Abstract
Background: Nut consumption has been associated with decreased risk of colorectal, endometrial, lung, and pancreatic cancers. Polyphenols, fiber, vitamins, and minerals in nuts may confer this observed protective effect. To our knowledge, no prospective study has evaluated the effect of nut consumption on esophageal and gastric cancers.
Objective: The objective was to evaluate the associations between nut and peanut butter consumption and the risk of esophageal and gastric cancers and their different subtypes.
Design: In this study we used data from the NIH-AARP Diet and Health Study, which enrolled 566,407 persons who were 50–71 y old at baseline (1995–1996). The median follow-up time was 15.5 y. Intakes of nuts and peanut butter were assessed through the use of a validated food-frequency questionnaire. We used Cox proportional hazard models to estimate HRs and 95% CIs for esophageal and gastric cancers and their subtypes.
Results: We identified 966 incident cases of esophageal adenocarcinomas, 323 cases of esophageal squamous cell carcinoma, 698 cases of gastric cardia adenocarcinoma, and 732 cases of gastric noncardia adenocarcinoma. Compared with those who did not consume nuts or peanut butter [lowest category of consumption (C0)], participants in the highest category of nut consumption (C3) had a lower risk of developing gastric noncardia adenocarcinoma [C3 compared with C0, HR: 0.73 (95% CI: 0.57, 0.94)]. This inverse association was also seen for peanut butter consumption [C3 compared with C0, HR: 0.75 (95% CI: 0.60, 0.94)]. We observed no significant associations between the highest and lowest intakes of nuts or peanut butter and the risk of gastric cardia adenocarcinoma, esophageal adenocarcinoma, or esophageal squamous cell carcinoma.
Conclusions: Among older American adults, both nut and peanut butter consumption were inversely associated with the risk of gastric noncardia adenocarcinoma. This trial was registered at clinicaltrials.gov as NCT00340015.
Keywords: nut, peanut butter, esophageal cancer, gastric cancer, Mediterranean diet, prevention, adenocarcinoma, squamous cell carcinoma
INTRODUCTION
Gastric cancer is the third and esophageal cancer the sixth most common cause of death from cancer worldwide (1). Gastric cancer has 2 subtypes, gastric cardia cancer and gastric noncardia cancer. Esophageal cancer also has 2 subtypes, adenocarcinoma and squamous cell carcinoma, which have distinct causes (2, 3). The incidence of gastric cardia adenocarcinoma and esophageal adenocarcinoma is increasing in the United States and some other Western countries, whereas the incidence of gastric noncardia adenocarcinoma and esophageal squamous cell carcinoma is decreasing (4); this warrants separate evaluations of the risk factors for these subtypes.
Nuts have been shown to have anti-inflammatory effects by affecting the production of prostaglandins and cytokines (5), and antioxidative effects by decreasing lipid peroxidation and protecting against oxidative DNA damage (6). These protective effects could be due to some components of nuts, including polyphenols, fibers, vitamins, and minerals (7). In some studies, nut consumption has been associated with decreased total and cancer-specific mortalities (6, 8–11). Also, a few studies have reported inverse associations with specific cancer sites, including the colon and rectum, endometrium, lungs, and pancreas (12, 13). Case-control studies have reported conflicting results (14, 15), and to our knowledge no study has prospectively evaluated the effect of nut consumption on upper gastrointestinal cancers. In addition, while 50% of the peanuts consumed in the United States take the form of peanut butter (16), just one previous publication reported the association of peanut butter with colorectal cancer incidence in the Nurses’ Health Study (17).
In this analysis, we investigated the association between nut and peanut butter consumption and the risk of esophageal and gastric cancers in the prospective NIH-AARP Diet and Health Study, which allowed us to evaluate the association of consumption with different subtypes of these 2 cancers.
METHODS
Study setting and population
The NIH-AARP Diet and Health Study was established in 1995–1996; 3.5 million members of AARP were mailed a baseline questionnaire that queried demographic characteristics, dietary intake, and other health-related factors. Eligible participants were 50–71 y old and resided in California, Florida, Louisiana, New Jersey, North Carolina, Pennsylvania, or either of 2 metropolitan areas (Atlanta, Georgia, or Detroit, Michigan). The details of the NIH-AARP Diet and Health Study have been described previously (18).
In total, 617,119 baseline questionnaires were returned (18% response rate). After excluding subjects who moved out of the registry ascertainment area, those who later requested to be removed from the study, and those who returned incomplete questionnaires, the cohort comprised 566,398 participants. For this analysis we excluded 1) proxy respondents (n = 15,760); 2) participants with a self-reported history of cancer other than nonmelanoma skin cancer, prevalent cancer, or cancer diagnosed only on a death certificate (n = 51,334); 3) participants outside 2 times the interquartile range (25th to 75th percentile) of the normalized distribution for energy intake (n = 4415); 4) those who self-reported end-stage renal disease at baseline (n = 997); and 5) participants with missing information on nut consumption (n = 8267) or peanut butter consumption (n = 6815). This analysis contains 2 substantially overlapping data sets: one for nut consumption, which consists of 485,625 participants, and one for peanut butter consumption, which consists of 487,077 participants.
The NIH-AARP Diet and Health study was approved by the Special Studies Institutional Review Board of the US National Cancer Institute, and all participants provided written informed consent.
Exposure assessment
A baseline food-frequency questionnaire (FFQ) was self-administered and queried intake of 124 items over the previous 12 mo. The NIH-AARP Diet and Health Study FFQ has been shown to be valid and reproducible (19). Nut consumption was defined as consumption of “peanuts, walnuts, seeds or other nuts” and was assessed through the use of 10 frequency categories, ranging from never to ≥2 times/d, and 3 portion size categories, including less than one-quarter cup, a quarter to a half-cup, and “more than a half cup.” According to the USDA nutrient database, a quarter cup of nuts weighs 32.75 g. Peanut butter consumption was assessed with the same frequency categories, but the portion sizes of peanut butter consisted of 3 different options: <1, 1–2, and >2 tablespoons. According to the USDA nutrient database, 1 tablespoon of peanut butter weighs 16 g. Grams of nuts and peanut butter consumed daily were computed through the use of frequency and portion size data. Energy and nutrient intakes were calculated by linking each item in the FFQ to the 1994–1996 USDA Continuing Survey of Food Intakes by Individuals (20).
Ascertainment of endpoints
Incident cases of gastroesophageal cancers were identified by linking to 11 state cancer registry databases (the 8 original states and 3 additional ones: Arizona, Nevada, and Texas). A previous study found that the sensitivity and specificity of case identification in the NIH-AARP Diet and Health Study cohort were 89.2% and 99.5%, respectively (21). Addresses of subjects were linked annually to the US Postal Service National Change of Address database to aid follow-up. Vital status was identified through linkages to the National Death Index, linkages to state cancer registries, and responses to annual mailings.
Cancer subtypes were determined based on the histologic codes of the International Classification of Diseases for Oncology (ICD-O), Third Edition, as well as anatomic site (22). We identified subjects with esophageal cancer with site codes C150–C159. Then we classified them by histology type and identified 966 incident cases of esophageal adenocarcinomas and 323 incident cases of esophageal squamous cell carcinoma. We excluded cases with nonadenocarcinoma or nonsquamous (n = 83) and unspecified (n = 39) histologies. We identified 698 incident cases of gastric cardia adenocarcinoma (site code 160) and 732 incident cases of gastric noncardia adenocarcinoma (site codes 161–169). Cases with nonadenocarcinoma histology (n = 183) were excluded because the risk factors for these malignancies are different.
Statistical analysis
Nut and peanut butter consumptions were tabulated for known and suspected risk factors for gastric and esophageal cancers. The correlation between nut consumption and peanut butter consumption was determined with Spearman correlation tests. Cox proportional hazards regression models were used to estimate HRs and 95% CIs. The proportional hazards assumption was verified through the use of the Schoenfeld residuals test for either nut or peanut butter consumption. Person-years were used as the underlying time metric and were computed from the date of receipt of a valid baseline questionnaire to the date of diagnosis of gastric, esophageal, or head and neck cancer (individuals with these cancers are more susceptible to gastroesophageal cancers); the date of departing the catchment area or loss to follow-up; the date of death; or the end of the follow-up period (31 December 2011), whichever occurred first.
For adjusted models, we investigated all suspected confounders, including age, sex, smoking status and dose (never, former <20 cigarettes/d, former ≥20 cigarettes/d, current <20 cigarettes/d, or current ≥20 cigarettes/d), alcohol drinking (grams per day), BMI (kg/m2), education (less than high school, completed high school, post-high school or some college, or college and postgraduate), leisure physical activity (never, rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, ≥5 times/wk), fruit and vegetable intake (servings per day), and calorie intake (kilocalories per day). For energy adjustment, we used a nutrition density model based on grams per 1000-kcal intake, in addition to including energy in the models (23). Ethnicity (non-Hispanic white; non-Hispanic black; Hispanic; or Asian, Pacific Islander, or Native American) was included in the model only for noncardia gastric adenocarcinoma because >95% of participants with all other cancers were non-Hispanic white.
Multivariate HRs were calculated for categories of nut and peanut butter consumption. Because a large percentage of the population reported never consuming nuts or peanut butter during the previous 12 mo, we used never consumers as the reference category and categorized ever-consumers into tertiles based on the grams of nuts they consumed per 1000 kcal. The median value of each category was used to assess linear trend. HRs for the continuous scale were reported for each one-quarter-cup increase in nut consumption and each 1-tablespoon increase in peanut butter consumption per 1000 kcal/d.
We performed several secondary analyses. We used the frequency of nut and peanut butter consumption to compare the risk of each cancer between participants who never or almost never (1–6 times/y) ate nuts or peanut butter, and participants who reported frequently consuming nuts or peanut butter. We also examined total nut consumption (grams) by adding the amount of nuts to the amount of peanut butter reported on the FFQ by each participant (24). To minimize the effect of healthy eating, an additional analysis further adjusted for the Healthy Eating Index 2010 (HEI 2010), because nut consumption could be a marker of healthy eating. To control for the effects of MUFAs and PUFAs, additional analyses further adjusted for MUFA and PUFA intake. Mutual adjustment was done for nut and peanut butter consumption. To accommodate the possible confounding of socioeconomic status, we performed a stratified analysis. We excluded from sensitivity analyses cases diagnosed in the first 2 y of follow-up in order to assess potential reverse causality. Finally, to minimize misclassification, we excluded participants with ICD-O codes C168 and C169; the C168 code designates lesions of the stomach that overlap different anatomic areas, and the C169 code designates gastric adenocarcinoma without a specific location within the stomach.
Using likelihood ratio tests, we also evaluated interactions between never or almost never nut or peanut butter consumption with age, sex, BMI, smoking status, smoking dose, alcohol drinking, education, and total energy consumption. We used the median of continuous variables to categorize them for evaluating interactions. The significance cutoff for the interaction likelihood ratio test was P < 0.05. Statistical analyses were done with STATA software (version 13; STATA Corp). All P values were 2-sided, and P < 0.05 was considered significant.
RESULTS
The median (IQR) follow-up time was 15.5 y (13.5–15.6 y). The median age of the participants was 62.6 y, 59.6% were male, and 91.2% were non-Hispanic white. The mean intake of nut consumption was 2.8 ± 8.4 g/d, and the mean intake of peanut butter consumption was 3.0 ± 7.4 g/d. The correlation between nut consumption and peanut butter consumption was 0.20 (P < 0.001).
In general, nut consumers were more likely to be male, drink more alcohol, have more education, be more active, and consume more calories; they were less likely to smoke or to have a history of cardiovascular disease, diabetes, or hypertension (Table 1). Suspected risk factors were less different among subjects in the categories of peanut butter consumption. However, peanut butter consumers were more likely to be smokers, be more active, have a higher BMI, consume more calories, drink less alcohol, have less education, and be less likely to have a history of heart disease and hypertension, according to their reports (Table 1).
TABLE 1.
Baseline characteristics of participants by categories of dietary nut and peanut butter intake in the NIH-AARP Diet and Health Study1
| Nut intake categories |
Peanut butter intake categories |
|||||||
| C0 (n = 47,501) | C1 (n = 143,796) | C2 (n = 144,195) | C3 (n = 150,133) | C0 (n = 105,739) | C1 (n = 126,954) | C2 (n = 124,796) | C3 (n = 129,588) | |
| Nut or peanut butter intake, g/1000 kcal | 0 | 0.11 (0.05, 0.16) | 0.51 (0.36, 0.68) | 2.20 (1.35, 4.12) | 0 | 0.14 (0.08, 0.21) | 0.64 (0.45, 0.90) | 3.04 (1.99, 5.90) |
| Age, y | 63.5 (58.5, 67.1) | 62.6 (57.7, 66.6) | 62.3 (57.4, 66.3) | 62.5 (57.7, 66.5) | 62.7 (57.8, 66.7) | 62.1 (57.3, 66.3) | 62.5 (57.7, 66.5) | 62.9 (58.1, 66.7) |
| Male sex, % | 55.6 | 54.7 | 60.4 | 65.0 | 62.2 | 53.4 | 57.0 | 66.5 |
| BMI, kg/m2 | 26.4 (23.7, 29.7) | 26.3 (23.7, 29.3) | 26.6 (24.1, 29.6) | 26.5 (23.9, 29.3) | 25.9 (23.6, 28.9) | 26.5 (23.9, 29.5) | 26.6 (24.0, 29.7) | 26.6 (24, 29.7) |
| Smoking, % | ||||||||
| Never | 33.8 | 37.6 | 36.6 | 36.1 | 35.9 | 37.8 | 36.9 | 35.1 |
| Former | ||||||||
| <20 cigarettes/d | 26.7 | 28.5 | 29.0 | 29.7 | 29.0 | 29.1 | 28.9 | 28.5 |
| ≥20 cigarettes/d | 23.6 | 20.9 | 22.3 | 23.0 | 24.3 | 21.1 | 21.3 | 22.7 |
| Current | ||||||||
| <20 cigarettes/d | 10.2 | 8.5 | 7.8 | 7.1 | 7.2 | 7.9 | 8.4 | 8.5 |
| ≥20 cigarettes/d | 5.6 | 4.4 | 4.2 | 4.1 | 3.6 | 4.2 | 4.5 | 5.1 |
| Alcohol, g/d | 0.66 (0, 5.7) | 1.5 (0, 10.3) | 2.0 (0.2, 11.7) | 2.7 (0.4, 14.2) | 2.24 (0.2, 4.6) | 1.9 (0.2, 12.8) | 1.8 (0.2, 10.4) | 1.6 (0, 9.2) |
| Education, % | ||||||||
| Less than high school | 38.3 | 28.9 | 25.1 | 21.3 | 25.2 | 26.0 | 26.7 | 26.6 |
| Completed high school | 10.1 | 10.8 | 10.2 | 9.3 | 9.1 | 10.3 | 10.4 | 10.7 |
| Post–high school or some college | 22.5 | 24.1 | 24.1 | 23.9 | 23.3 | 24.2 | 24.2 | 23.9 |
| College and postgraduate | 29.1 | 36.1 | 40.6 | 45.6 | 42.4 | 39.5 | 38.8 | 38.8 |
| Physical activity, % | ||||||||
| Never | 9.3 | 4.8 | 3.8 | 3.3 | 5.5 | 4.3 | 4.1 | 4.1 |
| Rarely | 17.1 | 14.4 | 13.5 | 12.1 | 13.6 | 14.0 | 13.8 | 13.2 |
| 1–3 times/mo | 12.0 | 13.9 | 14.3 | 13.4 | 12.2 | 14.4 | 14.4 | 13.5 |
| 1–2 times/wk | 17.7 | 21.3 | 22.6 | 22.5 | 19.8 | 22.0 | 22.6 | 22.3 |
| 3–4 times/wk | 23.4 | 26.5 | 27.2 | 28.3 | 26.6 | 26.7 | 27.3 | 27.5 |
| ≥5 times/wk | 20.4 | 19.1 | 18.6 | 20.3 | 22.2 | 18.7 | 17.8 | 19.3 |
| Vegetable intake, servings/d | 3.12 (1.9, 4.8) | 3.3 (2.2, 4.8) | 3.4 (2.3, 4.9) | 3.6 (2.4, 5.2) | 3.4 (2.2, 5.0) | 3.5 (2.3, 5.1) | 3.4 (2.3, 4.7) | 3.4 (2.3, 4.9) |
| Fruit intake, servings/d | 2.3 (1.2, 3.8) | 2.4 (1.4, 3.8) | 2.4 (1.4, 3.7) | 2.5 (1.5, 3.8) | 2.5 (1.4, 4.0) | 2.5 (1.5, 4.0) | 2.4 (1.4, 3.7) | 2.4 (1.4, 3.7) |
| Calories, kcal/d | 1548 (1142, 2078) | 1629 (1285, 2121) | 1668 (1252, 2237) | 1825 (1370, 2328) | 1582 (1190, 2081) | 1710 (1250, 2284) | 1651 (1322, 2054) | 1830 (1345, 2413) |
| HEI 2010 | 65.7 (56.6, 73.0) | 66.5 (58.2, 73.4) | 66.7 (58.9, 73.3) | 68.6 (61.2, 74.9) | 67.1 (58.7, 74.0) | 66.6 (58.3, 73.4) | 66.8 (59.1, 73.4) | 68.0 (60.4, 74.5) |
| MUFA intake, g/d | 17.3 (11.3, 25.8) | 19.2 (13.4, 27.0) | 20.7 (14.4, 29.6) | 23.9 (16.6, 33.3) | 17.3 (11.6, 25.2) | 20.0 (13.5, 29.1) | 20.9 (15.1, 28.2) | 24.9 (17.2, 35.1) |
| PUFA intake, g/d | 10.5 (7.1, 15.3) | 11.6 (8.3, 16.2) | 12.5 (8.8, 17.6) | 14.5 (10.2, 20) | 10.7 (7.4, 15.4) | 12.2 (8.4, 17.5) | 12.5 (9.1, 17.0) | 14.7 (10.3, 20.6) |
| Self-reported history, % | ||||||||
| Heart disease | 19.3 | 14.4 | 13.1 | 12.2 | 15.9 | 12.6 | 13.0 | 14.0 |
| Diabetes | 12.5 | 8.3 | 8.4 | 9.0 | 8.9 | 7.8 | 8.6 | 10.7 |
| Stroke | 3.4 | 2.2 | 1.8 | 1.7 | 2.2 | 2.0 | 1.9 | 2.2 |
| Hypertension | 49.1 | 44.1 | 42.9 | 41.2 | 44.7 | 43.4 | 42.9 | 42.6 |
Values are medians (IQRs) or percentages. Intake density is based on grams per 1000 kcal. All risk factors were associated with nut or peanut butter consumption, P < 0.001. C, category; HEI 2010, Health Eating Index 2010.
The associations between nut consumption and each outcome are summarized in Table 2. Being in the highest category of nut consumption was associated with a statistically significant lower risk of developing gastric noncardia adenocarcinoma compared with non-consumers [HR: 0.73 (95% CI: 0.57, 0.94)]. However, no significant association was found between every quarter-cup (per 1000 kcal/d) increase in the consumption of nuts and risk of gastric noncardia adenocarcinoma [HR: 0.73 (95% CI: 0.32, 1.64)] (Table 2). We observed no significant associations between nut consumption and risk of gastric cardia cancer, esophageal adenocarcinoma, or esophageal squamous cell carcinoma (Table 2).
TABLE 2.
Crude and adjusted HRs of upper gastroesophageal cancers, by categories of nut intake among participants (n = 485,625)1
| Categories of nut intake |
||||||
| C0 | C1 | C2 | C3 | P-trend | Continuous (every quarter cup, or 32.75 g/1000 kcal) | |
| Person-years, n | 612,763 | 2,002,495 | 1,973,416 | 2,065,554 | ||
| Esophageal adenocarcinoma (n = 951) | ||||||
| Cases, n | 93 | 262 | 298 | 298 | ||
| Crude HR (95% CI) | 1 | 0.88 (0.70, 1.12) | 1.00 (0.79, 1.26) | 0.96 (0.76, 1.22) | 0.77 | 0.95 (0.52, 1.74) |
| Adjusted HR (95% CI)2 | 1 | 0.94 (0.73, 1.21) | 0.97 (0.76, 1.24) | 0.90 (0.70, 1.16) | 0.32 | 0.84 (0.43, 1.62) |
| Esophageal squamous cell carcinoma (n = 313) | ||||||
| Cases, n | 34 | 101 | 80 | 98 | ||
| Crude HR (95% CI) | 1 | 0.92 (0.63, 1.35) | 0.74 (0.50, 1.10) | 0.84 (0.57, 1.24) | 0.71 | 1.52 (0.65, 3.55) |
| Adjusted HR (95% CI)2 | 1 | 1.00 (0.66, 1.53) | 0.93 (0.61, 1.44) | 1.01 (0.66, 1.55) | 0.78 | 0.82 (0.25, 2.73) |
| Gastric cardia adenocarcinoma (n = 688) | ||||||
| Cases, n | 78 | 175 | 194 | 241 | ||
| Crude HR (95% CI) | 1 | 0.70 (0.54, 0.90) | 0.74 (0.57, 0.96) | 0.90 (0.70, 1.16) | 0.03 | 0.74 (0.34, 1.61) |
| Adjusted HR (95% CI)2 | 1 | 0.71 (0.54, 0.94) | 0.76 (0.58, 1.00) | 0.88 (0.67, 1.15) | 0.12 | 0.72 (0.32, 1.65) |
| Gastric noncardia adenocarcinoma (n = 719)3 | ||||||
| Cases, n | 85 | 240 | 197 | 197 | ||
| Crude HR (95% CI) | 1 | 0.87 (0.69, 1.09) | 0.71 (0.56, 0.89) | 0.69 (0.54, 0.87) | 0.003 | 0.68 (0.31, 1.47) |
| Adjusted HR (95% CI)2 | 1 | 0.93 (0.73, 1.18) | 0.77 (0.60, 0.99) | 0.73 (0.57, 0.94) | 0.004 | 0.73 (0.32, 1.64) |
Intake density is based on grams per 1000 kcal. C, category.
Adjusted for age (years), sex (male or female), smoking status and dose (never, former <20 cigarettes/d, former ≥20 cigarettes/d, current <20 cigarettes/d, or current ≥20 cigarettes/d), alcohol drinking (grams per day), BMI (kg/m2), education (less than high school, completed high school, post-high school or some college, or college and postgraduate), physical activity (never, rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, or ≥5 times/wk), fruit and vegetable intake (servings per day), and calories (kilocalories per day). HRs (95% CIs) were calculated through the use of a Cox regression model.
Additionally adjusted for ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, or Asian/Pacific Islander/Native American). HRs (95% CIs) were calculated through the use of a Cox regression model.
The associations between peanut butter consumption and each outcome are summarized in Table 3. Being in the highest peanut butter consumption group was associated with a statistically significant lower risk of gastric noncardia adenocarcinoma compared with those who did not consume peanut butter [HR: 0.75 (95% CI: 0.60, 0.94)]. We also observed a significant 48% decreased risk of gastric noncardia adenocarcinoma for every 1 tablespoon (per 1000 kcal/d) of peanut butter consumed [HR: 0.52 (95% CI: 0.32, 0.84)] (Table 3). The P-trend for the risk of esophageal adenocarcinoma was statistically significant. However, neither the HR of the highest (C3) versus the lowest (C0) category of nut consumption nor the HR of continuous intake for esophageal adenocarcinoma was significant (Table 3).
TABLE 3.
Crude and adjusted HRs of upper gastroesophageal cancers, by categories of peanut butter intake among participants (n = 487,077)1
| Categories of peanut butter intake |
||||||
| C0 | C1 | C2 | C3 | P-trend | Continuous (every 1 tablespoon, or 16 g/1000 kcal) | |
| Person-years, n | 1,427,476 | 1,726,574 | 1,684,746 | 1,723,520 | ||
| Esophageal adenocarcinoma (n = 951) | ||||||
| Cases, n | 196 | 206 | 229 | 320 | ||
| Crude HR (95% CI) | 1 | 0.87 (0.71, 1.05) | 0.99 (0.82, 1.20) | 1.35 (1.13, 1.61) | <0.001 | 1.62 (1.30, 2.00) |
| Adjusted HR (95% CI)2 | 1 | 0.95 (0.77, 1.17) | 0.98 (0.80, 1.20) | 1.17 (0.97, 1.42) | 0.01 | 1.25 (0.97, 1.60) |
| Esophageal squamous cell carcinoma (n = 313) | ||||||
| Cases, n | 83 | 92 | 70 | 68 | ||
| Crude HR (95% CI) | 1 | 0.91 (0.68, 1.23) | 0.71 (0.52, 0.98) | 0.69 (0.51, 0.95) | 0.04 | 0.93 (0.54, 1.62) |
| Adjusted HR (95% CI)2 | 1 | 0.95 (0.69, 1.30) | 0.69 (0.49, 0.98) | 0.75 (0.53, 1.05) | 0.14 | 0.92 (0.51, 1.66) |
| Gastric cardia adenocarcinoma (n = 679) | ||||||
| Cases, n | 152 | 166 | 151 | 210 | ||
| Crude HR (95% CI) | 1 | 0.90 (0.72, 1.12) | 0.85 (0.68, 1.06) | 1.13 (0.92, 1.40) | 0.02 | 1.39 (1.04, 1.86) |
| Adjusted HR (95% CI)2 | 1 | 1.00 (0.79, 1.26) | 0.91 (0.71, 1.16) | 1.09 (0.87, 1.37) | 0.24 | 1.09 (0.78, 1.53) |
| Gastric noncardia adenocarcinoma (n = 721)3 | ||||||
| Cases, n | 183 | 191 | 173 | 174 | ||
| Crude HR (95% CI) | 1 | 0.87 (0.71, 1.06) | 0.79 (0.64, 0.97) | 0.79 (0.65, 0.97) | 0.11 | 0.70 (0.46, 1.06) |
| Adjusted HR (95% CI)2 | 1 | 0.94 (0.76, 1.17) | 0.81 (0.64, 1.00) | 0.75 (0.60, 0.94) | 0.02 | 0.52 (0.32, 0.84) |
Intake density is based on grams per 1000 kcal. C, category.
Adjusted for age (years), sex (male or female), smoking status and dose (never, former <20 cigarettes/d, former ≥20 cigarettes/d, current <20 cigarettes/d, or current ≥20 cigarettes/d), alcohol drinking (grams per day), BMI (kg/m2), education (less than high school, completed high school, post-high school or some college, or college and postgraduate), physical activity (never, rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, or ≥5 times/wk), fruit and vegetable intake (servings per day), and calories (kilocalories per day). HRs (95% CIs) were calculated through the use of a Cox regression model.
Additionally adjusted for ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, or Asian/Pacific Islander/Native American). HRs (95% CIs) were calculated through the use of a Cox regression model.
Eating nuts was associated with an 18% lower risk of developing gastric noncardia adenocarcinoma than reporting never or almost never eating nuts [HR: 0.82 (95% CI: 0.70, 0.95)]. Eating peanut butter was associated with a 28% lower risk of developing esophageal squamous cell carcinoma [HR: 0.72 (95% CI: 0.56, 0.91)] and a 19% lower risk of developing gastric noncardia adenocarcinoma [HR: 0.81 (95% CI: 0.69, 0.94)] than never or almost never eating peanut butter.
We also examined the association between total nut intake (nuts and peanut butter together, grams per 1000 kcal/d) and risk of each cancer subtype. Total nut intake was inversely associated with the risk of gastric noncardia cancer [C3 compared with C0, HR: 0.66 (95% CI: 0.53, 0.82)]. Total nut intake was not statistically significantly associated with the other 3 cancer subtypes (data not shown).
In additional analyses, including the HEI 2010 in the Cox proportional hazards regression model shifted the results of gastric noncardia adenocarcinoma toward the null, but the protective effect remained significant for nut [C3 compared with C0, HR: 0.76 (95% CI: 0.59, 0.99)] and peanut butter [C3 compared with C0, HR: 0.79 (95% CI: 0.63, 0.99)] consumption. Including HEI 2010 in the model did not change the results for the other cancer subtypes. When including monounsaturated fat intake in the model, the association between peanut butter consumption and the risk of esophageal squamous cell carcinoma became significant [C3 compared with C0, HR: 0.69 (95% CI: 0.49, 0.98)], but this did not change the results for the other cancers. Including polyunsaturated fat intake in the model did not significantly change any of the results. Adjusting the models for self-reported history of heart disease, diabetes, and stroke did not significantly change the results of any subtype, and the inverse association with gastric noncardia adenocarcinoma remained significant for nut [C3 compared with C0, HR: 0.73 (95% CI: 0.56, 0.95)] and peanut butter [C3 compared with C0, HR: 0.76 (95% CI: 0.62, 0.96)] consumption. Mutual adjustment for nut and peanut butter consumption did not alter the results. In sensitivity analyses, excluding the first 2 y of follow-up did not change the results (Supplemental Table 1). Excluding cases with ICD-O codes C168 and C169 (∼36% of the noncardia gastric cancer cases) shifted the point estimates for nut consumption [C3 compared with C0, HR: 0.89 (95% CI: 0.64, 1.25)] and peanut butter consumption [C3 compared with C0, HR: 0.81 (95% CI: 0.61, 1.07)] toward the null and made these associations nonsignificant (Supplemental Table 1).
When stratifying by level of education, the protective effect on gastric noncardia adenocarcinoma seemed to be more prominent in those at the more extreme levels of education; however, the differences were not significant and should be interpreted as equal. No statistical interaction was found between level of education and nut intake (P-interaction = 0.33). We found no evidence of a statistical interaction between nut or peanut butter consumption and any of the other factors evaluated (data not shown).
DISCUSSION
In this large prospective cohort study, nut and peanut butter consumption were inversely associated with the risk of noncardia gastric cancer in a dose-dependent manner. To our knowledge, this is the first prospective study to assess the association between nut and peanut butter consumption and the risk of esophageal and stomach cancers.
A few studies have shown an inverse association between nut consumption and specific cancer sites, including the colon, endometrium, and pancreas (12). In a population-based case-control study in Japan, subjects who ate nuts ≥3 times/mo had a 50% lower risk of gastric cancer than never consumers [OR: 0.5 (95% CI: 0.3, 0.7)], and a significant trend was observed in the protective effect across categories of nut consumption (14), which is consistent with our results. In another case-control study in Greece investigating the role of diet in the cause of gastric cancer, cases reported significantly more frequent consumption of beans and nuts (15). In a case-control study investigating the Mediterranean diet and risk of upper aerodigestive tract cancers, including esophageal cancer, nut consumption was not significantly associated with the risk of these cancers (25). All of these studies were case-control studies and thus were more likely to be subject to recall bias than our prospective study.
To our knowledge, no previous study has evaluated the association between peanut butter and the risk of gastroesophageal cancer. Yang et al. (17) reported no association between peanut butter and colorectal cancer risk in the Nurses’ Health Study. Peanut butter consumption was not associated with cancer mortality in the Southern Community Cohort Study (n = 71,764) in the southeastern United States, although this study had a relatively short follow-up (5.4 y) (24). Peanut butter consumption was not related to total mortality or cancer mortality in the Netherlands Cohort Study (8). However, the mean consumption of peanut butter in the Netherlands cohort (1.3 g/d) was less than half the mean consumption in our population (3 g/d).
Although nut consumers and peanut butter consumers seemed to be different with regard to several baseline characteristics, including alcohol intake, smoking, and amount of education (which are all potential risk factors for gastric cancer), a protective effect of consumption was observed in both, which suggests that confounding by these other factors was not driving the observed associations. We found no significant differences by education level in a stratified analysis. However, as in all observational studies, we cannot rule out the possibility that residual confounding was caused by other factors, including socioeconomic status.
Evidence suggests that some components of nuts—namely, phytosterols, fiber, MUFAs, vitamins, and minerals—may protect against cancer (12, 26). Phytosterols, including polyphenols in walnuts and quercetin in almonds, can inhibit carcinogenesis and prevent oxidative stress (27). Polyphenols regulate the inflammatory response by affecting the production of prostaglandins and cytokines (5). Peanuts contain resveratrol, which has anticarcinogenic and antioxidative effects (28). Nuts are also a good source of vitamins such as folate, niacin, carotenoid, and vitamin E, and of minerals such as calcium, selenium, zinc, and magnesium (8), which may protect against gastroesophageal cancers (26, 29–31). Nuts contain high concentrations of MUFAs and variable amounts of PUFAs (32). A recent study showed that unsaturated fatty acids inhibit colorectal cancer (33). MUFAs also have anti-inflammatory properties. Yu et al. (34) showed that nut consumption was inversely associated with concentrations of C-reactive protein and IL-6 in the Nurses’ Health Study and the Health Professionals Follow-up Study.
A nonsignificant association was found between peanut butter consumption and an increased risk of esophageal adenocarcinoma. This association could be due to residual confounding by BMI; peanut butter consumers had a slightly higher BMI than non-consumers. However, although nuts contain large amounts of calories, nut consumption was not related to obesity in our study or those of other groups (6).
This study has several strengths, including its prospective design, the very large sample size, extensive information on dietary and lifestyle factors, and information on tumor location and histology, which allowed us to thoroughly assess the effect modification and variation in risk by cancer subtype. This study also has several limitations. We collected nut and peanut butter consumption data from only one FFQ; thus the study is susceptible to measurement error as a result of the exposure collection methods and possible changes in diet during the long follow-up period. Our population was also more educated and more likely to be non-Hispanic white than the total US population, which may limit the generalizability of the results. The consumption of nuts was asked about in just one question, and we therefore could not investigate the effects of different types of nuts, such as peanuts and walnuts, which contain different nutrients. Also, we did not collect any information on whether participants consumed salted or unsalted nuts, which could be an effect modifier. We also do not have any data on Helicobacter pylori infection status, which is a known risk factor for gastric cancer. Reverse casualty is possible because subjects with early undiagnosed cancer may have restricted nut and peanut butter consumption. However, after we excluded the first 2 y of follow-up, the results remained unchanged.
In conclusion, in this large cohort study of older American adults, we found inverse associations between both nut and peanut butter consumption and the risk of gastric noncardia cancer. These prospective inverse associations are novel findings, and because both nuts and peanut butter are widely consumed, more research into their health effects, including their associations with cancer, is warranted. If their associations with cancer were confirmed in other prospective studies, investigation of possible mediating mechanisms in future studies would be valuable.
Acknowledgments
The authors’ responsibilities were as follows—CCA, SMD, and LML: designed and conducted the research; MH and GM: analyzed the data; MH: wrote the manuscript; CCA, SMD, and AE: critically revised the manuscript for important intellectual content; MH and CCA: had primary responsibility for the final content; and all authors: read and approved the final manuscript. None of the authors reported a conflict of interest related to the study.
Footnotes
Abbreviations used: C0, lowest category of nut consumption; C3, highest category of nut consumption; FFQ, food frequency questionnaire; HEI-2010, Healthy Eating Index 2010; ICD-O, International Classification of Diseases for Oncology.
REFERENCES
- 1.Global Burden of Disease Cancer Collaboration; Fitzmaurice C, Allen C, Barber RM, Barregard L, Bhutta ZA, Brenner H, Dicker DJ, Chimed-Orchir O, Dandona R, Dandona L, et al. Global, regional, and national cancer incidence, mortality, years of life lost, years lived with disability, and disability-adjusted life-years for 32 cancer groups, 1990 to 2015: a systematic analysis for the Global Burden of Disease Study. JAMA Oncol 2017;3:524–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Brown LM, Devesa SS. Epidemiologic trends in esophageal and gastric cancer in the United States. Surg Oncol Clin N Am 2002;11:235–56. [DOI] [PubMed] [Google Scholar]
- 3.Murphy G, McCormack V, Abedi B, Arnold M, Camargo M, Dar N, Dawsey S, Etemadi A, Fitzgerald R, Fleischer D, et al. International cancer seminars: a focus on esophageal squamous cell carcinoma. Ann Oncol 2017. DOI: 10.1093/annonc/mdx279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Arnold M, Soerjomataram I, Ferlay J, Forman D. Global incidence of oesophageal cancer by histological subtype in 2012. Gut 2015;64:381–7. [DOI] [PubMed] [Google Scholar]
- 5.Grosso G, Estruch R. Nut consumption and age-related disease. Maturitas 2016;84:11–6. [DOI] [PubMed] [Google Scholar]
- 6.Grosso G, Yang J, Marventano S, Micek A, Galvano F, Kales SN. Nut consumption on all-cause, cardiovascular, and cancer mortality risk: a systematic review and meta-analysis of epidemiologic studies. Am J Clin Nutr 2015;101:783–93. [DOI] [PubMed] [Google Scholar]
- 7.Kris-Etherton PM, Hecker KD, Bonanome A, Coval SM, Binkoski AE, Hilpert KF, Griel AE, Etherton TD. Bioactive compounds in foods: their role in the prevention of cardiovascular disease and cancer. Am J Med 2002;113(Suppl 9B):71S–88S. [DOI] [PubMed] [Google Scholar]
- 8.van den Brandt PA, Schouten LJ. Relationship of tree nut, peanut and peanut butter intake with total and cause-specific mortality: a cohort study and meta-analysis. Int J Epidemiol 2015;44:1038–49. [DOI] [PubMed] [Google Scholar]
- 9.Eslamparast T, Sharafkhah M, Poustchi H, Hashemian M, Dawsey SM, Freedman ND, Boffetta P, Abnet CC, Etemadi A, Pourshams A, et al. Nut consumption and total and cause-specific mortality: results from the Golestan Cohort Study. Int J Epidemiol 2017;46:75–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bonaccio M, Di Castelnuovo A, De Curtis A, Costanzo S, Bracone F, Persichillo M, Donati MB, de Gaetano G, Iacoviello L; Moli-sani Project Investigators. Nut consumption is inversely associated with both cancer and total mortality in a Mediterranean population: prospective results from the Moli-sani study. Br J Nutr 2015;114:804–11. [DOI] [PubMed] [Google Scholar]
- 11.Bao Y, Han J, Hu FB, Giovannucci EL, Stampfer MJ, Willett WC, Fuchs CS. Association of nut consumption with total and cause-specific mortality. N Engl J Med 2013;369:2001–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wu L, Wang Z, Zhu J, Murad AL, Prokop LJ, Murad MH. Nut consumption and risk of cancer and type 2 diabetes: a systematic review and meta-analysis. Nutr Rev 2015;73:409–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lee JT, Lai GY, Liao LM, Subar AF, Bertazzi PA, Pesatori AC, Freedman ND, Landi MT, Lam TK. Nut consumption and lung cancer risk: results from two large observational studies. Cancer Epidemiol Biomarkers Prev 2017;26:826–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hoshiyama Y, Sasaba T. A case-control study of stomach cancer and its relation to diet, cigarettes, and alcohol consumption in Saitama Prefecture, Japan. Cancer Causes Control 1992;3:441–8. [DOI] [PubMed] [Google Scholar]
- 15.Trichopoulos D, Ouranos G, Day NE, Tzonou A, Manousos O, Papadimitriou C, Trichopoulos A. Diet and cancer of the stomach: a case-control study in Greece. Int J Cancer 1985;36:291–7. [PubMed] [Google Scholar]
- 16.Awad AB, Chan KC, Downie AC, Fink CS. Peanuts as a source of beta-sitosterol, a sterol with anticancer properties. Nutr Cancer 2000;36:238–41. [DOI] [PubMed] [Google Scholar]
- 17.Yang M, Hu FB, Giovannucci EL, Stampfer MJ, Willett WC, Fuchs CS, Wu K, Bao Y. Nut consumption and risk of colorectal cancer in women. Eur J Clin Nutr 2016;70:333–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Schatzkin A, Subar AF, Thompson FE, Harlan LC, Tangrea J, Hollenbeck AR, Hurwitz PE, Coyle L, Schussler N, Michaud DS, et al. Design and serendipity in establishing a large cohort with wide dietary intake distributions: the National Institutes of Health-American Association of Retired Persons Diet and Health Study. Am J Epidemiol 2001;154:1119–25. [DOI] [PubMed] [Google Scholar]
- 19.Thompson FE, Kipnis V, Midthune D, Freedman LS, Carroll RJ, Subar AF, Brown CC, Butcher MS, Mouw T, Leitzmann M, et al. Performance of a food-frequency questionnaire in the US NIH-AARP (National Institutes of Health-American Association of Retired Persons) Diet and Health Study. Public Health Nutr 2008;11:183–95. [DOI] [PubMed] [Google Scholar]
- 20.Subar AF, Midthune D, Kulldorff M, Brown CC, Thompson FE, Kipnis V, Schatzkin A. Evaluation of alternative approaches to assign nutrient values to food groups in food frequency questionnaires. Am J Epidemiol 2000;152:279–86. [DOI] [PubMed] [Google Scholar]
- 21.Michaud D, Midthune D, Hermansen S, Leitzmann M, Harlan L, Kipnis V, Schatzkin A. Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study. J Registry Manag 2005;32:70–5. [Google Scholar]
- 22.Fritz AG. International classification of diseases for oncology: ICD-O. 3rd ed. Geneva (Switzerland): WHO; 2000. [Google Scholar]
- 23.Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr 1997;65(4 Suppl):1220S–8S; discussion 9S–31S. [DOI] [PubMed] [Google Scholar]
- 24.Luu HN, Blot WJ, Xiang YB, Cai H, Hargreaves MK, Li H, Yang G, Signorello L, Gao YT, Zheng W, et al. Prospective evaluation of the association of nut/peanut consumption with total and cause-specific mortality. JAMA Intern Med 2015;175:755–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Samoli E, Lagiou A, Nikolopoulos E, Lagogiannis G, Barbouni A, Lefantzis D, Trichopoulos D, Brennan P, Lagiou P. Mediterranean diet and upper aerodigestive tract cancer: the Greek segment of the Alcohol-Related Cancers and Genetic Susceptibility in Europe study. Br J Nutr 2010;104:1369–74. [DOI] [PubMed] [Google Scholar]
- 26.Hashemian M, Poustchi H, Abnet CC, Boffetta P, Dawsey SM, Brennan PJ, Pharoah P, Etemadi A, Kamangar F, Sharafkhah M, et al. Dietary intake of minerals and risk of esophageal squamous cell carcinoma: results from the Golestan Cohort Study. Am J Clin Nutr 2015;102:102–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yang CS, Landau JM, Huang MT, Newmark HL. Inhibition of carcinogenesis by dietary polyphenolic compounds. Annu Rev Nutr 2001;21:381–406. [DOI] [PubMed] [Google Scholar]
- 28.Sales JM, Resurreccion AV. Resveratrol in peanuts. Crit Rev Food Sci Nutr 2014;54:734–70. [DOI] [PubMed] [Google Scholar]
- 29.World Cancer Research Fund/American Institute for Cancer Research (AICR). Food nutrition, physical activity, and the prevention of cancer: a global perspective. Washington (DC): AICR; 2007. [Google Scholar]
- 30.Hashemian M, Hekmatdoost A, Poustchi H, Mohammadi Nasrabadi F, Abnet CC, Malekzadeh R. Systematic review of zinc biomarkers and esophageal cancer risk. Middle East J Dig Dis 2014;6:177–85. [PMC free article] [PubMed] [Google Scholar]
- 31.Abnet CC, Lai B, Qiao YL, Vogt S, Luo XM, Taylor PR, Dong ZW, Mark SD, Dawsey SM. Zinc concentration in esophageal biopsy specimens measured by x-ray fluorescence and esophageal cancer risk. J Natl Cancer Inst 2005;97:301–6. [DOI] [PubMed] [Google Scholar]
- 32.Brufau G, Boatella J, Rafecas M. Nuts: source of energy and macronutrients. Br J Nutr 2006;96(Suppl 2):S24–8. [DOI] [PubMed] [Google Scholar]
- 33.Yan G, Li L, Zhu B, Li Y. Lipidome in colorectal cancer. Oncotarget 2016;7:33429–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Yu Z, Malik VS, Keum N, Hu FB, Giovannucci EL, Stampfer MJ, Willett WC, Fuchs CS, Bao Y. Associations between nut consumption and inflammatory biomarkers. Am J Clin Nutr 2016;104:722–8. [DOI] [PMC free article] [PubMed] [Google Scholar]