary risk factors for upper aerodigestive tract cancers
Article first published online: 11 MAR 2002
Copyright © 2002 Wiley-Liss, Inc.
International Journal of Cancer
Volume 99, Issue 2, pages 267–272, 10 May 2002
How to Cite
Kasum, C. M., Jacobs, D. R., Nicodemus, K. and Folsom, A. R. (2002), Dietary risk factors for upper aerodigestive tract cancers. Int. J. Cancer, 99: 267–272. doi: 10.1002/ijc.10341
- Issue published online: 11 APR 2002
- Article first published online: 11 MAR 2002
- Manuscript Accepted: 19 JAN 2002
- Manuscript Revised: 14 JAN 2002
- Manuscript Received: 25 AUG 2001
- whole grain;
- upper aerodigestive tract cancer;
We examined the association between whole-grain intake and incident upper aerodigestive tract cancer in a cohort of 34,651 postmenopausal, initially cancer-free women. We also studied established risk factors for upper aerodigestive cancers, including fruit and vegetable intake, smoking and alcohol intake. A mailed questionnaire at baseline in 1986 included a food-frequency questionnaire and assessment of other cancer risk factors. During the 14-year follow-up period, 169 women developed cancer of the upper aerodigestive tract. For all upper aerodigestive cancers together, significant inverse associations were observed for the highest compared to the lowest tertile of whole grains [relative risk (RR) = 0.53, 95% confidence interval (CI) 0.34–0.81] and yellow/orange vegetables (RR = 0.58, 95% CI 0.39–0.87). In addition, those in the highest compared to lowest tertile of fiber intake from whole grain were less likely to develop upper aerodigestive tract cancer (RR = 0.56, 95% CI 0.37–0.84); fiber intake from refined grain was not significantly associated with upper aerodigestive tract cancer. Findings were generally similar for oropharyngeal (n = 53), laryngeal (n = 21), nasopharyngeal/salivary (n = 18), esophageal (n = 21) and gastric (n = 56) cancers, though numbers of cases were too small for statistical testing within individual cancers. These findings confirm previous observations that high intake of fruits and vegetables and that intake of whole grains and the fiber derived from them may reduce risk of upper aerodigestive tract cancers. © 2002 Wiley-Liss, Inc.
Studies that have considered the role of dietary factors with regard to cancers of the upper aerodigestive tract have generally reported inverse associations with intake of fruit and vegetables,1–7 and whole grains.8, 9 Several studies support the idea that whole-grain intake is associated with reduced risk of gastric cancers.10 Studies that included grain fiber as a predictor11, 12 further bolster the idea that whole grain may protect against upper aerodigestive tract cancers. Previous studies of upper aerodigestive cancers have been case-control in design and, therefore, potentially subject to recall bias.
The mechanism of the apparent protective effect of fruits, vegetables and whole grains against upper aerodigestive cancers remains unclear. It may be a result of their fiber content, antioxidant content or some other phytochemicals. Clarification of the protective mechanisms of fruits, vegetables and whole grains is important to our understanding of cancers of the upper aerodigestive tract. In an analysis of the Iowa Women's Health Study (IWHS), a prospective cohort study of initially 55- to 69-year old women living in Iowa, Zheng et al.13 reported lower dietary intake of yellow/orange vegetables and of carotene, vitamins C and E and retinol in the 59 women who developed upper digestive tract cancers between 1986 and 1992.
Our primary focus was to study whether, after 14 years of follow-up in the IWHS, whole-grain intake is related to reduced risk of upper aerodigestive tract cancers. At the same time, we reviewed evidence for the established lifestyle risk factors: low fruit and vegetable intake, smoking and alcohol intake. To address potential mechanisms, we considered the roles of fiber and specific micronutrients in the risk of upper digestive tract neoplasms.
MATERIAL AND METHODS
A detailed description of the methodology of the IWHS has been published elsewhere.14 Briefly, participants in the IWHS were a random sample of women aged 55–69 years from the 1985 Iowa driver's license list. A total of 41,836 women responded to a mail survey in January 1986 (42.7% response rate). Using information from the driver's license list and the 1980 U.S. census, it was determined that responders were about 3 months older than nonresponders, 0.4 kg/m2 lighter and slightly more likely to live in rural, less affluent counties. Nonresponders had somewhat higher rates of smoking-related diseases.14 Women were excluded from the analysis if they left 30 or more items blank on the food-frequency questionnaire or reported implausibly high or low energy intake (<600 or ≥5,000 kcal/day, n = 3,096). Also excluded from analysis were premenopausal women (n = 535) and women who at baseline reported cancer of any site other than skin (n = 3,554). After these exclusions, 34,651 women remained for analysis. To retain all available incident cases of upper aerodigestive cancers, pack-years of cigarette smoking were imputed to be the mean value among those for whom ages at starting and quitting smoking were given (21.7 pack-years in 277 ex-smokers and 36.2 pack-years in 102 current smokers). However, for 300 women, no information was provided about smoking status. Therefore, 34,351 women were included in multivariate analyses.
A self-administered questionnaire was used in the 1986 baseline mail survey to obtain information on diet and other major risk factors for cancer. A 127-item food-frequency questionnaire assessed usual intake of specified portions for each food included. The reliability of the questionnaire in this cohort has been described elsewhere.15 A commonly used serving size was specified for each item and participants were asked to report their frequency of consumption by choosing from 9 categories ranging from “never or less than 1 serving per month” to “6+ servings per day.” Servings per week were calculated from these categories. These food items were then combined into food groups based on similarity of nutrient content or food type. Whole-grain consumption was assessed by 8 items: dark bread, whole-grain breakfast cereal, popcorn, cooked oatmeal, wheat germ, brown rice, bran and other grains. Women were asked to name the breakfast cereal that they usually ate and the cold breakfast cereals were coded for whole/refined-grain content, as described in detail elsewhere.16 Briefly, breakfast cereals mentioned by respondents were evaluated for whole-grain and bran content, using either the package label or records shared by General Mills (Minneapolis, MN). Whole-grain breakfast cereals were defined as having ≥25% whole-grain or bran content by weight. Refined-grain consumption was assessed by the following items: white bread including pita bread, cold breakfast cereals containing <25% whole grain or bran by weight, English muffins, bagels or rolls, pancakes or waffles, white rice, pasta and pizza. Nutrient intake was estimated using the computer program and nutrient database developed for the Nurses Health Study.15
Follow-up for occurrence of cancer was performed using the Health Registry of Iowa, part of the National Cancer Institute's Surveillance, Epidemiology and End Results Program. Incident cases were identified through computer matching on name, zip code, birth date and social security number between the 1986–1999 registry cases and study participants. Data from follow-up surveys indicated that the migration rate from Iowa among cohort members was <1% annually. A total of 167 women in the at-risk cohort were identified as having cancers of the upper digestive tract during the 14 years of follow-up. Upper aerodigestive tract cancers included 53 oropharyngeal (ICD-O C01–C06, C12–C14), 18 nasopharyngeal/salivary (ICD-O C07, C08, C11), 21 esophageal (ICD-O C15), 56 gastric (ICD-O C16) and 21 laryngeal (ICD-O C32) cancers.
The distribution of baseline characteristics according to tertiles of whole-grain intake was calculated; analysis of variance procedures were used to determine whether characteristics differed significantly across tertiles.
Analyses examined the associations between fruits, vegetables, grains, fibers and certain micronutrients and incident cancers of the upper aerodigestive tract, primarily using proportional hazards regression. Dietary variables were categorized by tertiles and the incidence rate of upper aerodigestive tract cancers for each tertile was compared with that in the lowest tertile of intake. Initial analyses were adjusted for age and energy intake. Subsequent analyses were adjusted for other risk factors for cancers of the upper aerodigestive tract, including alcohol use (never, <2 drinks/day, ≥2 drinks/day) and smoking, both as a categorical variable (current/former/never) and as a continuous variable (pack-years of smoking). Body mass index (kg/m2) was also studied; it did not contribute in any model and was omitted.
For each analysis, the hazard rate ratio (HRR) of a given category of intake was estimated by exponentiation of the proportional hazards regression coefficient for that level of intake. A Wald χ2 test was used to determine whether the estimated HRR for each tertile of intake differed significantly from unity. HRRs and confidence intervals (CIs) are presented for the group of 169 upper aerodigestive tract cancers as dependent variables. To check for consistency of findings across cancer types, relative risks (RRs) are also presented for a series of regressions in which each cancer type was the dependent variable; numbers of cases of specific cancers were small, so CIs are not presented for the individual cancers. The SAS (Cary, NC) statistical analysis package (version 6.9) was used.
The distribution of baseline characteristics according to tertiles of whole-grain intake is presented in Table I. Intake of whole grains was positively associated with level of education, total energy intake, total fruit intake, total vegetable intake, intake of yellow/orange vegetables and intake of total fiber, including whole-grain fiber, fruit fiber and vegetable fiber. In addition, intake of whole grains was associated positively with intake of vitamin C, vitamin E, carotene and retinol. More frequent intake of whole grains was inversely associated with alcohol intake and refined-grain intake.
|Range of whole-grain intake||Whole-grain intake tertiles (servings/week)|
|0–6.5 (n = 11,961)||6.9–12.5 (n = 10,981)||13.0–108.5 (n = 11,709)|
|Education (% >high school)||34.3||40.6||44.6|
|Never smoker (%)||59.1||68.9||68.3|
|Former smoker (%)||19.6||18.8||20.4|
|Pack-years of smoking (%)||12.6||8.1||8.1|
|Body mass index (kg/m2)||27.2||26.9||26.8|
|Energy intake (kcal/day)||1,632.7||1,772.9||1,993.0|
|Alcohol intake (g/day)||2.4||1.9||1.7|
|≥2 alcoholic drinks/day (%)||24.7||22.7||21.4|
|Total fruit intake (servings/week)||15.5||18.9||21.1|
|Total vegetable intake (servings/week)||22.4||25.8||28.9|
|Yellow/orange vegetable intake (servings/week)||2.4||3.0||3.6|
|Total fiber (g/day)||15.6||19.6||24.1|
|Whole-grain fiber (g/day)||1.0||3.3||6.0|
|Refined-grain fiber (g/day)||2.2||2.1||2.0|
|Fruit fiber (g/day)||4.2||5.2||6.0|
|Vegetable fiber (g/day)||7.3||8.1||9.0|
|Vitamin C (mg/day)||250.6||296.2||341.5|
|Vitamin E (mg/day)||52.3||66.5||81.0|
|Total carotene (IU/day)||7,689.9||9,413.4||11,100.8|
Table II presents the HRRs for cancers of the upper aerodigestive tract according to tertiles of intake for whole grains, refined grains and yellow/orange vegetables; for categories of alcohol; and for smoking pack-years, age and energy intake. For all upper aerodigestive cancers together, significant inverse associations were observed for intake of whole grains and yellow/orange vegetables. Those in the highest tertile of whole-grain intake were 0.53 times as likely to develop cancer of the upper aerodigestive tract as those in the lowest tertile (95% CI 0.34–0.81, HRR adjusted only for age and energy intake = 0.43, 95% CI 0.29–0.64). Those in the highest tertile of yellow/orange vegetable intake were 0.58 times as likely to develop an upper aerodigestive tract cancer as those in the lowest tertile (95% CI 0.39–0.87, HRR adjusted only for age and energy intake = 0.49, 95% CI 0.33–0.73). A powerful and significant association between all upper aerodigestive cancers together and pack-years of smoking was observed (HRR/20 pack-years = 1.54, 95% CI 1.38–1.73; HRR adjusted only for age and energy intake = 1.62, 95% CI 1.45–1.80). Alcohol intake was positively, but not significantly after adjustment, associated with upper aerodigestive tract cancers in this analysis (HRR = 1.34, 95% CI 0.91–1.81, HRR adjusted only for age and energy intake = 1.75, 95% CI 1.24–2.48). Age was strongly related and energy intake weakly related to incidence of upper aerodigestive cancers. Generally, similar associations were observed for most variables across most cancer types, though this statement is weakened by the small numbers of incident cases for each specific cancer type.
|Upper aerodigestive||Trend test||Oral/pharyngeal||Esophageal||Gastric||Nasopharyngeal/ salivary||Laryngeal|
|Cases||HRR (95% CI)||Cases||HRR||Cases||HRR||Cases||HRR||Cases||HRR||Cases||HRR|
|Whole grains (servings/week)2|
|Refined grains (servings/week)2|
|Yellow/orange vegetables (servings/week)2|
|Smoking (20 pack-years)||436,266||169||1.54 (1.38–1.73)||<0.001||53||1.73||21||1.67||56||1.26||18||1.21||21||1.80|
|Age (10 years)||440,122||169||2.33 (1.62–3.35)||<0.001||53||1.98||21||3.64||56||3.91||18||2.10||21||0.62|
|Energy intake (250 kcal/day)||440,122||169||1.08 (1.01–1.15)||0.04||53||1.09||21||0.98||56||1.10||18||0.98||21||1.13|
Table III presents the HRRs for cancers of the upper aerodigestive tract according to tertiles of fiber intake. Compared to those in the lowest tertile of total fiber intake, those in the highest tertile were 0.57 times as likely to develop cancer of the upper aerodigestive tract (95% CI 0.36–0.92). This association appeared to be driven by the strong inverse association for whole-grain fiber intake; those in the highest tertile of whole-grain fiber intake were 0.56 times as likely to develop upper aerodigestive tract cancer as those in the lowest tertile (95% CI 0.37–0.84). Fiber intake from refined grain was not significantly associated with upper aerodigestive tract cancer; the HRR for those in the highest tertile of intake was 1.02 (95% CI 0.66–1.58). Inverse associations were observed for fruit fiber, vegetable fiber and total grain fiber, though none of these associations was statistically significant.
|Daily intake (g/day)||Person-years||Cancer sites|
|Upper aerodigestive||Trend test||Oral/pharyngeal||Esophageal||Gastric||Nasopharyngeal/ salivary||Laryngeal|
|Cases||HRR (95% CI)||Cases||HRR||Cases||HRR||Cases||HRR||Cases||HRR||Cases||HRR|
|Total grain fiber|
Inverse associations were also observed for total fruit intake and vegetable intake, though these associations were not statistically significant (data not shown). Neither red meat nor tomatoes was related to incidence of upper aerodigestive cancers (data not shown). To identify phytochemicals that might underlie the relationships of whole grains and yellow/orange vegetables with upper aerodigestive tract cancers, we performed exploratory analyses studying the associations of these cancers with several nutrients. An inverse association of the cancers was seen with α-carotene (highest compared to lowest tertile HRR = 0.62, 95% CI 0.41–0.94). α-Carotene was added to the model for all upper aerodigestive cancers presented in Table II; the coefficients relating to yellow/orange vegetables shown in Table II were little changed (HRR for highest vs. lowest tertile = 0.65, 95% CI 0.35–1.21), while the coefficients for tertiles of α-carotene moved close to null (HRR for highest vs. lowest tertile = 0.90, 95% CI 0.48–1.71). No associations were observed between upper aerodigestive cancers and dietary retinol, vitamin C, vitamin E, several other carotenoids (β-carotene, lycopene, β-cryptoxanthin and lutein) or several flavonoids (myricetin, kampesterol and quercetin).
In our prospective study of over 34,000 women, cancers of the upper aerodigestive tract were inversely associated with intake of whole grains, yellow/orange vegetables, total fiber and whole-grain fiber. Our findings confirm previous observations that high intake of fruits and vegetables may reduce the risk of upper aerodigestive tract cancers and that whole grains and whole-grain fiber may be associated with this reduced risk. In contrast, intake of refined grains and refined-grain fiber was not associated with upper aerodigestive tract cancers.
Most previous studies that have considered the role of dietary factors in cancers of the upper aerodigestive tract have reported inverse associations with intake of fruits, vegetables and whole grains.1–9 An earlier report from the IWHS13 stated that higher intake of carotene and vitamins C and E was associated with decreased risk of both oral/pharyngeal/esophageal and gastric cancers. The dose-response association between gastric cancer and intake of carotene was clear and statistically significant (RR = 0.6 and 0.3, respectively, for women in the upper 2 vs. the lowest tertiles of intake). A case-control study by Negri et al.11 reported statistically significant associations between carotene, vitamin E, vitamin C, thiamine, folic acid, niacin, potassium and iron with oral and pharyngeal cancers. In our data, α-carotene intake was inversely associated with upper aerodigestive tract cancers. In contrast to these previous reports, however, our prospective analysis did not find an association between retinol, vitamin C, vitamin E, several carotenoids or several flavonoids and any upper aerodigestive cancer. Furthermore, in simultaneous regression analysis of yellow/orange vegetable intake with α-carotene intake, the food retained predictive power, while the nutrient retained little predictive power.
Some studies report associations of upper aerodigestive cancers with fiber rather than with the foods (particularly whole grain) containing the dietary fiber. One case-control study reported inverse associations with total fiber intake [odds ratio (OR) = 0.40], vegetable fiber (OR = 0.51), fruit fiber (OR = 0.60) and grain fiber (OR = 0.56).12 Similarly, an Indian study found reduced risk of oral leukoplakia, a precursor of oral cancer, in those who consumed more fiber.17 In the current data, whole-grain fiber, but not refined-grain fiber, was associated with reduced risk of upper aerodigestive cancers.18 Risk for upper aerodigestive cancer was nonsignificantly reduced for fiber from fruit or vegetables. While the inverse associations between fruits, vegetables and whole grains and these cancers could be determined by their fiber content, these foods are also rich in micronutrients and other phytochemicals that may be beneficial to health independent of the direct effects of fiber, such as improved laxation.19 Thus, our findings with respect to dietary fiber further support the concept that it is the combination of constituents of certain plant foods that is protective against these cancers.
Plant foods contain many constituents that might help to prevent upper aerodigestive cancers. Of current interest are dietary antioxidants to combat oxidative stress that might lead to DNA damage19 and retinoids and other compounds in the reversal of premalignancy.20 Among plant foods, whole grain is particularly interesting. The bran and the germ, which comprises <20% of the grain and is removed in refining, contain most of the biologically active constituents in the grain. These constituents include unsaturated fatty acids, copper, zinc, magnesium, molybdenum, potassium, manganese, phosphorous, niacin, B6, pantothenate, riboflavin and thiamin tocotrienols, lignans, phytate, tannins, enzyme inhibitors and phenolic compounds such as ferulic acid, syringic acid and p-coumaric acid.21, 22 Thus, the finding that whole grain and other plant foods are inversely related to the incidence of upper aerodigestive cancer suggests many possible avenues for research into cancer mechanisms. As in other studies,23–27 we found that cigarette smoking was strongly positively associated with upper aerodigestive tract cancers. In our study, alcohol use was positively associated with upper aerodigestive cancers but lost statistical significance after adjustment for age, smoking and dietary factors. This is in contrast to previous observations, which have suggested that alcohol is a risk factor for oral and pharyngeal cancers.27 However, alcohol intake was extremely low in these Iowa women.
Our prospective study included a large sample of women who suffered 169 incident cases of upper aerodigestive tract cancers over 14 years of follow-up. The prospective nature of the study avoided problems such as differential recall bias, typically associated with case-control studies. Other strengths of the study are that it allowed for quantitative estimation of whole- and refined-grain intake and for control of potential dietary and nondietary confounders. Several limitations of our study must also be acknowledged. The number of cases was small compared to many case-control studies of these cancers.2, 5–9, 11, 12 Our study was observational, not experimental, so it is possible that selection bias and residual confounding are present. It is also possible that the relationships studied were underestimated because the dietary variables were imperfectly measured.
In conclusion, we found that intake of several plant foods, namely, whole grains and yellow/orange vegetables, was associated with reduced risk of upper aerodigestive tract cancers, including oropharyngeal, laryngeal, nasopharyngeal/salivary, esophageal and gastric cancers. Cigarette smoking was strongly positively related to incidence of these cancers.
- 19Whole grains, cereal fiber and chronic disease: possible biological mechanisms. In: SpillerGA, ed. CRC handbook of dietary fiber in human disease, 3rd ed. New York: CRC Press, 2001. 481–98., , , , .
- 22Nutritive value of cereal products with emphasis on the effect of milling. World Rev Nutr Diet 1989;60: 1–91., , .
- 26Smoking, alcohol, dentition and diet in the epidemiology of oral cancer. Eur J Cancer B Oral Oncol 1992;28: 9–15., , , , , , , .