Abstract
Infection with Helicobacter pylori is linked to inflammation and is the main cause of peptic ulcer, gastritis, and gastric malignancies. To examine associations between gastric cancer risk and the erythrocyte composition of docosahexaenoic acid (DHA), a fatty acid with anti-inflammatory and apoptosis-inducing effects, here we conducted a case-control study of 179 incident gastric cancer cases and 357 noncancer controls (matched by age, sex, and season of sample collection). Dietary information and blood samples were collected from all subjects, and erythrocyte fatty acid levels were measured using accelerated solvent extraction and gas-liquid chromatography. Gastric cancer risk did not seem to be directly associated with dietary intake of fish and n-3 highly unsaturated fatty acids (HUFAs), such as DHA, derived from fish. However, risk was inversely associated with erythrocyte compositions of n-3 HUFAs [the highest to the lowest tertile, odds ratio (OR), 0.39; 95% confidence interval (95% CI), 0.23-0.68; Ptrend < 0.005] and DHA (OR, 0.47; 95% CI, 0.28-0.79; Ptrend < 0.01). Particularly strong associations were noted for well-differentiated type lesions and n-3 HUFAs (OR, 0.10; 95% CI, 0.03-0.35; Ptrend = 0.0005) as well as DHA (OR, 0.20; 95% CI, 0.07-0.58; Ptrend < 0.01) values. In conclusion, the erythrocyte composition of DHA was found to be negatively linked to risk of gastric cancer, especially of well-differentiated adenocarcinoma. Further studies are needed to investigate mechanisms of action of DHA relevant to antitumor effects in the stomach. (Cancer Epidemiol Biomarkers Prev 2007;16(11):2406–15)
Introduction
Infection with Helicobacter pylori plays a key role in the etiology of peptic ulcers, gastritis, and gastric malignancies related to chronic inflammation and is regarded as a trigger for the sequence leading to carcinoma development in the stomach (Fig. 1; ref. 1). In gastric cancer tissues and cell lines, overexpression of cyclooxygenase-2 and down-regulation of 15-lipoxygenase-1, linked to inflammation, angiogenesis, and apoptosis, have been shown (2, 3). A recent laboratory study also provided evidence of carcinogenesis related to H. pylori infection and cyclooxygenase-2 (4). Risk of development and mortality from gastric cancer among nonsteroidal anti-inflammatory drugs (NSAIDs; in here, as cyclooxygenase inhibitors) users also seems to be lower than among nonusers (5-7). Some studies have reported that NSAIDs have a protective role in patients with peptic ulcers, but others have shown elevation of the risk among NSAIDs users infected with H. pylori (8, 9). NSAIDs are furthermore associated with serious upper gastrointestinal bleeding and cardiovascular events, and such harmful affects are a serious public health concern in American-European populations (10-13). Eradication of H. pylori has been reported to markedly reduce the likelihood of gastric diseases, but evidence for gastric cancer is inconclusive (14-17).
Hypothetical mechanism for suppressing gastric cancer related to chronic inflammation according to DHA. Infection with H. pylori to normal gastric mucosa is suggested to induce chronic inflammation in the early life of carriers and cause diseases such as peptic ulceration, gastritis, and gastric cancer after years or decades (1). Infection, therefore, is regarded as a trigger for the sequence of carcinoma development in the stomach. In gastric cancer tissues and cell lines, overexpression of cyclooxygenase-2 linked to the inflammatory responses, angiogenesis, and apoptosis has been observed (2-4, 42, 43). DHA, which is rich in blue skin fish and is n-3 HUFAs, is related to inhibition for the tumorigenesis, like NSAIDs (i.e., here, cyclooxygenase-2 inhibitors). Against the promotion and progression stages of tumor development, DHA is thought to play roles in (a) competitive inhibition of incorporation of arachidonic acid into membrane phospholipids, (b) competitive inhibition of metabolism of arachidonic acid to prostaglandin E2, and (c) apoptosis induction according to the up-regulation of Bax expression (18-20).
Hypothetical mechanism for suppressing gastric cancer related to chronic inflammation according to DHA. Infection with H. pylori to normal gastric mucosa is suggested to induce chronic inflammation in the early life of carriers and cause diseases such as peptic ulceration, gastritis, and gastric cancer after years or decades (1). Infection, therefore, is regarded as a trigger for the sequence of carcinoma development in the stomach. In gastric cancer tissues and cell lines, overexpression of cyclooxygenase-2 linked to the inflammatory responses, angiogenesis, and apoptosis has been observed (2-4, 42, 43). DHA, which is rich in blue skin fish and is n-3 HUFAs, is related to inhibition for the tumorigenesis, like NSAIDs (i.e., here, cyclooxygenase-2 inhibitors). Against the promotion and progression stages of tumor development, DHA is thought to play roles in (a) competitive inhibition of incorporation of arachidonic acid into membrane phospholipids, (b) competitive inhibition of metabolism of arachidonic acid to prostaglandin E2, and (c) apoptosis induction according to the up-regulation of Bax expression (18-20).
Similar to NSAIDs, n-3 (or omega-3) highly unsaturated fatty acids (HUFAs), such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), may inhibit the promotion and progression stages of neoplasia through modulating gene expression or signal transduction involved in the control of cell growth, differentiation, apoptosis, angiogenesis, and metastasis (18, 19). These fatty acids also competitively inhibit the biosynthesis of eicosanoids (e.g., prostaglandin E2) via the arachidonic acid cascade, linked to inflammation, tumorigenesis, and cell proliferation (18, 20). Eicosanoids from n-3 HUFAs also decrease the formation of arachidonic acid–derived eicosanoids (18). Although ecological studies have suggested that high intake of fish rich in n-3 HUFAs is correlated with low cancer incidences in the lung, colon, and prostate for men and the breast, cervix, lung, and colon for women; such a negative relation for gastric cancer has not been observed (21, 22). EPA and DHA values in tumor parts of mucosa phospholipids among Korean gastric cancer patients, however, were significantly lower than those in noncancer parts (23).
The Japanese have the highest incidence of and the mortality from gastric cancer in the world (24, 25), and their dietary intake of fish is also the highest. A very recent large-scale cohort study in Japan showed the rate of H. pylori infection among the general population to be ∼75.0%, the risk of gastric cancer being at least 3-fold elevated among people with pepsinogen levels indicative of severe atrophic gastritis (26). As biomarkers for assessing dietary intakes of n-3 HUFAs, EPA, and DHA over the middle to long term, fatty acids in erythrocytes (120 days half-life) are useful because they are not biosynthesized in vivo (27, 28). Utilities of these biomarkers in intervention studies have been shown (27). A new analytic method for measuring fatty acids in biomaterials, including phospholipids of erythrocyte membranes, has been recently developed using an automatic solvent extractor and gas-liquid chromatography, with confirmed high precision and accuracy, making it feasible to use small volume multi-samples routinely, rapidly, and cheaply (28). The hypothesis tested in the present study is that DHA with anti-inflammatory effects (and possible induction of apoptosis in the stomach) might have a negative relation to gastric cancer risk (Fig. 1).
Materials and Methods
Procedure
Between December 2002 and May 2005, the subjects (20-79 years) of the present study were recruited within the framework of the Hospital-Based Epidemiologic Research Program at Aichi Cancer Center (HERPACC; ref. 29). The study design has been elsewhere described (29), and the investigation was executed in series within the Aichi Fatty Acid (AiFat) Research project to clarify associations between a large number of blood parameters and cancers in several sites (28, 30, 31). Briefly, all first-visit outpatients (n = 18,103), including all cancer cases (n = 3,972), were asked to fill out a questionnaire regarding their lifestyle and to provide 7-mL blood samples. All subjects are provided with an explanatory document and requested to give their written informed consent for participation in the study, which was approved by the Ethics Committee of the Aichi Cancer Center.
Participants
Dietary information was provided by 97.5% of 15,650 eligible subjects and systematically checked by trained interviewers. Blood samples were provided by 6,464 subjects (42.4%). The participation rates for all cancer cases, gastric cancer cases, and controls were 43.6%, 42.7%, and 41.9%, in that order. Almost all of the participants provided their blood on their first visit day at Aichi Cancer Center Hospital. Outcome was confirmed using hospital-based cancer registries. In the present study, 35 cases with multiple, recurrent, and metastasis cancers were excluded, and 13 cases were also excluded due to the small volume of collected blood. One hundred and seventy-nine newly diagnosed (incident) cases completed the questionnaire and donated blood samples during the study period and were histologically diagnosed as having gastric cancers at our hospital. Only two cases, who were diagnosed at other hospitals within 1 and 9 months, were included. With reference to the Japanese Classification of Gastric Carcinoma (2nd English edition), case subjects were divided into two groups regarding the histologic classification of the gastric adenocarcinoma; i.e., poorly differentiated adenocarcinoma (including signet-ring cell carcinoma) and well-differentiated types (including papillary adenocarcinoma and well- and moderately differentiated types of tubular adenocarcinoma).
The control subjects were randomly selected from first-visit outpatients who visited Aichi Cancer Center Hospital at the same period. They were confirmed to have no cancer or any prior history of cancer according to both a questionnaire and the cancer registry system in Aichi Prefecture. Approximately 57.9% of study subjects had no current or past diseases according to the questionnaire, and most of them visited for a health check or cancer screening at our hospital. Subjects with or having a current or past history of the following diseases were excluded: hepatitis (3.2% of all subjects), liver cirrhosis (1.9%), chronic nephritis (2.3%), diabetes (5.4%), angina (2.9%), stroke (1.1%), ovary resection (3.7%) and uterus resection (4.9%). Considering seasonal differences in biomarkers for dietary intakes of fish and n-3 HUFAs (30, 31), 357 controls were individually matched for age (±5 years), sex, and season of sample collection to cases with a 1:2 case-control ratio.
All subjects for the present study were Japanese, living in and around Aichi Prefecture, central Japan. We earlier showed that it is feasible to use noncancer outpatients at the Aichi Cancer Center Hospital as controls in epidemiologic studies because their general lifestyles are accordant with those of the general population randomly selected from the electoral roll in Nagoya City, Aichi Prefecture (32).
Questionnaire
Our questionnaire covered the following personal information: height, weight, dietary habits, habitual exercise, drinking habit, smoking status, and thorough medical information such as family history of cancer and current and prevalent history of diseases. Dietary assessment was elsewhere described in detail (30, 31). In brief, average daily intakes of various foods and nutrients were assessed using a semiquantitative food frequency questionnaire with 47 food items, which had an acceptable relative validity and reproducibility for their consumption (33-35). Spearman's correlation coefficients between dietary intakes of fatty acids and the corresponding plasma concentrations (mmol/L) as biomarkers were 0.38 and 0.26 for n-3 HUFAs and 0.43 and 0.37 for the ratio of n-6 PUFAs/n-3 PUFAs in men and women, respectively (36).
For the control and case subjects, we asked for information about their lifestyle at the study enrollment and before the onset of disease, respectively. Habitual dietary intake during the latest 1 year was inquired with eight category frequencies of 47 foods/food groups: i.e., never or seldom, 1 to 3 times/month, 1 to 2 times/week, 3 to 4 times/week, 5 to 6 times/week, 1 time/day, 2 times/day and ≥3 times/day. Dietary intake (g/day) of individual fatty acids was computed by multiplying the standard portion size (in grams), frequency of consumption, and the content (per 100 g) of each fatty acid in foods as listed in the national Standard Tables of Food Consumption. Dietary intakes of seafood (composed of fish and other seafood), individual fatty acids, and sodium were also calculated. Fatty acids were summed up and categorized as described in Selected Fatty Acids and Grouping. Salt (sodium chloride) from traditional Japanese condiments such as soy sauce was not included in sodium because they were not listed in our questionnaire due to very low validities and reliabilities for their intakes. Dietary intakes of foods and nutrients were adjusted for total energy intake of each person and tertile cut-points (mg or g/1,000 kcal) in control subjects were used to designate low, moderate, and high intakes. Green tea consumption (times/day) was defined as follows: 0.1 for never or seldom, 1 to 3 times/month; 0.5 for 1 to 6 times/week; 1.5 for 1 or 2 times/day; and 3.0 for ≥3 times/day.
Lifestyle factors were classified into three groups as follows: (a) <1 time/week as low, 1 to 2 times/week as moderate, and ≥3 times/week as high for habitual exercise other than work; (b), <1 time/week as low, 1 to 4 times/week as moderate, and ≥5 times/week as high for drinking habits; and (c) current, former, and never-smokers for smoking status. We defined former smokers as those who quit smoking more than 2 years before the questionnaire study. Body mass index (BMI, kg/m2) was calculated from the self-reported height (m) and weight (kg).
Analysis of Fatty Acids in Erythrocytes
We earlier established a new analytic method for measuring fatty acids in biomaterials using an accelerated solvent extractor and a gas-liquid chromatography (28). In short, membranes (white ghosts) from 50 μL of erythrocytes were prepared with sodium phosphate buffer. Using an accelerated solvent extractor (ASE) 200 (Nippon Dionex), lipids (first extraction) and fatty acid methyl esters (second extraction) were extracted with chloroform-methanol, 1:2, by volume and petroleum ether, respectively. Heptadecanoic acid and butylated hydroxytoluene were applied as an internal standard and an antioxidant, respectively. The two extraction processes were automatically achieved with computerized programs. The extracted lipids were treated with hydrochloride-methanol reagents for fatty acid conversion and subsequent methyl-transformation of fatty acids. Under the conditions described elsewhere (28), the fatty acid methyl esters were analyzed by Shimadzu GC-2010 gas chromatography (Shimadzu) on a capillary column DB-225 (J&W Scientific), equipped with an auto-injector, an auto-sampler, and a flame ionization detector. With the use of commercial standards of known retention time, individual fatty acids were identified, and then integration of the peak areas was done with GC solution software, version 2 (Shimadzu). Regarding fatty acid measurement, all information for study subjects was completely blind to the laboratory staff.
In line with our previous studies (28, 30, 31), we selected 13 fatty acids, predominating in both dietary intake and erythrocyte contents. The composition of each in erythrocyte membranes was determined as mol percentage (mol%) of the total fatty acid concentration (mmol/L). Based on the analysis of a series of 10 samples measured within 1 day, intra-assay coefficients of variation were <4.0%, except for a minor group of myristic acid, γ-linolenic acid, and α-linolenic acid (≤0.5% of total fatty acids for each; ref. 28). Inter-assay coefficients of variation were based on replicate analyses of pooled erythrocytes (a total 100 samples) over a period of 10 days and also were <4.0%, except for a minor group of the three fatty acids (28).
Selected Fatty Acids and Grouping
Myristic acid, palmitic acid, palmitoleic acid, stearic acid, oleic acid, linoleic acid, γ-linolenic acid, α-linolenic acid, dihomo-γ-linolenic acid, arachidonic acid, EPA, docosapentaenoic acid, and DHA were analyzed. Using total fatty acids as the denominator, the following seven groups of fatty acids in erythrocyte membranes were summarized: saturated fatty acids (SFAs; myristic acid, palmitic acid, and stearic acid); monounsaturated fatty acids (MUFAs; palmitoleic acid and oleic acid); polyunsaturated fatty acids (PUFAs; n-6 PUFAs and n-3 PUFAs); n-6 PUFAs (linoleic acid, γ-linolenic acid, dihomo-γ-linolenic acid, and arachidonic acid); n-3 PUFAs (α-linolenic acid and n-3 HUFAs); and n-3 HUFAs (EPA, docosapentaenoic acid, and DHA). We also defined the ratios of specific fatty acids as follows: saturation indexn-7 (palmitic acid/palmitoleic acid), saturation indexn-9 (stearic acid/oleic acid), SFAs/PUFAs, SFAs/n-6 PUFAs, SFAs/n-3 PUFAs, SFAs/n-3 HUFAs, n-6 PUFAs/n-3 PUFAs, n-6 PUFAs/n-3 HUFAs, arachidonic acid/EPA, and arachidonic acid/DHA. As indicators of membrane fluidity, two saturation indices were used considering activity of the rate-limiting delta 9-desaturases (stearoyl-CoA desaturase) that transform SFAs into the corresponding MUFAs (30, 31). The ratio of n-6 PUFAs/n-3 PUFAs has been suggested to be particularly important for human health (30, 31). The ratios of SFAs/PUFAs, arachidonic acid/EPA, and arachidonic acid/DHA were applied as indicators of competitive incorporation of fatty acids into phospholipids in erythrocyte membranes (30, 31).
Statistical Methods
The Student's t test and the χ2 test were used to assess the significance of differences in means and proportions between cases and controls, respectively. In control subjects, partial Spearman's correlation coefficients between fatty acids in diet (g/1,000 kcal) and in erythrocyte membranes (mol%) were adjusted for age, BMI, and season of sample collection. Case subjects were categorized according to the tertile levels of erythrocyte fatty acids among control subjects, and the odds ratios (OR) for the middle and the highest to the lowest tertile were estimated. Using conditional logistic regression models, ORs and the 95% confidence intervals (95% CI) were calculated after adjustment for BMI (continuous), smoking habit, family history of gastric cancer (yes or no) in parents and/or siblings, green-yellow vegetables intake (g/1,000 kcal), other vegetables intake (g/1,000 kcal), fruit intake (g/1,000 kcal), sodium intake (g/1,000 kcal) and green tea consumption (times/day). These possible confounding factors were considered with reference to previously published reports for gastric cancer risk (26, 37-39). Considering the observed colinearity of dietary intakes of vitamin C and those foods, vitamin C as a possible protective factor was not included in the multivariate models. In Japanese, NSAIDs users are very few. Tests for the trend with each variable were conducted by assigning the median values in control subjects. All statistical analyses were conducted with SAS version 9.1 (SAS Institute Inc.), and P < 0.05 was considered statistically significant.
Results
Table 1 shows characteristics of the subjects. Sodium intake was higher in cases than controls (P < 0.05), but no difference was found for other foods and nutrients. Proportions of habitual exercise and drinking habit also did not differ (data not shown). The proportion of poorly differentiated adenocarcinoma was higher than that of well-differentiated type (62.0% and 38.0%, respectively), with average patient ages of 57.8 ± 10.3 and 61.6 ± 9.2 years (P < 0.05), respectively. The proportions of women were 30.6% and 13.2% for poorly and well-differentiated adenocarcinoma (P < 0.01), respectively. With the exception of MUFAs and arachidonic acid, the values for selected fatty acids in erythrocyte membranes significantly differed between case and control subjects (Table 2). In cases, values for SFAs, palmitic acid, the saturation indexn-7, and a series of SFAs/PUFAs and n-6 PUFAs/n-3 PUFAs ratios were higher than in controls, whereas those of DHA and n-3 HUFAs were lower (at least P < 0.05 for all).
Mean (SD) of some variables possibly related to gastric cancer in case and control subjects
. | Case subjects (n = 179) . | Control subjects (n = 357) . | P value for χ2 or t test . | |||
---|---|---|---|---|---|---|
Age (y), mean (SD) | 59.3 (10.0) | 58.9 (10.1) | NS | |||
Body mass index (kg/m2), mean (SD) | 23.0 (3.0) | 22.9 (2.6) | NS | |||
Women, n (%) | 43 (24.0) | 86 (24.1) | NS | |||
Smoking habit, n (%) | ||||||
Smokers | 70 (39.1) | 112 (31.4) | NS | |||
Ex-smokers | 50 (27.9) | 110 (30.8) | ||||
Nonsmokers | 59 (33.0) | 135 (37.8) | ||||
Family history of gastric cancer in parents and/or siblings, n (%) | 41 (22.9) | 62 (17.4) | NS | |||
Histologic classification of gastric cancer*, n (%) | ||||||
Poorly differentiated adenocarcinoma | 111 (62.0) | |||||
Well-differentiated adenocarcinoma | 68 (38.0) | |||||
Dietary intake, mean (SD) | ||||||
Seafood (g/1,000 kcal) | 36.4 (20.5) | 34.4 (19.1) | NS | |||
Fish (g/1,000 kcal) | 24.4 (15.6) | 23.3 (14.2) | NS | |||
Other seafood (g/1,000 kcal) | 12.0 (8.7) | 11.1 (8.0) | NS | |||
Green-yellow vegetables (g/1,000 kcal) | 34.2 (23.8) | 36.5 (24.5) | NS | |||
Other vegetables (g/1,000 kcal) | 33.5 (20.5) | 34.7 (21.7) | NS | |||
Fruit (g/1,000 kcal) | 38.0 (36.9) | 36.6 (36.4) | NS | |||
Green tea (times/d) | 1.46 (1.15) | 1.42 (1.17) | NS | |||
n-3 HUFAs† (g/1,000 kcal) | 0.50 (0.25) | 0.48 (0.23) | NS | |||
Sodium (mg/1,000 kcal) | 1,112 (357) | 1,042 (319) | <0.05 |
. | Case subjects (n = 179) . | Control subjects (n = 357) . | P value for χ2 or t test . | |||
---|---|---|---|---|---|---|
Age (y), mean (SD) | 59.3 (10.0) | 58.9 (10.1) | NS | |||
Body mass index (kg/m2), mean (SD) | 23.0 (3.0) | 22.9 (2.6) | NS | |||
Women, n (%) | 43 (24.0) | 86 (24.1) | NS | |||
Smoking habit, n (%) | ||||||
Smokers | 70 (39.1) | 112 (31.4) | NS | |||
Ex-smokers | 50 (27.9) | 110 (30.8) | ||||
Nonsmokers | 59 (33.0) | 135 (37.8) | ||||
Family history of gastric cancer in parents and/or siblings, n (%) | 41 (22.9) | 62 (17.4) | NS | |||
Histologic classification of gastric cancer*, n (%) | ||||||
Poorly differentiated adenocarcinoma | 111 (62.0) | |||||
Well-differentiated adenocarcinoma | 68 (38.0) | |||||
Dietary intake, mean (SD) | ||||||
Seafood (g/1,000 kcal) | 36.4 (20.5) | 34.4 (19.1) | NS | |||
Fish (g/1,000 kcal) | 24.4 (15.6) | 23.3 (14.2) | NS | |||
Other seafood (g/1,000 kcal) | 12.0 (8.7) | 11.1 (8.0) | NS | |||
Green-yellow vegetables (g/1,000 kcal) | 34.2 (23.8) | 36.5 (24.5) | NS | |||
Other vegetables (g/1,000 kcal) | 33.5 (20.5) | 34.7 (21.7) | NS | |||
Fruit (g/1,000 kcal) | 38.0 (36.9) | 36.6 (36.4) | NS | |||
Green tea (times/d) | 1.46 (1.15) | 1.42 (1.17) | NS | |||
n-3 HUFAs† (g/1,000 kcal) | 0.50 (0.25) | 0.48 (0.23) | NS | |||
Sodium (mg/1,000 kcal) | 1,112 (357) | 1,042 (319) | <0.05 |
Abbreviation: NS, not significant.
Poorly differentiated adenocarcinoma included signet-ring cell carcinoma, whereas well-differentiated adenocarcinoma included papillary adenocarcinoma and well- and moderately differentiated types of tubular adenocarcinoma.
n-3 HUFAs were summed up as eicosapentaenoic acid, docosapentaenoic acid, and docosahexaenoic acid.
Percentage fatty acid compositions in erythrocyte membranes in case and control subjects
Fatty acids and the indices . | Case subjects (n = 179), mean (SD) . | Control subjects (n = 357), mean (SD) . | P value for t test . |
---|---|---|---|
SFAs | 55.1 (2.9) | 54.0 (3.0) | 0.0001 |
Palmitic acid | 32.9 (2.4) | 31.7 (2.3) | <0.0001 |
MUFAs | 19.2 (1.6) | 19.0 (2.1) | NS |
PUFAs | 25.8 (3.6) | 27.1 (3.7) | 0.0001 |
n-6 PUFAs | 19.3 (2.7) | 19.7 (2.1) | <0.05 |
Arachidonic acid | 8.4 (1.7) | 8.7 (1.6) | NS |
n-3 PUFAs | 6.5 (2.2) | 7.3 (2.4) | <0.0001 |
n-3 HUFAs | 6.2 (2.2) | 7.0 (2.3) | <0.0005 |
EPA | 1.2 (0.6) | 1.3 (0.6) | <0.05 |
Docosapentaenoic acid | 0.9 (0.4) | 1.1 (0.4) | <0.0001 |
DHA | 4.1 (1.4) | 4.6 (1.5) | 0.0001 |
SFAs/PUFAs | 2.20 (0.44) | 2.05 (0.39) | <0.0001 |
SFAs/n-3 HUFAs | 10.69 (6.45) | 8.90 (3.80) | <0.001 |
n-6 PUFAs/n-3 PUFAs | 3.42 (1.64) | 2.99 (1.01) | <0.005 |
n-6 PUFAs/n-3 HUFAs | 3.62 (1.77) | 3.17 (1.14) | <0.005 |
Arachidonic acid/DHA | 2.31 (0.88) | 2.06 (0.67) | <0.001 |
Saturation index n-7 | 35.20 (8.20) | 29.23 (10.73) | <0.0001 |
Saturation index n-9 | 1.18 (0.15) | 1.21 (0.15) | <0.05 |
Fatty acids and the indices . | Case subjects (n = 179), mean (SD) . | Control subjects (n = 357), mean (SD) . | P value for t test . |
---|---|---|---|
SFAs | 55.1 (2.9) | 54.0 (3.0) | 0.0001 |
Palmitic acid | 32.9 (2.4) | 31.7 (2.3) | <0.0001 |
MUFAs | 19.2 (1.6) | 19.0 (2.1) | NS |
PUFAs | 25.8 (3.6) | 27.1 (3.7) | 0.0001 |
n-6 PUFAs | 19.3 (2.7) | 19.7 (2.1) | <0.05 |
Arachidonic acid | 8.4 (1.7) | 8.7 (1.6) | NS |
n-3 PUFAs | 6.5 (2.2) | 7.3 (2.4) | <0.0001 |
n-3 HUFAs | 6.2 (2.2) | 7.0 (2.3) | <0.0005 |
EPA | 1.2 (0.6) | 1.3 (0.6) | <0.05 |
Docosapentaenoic acid | 0.9 (0.4) | 1.1 (0.4) | <0.0001 |
DHA | 4.1 (1.4) | 4.6 (1.5) | 0.0001 |
SFAs/PUFAs | 2.20 (0.44) | 2.05 (0.39) | <0.0001 |
SFAs/n-3 HUFAs | 10.69 (6.45) | 8.90 (3.80) | <0.001 |
n-6 PUFAs/n-3 PUFAs | 3.42 (1.64) | 2.99 (1.01) | <0.005 |
n-6 PUFAs/n-3 HUFAs | 3.62 (1.77) | 3.17 (1.14) | <0.005 |
Arachidonic acid/DHA | 2.31 (0.88) | 2.06 (0.67) | <0.001 |
Saturation index n-7 | 35.20 (8.20) | 29.23 (10.73) | <0.0001 |
Saturation index n-9 | 1.18 (0.15) | 1.21 (0.15) | <0.05 |
NOTE: SFAs were summed up as myristic acid, palmitic acid, and stearic acid. MUFAs were summed up as palmitoleic acid and oleic acid. PUFAs were summed up as the following: n-6 PUFAs and n-3 PUFAs. n-6 PUFAs were summed up as linoleic acid, γ-linolenic acid, dihomo-γ-linolenic acid, and arachidonic acid. n-3 PUFAs were summed up as α-linolenic acid and n-3 HUFAs. n-3 HUFAs were summed up as EPA, docosapentaenoic acid, and DHA. Saturation indices n-7 and n-9 were meant as the ratios of palmitic acid/palmitoleic acid and stearic acid/oleic acid, respectively.
Abbreviation: NS, not significant.
In control subjects, dietary fish intake was positively correlated with erythrocyte compositions of EPA [partial Spearman's correlation coefficients (r) = 0.28; P < 0.0001] and n-3 HUFAs (r = 0.15; P < 0.005). Dietary n-3 HUFAs intake was also positively correlated with erythrocyte values of EPA (r = 0.27; P < 0.0001), DHA (r = 0.10; P = 0.05) and n-3 HUFAs (r = 0.16; P < 0.005). Dietary intake of sodium was positively correlated with seafood (r = 0.45; P < 0.0001), fish (r = 0.41; P < 0.0001), and other seafood (r = 0.37; P < 0.0001) consumption, but with no relation to erythrocyte levels of EPA, DHA, and n-3 HUFAs. Sodium intake was related to increased risk of gastric cancer: the OR for the highest to the lowest tertile, 1.97; 95% CI, 1.19-3.25; Ptrend < 0.01 for both histologic types; OR, 2.02; 95% CI, 1.07-3.80; Ptrend < 0.05 for poorly differentiated adenocarcinoma; and OR, 2.37; 95% CI, 0.97-5.79; Ptrend = 0.06 for well-differentiated type. Dietary intakes of green-yellow vegetables, other vegetables, fruit and green tea consumption were positively correlated with that of vitamin C as a possible protective factor for gastric cancer (at least r > 0.59; P < 0.0001 for all). Dietary intake of green-yellow vegetable was also associated with a decreased risk of the poorly differentiated type gastric cancer (OR for the highest tertile, 0.42; 95% CI, 0.19-0.85; Ptrend < 0.05). However, no associations were found between gastric cancer risk and dietary intakes of other foods and nutrients.
Table 3 shows the ORs for gastric cancer according to erythrocyte fatty acid compositions among all cases (n = 179) and all the corresponding controls (n = 357). The risk was positively associated with erythrocyte compositions of SFAs (the highest to the lowest tertile, OR, 2.33; 95% CIs, 1.39-3.90; Ptrend < 0.005) and palmitic acid (OR, 3.14; 95% CI, 1.77-5.70; Ptrend < 0.001) and negatively with that of PUFAs (OR, 0.47; 95% CI, 0.28-0.80; Ptrend < 0.01). In particular, decreased risk was linked to the DHA value (OR, 0.47; 95% CI, 0.28-0.79; Ptrend < 0.01) and n-3 HUFAs (OR, 0.39; 95% CI, 0.23-0.68; Ptrend < 0.005). Increased risk was strongly related to high specific ratios of SFAs/n-3 HUFAs (OR, 3.10; 95% CI, 1.78-5.39; Ptrend < 0.001) and arachidonic acid/DHA (OR, 2.38; 95% CIs, 1.34-4.20; Ptrend < 0.05). Risk was positively associated with saturation indexn-7 (OR, 12.07; 95% CI, 5.85-24.94; Ptrend < 0.0001), but negatively with saturation indexn-9 (OR, 0.59; 95% CI, 0.36-0.96; Ptrend < 0.05).
ORs for gastric cancer and the 95% CIs according to tertile of fatty acid compositions in erythrocyte membranes among all cases (n = 179) and the corresponding controls (n = 357)
ORs (95% CI)* . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|
Fatty acids and the indices . | T1 (reference) . | T2 . | T3 . | Ptrend . | ||||
SFAs | <52.45 | 52.45-55.76 | >55.76 | |||||
1.00 | 1.86 (1.10-3.15) | 2.33 (1.39-3.90) | <0.005 | |||||
Palmitic acid | <30.55 | 30.55-32.88 | >32.88 | |||||
1.00 | 3.46 (1.95-6.14) | 3.14 (1.77-5.70) | <0.001 | |||||
MUFAs | <18.24 | 18.24-19.58 | >19.58 | |||||
1.00 | 0.90 (0.55-1.48) | 1.19 (0.74-1.91) | NS | |||||
PUFAs | <24.90 | 24.90-28.99 | >28.99 | |||||
1.00 | 0.94 (0.61-1.45) | 0.47 (0.28-0.80) | <0.01 | |||||
n-6 PUFAs | <18.74 | 18.74-20.66 | >20.66 | |||||
1.00 | 0.90 (0.58-1.41) | 0.68 (0.43-1.10) | NS | |||||
Arachidonic acid | <7.83 | 7.83-9.10 | >9.10 | |||||
1.00 | 0.65 (0.40-1.04) | 0.79 (0.50-1.26) | NS | |||||
n-3 PUFAs | <5.92 | 5.92-8.29 | >8.29 | |||||
1.00 | 0.89 (0.58-1.37) | 0.37 (0.21-0.64) | <0.001 | |||||
n-3 HUFAs | <5.61 | 5.61-7.98 | >7.98 | |||||
1.00 | 0.95 (0.62-1.47) | 0.39 (0.23-0.68) | <0.005 | |||||
EPA | <0.98 | 0.98-1.47 | >1.47 | |||||
1.00 | 0.63 (0.39-1.00) | 0.64 (0.40-1.03) | NS | |||||
Docosapentaenoic acid | <0.85 | 0.85-1.29 | >1.29 | |||||
1.00 | 1.12 (0.74-1.69) | 0.32 (0.18-0.56) | 0.0005 | |||||
DHA | <3.70 | 3.70-5.23 | >5.23 | |||||
1.00 | 0.90 (0.59-1.39) | 0.47 (0.28-0.79) | <0.01 | |||||
SFAs/PUFAs | <1.80 | 1.80-2.24 | >2.24 | |||||
1.00 | 1.99 (1.19-3.33) | 2.12 (1.27-3.54) | <0.05 | |||||
SFAs/n-3 HUFAs | <6.30 | 6.30-9.20 | >9.20 | |||||
1.00 | 2.63 (1.54-4.50) | 3.10 (1.78-5.39) | <0.001 | |||||
n-6 PUFAs/n-3 PUFAs | <2.44 | 2.44-3.32 | >3.32 | |||||
1.00 | 2.48 (1.32-3.81) | 2.15 (1.25-3.70) | <0.05 | |||||
n-6 PUFAs/n-3 HUFAs | <2.55 | 2.55-3.51 | >3.51 | |||||
1.00 | 1.74 (1.04-2.94) | 1.87 (1.10-3.19) | <0.05 | |||||
Arachidonic acid/DHA | <1.68 | 1.68-2.24 | >2.24 | |||||
1.00 | 2.52 (1.47-4.34) | 2.38 (1.34-4.20) | <0.05 | |||||
Saturation index n-7 | <23.74 | 23.74-30.89 | >30.89 | |||||
1.00 | 3.09 (1.52-6.31) | 12.07 (5.85-24.94) | <0.0001 | |||||
Saturation index n-9 | <1.15 | 1.15-1.25 | >1.25 | |||||
1.00 | 0.40 (0.24-0.65) | 0.59 (0.36-0.96) | <0.05 |
ORs (95% CI)* . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|
Fatty acids and the indices . | T1 (reference) . | T2 . | T3 . | Ptrend . | ||||
SFAs | <52.45 | 52.45-55.76 | >55.76 | |||||
1.00 | 1.86 (1.10-3.15) | 2.33 (1.39-3.90) | <0.005 | |||||
Palmitic acid | <30.55 | 30.55-32.88 | >32.88 | |||||
1.00 | 3.46 (1.95-6.14) | 3.14 (1.77-5.70) | <0.001 | |||||
MUFAs | <18.24 | 18.24-19.58 | >19.58 | |||||
1.00 | 0.90 (0.55-1.48) | 1.19 (0.74-1.91) | NS | |||||
PUFAs | <24.90 | 24.90-28.99 | >28.99 | |||||
1.00 | 0.94 (0.61-1.45) | 0.47 (0.28-0.80) | <0.01 | |||||
n-6 PUFAs | <18.74 | 18.74-20.66 | >20.66 | |||||
1.00 | 0.90 (0.58-1.41) | 0.68 (0.43-1.10) | NS | |||||
Arachidonic acid | <7.83 | 7.83-9.10 | >9.10 | |||||
1.00 | 0.65 (0.40-1.04) | 0.79 (0.50-1.26) | NS | |||||
n-3 PUFAs | <5.92 | 5.92-8.29 | >8.29 | |||||
1.00 | 0.89 (0.58-1.37) | 0.37 (0.21-0.64) | <0.001 | |||||
n-3 HUFAs | <5.61 | 5.61-7.98 | >7.98 | |||||
1.00 | 0.95 (0.62-1.47) | 0.39 (0.23-0.68) | <0.005 | |||||
EPA | <0.98 | 0.98-1.47 | >1.47 | |||||
1.00 | 0.63 (0.39-1.00) | 0.64 (0.40-1.03) | NS | |||||
Docosapentaenoic acid | <0.85 | 0.85-1.29 | >1.29 | |||||
1.00 | 1.12 (0.74-1.69) | 0.32 (0.18-0.56) | 0.0005 | |||||
DHA | <3.70 | 3.70-5.23 | >5.23 | |||||
1.00 | 0.90 (0.59-1.39) | 0.47 (0.28-0.79) | <0.01 | |||||
SFAs/PUFAs | <1.80 | 1.80-2.24 | >2.24 | |||||
1.00 | 1.99 (1.19-3.33) | 2.12 (1.27-3.54) | <0.05 | |||||
SFAs/n-3 HUFAs | <6.30 | 6.30-9.20 | >9.20 | |||||
1.00 | 2.63 (1.54-4.50) | 3.10 (1.78-5.39) | <0.001 | |||||
n-6 PUFAs/n-3 PUFAs | <2.44 | 2.44-3.32 | >3.32 | |||||
1.00 | 2.48 (1.32-3.81) | 2.15 (1.25-3.70) | <0.05 | |||||
n-6 PUFAs/n-3 HUFAs | <2.55 | 2.55-3.51 | >3.51 | |||||
1.00 | 1.74 (1.04-2.94) | 1.87 (1.10-3.19) | <0.05 | |||||
Arachidonic acid/DHA | <1.68 | 1.68-2.24 | >2.24 | |||||
1.00 | 2.52 (1.47-4.34) | 2.38 (1.34-4.20) | <0.05 | |||||
Saturation index n-7 | <23.74 | 23.74-30.89 | >30.89 | |||||
1.00 | 3.09 (1.52-6.31) | 12.07 (5.85-24.94) | <0.0001 | |||||
Saturation index n-9 | <1.15 | 1.15-1.25 | >1.25 | |||||
1.00 | 0.40 (0.24-0.65) | 0.59 (0.36-0.96) | <0.05 |
NOTE: SFAs were summed up as myristic acid, palmitic acid, and stearic acid. MUFAs were summed up as palmitoleic acid and oleic acid. PUFAs were summed up as the following: n-6 PUFAs and n-3 PUFAs. n-6 PUFAs were summed up as linoleic acid, γ-linolenic acid, dihomo-γ-linolenic acid, and arachidonic acid. n-3 PUFAs were summed up as α-linolenic acid and n-3 HUFAs. n-3 HUFAs were summed up as EPA, docosapentaenoic acid, and DHA. Saturation indices n-7 and n-9 were meant as the ratios of palmitic acid/palmitoleic acid and stearic acid/oleic acid, respectively.
Abbreviation: NS, not significant.
Adjusted for BMI, family history of gastric cancer, smoking habit, green-yellow vegetables intake (g/1,000 kcal), other vegetables intake (g/1,000 kcal), fruit intake (g/1,000 kcal), sodium intake (mg/1,000 kcal), and green tea consumption (times/d).
Table 4 shows ORs for gastric cancer and the 95% CIs among cases (n = 111) with poorly differentiated adenocarcinoma and the corresponding controls (matched age, sex, and season of sample collection, n = 221). The risk was positively associated with erythrocyte compositions of palmitic acid (the highest to the lowest tertile, OR, 2.82; 95% CI, 1.31-6.07; Ptrend < 0.05), but negatively with those of n-3 PUFAs, including α-linolenic acid (OR, 0.42; 95% CI, 0.20-0.86; Ptrend < 0.05). A tendency for decreased risk was observed with DHA and n-3 HUFAs, but without statistical significance. There was increased risk with a high erythrocyte ratio of SFAs/n-3 HUFAs (OR, 2.12; 95% CI, 1.05-4.09; Ptrend < 0.05), but not arachidonic acid/DHA. The risk was positively related to saturation indexn-7 (OR, 10.65; 95% CI, 4.30-26.37; Ptrend < 0.0001), but negatively to saturation indexn-9 (OR, 0.48; 95% CI, 0.24-0.96; Ptrend < 0.05).
ORs for gastric cancer and the 95% CIs according to tertile of fatty acid compositions in erythrocyte membranes among 111 cases with poorly differentiated adenocarcinoma and the corresponding 221 controls
ORs (95% CIs)* . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|
Fatty acids and the indices . | T1 (reference) . | T2 . | T3 . | Ptrend . | ||||
SFAs | <53.17 | 53.17-55.97 | >55.97 | |||||
1.00 | 0.66 (0.34-1.26) | 1.25 (0.67-2.32) | NS | |||||
Palmitic acid | <30.61 | 30.61-32.94 | >32.94 | |||||
1.00 | 3.06 (1.41-6.62) | 2.82 (1.31-6.07) | <0.05 | |||||
MUFAs | <18.31 | 18.31-19.55 | >19.55 | |||||
1.00 | 0.59 (0.32-1.12) | 0.96 (0.51-1.82) | NS | |||||
PUFAs | <24.76 | 24.76-28.49 | >28.49 | |||||
1.00 | 0.65 (0.36-1.16) | 0.57 (0.29-1.13) | NS | |||||
n-6 PUFAs | <18.78 | 18.78-20.49 | >20.49 | |||||
1.00 | 0.59 (0.32-1.07) | 0.92 (0.49-1.71) | NS | |||||
Arachidonic acid | <7.79 | 7.79-8.89 | >8.89 | |||||
1.00 | 0.50 (0.26-0.94) | 1.04 (0.57-1.88) | NS | |||||
n-3 PUFAs | <6.00 | 6.00-7.97 | >7.97 | |||||
1.00 | 0.83 (0.47-1.48) | 0.42 (0.20-0.86) | <0.05 | |||||
n-3 HUFAs | <5.68 | 5.68-7.60 | >7.60 | |||||
1.00 | 0.81 (0.45-1.45) | 0.54 (0.27-1.07) | NS | |||||
EPA | <0.98 | 0.98-1.42 | >1.42 | |||||
1.00 | 0.54 (0.29-0.99) | 0.74 (0.41-1.34) | NS | |||||
Docosapentaenoic acid | <0.81 | 0.81-1.21 | >1.21 | |||||
1.00 | 1.26 (0.75-2.11) | 0.37 (0.18-0.78) | 0.05 | |||||
DHA | <3.72 | 3.72-5.05 | >5.05 | |||||
1.00 | 0.74 (0.42-1.32) | 0.58 (0.30-1.13) | NS | |||||
SFAs/PUFAs | <1.87 | 1.87-2.28 | >2.28 | |||||
1.00 | 1.16 (0.59-2.26) | 1.48 (0.76-2.88) | NS | |||||
SFAs/n-3 HUFAs | <7.03 | 7.03-9.74 | >9.74 | |||||
1.00 | 1.54 (0.79-3.02) | 2.12 (1.05-4.09) | <0.05 | |||||
n-6 PUFAs/n-3 PUFAs | <2.50 | 2.50-3.34 | >3.34 | |||||
1.00 | 1.76 (0.90-3.47) | 1.66 (0.83-3.31) | NS | |||||
n-6 PUFAs/n-3 HUFAs | <2.63 | 2.63-3.55 | >3.55 | |||||
1.00 | 1.59 (0.81-3.11) | 1.56 (0.78-3.09) | NS | |||||
Arachidonic acid/DHA | <1.71 | 1.71-2.25 | >2.25 | |||||
1.00 | 1.63 (0.83-3.20) | 1.69 (0.85-3.36) | NS | |||||
Saturation indexn-7 | <23.93 | 23.93-31.61 | >31.61 | |||||
1.00 | 3.52 (1.46-8.53) | 10.65 (4.30-26.37) | <0.0001 | |||||
Saturation indexn-9 | <1.15 | 1.15-1.27 | >1.27 | |||||
1.00 | 0.46 (0.24-0.86) | 0.48 (0.24-0.96) | <0.05 |
ORs (95% CIs)* . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|
Fatty acids and the indices . | T1 (reference) . | T2 . | T3 . | Ptrend . | ||||
SFAs | <53.17 | 53.17-55.97 | >55.97 | |||||
1.00 | 0.66 (0.34-1.26) | 1.25 (0.67-2.32) | NS | |||||
Palmitic acid | <30.61 | 30.61-32.94 | >32.94 | |||||
1.00 | 3.06 (1.41-6.62) | 2.82 (1.31-6.07) | <0.05 | |||||
MUFAs | <18.31 | 18.31-19.55 | >19.55 | |||||
1.00 | 0.59 (0.32-1.12) | 0.96 (0.51-1.82) | NS | |||||
PUFAs | <24.76 | 24.76-28.49 | >28.49 | |||||
1.00 | 0.65 (0.36-1.16) | 0.57 (0.29-1.13) | NS | |||||
n-6 PUFAs | <18.78 | 18.78-20.49 | >20.49 | |||||
1.00 | 0.59 (0.32-1.07) | 0.92 (0.49-1.71) | NS | |||||
Arachidonic acid | <7.79 | 7.79-8.89 | >8.89 | |||||
1.00 | 0.50 (0.26-0.94) | 1.04 (0.57-1.88) | NS | |||||
n-3 PUFAs | <6.00 | 6.00-7.97 | >7.97 | |||||
1.00 | 0.83 (0.47-1.48) | 0.42 (0.20-0.86) | <0.05 | |||||
n-3 HUFAs | <5.68 | 5.68-7.60 | >7.60 | |||||
1.00 | 0.81 (0.45-1.45) | 0.54 (0.27-1.07) | NS | |||||
EPA | <0.98 | 0.98-1.42 | >1.42 | |||||
1.00 | 0.54 (0.29-0.99) | 0.74 (0.41-1.34) | NS | |||||
Docosapentaenoic acid | <0.81 | 0.81-1.21 | >1.21 | |||||
1.00 | 1.26 (0.75-2.11) | 0.37 (0.18-0.78) | 0.05 | |||||
DHA | <3.72 | 3.72-5.05 | >5.05 | |||||
1.00 | 0.74 (0.42-1.32) | 0.58 (0.30-1.13) | NS | |||||
SFAs/PUFAs | <1.87 | 1.87-2.28 | >2.28 | |||||
1.00 | 1.16 (0.59-2.26) | 1.48 (0.76-2.88) | NS | |||||
SFAs/n-3 HUFAs | <7.03 | 7.03-9.74 | >9.74 | |||||
1.00 | 1.54 (0.79-3.02) | 2.12 (1.05-4.09) | <0.05 | |||||
n-6 PUFAs/n-3 PUFAs | <2.50 | 2.50-3.34 | >3.34 | |||||
1.00 | 1.76 (0.90-3.47) | 1.66 (0.83-3.31) | NS | |||||
n-6 PUFAs/n-3 HUFAs | <2.63 | 2.63-3.55 | >3.55 | |||||
1.00 | 1.59 (0.81-3.11) | 1.56 (0.78-3.09) | NS | |||||
Arachidonic acid/DHA | <1.71 | 1.71-2.25 | >2.25 | |||||
1.00 | 1.63 (0.83-3.20) | 1.69 (0.85-3.36) | NS | |||||
Saturation indexn-7 | <23.93 | 23.93-31.61 | >31.61 | |||||
1.00 | 3.52 (1.46-8.53) | 10.65 (4.30-26.37) | <0.0001 | |||||
Saturation indexn-9 | <1.15 | 1.15-1.27 | >1.27 | |||||
1.00 | 0.46 (0.24-0.86) | 0.48 (0.24-0.96) | <0.05 |
NOTE: SFAs were summed up as myristic acid, palmitic acid and stearic acid. MUFAs were summed up as palmitoleic acid and oleic acid. PUFAs were summed up as the following: n-6 PUFAs and n-3 PUFAs. n-6 PUFAs were summed up as linoleic acid, γ-linolenic acid, dihomo-γ-linolenic acid, and arachidonic acid. n-3 PUFAs were summed up as α-linolenic acid and n-3 HUFAs. n-3 HUFAs were summed up as EPA, docosapentaenoic acid, and DHA. Saturation indices n-7 and n-9 were meant as the ratios of palmitic acid/palmitoleic acid and stearic acid/oleic acid, respectively.
Abbreviation: NS, not significant.
Adjusted for BMI, family history of gastric cancer, smoking habit, green-yellow vegetables intake (g/1,000 kcal), other vegetables intake (g/1,000 kcal), fruit intake (g/1,000 kcal), sodium intake (mg/1,000 kcal), and green tea consumption (times/d).
Table 5 shows ORs and the 95% CIs for gastric cancer cases (n = 68) with well-differentiated adenocarcinomas and the corresponding controls (n = 136). The risk was positively linked to erythrocyte compositions of palmitic acid (the highest to the lowest tertile, OR, 4.82; 95% CI, 1.86-15.51; Ptrend < 0.005) and SFAs (OR, 5.35; 95% CI, 2.18-13.11; Ptrend < 0.0001), whereas negatively with those of EPA (OR, 0.40; 95% CI, 0.17-0.98; Ptrend < 0.05), DHA (OR, 0.20; 95% CI, 0.07-0.58; Ptrend < 0.01), and n-3 HUFAs (OR, 0.10; 95% CI, 0.03-0.35; Ptrend = 0.0005,). The increased risk was associated with erythrocyte ratios of SFAs/n-3 HUFAs (OR, 7.86; 95% CI, 2.47-25.02; Ptrend < 0.005) and n-6 PUFAs/n-3 HUFAs (OR, 5.64; 95% CI, 1.88-16.93; Ptrend < 0.005). A tendency for increased risk was observed, with the ratio of arachidonic acid/DHA without statistical significance. The saturation indexn-7 was positively related to the risk (OR, 12.37; 95% CI, 3.67-41.47; Ptrend < 0.0001).
ORs for gastric cancer and the 95% CIs according to tertile of fatty acid compositions in erythrocyte membranes among 68 cancer cases with well-differentiated adenocarcinoma and the corresponding 136 controls
ORs (95% CIs)* . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|
Fatty acids and the indices . | T1 (reference) . | T2 . | T3 . | P for trend . | ||||
SFAs | <51.81 | 51.81-54.30 | >54.30 | |||||
1.00 | 1.76 (0.65-4.74) | 5.35 (2.18-13.11) | <0.0001 | |||||
Palmitic acid | <30.51 | 30.51-32.74 | >32.74 | |||||
1.00 | 3.79 (1.44-9.93) | 4.82 (1.86-15.51) | <0.005 | |||||
MUFAs | <18.01 | 18.01-19.69 | >19.69 | |||||
1.00 | 1.89 (0.77-4.63) | 2.02 (0.85-4.83) | NS | |||||
PUFAs | <25.46 | 25.46-29.62 | >29.62 | |||||
1.00 | 0.70 (0.33-1.47) | 0.26 (0.11-0.63) | <0.005 | |||||
n-6 PUFAs | <18.60 | 18.60-20.74 | >20.74 | |||||
1.00 | 1.18 (0.57-2.44) | 0.50 (0.22-1.23) | NS | |||||
Arachidonic acid | <7.99 | 7.99-9.57 | >9.57 | |||||
1.00 | 0.85 (0.37-1.98) | 0.38 (0.16-0.87) | <0.05 | |||||
n-3 PUFAs | <5.85 | 5.85-9.25 | >9.25 | |||||
1.00 | 0.85 (0.41-1.75) | 0.11 (0.03-0.36) | <0.001 | |||||
n-3 HUFAs | <5.54 | 5.54-8.86 | >8.86 | |||||
1.00 | 0.83 (0.40-1.70) | 0.10 (0.03-0.35) | 0.0005 | |||||
EPA | <0.98 | 0.98-1.57 | >1.57 | |||||
1.00 | 0.85 (0.38-1.89) | 0.40 (0.17-0.98) | <0.05 | |||||
Docosapentaenoic acid | <0.93 | 0.93-1.45 | >1.45 | |||||
1.00 | 0.80 (0.38-1.68) | 0.12 (0.04-0.39) | <0.001 | |||||
DHA | <3.66 | 3.66-5.76 | >5.76 | |||||
1.00 | 1.02 (0.50-2.11) | 0.20 (0.07-0.58) | <0.01 | |||||
SFAs/PUFAs | <1.74 | 1.74-2.15 | >2.15 | |||||
1.00 | 4.20 (1.49-11.87) | 5.57 (2.11-14.74) | <0.005 | |||||
SFAs/n-3 HUFAs | <5.62 | 5.62-9.35 | >9.35 | |||||
1.00 | 5.84 (2.06-16.53) | 7.86 (2.47-25.02) | <0.005 | |||||
n-6 PUFAs/n-3 PUFAs | <2.15 | 2.15-3.30 | >3.30 | |||||
1.00 | 4.34 (1.66-11.39) | 5.02 (1.71-14.74) | <0.05 | |||||
n-6 PUFAs/n-3 HUFAs | <2.24 | 2.24-3.46 | >3.46 | |||||
1.00 | 4.30 (1.63-11.38) | 5.64 (1.88-16.93) | <0.005 | |||||
Arachidonic acid/DHA | <1.59 | 1.59-2.22 | >2.22 | |||||
1.00 | 2.25 (0.96-5.26) | 2.27 (0.87-5.92) | NS | |||||
Saturation indexn-7 | <23.64 | 23.64-30.80 | >30.80 | |||||
1.00 | 1.78 (0.55-5.78) | 12.37 (3.67-41.47) | <0.0001 | |||||
Saturation indexn-9 | <1.15 | 1.15-1.23 | >1.23 | |||||
1.00 | 0.37 (0.16-0.87) | 0.69 (0.33-1.48) | NS |
ORs (95% CIs)* . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|
Fatty acids and the indices . | T1 (reference) . | T2 . | T3 . | P for trend . | ||||
SFAs | <51.81 | 51.81-54.30 | >54.30 | |||||
1.00 | 1.76 (0.65-4.74) | 5.35 (2.18-13.11) | <0.0001 | |||||
Palmitic acid | <30.51 | 30.51-32.74 | >32.74 | |||||
1.00 | 3.79 (1.44-9.93) | 4.82 (1.86-15.51) | <0.005 | |||||
MUFAs | <18.01 | 18.01-19.69 | >19.69 | |||||
1.00 | 1.89 (0.77-4.63) | 2.02 (0.85-4.83) | NS | |||||
PUFAs | <25.46 | 25.46-29.62 | >29.62 | |||||
1.00 | 0.70 (0.33-1.47) | 0.26 (0.11-0.63) | <0.005 | |||||
n-6 PUFAs | <18.60 | 18.60-20.74 | >20.74 | |||||
1.00 | 1.18 (0.57-2.44) | 0.50 (0.22-1.23) | NS | |||||
Arachidonic acid | <7.99 | 7.99-9.57 | >9.57 | |||||
1.00 | 0.85 (0.37-1.98) | 0.38 (0.16-0.87) | <0.05 | |||||
n-3 PUFAs | <5.85 | 5.85-9.25 | >9.25 | |||||
1.00 | 0.85 (0.41-1.75) | 0.11 (0.03-0.36) | <0.001 | |||||
n-3 HUFAs | <5.54 | 5.54-8.86 | >8.86 | |||||
1.00 | 0.83 (0.40-1.70) | 0.10 (0.03-0.35) | 0.0005 | |||||
EPA | <0.98 | 0.98-1.57 | >1.57 | |||||
1.00 | 0.85 (0.38-1.89) | 0.40 (0.17-0.98) | <0.05 | |||||
Docosapentaenoic acid | <0.93 | 0.93-1.45 | >1.45 | |||||
1.00 | 0.80 (0.38-1.68) | 0.12 (0.04-0.39) | <0.001 | |||||
DHA | <3.66 | 3.66-5.76 | >5.76 | |||||
1.00 | 1.02 (0.50-2.11) | 0.20 (0.07-0.58) | <0.01 | |||||
SFAs/PUFAs | <1.74 | 1.74-2.15 | >2.15 | |||||
1.00 | 4.20 (1.49-11.87) | 5.57 (2.11-14.74) | <0.005 | |||||
SFAs/n-3 HUFAs | <5.62 | 5.62-9.35 | >9.35 | |||||
1.00 | 5.84 (2.06-16.53) | 7.86 (2.47-25.02) | <0.005 | |||||
n-6 PUFAs/n-3 PUFAs | <2.15 | 2.15-3.30 | >3.30 | |||||
1.00 | 4.34 (1.66-11.39) | 5.02 (1.71-14.74) | <0.05 | |||||
n-6 PUFAs/n-3 HUFAs | <2.24 | 2.24-3.46 | >3.46 | |||||
1.00 | 4.30 (1.63-11.38) | 5.64 (1.88-16.93) | <0.005 | |||||
Arachidonic acid/DHA | <1.59 | 1.59-2.22 | >2.22 | |||||
1.00 | 2.25 (0.96-5.26) | 2.27 (0.87-5.92) | NS | |||||
Saturation indexn-7 | <23.64 | 23.64-30.80 | >30.80 | |||||
1.00 | 1.78 (0.55-5.78) | 12.37 (3.67-41.47) | <0.0001 | |||||
Saturation indexn-9 | <1.15 | 1.15-1.23 | >1.23 | |||||
1.00 | 0.37 (0.16-0.87) | 0.69 (0.33-1.48) | NS |
NOTE: SFAs was summed up as myristic acid, palmitic acid, and stearic acid. MUFAs was summed up as palmitoleic acid and oleic acid. PUFAs were summed up as the following: n-6 PUFAs and n-3 PUFAs. n-6 PUFAs were summed up as linoleic acid, γ-linolenic acid, dihomo-γ-linolenic acid, and arachidonic acid. n-3 PUFAs were summed up as α-linolenic acid and n-3 HUFAs. n-3 HUFAs was summed up as EPA, docosapentaenoic acid, and DHA. Saturation indices n-7 and n-9 were meant as the ratios of palmitic acid/palmitoleic acid and stearic acid/oleic acid, respectively.
Abbreviation: NS, not significant.
Adjusted for BMI, family history of gastric cancer, smoking habit, green-yellow vegetables intake (g/1,000 kcal), other vegetables intake (g/1,000 kcal), fruit intake (g/1,000 kcal), sodium intake (mg/1,000 kcal), and green tea consumption (times/d).
Discussion
In the present study, no association was found between the risk of gastric cancer and dietary intakes of fish and other seafood. However, inverse links with erythrocyte compositions of n-3 HUFAs, especially DHA, and a positive association with the arachidonic acid/DHA erythrocyte ratio were evident. We found that the decreased risk with high erythrocyte values for DHA and n-3 HUFAs was particularly strong for well-differentiated type adenocarcinomas.
Japanese traditional foods such as soy paste (miso), soy sauce, pickled vegetables, salted fish roe, and preserves rich in salt are risk factors for gastric cancer development and death (40, 41). In Japan, a combination of infection with H. pylori and high intake of salt and salty foods is strongly associated with risk (38). Eradication of H. pylori is thought to reduce the risk, but there has not yet been enough evidence in epidemiologic and clinical studies (14-17). In a gastric mucosal cell line, H. pylori infection has been reported to be linked to the up-regulation of cyclooxygenase-2 mRNA expression and high levels of prostaglandin E2 (42). A recent study has suggested that interactions between CagA (which is encoded by the cytokine-associated gene A in H. pylori and is linked to peptic ulcer disease and gastric carcinoma), SHP2 (a tyrosine phosphatase as an intracellular target of CagA protein), partitioning-defective 1 (PAR1), and microtubule affinity-regulating kinase (MARK) are related to the disruption of gastric epithelial cell polarity caused by mucosal damage, inflammation, and carcinogenesis in the stomach (43). There is a critical issue concerning the administration of NSAIDs to patients with peptic ulcers due to serious adverse effects such as upper gastrointestinal bleeding (10, 11).
In Japan, no negative association has been observed between gastric cancer risk and fish intake in standard epidemiologic studies. In our diet, non-salted fish is usually consumed along with salt (and our traditional condiment, i.e., soy sauce rich in salt), which is a risk factor of gastric cancer. One possible complication is the existence of carcinogenic nitrosamine precursors in salted, but not non-salted fish (39). In our previous study among women, however, an inverse risk tendency was shown for cooked fish (rich in n-3 HUFAs, EPA, and DHA; ref. 39). EPA and DHA are suggested to play critical roles in competitive inhibition against tumor development through effects on arachinoid acid metabolism (31). Moreover, DHA has been shown to be linked to up-regulation of Bax expression and apoptosis induction according to the increment of Bax/Bcl-2 ratio (19). In the present study, we found not only a positive association between gastric cancer risk and erythrocyte ratio of arachidonic acid/DHA, but also a potent negative partial Spearman's correlation coefficient between erythrocyte values for arachidonic acid and DHA (r = −0.62, P, <0.0001, adjusted for age, BMI, and season of sample collection). Our findings are supported by a report that mucosa EPA and DHA values in tumor parts among Korean gastric cancer patients were lower than those in noncancer parts, whereas mucosa SFAs and palmitic acid levels in tumor parts were higher (23).
Here, we showed positive associations between gastric cancer risk and erythrocyte compositions of SFAs, especially palmitic acid, as our previous findings for colorectal and breast cancers (30, 31). We therefore suggest the hypothesis that palmitic acid is universally related to tumor development in the colorectum, breast, and stomach. In contrast to colorectal and breast cancers, however, saturation indexn-7 (i.e., the ratio of palmitic acid/palmitoleic acid as an index of activity of delta 9 desaturase) had a positive association with gastric cancer risk. Palmitic acid in our body is supplied via various food materials and also is biosynthesized from acetyl-CoA and degraded via β-oxidation pathway for energy production. With the reduction of substrates (e.g., palmitic acid) for β-oxidation, for instance, cellular transformation may be linked to a requirement for a constant supply of fatty acids for synthesis of new cellular membranes (44). Suppression of growth of cancer cell lines has been shown using inhibitors of enzymes involved in fatty acid synthesis. In the future, we need to clarify mechanisms of tumorigenesis related to palmitic acid.
Erythrocyte values for palmitic acid and DHA had particularly strong positive and negative relations, respectively, to the risk of well-differentiated type gastric cancer. Peroxisome proliferator–activated receptors (PPAR) plays roles in cellular differentiation and proliferation in the colon and stomach. Ligand activation results in growth inhibition and induction of apoptosis in laboratory studies, including those with gastric cancer cell lines (45, 46). A recent study showed distinctive histopathologic patterns of tumor differentiation and tumor development caused by PPARγ and PPARδ agonists, i.e., tumors in mice treated with a PPARγ agonist were predominantly ductal adenocarcinomas, whereas those treated with a PPARδ agonist were other types of carcinoma (47). In contrast to the situation with synthetic PPARγ agonists for cancer chemoprevention, few studies have been conducted regarding long-chain fatty acids as natural ligands of PPARγ (48-50). Activation of the PPARγ pathway might also be related to the regulation of hyperproliferation and apoptosis of gastric epithelial cells induced by H. pylori infection (51). We think that increased and decreased risks of well-differentiated adenocarcinomas in the stomach may be related to the modification of environmental factors for gastric epithelial cells.
In the present study, potential limitations should be considered. The sample size was not so large, and the proportion of poorly differentiated adenocarcinoma was higher in these study subjects than in the general population (52). However, we showed a negative association between gastric cancer and DHA composition in erythrocytes. To increase the statistical power, we randomly selected the control subjects, who were individually matched for age and season of sample collection with a 1:2 case-control ratio. No differences were earlier found regarding the general lifestyle, including dietary habits, between noncancer outpatients presented at the Aichi Cancer Center (i.e., control subjects) and the general population randomly selected from the electoral roll in the same area (32). Another potential source of bias is the medical background of control subjects, but we had already clarified that the majority did not have any specific medical conditions (30, 31). We excluded control subjects with any diseases related to fat and/or lipid metabolism. We therefore conclude that the use of noncancer outpatients as references for the present study was acceptable because it was reasonable to assume that our case subjects arose within this population base. In the present study, we could not show any relationships between gastric cancer risk and dietary intakes of seafood, fish, and DHA. We did not consider CagA and pepsinogen status linked to H. pylori infection because the infection rates have been reported to be 93.5% for gastric cancer patients and 75.0% for general population controls in Japan (26). In Japan, NSAIDs users are very few.
Moreover, this was a retrospective study. In the present study, we could not clarify causal relationships between gastric tumorigenesis and erythrocyte compositions of fatty acids due to the effects of the disease status on both nutrient intake and the metabolism. Fatty acid values in erythrocyte membranes are appropriate indices for evaluating dietary intakes of fish, fat, and fatty acids at least 3 months without recall bias regarding dietary intake, because erythrocytes have a half-life of 120 days. We did not use blinded quality controls during the period of fatty acid measurements, but those responsible for measurements were completely blinded to any information on study subjects. Compared with other studies, the values for SFAs and n-3 HUFAs were higher and lower, respectively, due to our methods, but this does not present a problem for the comparisons within our study series (30, 31). Likewise, plasma levels of fatty acids are available to use as biomarkers for intakes, but they are less useful as a measure of middle- to long-term intake. Plasma samples should be also collected under fasted conditions (27).
In conclusion, here we showed that the risk of gastric cancer was negatively linked to the erythrocyte composition of DHA and positively linked to that of palmitic acid. Increased risk was also related to high erythrocyte ratios of arachidonic acid/DHA and SFAs/n-3 HUFAs. Using a newly established method for measuring fatty acid values in erythrocyte membranes, we could show potent associations with the risk for the type of well-differentiated adenocarcinoma relative to that of poorly differentiated adenocarcinoma. Further studies, however, are needed to investigate mechanisms linked to decreased and increased risks of gastric cancer, especially with well-differentiated adenocarcinoma.
Grant support: Grant-in-Aid for Scientific Research on Priority Areas for Cancer (grant 12218242 and 17015052) from the Ministry of Education, Science, Sports, Culture and Technology of Japan, a Grant-in-Aid for a Research Fellow of the Japan Society for the Promotion of Science, and a Grant-in-Aid for the Third Term Comprehensive 10-Year-Strategy for Cancer Control. K. Kuriki was supported by a Research Resident Fellowship from the Foundation for Promotion of Cancer Research (Japan) for the Third Term Comprehensive 10-Year-Strategy for Cancer Control during the performance of much of this work.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Application for a patent for the method to quantitatively measure fatty acids in biomaterials has been made (Japanese Patent applied 2005-080461).
Acknowledgments
The authors are grateful to the following research staffs. Dr. M.A. Moore contributed to the drafting and revising of the manuscript and checking the English language. S. Inui, T. Saito, M. Watanabe, and T. Sato took part in the laboratory assays. H. Fujikura, K. Faukaya, M. Sato, and C. Yoshida took part in interviewing subjects and data collection using the questionnaire. K. Suganuma, T. Nishiwaki, C. Kanto, R. Saito, Y. Ishida, and S. Irikura took part in interviewing subjects and blood collection. Y. Yamauchi, H. Hasegawa, and S. Hiraiwa took part in inputting and cleaning the data.