Abstract
Background
Studies on diet and gastric intestinal metaplasia (GIM) risk are lacking in US populations.
Aim
To determine the associations of dietary factors and risk of GIM among a US population with typical American diet.
Methods
We analyzed data from a cross-sectional study of veterans attending primary care and endoscopy clinics at the Houston VA Medical Center. Patients completed a 110-item Block Food Frequency Questionnaire then underwent upper endoscopy with gastric mapping biopsies. We compared cases defined by GIM on ≥ 1 non-cardia gastric biopsy to controls without GIM. Associations of dietary factors and GIM were estimated using logistic regression models as odds ratios (OR) and 95% confidence intervals (CI).
Results
Among 423 GIM cases and 1796 controls, cases were older (62.1 vs. 59.9 years) and more likely to be male (97.2% vs. 90.8%) and non-White (58.6% vs. 39.0%). GIM cases had lower fat intake (percent kcal from fat tertile 1: 43.6% vs. 33.4%) and higher carbohydrate intake (percent kcal from carbohydrate T3: 41.8% vs. 33.3%) than controls. Adjusting for age, gender, race, smoking, and Helicobacter pylori, percent kcal from carbohydrates (T3 vs. T1: OR 1.35, 95% CI 1.08–1.67), fruit intake (T3 vs. T1: OR 1.28, 95% CI 1.02–1.61), and fiber intake (T3 vs. T1: OR 1.37, 95% CI 1.04–1.80) were associated with GIM. In subgroup analyses, these associations were primarily seen in non-White patients.
Conclusions
Few dietary factors, including high carbohydrate intake, are associated with increased risk of GIM in US populations, independent of H. pylori or smoking.
Keywords: Diet, Dietary factors, Nutrition, Gastric intestinal metaplasia, Gastric cancer
Introduction
Non-cardia gastric adenocarcinoma (gastric cancer) is the fifth-leading cause of cancer-related death worldwide [1]. Despite a decreasing overall incidence of gastric cancer in the USA [2], the incidence rates are increasing among adults < 50 years [3]. The majority of gastric cancer occurs in the setting of mucosal inflammation and is preceded by the development of gastric intestinal metaplasia (GIM) and atrophy. In the US population, Helicobacter pylori infection [4–8], smoking [8], and family history of gastric cancer [7, 9–12] have been described as independent risk factors for GIM.
There has been much interest in identifying potentially modifiable dietary factors and determining the role that diet and food preparation may have in the development of GIM and gastric cancer. In countries with high gastric cancer rates (e.g., Asia, Central/South America), spicy foods, salt, nitrite-rich foods, and low intake of fruits and vegetables have been associated with increased risk of GIM [5, 6, 13, 14]. However, few studies have focused on dietary factors in countries with Western diets (e.g., USA), and these studies produced conflicting results on the effect of salt, fruits, vegetables, and high fat foods on the risk of GIM [10, 15, 16].
The few previous US studies of diet and gastric cancer primarily consisted of populations with high numbers of immigrants, who usually maintain similar diets as their home country [17–19], and have focused on gastric cancer [20] rather than GIM, the precancerous stage with greatest potential benefit from dietary modification. Additionally, few studies on Western dietary factors have adjusted for the presence of H. pylori infection, which is the strongest risk factor for gastric cancer. The purpose of this study was to examine dietary factors associated with the development of GIM among a US veteran population with diets representative of typical American diets.
Methods
Study Design
We used data from a previously completed cross-sectional study of patients attending primary care and endoscopy clinics at the Michael E. DeBakey Veterans Affairs Medical Center (MEDVAMC) in Houston, Texas, from February 2008 to August 2013 [21]. Study participants were selected and recruited to undergo study esophagogastroduodenoscopy (EGD) from 2 sources: (1) randomly selected patients eligible for screening colonoscopy from 1 of 7 primary care facilities and also agreed to undergo an EGD for the research study, and (2) consecutive patients previously scheduled to undergo EGD for any indication and agreed to research study enrollment. The first group of patients were patients eligible for colon cancer screening colonoscopy and who consented to additional EGD with gastric biopsies at the time of their colonoscopy; the second group of patients were ones already scheduled for EGD due to gastrointestinal symptoms and who consented to additional gastric biopsies as part of the research study. Inclusion criteria included any patient between 40 and 80 years old (50–80 years among the patients recruited from primary care clinics given colonoscopy recommendation starting at age 50). Exclusion criteria included: (1) previous gastroesophageal surgery; (2) previous cancer; (3) the use of anticoagulants; (4) platelet counts < 70,000, ascites, or gastroesophageal varices; or (5) history of major stroke or mental condition inhibiting interview ability. All participants provided written informed consent to take part in the study. The study was approved by the Institutional Review Board for Human Subjects Research for Baylor College of Medicine and the VA Research and Development Committee of the MEDVAMC.
All study participants answered questionnaires administered by trained research assistants reporting demographics (age, sex, race/ethnicity), lifestyle factors (alcohol, smoking), and medical history, including the use of medications (proton-pump inhibitor [PPI], histamine-2 receptor antagonist [H2RA], aspirin, and non-steroidal anti-inflammatory drug [NSAID]). Each individual had measurements of height, weight, and waist and hip circumference to calculate body mass index (BMI) and waist-to-hip ratio (WHR).
Dietary Questionnaire and Endoscopy
All subjects completed a 110-item Block Food Frequency Questionnaire (FFQ) version 2005 which ascertained dietary intake in the past year. The Block FFQ has previously been used and validated to estimate absolute intake [22]. The Block FFQ takes about 40–50 min to complete, and subjects who did not complete it at the endoscopy visit were asked to complete the questionnaire after the visit. The Block FFQ includes estimations of frequency (“never or less than once per month” to “2 + times per day”) and portion sizes (“small”, “medium”, “large”) for various types of food. Pictures were provided to help approximate portion size. All raw food items data from the Block FFQ were analyzed by NutritionQuest (Nutrition Data Systems for Research database version 2005, Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN, 2005) software. The amount consumed for each FFQ food item in the Nutrition Data Systems for Research was translated to cup or ounce equivalents.
Upper endoscopy with gastric mapping biopsies (at least 10 total biopsies from 5 to 7 biopsy sites) from the antrum (both greater and lesser curvature), corpus (distal greater curvature, distal lesser curvature, proximal greater curvature, and proximal lesser curvature), and cardia was performed on all study participants. The presence of GIM and H. pylori was determined according to updated Sydney classification [23] by two independent gastrointestinal pathologists, blinded to dietary data and endoscopic findings. A third gastrointestinal pathologist determined the readings in cases where there were discrepancies. H. pylori positivity was determined as H. pylori bacteria seen on histopathology of any gastric biopsy site (using hematoxylin and eosin, alcian blue at pH 2.5, a modified silver stain, or alcian blue–periodic acid Schiff stain) or positive H. pylori culture of gastric biopsy tissue. We have previously described the details of gastric tissue processing and storage for H. pylori culture [24].
Definition of Cases and Controls
Cases were defined as those patients with evidence of GIM on histopathology of ≥ 1 non-cardia gastric biopsy obtained during the study endoscopy. Cases were compared to controls without GIM present on histopathology of any non-cardia gastric biopsy.
Statistical Analysis
We compared characteristics between cases and controls using Chi-square for categorical variables and Student’s t test for continuous variables. Tertiles for each dietary factor were generated based on distribution of intake in the control group (3 highest to 1 lowest). Odds ratios (OR) and 95% confidence intervals (95% CI) were estimated for associations of dietary factors with risk of GIM using multivariable logistic regression models adjusted for age, gender, race/ethnicity, smoking, alcohol use, BMI, H. pylori infection, and total energy intake. Percent of daily kcal intake from fat, protein, carbohydrates, and sweets was ascertained, and associations of each factor with risk of GIM were estimated from multivariable models without a term for total energy intake.
All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC), and a 2-tailed p value of < 0.05 was considered statistically significant.
Results
A total of 564 patients were recruited from primary care clinics, and 1655 patients were recruited from endoscopy clinic. We analyzed 423 cases with GIM and 1796 controls without GIM. The overall prevalence of GIM was 19.1% in this cohort, ranging from 18.3% among asymptomatic patients recruited from primary care clinics to 21.3% among symptomatic patients recruited from endoscopy clinic. Of the control group, 21.9% had H. pylori infection, 22.0% had active gastritis, and 43.6% had chronic gastritis. Cases were older on average (mean age 62.1 years, SD 7.4 years) than controls (59.9 years, SD 8.2 years) and more likely to be male (97.2% vs. 90.8%), non-White ethnicity/race (58.6% vs. 39.0%), and ever smokers (76.4% vs. 66.3%) (Table 1). Of 40 patients with other/unknown race, 29 were of Asian ancestry (5 cases, 24 controls). Due to the small numbers, we were unable to examine associations between dietary factors and GIM among Asian race subgroup.
Table 1.
Demographic and clinical characteristics of 423 cases with gastric intestinal metaplasia and 1796 controls at the Michael E. DeBakey VA Medical Center who answered dietary questionnaires then underwent gastric mapping biopsies
| Cases (n = 423) | Controls (n = 1796) | p value | |
|---|---|---|---|
| Recruitment source | 0.121 | ||
| Endoscopy | 303 (71.63) | 1352 (75.28) | |
| Primary care | 120 (28.37) | 444 (24.72) | |
| Age | |||
| < 60 | 138 (32.62) | 761 (42.37) | < 0.001 |
| 60-69 | 218 (51.54) | 842 (46.88) | |
| ≥ 70 | 67 (15.84) | 193 (10.75) | |
| Sex | < 0.001 | ||
| Male | 411 (97.16) | 1630 (90.76) | |
| Female | 12 (2.84) | 166 (9.24) | |
| Race/ethnicity | < .0001 | ||
| White | 175 (41.37) | 1095 (60.97) | |
| Hispanic | 62 (14.66) | 148 (8.24) | |
| Black | 178 (42.08) | 521 (29.01) | |
| Other/unknown | 8 (1.89) | 32 (1.78) | |
| BMI (kg/m2) | 0.080 | ||
| < 25 | 89 (21.04) | 318 (17.71) | |
| 25-29 | 163 (38.53) | 634 (35.03) | |
| ≥ 30 | 171 (40.43) | 840 (46.77) | |
| Unknown/missing | 0 (0.00) | 4 (0.22) | |
| Waist-to-hip ratio | |||
| Low | 60 (14.18) | 248 (13.81) | 0.712 |
| High | 349 (82.51) | 1501 (83.57) | |
| Unknown/missing | 14 (3.31) | 47 (2.62) | |
| Smoking status | < 0.001 | ||
| Never smoked | 82 (19.39) | 501 (27.90) | |
| Current smoker | 138 (32.62) | 473 (26.34) | |
| Former smoker | 185 (43.74) | 718 (39.98) | |
| Unknown/missing | 18 (4.26) | 104 (5.79) | |
| Alcohol status | 0.169 | ||
| Never drinker | 26 (6.15) | 151 (8.41) | |
| Current drinker | 211 (49.88) | 891 (49.61) | |
| Former drinker | 166 (39.24) | 639 (35.58) | |
| Unknown/missing | 20 (4.73) | 115 (6.40) | |
| GERD symptoms | 0.047 | ||
| No | 224 (52.96) | 832 (46.33) | |
| Yes | 180 (42.55) | 865 (48.16) | |
| Unknown/missing | 19 (4.49) | 99 (5.51) | |
| Helicobacter pylori | < 0.001 | ||
| No | 199 (47.04) | 1377 (76.67) | |
| Yes | 219 (51.77) | 394 (21.94) | |
| Unknown/missing | 5 (1.18) | 25 (1.39) | |
| Active gastritis | < 0.001 | ||
| No | 193 (45.63) | 1395 (77.67) | |
| Yes | 225 (53.19) | 395 (21.99) | |
| Missing | 5 (1.18) | 6 (0.33) | |
| Chronic gastritis | < 0.001 | ||
| No | 77 (18.20) | 1011 (56.29) | |
| Yes | 344 (81.32) | 783 (43.60) | |
| Missing | 2 (0.47) | 2 (0.11) | |
| PPI/H2RA use | 0.076 | ||
| No | 180 (42.55) | 659 (36.69) | |
| Yes | 218 (51.54) | 1010 (56.24) | |
| Unknown/missing | 25 (5.91) | 127 (7.07) | |
| NSAID use | 0.267 | ||
| No | 153 (36.17) | 717 (39.92) | |
| Less than daily | 18 (4.26) | 83 (4.62) | |
| At least daily | 188 (44.44) | 704 (39.20) | |
| Unknown/missing | 64 (15.13) | 292 (16.26) |
BMI body mass index, GERD gastroesophageal reflux disease, PPI proton-pump inhibitor, H2RA histamine-2 receptor antagonist, NSAID non-steroidal anti-inflammatory drug
Cases with GIM had a median dietary energy consumption of 1683.5 kcal (25–75 percentile: 1161.2–2317.0 kcal) which was not different from controls (1613.5 kcal; 25–75 percentile: 1159.3–2161.4 kcal) (Table 2). A greater proportion of GIM cases than controls were in the lowest tertile of percent kcal from fat (tertile 1: 43.6% vs. 33.4%). However, higher proportions of GIM cases than controls were in the highest tertile of percent kcal from carbohydrates (tertile 3: 41.8% vs. 33.3%). Total fruit intake was the only dietary factor positively associated with GIM on unadjusted analysis (tertile 3 vs. 1: OR 1.51, 95% CI 1.04–2.18) (Table 3). There were no significant differences between cases and controls in the distribution of dietary consumption of total vegetables, dietary fiber, dietary fat, saturated fat, monounsaturated fat, polyunsaturated fat, cholesterol, trans fat, omega 3 fatty acids, sodium, or vitamin C. Similarly, the distribution of serving frequency of grains, meat, dairy, fat, and whole grains was not different between cases and controls.
Table 2.
Comparison of dietary intake between 423 cases with and 1796 controls without gastric intestinal metaplasia who answered dietary questionnaires then underwent gastric mapping biopsies
| Cases (n = 423) | Controls (n = 1796) | p value | |
|---|---|---|---|
| Dietary kcal (kcal), median (P25–P75) | 1683.45 (1161.22–2317.01) | 1613.5 (1159.31–2161.35) | 0.221 |
| Dietary protein (gm), median (P25–P75) | 62.41 (42.46–88.64) | 60.8 (43.31–84.48) | 0.441 |
| Dietary carbohydrate (gm), median (P25–P75) | 192.42 (130.81–269.69) | 180.36 (127.29–257.22) | 0.283 |
| Total fruit | 0.087 | ||
| Tertile 1 | 59 (26.82) | 291 (34.11) | |
| Tertile 2 | 74 (33.64) | 278 (32.59) | |
| Tertile 3 | 87 (39.55) | 284 (33.29) | |
| Total vegetables | 0.487 | ||
| Tertile 1 | 71 (32.42) | 286 (33.53) | |
| Tertile 2 | 66 (30.14) | 283 (33.18) | |
| Tertile 3 | 82 (37.44) | 284 (33.29) | |
| Fiber | 0.151 | ||
| Tertile 1 | 59 (26.94) | 284 (33.29) | |
| Tertile 2 | 75 (34.25) | 285 (33.41) | |
| Tertile 3 | 85 (38.81) | 284 (33.29) | |
| Fat | 0.511 | ||
| Tertile 1 | 74 (34.10) | 284 (33.29) | |
| Tertile 2 | 64 (29.49) | 285 (33.41) | |
| Tertile 3 | 79 (36.41) | 284 (33.29) | |
| Saturated fat | 0.385 | ||
| Tertile 1 | 73 (33.64) | 283 (33.18) | |
| Tertile 2 | 63 (29.03) | 286 (33.53) | |
| Tertile 3 | 81 (37.33) | 284 (33.29) | |
| Monounsaturated fat | 0.607 | ||
| Tertile 1 | 71 (32.57) | 284 (33.29) | |
| Tertile 2 | 67 (30.73) | 285 (33.41) | |
| Tertile 3 | 80 (36.70) | 284 (33.29) | |
| Polyunsaturated fat | 0.684 | ||
| Tertile 1 | 68 (31.78) | 284 (33.29) | |
| Tertile 2 | 68 (31.78) | 285 (33.41) | |
| Tertile 3 | 78 (36.45) | 284 (33.29) | |
| Cholesterol | 0.639 | ||
| Tertile 1 | 67 (30.59) | 284 (33.29) | |
| Tertile 2 | 80 (36.53) | 285 (33.41) | |
| Tertile 3 | 72 (32.88) | 284 (33.29) | |
| Trans fat | 0.363 | ||
| Tertile 1 | 72 (32.88) | 284 (33.29) | |
| Tertile 2 | 64 (29.22) | 285 (33.41) | |
| Tertile 3 | 83 (37.90) | 284 (33.29) | |
| Omega 3 fatty acids | 0.53 | ||
| Tertile 1 | 66 (30.41) | 284 (33.29) | |
| Tertile 2 | 70 (32.26) | 284 (33.29) | |
| Tertile 3 | 81 (37.33) | 285 (33.41) | |
| Sodium | 0.629 | ||
| Tertile 1 | 67 (31.16) | 284 (33.29) | |
| Tertile 2 | 69 (32.09) | 285 (33.41) | |
| Tertile 3 | 79 (36.74) | 284 (33.29) | |
| Vitamin C | 0.413 | ||
| Tertile 1 | 64 (29.09) | 284 (33.29) | |
| Tertile 2 | 74 (33.64) | 285 (33.41) | |
| Tertile 3 | 82 (37.27) | 284 (33.29) | |
| Grain servings | 0.87 | ||
| Tertile 1 | 0.87 | ||
| Tertile 2 | 74 (33.79) | 285 (33.41) | |
| Tertile 3 | 76 (34.70) | 284 (33.29) | |
| Meat servings | 0.745 | ||
| Tertile 1 | 79 (35.91) | 283 (33.18) | |
| Tertile 2 | 70 (31.82) | 285 (33.41) | |
| Tertile 3 | 71 (32.27) | 285 (33.41) | |
| Dairy servings | 0.296 | ||
| Tertile 1 | 81 (36.99) | 285 (33.41) | 0.296 |
| Tertile 2 | 61 (27.85) | 284 (33.29) | |
| Tertile 3 | 77 (35.16) | 284 (33.29) | |
| Fat servings | 0.282 | ||
| Tertile 1 | 76 (34.70) | 284 (33.29) | |
| Tertile 2 | 61 (27.85) | 284 (33.29) | |
| Tertile 3 | 82 (37.44) | 285 (33.41) | |
| Whole grain servings | 0.609 | ||
| Tertile 1 | 70 (31.96) | 287 (33.65) | |
| Tertile 2 | 68 (31.05) | 281 (32.94) | |
| Tertile 3 | 81 (36.99) | 285 (33.41) | |
| Percent kcal from fat | 0.003 | ||
| Tertile 1 | 95 (43.58) | 285 (33.41) | |
| Tertile 2 | 49 (22.48) | 284 (33.29) | |
| Tertile 3 | 74 (33.94) | 284 (33.29) | |
| Percent kcal from protein | 0.502 | ||
| Tertile 1 | 80 (36.87) | 285 (33.41) | |
| Tertile 2 | 73 (33.64) | 284 (33.29) | |
| Tertile 3 | 64 (29.49) | 284 (33.29) | |
| Percent kcal from carbohydrates | 0.004 | ||
| Tertile 1 | 79 (35.91) | 284 (33.29) | |
| Tertile 2 | 49 (22.27) | 285 (33.41) | |
| Tertile 3 | 92 (41.82) | 284 (33.29) | |
| Percent kcal from sweets | 0.412 | ||
| Tertile 1 | 63 (28.64) | 284 (33.29) | |
| Tertile 2 | 80 (36.36) | 285 (33.41) | |
| Tertile 3 | 77 (35.00) | 284 (33.29) |
p values < 0.05 are in bold
Kcal kilocalorie; P25–P75: 25–75 percentile; gm grams
Table 3.
Associations of dietary factors with risk of gastric intestinal metaplasia
| Unadjusted OR (95% CI) | Adjusted OR (95% CI) | |
|---|---|---|
| Total fruit | ||
| Tertile 2 versus 1 | 1.31 (0.90–1.92) | 0.96 (0.76–1.20) |
| Tertile 3 versus 1 | 1.51 (1.04–2.18) | 1.28 (1.02–1.61) |
| Total vegetables | ||
| Tertile 2 versus 1 | 0.94 (0.65–1.36) | 0.89 (0.71–1.12) |
| Tertile 3 versus 1 | 1.16 (0.81–1.66) | 1.19 (0.93–1.51) |
| Fiber | ||
| Tertile 2 versus 1 | 1.27 (0.87–1.85) | 0.99 (0.79–1.25) |
| Tertile 3 versus 1 | 1.44 (0.99–2.09) | 1.37 (1.04–1.80) |
| Fat | ||
| Tertile 2 versus 1 | 0.86 (0.59–1.25) | 0.91 (0.72–1.16) |
| Tertile 3 versus 1 | 1.07 (0.75–1.53) | 1.00 (0.74–1.36) |
| Saturated fat | ||
| Tertile 2 versus 1 | 0.85 (0.59–1.24) | 0.84 (0.67–1.07) |
| Tertile 3 versus 1 | 1.11 (0.77–1.58) | 1.11 (0.82–1.50) |
| Monounsaturated fat | ||
| Tertile 2 versus 1 | 0.94 (0.65–1.36) | 0.96 (0.76–1.20) |
| Tertile 3 versus 1 | 1.13 (0.79–1.61) | 1.06 (0.78–1.44) |
| Polyunsaturated fat | ||
| Tertile 2 versus 1 | 1.00 (0.69–1.45) | 0.96 (0.76–1.21) |
| Tertile 3 versus 1 | 1.15 (0.80–1.65) | 1.06 (0.79–1.43) |
| Cholesterol | ||
| Tertile 2 versus 1 | 1.19 (0.83–1.71) | 1.10 (0.88–1.37) |
| Tertile 3 versus 1 | 1.07 (0.74–1.56) | 0.88 (0.67–1.16) |
| Trans fat | ||
| Tertile 2 versus 1 | 0.89 (0.61–1.29) | 0.85 (0.67–1.07) |
| Tertile 3 versus 1 | 1.15 (0.81–1.65) | 1.17 (0.88–1.56) |
| Omega 3 fatty acids | ||
| Tertile 2 versus 1 | 1.06 (0.73–1.54) | 0.99 (0.79–1.25) |
| Tertile 3 versus 1 | 1.22 (0.85–1.76) | 1.09 (0.82–1.44) |
| Sodium | ||
| Tertile 2 versus 1 | 1.03 (0.71–1.49) | 0.99 (0.78–1.24) |
| Tertile 3 versus 1 | 1.18 (0.82–1.70) | 1.09 (0.80–1.48) |
| Vitamin C | ||
| Tertile 2 versus 1 | 1.15 (0.79–1.67) | 1.08 (0.86–1.35) |
| Tertile 3 versus 1 | 1.28 (0.89–1.85) | 1.11 (0.87–1.42) |
| Grain servings | ||
| Tertile 2 versus 1 | 1.07 (0.74–1.54) | 0.99 (0.79–1.24) |
| Tertile 3 versus 1 | 1.10 (0.76–1.59) | 1.04 (0.79–1.37) |
| Meat servings | ||
| Tertile 2 versus 1 | 0.88 (0.61–1.26) | 0.96 (0.77–1.21) |
| Tertile 3 versus 1 | 0.89 (0.62–1.28) | 0.82 (0.62–1.10) |
| Dairy servings | ||
| Tertile 2 versus 1 | 0.76 (0.52–1.09) | 0.82 (0.65–1.03) |
| Tertile 3 versus 1 | 0.95 (0.67–1.36) | 1.04 (0.82–1.32) |
| Fat servings | ||
| Tertile 2 versus 1 | 0.80 (0.55–1.17) | 0.83 (0.65–1.05) |
| Tertile 3 versus 1 | 1.08 (0.76–1.53) | 1.03 (0.80–1.33) |
| Whole grain servings | ||
| Tertile 2 versus 1 | 0.99 (0.68–1.44) | 0.93 (0.74–1.17) |
| Tertile 3 versus 1 | 1.17 (0.81–1.67) | 1.16 (0.92–1.47) |
| Percent kcal from fat a | ||
| Tertile 2 versus 1 | 0.52 (0.35–0.76) | 0.68 (0.53–0.87) |
| Tertile 3 versus 1 | 0.78 (0.55–1.10) | 1.05 (0.83–1.31) |
| Percent kcal from protein a | ||
| Tertile 2 versus 1 | 0.92 (0.64–1.31) | 1.03 (0.82–1.29) |
| Tertile 3 versus 1 | 0.80 (0.56–1.16) | 0.88 (0.70–1.11) |
| Percent kcal from carbohydrates a | ||
| Tertile 2 versus 1 | 0.62 (0.42–0.92) | 0.68 (0.53–0.87) |
| Tertile 3 versus 1 | 1.16 (0.83–1.64) | 1.35 (1.08–1.67) |
| Percent kcal from sweets a | ||
| Tertile 2 versus 1 | 1.27 (0.88–1.83) | 1.12 (0.90–1.40) |
| Tertile 3 versus 1 | 1.22 (0.84–1.77) | 0.98 (0.79–1.23) |
Dietary factors that demonstrate statistically significant associations with GIM are in bold
Models adjusted for age, gender, race/ethnicity, smoking status, alcohol status, body mass index, H. pylori infection, and total energy intake
OR odds ratio, CI confidence interval, Kcal kilocalorie
adjusted without total energy intake
After adjusting for age, gender, race/ethnicity, smoking, alcohol, BMI, H. pylori, and total energy intake, percent kcal from carbohydrates (OR 1.35, 95% CI 1.08–1.67) showed a positive association with GIM but also total fruit intake (tertile 3 vs. 1: OR 1.28, 95% CI 1.02–1.61) and total fiber intake (OR 1.37, 95% CI 1.04–1.80). Examining these associations among race/ethnic subgroups, the point estimates were generally higher and only statistically significant among non-Whites; however, the CIs overlapped with those for associations among Whites (Supplementary Table 1).
Discussion
Total fruit intake was the dietary factor consistently associated with GIM in this study of a non-immigrant US population. High carbohydrate intake but also fiber was associated with GIM on adjusted analysis. Otherwise, there were no significant associations between the presence of GIM and other self-reported dietary factors, including intake of vegetables, fat (including saturated, monounsaturated, polyunsaturated, cholesterol, trans fat, omega 3 fatty acids), sodium, vitamin C, grains, meat, dairy, and whole grains. The results from our stratified analyses suggest that these dietary associations may be strongest among non-Whites. However, we caution about drawing strong inferences from these small stratified analyses.
Several Western dietary studies found a significant relationship between high consumption levels of refined carbohydrates specifically and risk of gastric cancer [25–27]. One meta-analysis included 4 previous studies on carbohydrates and found a non-significant trend toward increased risk of gastric cancer with carbohydrate intake (RR 1.17, 95% CI 0.92–1.49) [28]. However, to our knowledge our study is the first to have described an association between carbohydrate consumption and development of GIM. High carbohydrate foods, such as rice, bread, and potatoes, often have a high glycemic index and are low in antioxidants [29] that may protect against gastric carcinogenesis. In mouse models, excess carbohydrates when combined with salt caused hypertrophy of the forestomach to digest carbohydrates with compensatory atrophic changes in the glandular stomach [30].
We could not replicate most of the positive findings shown in previous studies from countries with high prevalence of gastric cancer. In these studies, dietary factors associated with GIM included excessive consumption of spicy foods, salt, and nitrite-rich proteins, such as canned or smoked meat or fish, and low consumption of fruits and vegetables [5, 6, 13, 14]. In a study of Japanese immigrants living in Hawaii, nitrite-rich salty foods (cured meats) were associated with GIM, while vitamin C consumption was not [31]. However, in studies on Western diets, there have been conflicting data on the benefit of fruits and vegetables and harm of salt and high fat foods (e.g., butter, milk) on GIM risk. One Italian study compared H. pylori-positive patients with GIM to H. pylori-positive patients without GIM and found that butter was a risk factor for GIM development [10]. Similar to our findings, this study did not find a significant association between consumption of vegetables, milk, and processed meat and risk of GIM. Similar to our study, another Brazilian study found no association with salt intake but found fruit and vegetable intake to be inversely associated with GIM, while canned and smoked foods were associated with increased risk [14]. A US study found that higher milk consumption was associated with increased risk for GIM but inconsistent findings with fruit, vegetable, and vitamin C consumption [16]. One meta-analysis of 17 studies (that included 10 Western studies) found a trend toward increased risk of GIM with higher intake of salt and salted meats (combined OR 1.68, 95% CI 0.98–2.90) [15]. These studies, however, did not consistently control for important demographic and clinical risk factors in GIM (e.g., age, smoking, race, and H. pylori infection). Our study controlled for these demographic and clinical risk factors of GIM in addition to total energy intake and represents a more comprehensive risk assessment for GIM.
Recent studies have examined the interactions of diet, H. pylori, and the gastric microbiome on gastric cancer and GIM risk. Gastric atrophy and resulting hypochlorhydria due to H. pylori infection result in increased colonization by acid-intolerant bacteria [32, 33] and increased carcinogen production by ingested foods and nitrates [34–36]. Additionally, dietary factors (e.g., high salt) may increase the virulence of H. pylori organisms [37, 38]. Intake of starchy vegetables and fiber has been shown to increase precancerous gastric lesions in a Venezuelan study [39]. Although one study found high fiber intake increases mucosal cell apoptosis in rats treated with carcinogen [40], fiber has not conclusively been associated with gastric cancer risk [41]. Intake of antioxidants, such as vitamin E, C, beta-carotene found in fruits and vegetables, can promote apoptosis of cancerous cells in people with normal levels of reactive oxygen species. However, excessive consumption of these antioxidants could overly suppress reactive oxygen species and allow proliferation of cancerous cells, especially in those with low baseline levels of reactive oxygen species [42]. Vitamin E, contained in some nuts, vegetables, and fruits, may inhibit apoptosis and accelerate carcinogenesis [42] and has been shown to increase risk of gastric cancer [43].
Given the inconsistent and generally weak associations of dietary factors with GIM in this and other studies conducted in Western countries, there may not be an important role for diet in GIM development. Demographic and clinical risk factors (i.e., H. pylori, race/ethnicity, smoking) have been extensively described in GIM [8] and remain the strongest risk factors in GIM development. Diet may play a bigger role in neoplastic progression after GIM development. A subset of patients with GIM who continue to consume high-risk foods (nitrite, salty, low intake of fruits, vegetables, vitamin C) may progress to gastric cancer as there are stronger associations of dietary factors with gastric cancer than with GIM. Our findings should be replicated in future studies before recommending to decrease carbohydrate, fruit, and fiber consumption to attenuate the risk of GIM, especially as fruits and fiber may be beneficial to other conditions, such as cardiovascular disease and colon cancer [44–46]. Modifiable risk factors, such as smoking and H. pylori, that have shown consistent associations with GIM should be primarily targeted rather than diet for gastric cancer prevention.
Strengths of this study include the use of prospective subject enrollment and the use of the FFQ to gather dietary data prior to participants and study team knowing GIM case–control status. Patients were recruited from both an asymptomatic primary care clinic population and from those scheduled for upper endoscopy for gastrointestinal complaints and thus are representative of the overall VA population. Our high-risk cohort (GIM prevalence 19.1%) adequately allowed examination of significant dietary factors. Previous US studies that included only symptomatic patients undergoing biopsies in the setting of abnormal endoscopic findings estimated GIM prevalence to be 3.1–15.0% [47–49]. However, these studies did not perform systematic biopsies and may have under-reported true prevalence of GIM as GIM detection using white-light endoscopy is poor (51.0–71.2%) [50–52]. A previous US study performed systematic biopsies in symptomatic patients and found prevalence of GIM to be 11% among those who underwent 2 gastric biopsies and 25% among those who underwent 4 gastric biopsies [53]. Our cohort underwent systematic gastric mapping biopsies irrespective of symptoms or endoscopic findings, therefore minimizing bias by indication for endoscopy and abnormal endoscopic findings. Additionally, we systematically gathered H. pylori data on all patients and adjusted for multiple important risk factors for GIM risk (age, gender, race/ethnicity, smoking, and alcohol use).
One of the limitations of this study is the use of the Block FFQ, which requires participants to recall their dietary habits within the last 12 months. This may present recall bias associated with survey data, although any bias is unlikely to be differential between cases and controls as participants completed the FFQ prior to the study EGD. Furthermore, the median total calories are low in our study population (1683 kcal), possibly due to missing/not accounted for dietary components in the FFQ that are relevant to a veteran population. To our knowledge, the Block FFQ has not been validated in a veteran population. Nonetheless, this would not have impacted the study findings as reporting was consistent between cases and controls. The Block FFQ was not able to quantify nitrite-rich and processed foods, which have been associated with GIM and gastric cancer carcinogenesis. However, measurement of nitrites using FFQ is prone to errors as nitrite intake is difficult to directly quantify. Most nitrite exposure occurs when ingested nitrate is excreted in saliva and reduced to nitrite by oral bacteria and re-ingested [54]. Although we adjusted for several important risk factors (i.e., age, gender, race/ethnicity, smoking, alcohol, H. pylori), we were not able to adjust for all risk factors (e.g., family history of gastric cancer) which were not reliably available. As such, we cannot completely exclude residual confounding by these factors. Due to the observational, cross-sectional design of this study, there may be remaining bias due to unmeasured confounders, and the associations found in our study may not explain a causal relationship between diet and GIM risk. Additionally, we excluded certain populations from our recruitment (e.g., patients with cirrhosis and stroke) which introduces a selection bias, but these excluded populations with significant co-morbid conditions would have little benefit from reducing risk of GIM. The generalizability of our findings may be limited to older men as our population was a US veteran population. But given the high prevalence of GIM found in our cohort, our findings may apply to high-risk populations in the USA (e.g., immigrant, Hispanic).
We found higher intake of carbohydrates but also fruits and fiber were associated with GIM, but did not find an association for vegetable, fat, sodium, vitamin C, grain, meat, or dairy intake in a US population with primarily Western diets. The main actionable risk factors remain H. pylori and smoking, and these need to be targeted to reduce gastric cancer risk in US populations.
Supplementary Material
Acknowledgments
This work was supported in part by National Institutes of Health grant P30 DK056338 (Study Design and Clinical Research Core), which supports the Texas Medical Center Digestive Diseases Center. This research was supported in part with resources at the VA HSR&D Center for Innovations in Quality, Effectiveness and Safety (#CIN 13-413), at the Michael E. DeBakey VA Medical Center, Houston, TX. The opinions expressed reflect those of the authors and not necessarily those of the Department of Veterans Affairs, the US government or Baylor College of Medicine. Mimi C. Tan, MD, MPH is guarantor of the article.
Footnotes
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10620-020-06399-9) contains supplementary material, which is available to authorized users.
Conflict of interest The authors report no competing interests for this publication.
Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
References
- 1.Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018;68:394–424. [DOI] [PubMed] [Google Scholar]
- 2.Howlader NNA, Krapcho M, Miller D, Brest A, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA (eds). SEER Cancer Statistics Review, 1975–2016, National Cancer Institute. Bethesda, MD, https://seer.cancer.gov/csr/1975_2016/ based on November 2018 SEER data submission, posted to the SEER web site, April 2019. [Google Scholar]
- 3.Wang Z, Graham DY, Khan A, et al. Incidence of gastric cancer in the USA during 1999 to 2013: a 50-state analysis. Int J Epidemiol. 2018;47:966–975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Graham DY. Helicobacter pylori update: gastric cancer, reliable therapy, and possible benefits. Gastroenterology. 2015;148:719–31.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Oh S, Kim N, Yoon H, et al. Risk factors of atrophic gastritis and intestinal metaplasia in first-degree relatives of gastric cancer patients compared with age-sex matched controls. J Cancer Prev. 2013;18:149–160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kim N, Park YS, Cho SI, et al. Prevalence and risk factors of atrophic gastritis and intestinal metaplasia in a Korean population without significant gastroduodenal disease. Helicobacter. 2008;13:245–255. [DOI] [PubMed] [Google Scholar]
- 7.Leung WK, Ng EK, Chan WY, et al. Risk factors associated with the development of intestinal metaplasia in first-degree relatives of gastric cancer patients. Cancer Epidemiol Biomarkers Prev. 2005;14:2982–2986. [DOI] [PubMed] [Google Scholar]
- 8.Tan MC, Mallepally N, Liu Y, et al. Demographic and lifestyle risk factors for Gastric Intestinal Metaplasia among US Veterans. Am J Gastroenterol 2019; accepted manuscript (in press). [DOI] [PubMed] [Google Scholar]
- 9.Jedrychowski W, Popiela T, Drews M, et al. Effect of Helicobacter pylori infection, smoking and dietary habits on the occurrence of antrum intestinal metaplasia. Clinico-epidemiological study in Poland. Pol J Pathol Off J Pol Soc Pathol. 1999;50:289–295. [PubMed] [Google Scholar]
- 10.Russo A, Maconi G, Spinelli P, et al. Effect of lifestyle, smoking, and diet on development of intestinal metaplasia in H. pylori-positive subjects. Am J Gastroenterol. 2001;96:1402–1408. [DOI] [PubMed] [Google Scholar]
- 11.Marcos-Pinto R, Carneiro F, Dinis-Ribeiro M, et al. First-degree relatives of patients with early-onset gastric carcinoma show even at young ages a high prevalence of advanced OLGA/OLGIM stages and dysplasia. Aliment Pharmacol Ther. 2012;35:1451–1459. [DOI] [PubMed] [Google Scholar]
- 12.Rokkas T, Sechopoulos P, Pistiolas D, et al. Helicobacter pylori infection and gastric histology in first-degree relatives of gastric cancer patients: a meta-analysis. Eur J Gastroenterol Hepatol. 2010;22:1128–1133. [DOI] [PubMed] [Google Scholar]
- 13.Song JH, Kim YS, Heo NJ, et al. High salt intake is associated with atrophic gastritis with intestinal metaplasia. Cancer Epidemiol Biomark Prev. 2017;26:1133–1138. [DOI] [PubMed] [Google Scholar]
- 14.Taborda AG, Prolla JC. Alimentary factors in the development of gastric intestinal metaplasia in functional dyspeptic patients. Arquivos de Gastroenterologia. 2012;49:208–213. [DOI] [PubMed] [Google Scholar]
- 15.Dias-Neto M, Pintalhao M, Ferreira M, et al. Salt intake and risk of gastric intestinal metaplasia: systematic review and meta-analysis. Nutr Cancer. 2010;62:133–147. [DOI] [PubMed] [Google Scholar]
- 16.Fontham E, Zavala D, Correa P, et al. Diet and chronic atrophic gastritis: a case-control study. J Natl Cancer Inst. 1986;76:621–627. [DOI] [PubMed] [Google Scholar]
- 17.Garnweidner LM, Terragni L, Pettersen KS, et al. Perceptions of the host country’s food culture among female immigrants from Africa and Asia: aspects relevant for cultural sensitivity in nutrition communication. J Nutr Educ Behav. 2012;44:335–342. [DOI] [PubMed] [Google Scholar]
- 18.Park SY, Paik HY, Skinner JD, et al. Mothers’ acculturation and eating behaviors of Korean American families in California. J Nutr Educ Behav. 2003;35:142–147. [DOI] [PubMed] [Google Scholar]
- 19.Yang EJ, Chung HK, Kim WY, et al. Chronic diseases and dietary changes in relation to Korean Americans’ length of residence in the United States. J Am Diet Assoc. 2007;107:942–950. [DOI] [PubMed] [Google Scholar]
- 20.Thapa S, Fischbach LA, Delongchamp R, et al. Association between dietary salt intake and progression in the gastric precancerous process. Cancers. 2019;11:467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Nguyen TH, Thrift AP, Ramsey D, et al. Risk factors for Barrett’s esophagus compared between African Americans and non-Hispanic Whites. Am J Gastroenterol. 2014;109:1870–1880. [DOI] [PubMed] [Google Scholar]
- 22.Subar AF, Thompson FE, Kipnis V, et al. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires: the Eating at America’s Table Study. Am J Epidemiol. 2001;154:1089–1099. [DOI] [PubMed] [Google Scholar]
- 23.Dixon MF, Genta RM, Yardley JH, et al. Classification and grading of gastritis. The updated Sydney System. International workshop on the Histopathology of Gastritis, Houston 1994. Am J Surg Pathol. 1996;20:1161–1181. [DOI] [PubMed] [Google Scholar]
- 24.Shiota S, Thrift AP, Green L, et al. Clinical manifestations of Helicobacter pylori-negative gastritis. Clin Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Assoc. 2017;15:1037–46.e3. [DOI] [PubMed] [Google Scholar]
- 25.Jansen MC, Bueno-de-Mesquita HB, Rasanen L, et al. Consumption of plant foods and stomach cancer mortality in the seven countries study. Is grain consumption a risk factor? Seven Countries Study Research Group. Nutr Cancer. 1999;34:49–55. [DOI] [PubMed] [Google Scholar]
- 26.Jedrychowski W, Popiela T, Steindorf K, et al. Nutrient intake patterns in gastric and colorectal cancers. Int J Occup Med Environ Health. 2001;14:391–395. [PubMed] [Google Scholar]
- 27.Risch HA, Jain M, Choi NW, et al. Dietary factors and the incidence of cancer of the stomach. Am J Epidemiol. 1985;122:947–959. [DOI] [PubMed] [Google Scholar]
- 28.Fang X, Wei J, He X, et al. Landscape of dietary factors associated with risk of gastric cancer: a systematic review and dose-response meta-analysis of prospective cohort studies. Eur J Cancer. 2015;51:2820–2832. [DOI] [PubMed] [Google Scholar]
- 29.Larsson SC, Bergkvist L, Wolk A. Glycemic load, glycemic index and carbohydrate intake in relation to risk of stomach cancer: a prospective study. Int J Cancer. 2006;118:3167–3169. [DOI] [PubMed] [Google Scholar]
- 30.Kodama M, Kodama T, Suzuki H, et al. Effect of rice and salty rice diets on the structure of mouse stomach. Nutr Cancer. 1984;6:135–147. [DOI] [PubMed] [Google Scholar]
- 31.Stemmermann GN, Nomura AM, Chyou PH, et al. Impact of diet and smoking on risk of developing intestinal metaplasia of the stomach. Dig Dis Sci. 1990;35:433–438. [DOI] [PubMed] [Google Scholar]
- 32.Brawner KM, Morrow CD, Smith PD. Gastric microbiome and gastric cancer. Cancer J (Sudbury, Mass). 2014;20:211–216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Eun CS, Kim BK, Han DS, et al. Differences in gastric mucosal microbiota profiling in patients with chronic gastritis, intestinal metaplasia, and gastric cancer using pyrosequencing methods. Helicobacter. 2014;19:407–416. [DOI] [PubMed] [Google Scholar]
- 34.Stillwell WG, Glogowski J, Xu HX, et al. Urinary excretion of nitrate, N-nitrosoproline, 3-methyladenine, and 7-methylguanine in a Colombian population at high risk for stomach cancer. Cancer Res. 1991;51:190–194. [PubMed] [Google Scholar]
- 35.Tan MC, Graham DY. Proton pump inhibitor therapy after Helicobacter pylori eradication may increase the risk of gastric cancer. BMJ Evid Based Med. 2018;23:111–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Miftahussurur M, Yamaoka Y, Graham DY. Helicobacter pylori as an oncogenic pathogen, revisited. Expert Rev Mol Med. 2017;19:e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Gancz H, Jones KR, Merrell DS. Sodium chloride affects Helicobacter pylori growth and gene expression. J Bacteriol. 2008;190:4100–4105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Loh JT, Friedman DB, Piazuelo MB, et al. Analysis of Helicobacter pylori cagA promoter elements required for salt-induced upregulation of CagA expression. Infect Immun. 2012;80:3094–3106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Kato I, Vivas J, Plummer M, et al. Environmental factors in Helicobacter pylori-related gastric precancerous lesions in Venezuela. Cancer Epidemiol Biomark Prev. 2004;13:468–476. [PubMed] [Google Scholar]
- 40.Jenab M, Thompson LU. Phytic acid in wheat bran affects colon morphology, cell differentiation and apoptosis. Carcinogenesis. 2000;21:1547–1552. [PubMed] [Google Scholar]
- 41.World Cancer Research Fund American Institute for Cancer Research. Diet, nutrition, physical activity and stomach cancer: World Cancer Research Fund International; 2018. Retrieved from https://www.wcrf.org/sites/default/files/Stomach-cancer-report.pdf Available from: https://www.wcrf.org/sites/default/files/Stomach-cancer-report.pdf. [Google Scholar]
- 42.Salganik RI. The benefits and hazards of antioxidants: controlling apoptosis and other protective mechanisms in cancer patients and the human population. J Am Coll Nutr. 2001;20:464S–472S. discussion 73S-75S. [DOI] [PubMed] [Google Scholar]
- 43.Nouraie M, Pietinen P, Kamangar F, et al. Fruits, vegetables, and antioxidants and risk of gastric cancer among male smokers. Cancer Epidemiol Biomark Prev. 2005;14:2087–2092. [DOI] [PubMed] [Google Scholar]
- 44.Threapleton DE, Greenwood DC, Evans CE, et al. Dietary fibre intake and risk of cardiovascular disease: systematic review and meta-analysis. BMJ (Clin Res Ed). 2013;347:f6879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Aune D, Giovannucci E, Boffetta P, et al. Fruit and vegetable intake and the risk of cardiovascular disease, total cancer and all-cause mortality-a systematic review and dose-response meta-analysis of prospective studies. Int J Epidemiol. 2017;46:1029–1056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Ma Y, Hu M, Zhou L, et al. Dietary fiber intake and risks of proximal and distal colon cancers: a meta-analysis. Medicine (Baltimore). 2018;97:e11678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Da B, Jani N, Gupta N, et al. High-risk symptoms do not predict gastric cancer precursors. Helicobacter. 2019;24:e12548. [DOI] [PubMed] [Google Scholar]
- 48.Simko V, Anand N, Ginter E. Gastric intestinal metaplasia—age, ethnicity and surveillance for gastric cancer. Bratislavske lekarske listy. 2015;116:3–8. [DOI] [PubMed] [Google Scholar]
- 49.Almouradi T, Hiatt T, Attar B. Gastric intestinal metaplasia in an underserved population in the USA: prevalence, epidemiologic and clinical features. Gastroenterol Res Pract. 2013;2013:856256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Capelle LG, Haringsma J, de Vries AC, et al. Narrow band imaging for the detection of gastric intestinal metaplasia and dysplasia during surveillance endoscopy. Dig Dis Sci. 2010;55:3442–3448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Pimentel-Nunes P, Libanio D, Lage J, et al. A multicenter prospective study of the real-time use of narrow-band imaging in the diagnosis of premalignant gastric conditions and lesions. Endoscopy. 2016;48:723–730. [DOI] [PubMed] [Google Scholar]
- 52.Sobrino-Cossío S, Abdo Francis JM, Emura F, et al. Efficacy of narrow-band imaging for detecting intestinal metaplasia in adult patients with symptoms of dyspepsia. Revista de Gastroenterología de México (English Edition). 2018;83:245–252. [DOI] [PubMed] [Google Scholar]
- 53.Fennerty MB, Emerson JC, Sampliner RE, et al. Gastric intestinal metaplasia in ethnic groups in the southwestern United States. Cancer Epidemiol Biomark Prev. 1992;1:293–296. [PubMed] [Google Scholar]
- 54.Tannenbaum SR, Weisman M, Fett D. The effect of nitrate intake on nitrite formation in human saliva. Food Cosmet Toxicol. 1976;14:549–552. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.