| |
Home use of vegetable oils, markers of systemic inflammation, and endothelial dysfunction among women
|
| |
| |
American Journal of Clinical Nutrition, Vol. 88, No. 4, 913-921, October 2008
Ahmad Esmaillzadeh and Leila Azadbakht
1 From the Department of Nutrition, School of Public Health, and the Food Security and Nutrition Research Center, Isfahan University of Medical Sciences, Isfahan, Iran
2 Supported by the National Nutrition and Food Technology Research Institute of the Islamic Republic of Iran (contract number P. 25/47/2337).
"In conclusion, our findings indicated that higher intakes of PHVOs (partially hydrogenated vegetable oils) were associated with elevated concentrations of inflammatory biomarkers, whereas higher intakes of non-HVOs were associated with lower plasma concentrations of these biomarkers. These findings suggest a potential mechanism through which consumption of PHVOs and non-HVOs might affect insulin resistance, metabolic syndrome, coronary artery disease, and diabetes."
ABSTRACT
Background: Most knowledge about adverse health effects of trans fats was mainly derived from studies done in Western populations of European or American origins; few data are available in the understudied region of the Middle East.
Objective: We assessed the association between consumption of partially hydrogenated vegetable oils (PHVOs) and non-HVOs and circulating concentrations of inflammatory markers among Tehrani women aged 40-60 y.
Design: Usual dietary intakes were assessed with a food-frequency questionnaire among 486 apparently healthy women. PHVOs (commonly used for cooking in Iran) were considered as PHVOs category. Sunflower oil, corn oil, canola oil, soybean oil, and olive oil were defined as non-HVOs. Anthropometric measurements were done, and fasting blood samples were taken to measure inflammatory markers.
Results: The energy-adjusted daily intakes (mean ± SD) of PHVOs and non-HVOs were 23 ± 11 and 22 ± 10 g/d, respectively. After control for potential confounders, women in the highest quintile of PHVO intake had higher plasma concentrations of C-reactive protein (CRP; percentage difference from lowest quintile: 45%; P for trend: <0.01), tumor necrosis factor-a (TNF-a; 66%; P for trend: <0.01), interleukin-6 (72%; P for trend: <0.05), and soluble intercellular adhesion molecule-1 (sICAM-1; 22%; P for trend: <0.01) than did women in the lowest quintile. In contrast, higher consumption of non-HVOs was associated with lower circulating concentrations of CRP (percentage difference between top and bottom quintiles: -23%; P for trend: 0.05), TNF-a (-29%; P for trend: <0.01), serum amyloid A (-24%; P for trend: <0.01), and sICAM-1 (-19%; P for trend:<0.05). Adjustment for body mass index, fasting plasma glucose, and lipid profiles slightly attenuated the associations in some cases.
Conclusions: Higher intakes of PHVOs are associated with elevated concentrations of inflammatory biomarkers, whereas higher intakes of non-HVOs are associated with lower plasma concentrations of these biomarkers.
INTRODUCTION
Systemic inflammation was recently reported to be involved in the incidence of atherosclerosis (1), coronary heart disease (2), diabetes (3), and the metabolic syndrome (4). Elevated concentrations of C-reactive protein (CRP), interleukin-6 (IL-6), and interleukin-18 were postulated of particular importance in the pathogenesis of type 2 diabetes among women (3, 5). Some investigators have introduced inflammation as a possible mediating mechanism through which dietary intakes of trans fatty acids (TFAs) affect chronic disease risk (6, 7). Higher intakes of TFAs were shown to be associated with tumor necrosis factor (TNF) system activity among healthy (8) and overweight (9) women and also among patients with established heart disease (10). TFAs were associated with higher concentrations of CRP, soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble intercellular adhesion molecule-1 (sICAM-1), and E-selectin among overweight women (9) and with higher concentrations of IL-6 among patients with established heart disease (10). Such detrimental effects of TFAs were also confirmed in feeding trials (11).
TFAs exist in large amounts in partially hydrogenated vegetable oils (PHVOs). Dietary habits of Iranians provide a unique opportunity to identify adverse health effects of PHVOs, because the average per-person intake of PHVOs among Iranians is about 14 g/1000 kcal (12) with TFAs accounting for 33% of total fatty acids in these products. Previous investigations have shown that Iranians consume twice as much TFAs as the US population (12). Furthermore, most knowledge about adverse health effects of TFAs was mainly derived from studies done in Western populations of European or American origins, and few data are available in the understudied region of the Middle East. We have recently found that higher intakes of HVOs are associated with greater risk of insulin resistance and the metabolic syndrome among Iranian women (13). In the current study, we aimed to identify the association of PHVOs and nonhydrogenated vegetable oils (non-HVOs) intakes with circulating concentrations of markers of systemic inflammation and endothelial dysfunction.
DISCUSSION
Our findings suggest a significant association between consumption of PHVOs and non-HVOs and circulating concentrations of inflammatory markers among women in Iran. The associations we observed persisted in multivariate models accounting for known potential confounders. Furthermore, we controlled for BMI, fasting plasma glucose, and lipid profiles, which may be mediators as well as confounders, in our analysis. Most associations remained significant even after this tight control. Although the intake and potential adverse health effects of TFAs in Western countries have received considerable attention, little is known about intake of PHVOs, a rich source of TFAs, in Middle Eastern countries (12). To our knowledge, this study is among the first investigations in which consumption of PHVOs and non-HVOs was directly and separately related to systemic inflammation in a Middle Eastern country.
The main source of dietary fat intake among Iranians is PHVOs (18), such that >75% of total vegetable oils consumed in Iran are partially hydrogenated oils. The TFA content of PHVOs used in Iranian households is 25-35% (12). The TFA content rises to 50% in food industries (19). Higher intakes of PHVOs were shown to increase the risk of atherosclerosis, cardiovascular disease, metabolic syndrome, insulin resistance, and diabetes (13, 20-22). Findings of the current study suggest the possible mechanisms through which these kinds of fats can elevate the risk of these chronic diseases because inflammatory processes are central in the pathogenesis of these chronic diseases (23, 24).
Although limited data are available that directly relate the intake of PHVOs to inflammation, some investigators have reported a positive association between TFA intake and markers of systemic inflammation. High intakes of TFAs were associated with increased TNF system activity not only among healthy women (8) but also among overweight women (9) and among patients with established heart disease (10). TFA intake was also associated with higher concentrations of CRP, soluble adhesion molecules (sICAM-1 and sVCAM-1), and E-selectin among overweight women (9) and with higher IL-6 concentrations among patients with established heart disease (10). Others have suggested that the association of TFA intake with CRP and IL-6 concentrations is an obesity-dependent process (6). Besides observational studies, detrimental effects of TFA intake were shown by clinical trials. Consumption of a diet with 6.7% of energy from TFAs increased TNF production by cultured mononuclear cells compared with a diet with 0.6% of energy from TFAs (11). In another controlled feeding trial, a higher percentage of energy from TFAs increased plasma concentrations of CRP, IL-6, and E-selectin (7). A recent in vitro study showed that incorporating a greater extent (2-fold) of TFA in the phospholipid fraction of endothelial cells significantly increased the expression of endothelial adhesion molecules, including ICAM-1 (25). However, it seems that the effects of trans fats on systemic inflammation are isomer specific (26). Our findings are in line with these studies indicating the unfavorable positive association between PHVO (rich source of trans fats) intake and markers of systemic inflammation. This may also explain why Western dietary patterns (greatly loaded with dietary sources of TFAs) in this population were associated with elevated concentrations of inflammatory markers (14). Given the pivotal role of inflammation in the development of many chronic diseases, reducing PHVOs would help to curb the epidemic of noncommunicable diseases.
We found that higher intakes of non-HVOs are independently associated with lower plasma concentrations of some inflammatory markers. Higher intakes of n-6 fatty acids were suggested to be harmful to human health because of being the precursors of proinflammatory eicosanoids (27-29). Therefore, the ratio of n-6 to n-3 was suggested. However, other investigators argue that in humans higher intakes of n-6 fatty acids were not associated with elevated concentrations of inflammatory markers (30). They think that reductions of n-6 fatty acids to improve the ratio of n-6 to n-3 would probably increase rates of cardiovascular disease and diabetes (30). A recent report has shown that rapid declines in the rate of mortality from coronary heart disease in Eastern Europe are associated with increased consumption of a-linolenic acid (31). Published clinical trials showed beneficial effects of polyunsaturated fats on systemic inflammation. A diet high in polyunsaturated fats decreased CRP, IL-6, sVCAM-1, and E-selectin compared with the typical American diet among patients with hypercholesterolemia (32, 33). In a cell-culture study, polyunsaturated fats did not alter TNF-a-mediated up-regulation of the adhesion molecules ICAM-1, VCAM-1, and E-selectin (34). Putting all this information together, it seems that increased intakes of dietary polyunsaturated fats elicit antiinflammatory effects rather than inducing systemic inflammation. The view that all n-6 PUFAs are proinflammatory requires revision, in part, and their essential regulatory and developmental role in the immune system warrants appreciation.
Several potential mechanisms were proposed by which dietary fatty acids might affect systemic inflammation. It seems that saturated fats affect systemic inflammation through inducing nuclear factor kB activation and expression of Cox-2 through Toll-like receptor 4 (35). Unsaturated fatty acids were reported to inhibit SFA-induced nuclear factor kB activation and Cox-2 expression (35). Trans fats might affect systemic inflammation through incorporating into endothelial cell membranes (36), activating cell-specific pathways (37), and affecting macrophage membrane phospholipids (38). Further investigations are required to confirm the potential mechanisms of effect of dietary fatty acids on systemic inflammation.
Several limitations need to be considered in the interpretation of our findings. The main limitation of our study is its cross-sectional nature. Thus, the association between types of vegetable oil consumption and inflammatory biomarkers remains to be confirmed in prospective analyses. Our estimates were confined to in-home consumption of PHVOs and non-HVOs, which probably underestimates total intake in the population. We cannot generalize our findings to all Iranian populations, because teachers in our community have a higher socioeconomic status than do the general Iranian population. However, it should be noted that participants in our study were selected from 4 large socioeconomically diverse districts of Tehran, covering a broad range of dietary habits. We only included women aged 40-60 y that further decreases the ability to extrapolate our findings to the whole population. Although we controlled for several lifestyle factors associated with intakes of PHVOs and non-HVOs, residual confounding because of unknown confounding factors cannot be excluded. Because of the lack of dietary data on individual fatty acids in the Iranian food composition table, we were unable to compare dietary intakes of SFAs, TFAs, monounsaturated fatty acids, and PUFAs across quintiles of PHVOs and non-HVOs. Such data can help to look at the effects of these fatty acids on inflammation.
In conclusion, our findings indicated that higher intakes of PHVOs were associated with elevated concentrations of inflammatory biomarkers, whereas higher intakes of non-HVOs were associated with lower plasma concentrations of these biomarkers. These findings suggest a potential mechanism through which consumption of PHVOs and non-HVOs might affect insulin resistance, metabolic syndrome, coronary artery disease, and diabetes.
RESULTS
The energy-adjusted daily intakes (mean ± SD) of PHVOs and non-HVOs were 23 ± 11 and 22 ± 10 g/d, respectively. Characteristics of the study participants across quintile categories of PHVOs and non-HVOs are shown in Table 1. Participants in the upper quintile of HVOs were older, had higher BMI, and waist-to-hip ratio, and were more likely to be postmenopausal, whereas participants in the upper quintile of non-HVOs were younger and more likely to have a family history of diabetes than participants in the lowest quintile. No significant difference was observed about the distribution of current smokers and obese participants across quintile categories of either PHVOs or non-HVOs.
Dietary intakes of participants across quintiles of PHVOs and non-HVOs are provided in Table 2. Participants in the upper category of PHVOs had higher intakes of cholesterol, whereas participants in the top quintile of non-HVOs had lower intakes of energy. Other nutrient intakes were not significantly different across quintile categories of either PHVOs or non-HVOs. Consumption of PHVOs was associated with higher intakes of high-fat dairy and with lower intakes of low-fat dairy products, whereas non-HVOs intake was associated with higher intakes of vegetables and low-fat dairy and with lower intakes of high-fat dairy products.
Multivariate-adjusted geometric means of circulating concentrations of inflammatory markers across quintiles of PHVOs and non-HVOs intakes are shown in Table 3. After control for potential confounders (model 2), participants in the highest quintile of PHVOs intake had higher plasma concentrations of CRP (percentage difference from lowest quintile: 45%; P for trend: <0.01), TNF-a (66%; P for trend: <0.01),and IL-6 (72%; P for trend: <0.05) than did participants in the lowest quintile. Higher consumption of non-HVOs was associated with lower circulating concentrations of CRP (percentage difference between top and bottom quintiles: -23%; P for trend: 0.05), TNF-a (-29%; P for trend: <0.01), and serum amyloid A (-24%; P for trend: <0.01). Adjustment for BMI, fasting plasma glucose, and lipid profiles in the last model slightly attenuated the associations in some cases.
Consumption of PHVOs was associated with higher plasma concentrations of sICAM-1, either before (percentage difference between top and bottom quintiles in crude model: 28%; P for trend: <0.01) or after adjustment for confounders (percentage difference between top and bottom quintiles in model 2: 22%; P for trend: <0.01) (Table 4). In contrast, consumption of non-HVOs was related to lower circulating concentrations of sICAM-1 (percentage difference between top and bottom quintiles in model 2: -19%; P for trend: <0.05). Additional control for BMI, fasting plasma glucose, and lipid profiles did not affect the associations. The same findings were reached with the use of multiple linear regression analysis (data not shown).
SUBJECTS AND METHODS
Participants
The current cross-sectional study was done in the framework of a project approved by the research council of the National Nutrition and Food Technology Research Institute, Shaheed Beheshti University of Medical Sciences. The aim of the whole project was to identify the main dietary patterns among Tehrani women and to assess the association of these dietary patterns with metabolic syndrome and inflammation (14, 15). The project was performed among a representative sample of Tehrani female teachers aged 40-60 y selected by a multistage cluster random sampling method. From the 521 women who agreed to participate (response rate: 89%), we excluded women with a prior history of cardiovascular disease, diabetes, cancer, and stroke because of possible changes in diet. We also excluded women who left >70 items blank on the food-frequency questionnaire (FFQ), reported a total daily energy intake outside the range of 800-4200 kcal, and were taking medications (propranolol, lovastatin, furosemide, metformin) that would affect serum lipoprotein, blood pressure, and carbohydrate metabolism. This left 486 women for the present analysis. Informed written consent was obtained from each participant.
Assessment of dietary intake
As described previously (14, 15), usual dietary intakes were assessed with the use of a Willett-format 168-item semiquantitative FFQ. All the questionnaires were administered by a trained dietitian. The FFQ consisted of a list of foods with a standard serving size commonly consumed by Iranians. Participants were asked to report their frequency of consumption of a given serving of each food item during the previous year on a daily (eg, bread), weekly (eg, rice, meat), or monthly (eg, fish) basis. The reported frequency for each food item was then converted to a daily intake. Portion sizes of consumed foods were converted to grams with the use of household measures. Total energy intake was calculated by summing energy intakes from all foods. Unfortunately, the Iranian food composition table lacks the food contents of individual fatty acids. Therefore, we were unable to estimate the dietary intake of saturated fatty acids (SFAs), TFAs, monounsaturated fatty acids, and polyunsaturated fatty acids (PUFAs).
PHVO (commonly used for cooking in Iran) was considered as the PHVOs category. Sunflower oil, corn oil, canola oil, soybean oil, and olive oil were defined as the non-HVOs category. A previous validation study of this FFQ showed good correlations between dietary intakes assessed by similar FFQs and multiple days of 24-h dietary recalls completed during the year (16). The correlation coefficients for the repeatability of PHVOs and non-HVOs were 0.59 and 0.69, respectively. The validity of the FFQ for assessing consumption of PHVOs and non-HVOs was also good, such that, between the FFQ and detailed dietary recalls, correlation coefficients were 0.55 for PHVOs and 0.44 for non-HVOs.
Assessment of biomarkers
Description of measurement methods for inflammatory biomarkers was provided elsewhere with greater details (14). Briefly, a fasting (>12 h) blood sample was taken and centrifuged within 30-45 min of collection, and plasma was frozen at -70 °C until analysis. Ultrasensitive latex-enhanced immunoturbidimetric assay (Randox Laboratory Ltd, Crumlin, United Kingdom) was used for measuring CRP concentrations. Circulating concentrations of serum amyloid A, E-selectin, sICAM-1, and sVCAM-1 were measured by commercially available enzyme-linked immunoabsorbent assay and standards (Biosource International, Camarillo, CA; Bender MedSystems, Vienna, Austria). TNF-a and IL-6 concentrations were measured by enzyme-linked immunoassay (Bender MedSystems). Inter- and intraassay CVs for all biomarkers were <10%. Blood lipid concentrations were assessed according to standard methods (17).
Assessment of other variables
Anthropometric measures, including weight, height, and waist circumference, were measured and body mass index (BMI; in kg/m2) was calculated (17). Data on physical activity was obtained with the use of subjects' oral responses to a pretested questionnaire and expressed as metabolic equivalent hours per week (14). Additional covariate information about age, smoking habit, socioeconomic status, family history of diabetes and stroke, menopausal status, medical history, and current use of medications was obtained with the use of pretested questionnaires. Blood pressure was also assessed according to a standard protocol.
Statistical methods
Energy-adjusted intakes of PHVOs and non-HVOs were calculated with the use of the residual method. Then participants were categorized based on quintiles of energy-adjusted intakes of PHVOs or non-HVOs. General characteristics were compared with analysis of variance with Tukey's post hoc comparisons, and qualitative variables were compared with chi-square tests. Age- and energy-adjusted means for dietary variables were compared with analysis of covariance.
The distribution of inflammatory markers was highly skewed. Therefore, logarithmically transformed values of these markers were used in all analyses. Geometric means of inflammatory markers across quintiles of PHVO or non-HVO intake were computed with analysis of covariance in 3 different models. First, we adjusted for age (in y), cigarette smoking (yes or no), physical activity (continuous), current estrogen use (yes or no), menopausal status (yes or no), socioeconomic status (categorical), family history of diabetes and stroke (yes or no), and systolic and diastolic blood pressures (continuous). Dietary intakes, including cholesterol intake; consumption of fruit, vegetables, meats and fish, whole and refined grains, high- and low-fat dairy; percentage of energy from fat; and mutual effects of PHVOs and non-HVOs (all as continuous) were taken into account in the second model. Finally, we added BMI, fasting plasma glucose, and lipid profiles (continuous), which may be mediators as well as confounders, into the model.
Multiple linear regression analysis with log-transformed plasma concentrations of inflammatory markers as dependent variables and dietary intakes of PHVOs and non-HVOs as independent variables (all as continuous) were used. We looked at the mentioned associations in 3 different models with covariates as those mentioned above. Statistical significance was set at <0.05. SPSS (version 9.05; SPSS Inc, Chicago IL) was used for all statistical analyses.
|
|
| |
| |
|
|
|
