Healthy and Unhealthy Dietary Patterns are related to Lipid Parameters in Patients with Type 2 Diabetes Mellitus

Citation: Noorshahi N, Sotoudeh G, Djalali M, Eshraghian MR, Karimi Z, et al. (2016) Healthy and Unhealthy Dietary Patterns are related to Lipid Parameters in Patients with Type 2 Diabetes Mellitus. J Nutr Health Sci 3(1): 103. doi: 10.15744/2393-9060.3.103 Volume 3 | Issue 1 Journal of Nutrition and Health Sciences


Introduction
Today diabetes is a major health problem in the world due to its high prevalence [1].Lipid abnormalities (dyslipidemia) are one of diabetes complications which led to increased prevalence of cardiovascular disease (CVD) in these patients [2].Studies have suggested that impaired glucose regardless of its severity increases the risk of cardiovascular disease [3,4].Studies have shown a strong association between hypercholesterolemia and incidence of cardiovascular diseases and its mortality rate, so that 1 mg/dl reduction in LDL-c decreased 1-2% relative risk of cardiovascular disease.Several factors affect the amount and composition of plasma lipoproteins [5,6].
Nutrition is one of the most important environmental factors in initiation and treatment of diabetes mellitus [7].A healthy diet may reduce risk of dyslipidemia and its associated disorders [8].Most studies on the relationship between diabetic diet and lipid profile have been focused on single nutrients or foods [9][10][11][12][13].Since the relationship between diet and disease is very complicated, it has been suggested that using dietary pattern analysis or whole diet study, instead of analysis of a particular nutrient.Factor analysis [14,15] could eliminate the concerns relating to confounding factors and interactions between food and nutrient and taking nutritional behaviors.It could provide more detailed information about the relationship between diet and lipid profile as a main risk factor of cardiovascular diseases in diabetic patients [2,[16][17][18][19][20][21][22][23][24].
Lipid profiles have been studied as a part of metabolic syndrome components [3,4,12,22,23,[25][26][27][28][29][30][31][32][33][34][35][36].Most of these studies are held in western countries [15,[21][22][23]26,[32][33][34]37].Some studies have revealed the association between metabolic syndrome and dietary patterns suggesting that certain dietary patterns characterized by low total and saturated fat intake and high fruit, vegetable, whole grain, and low-fat dairy intake are protective against metabolic syndrome or its components [23,25,26,38].The studies on Asian population have also shown positive effects of dietary patterns characterized by high consumption of vegetables, fruits, whole grains and fish on the reduction risk of having high cholesterol or/and triglyceride [27][28][29][30]39].On the other hands, the dietary patterns rich in sugar, sweets and refined grains increase the risk of having low HDL-c and/or high triglyceride [27,28,30,39,40].According to our knowledge there is no study which investigated the association between lipid abnormalities and dietary patterns singly in diabetic patients.According to the similarity of food items, thirty-four food groups were created (Table 1).Dietary patterns determined by exploratory factor analysis.Extracted food groups were entered into the factor analysis and a principal components method was used to determine dietary patterns.Orthogonal varimax rotation method was used to generate a simple and distinction matrix for identifying major dietary patterns, based on number of food groups.The eigenvalue and Scree plot was used to determine the number of dietary patterns.The Scree plot is a plot of the eigenvalues of derived factors.The point shift in action on Scree plot was used to determine the number of dietary patterns.The factor score for each pattern was calculated by summing intakes of food groups weighted by their factor score matrix.The derived dietary patterns were labeled on the basis of our interpretation of the data and of the earlier literature.

Dietary assessment
To determine the adequacy of the sample, Kaiser-Meyer-Olkin (KMO) test and to determine the relationship between variables, Bartlett Test of Sphericity (BTS) was utilized.KMO and BTS for this study were 0.690 and 0.001 respectively.Given the correlations between the data, food groups with factor loadings over 0•3 were considered as making a significant contribution to the pattern.Two dietary patterns which have higher percent of variance in our population were chosen.Subjects were categorized into four groups (quartiles) in terms of dietary pattern scores.Each quartile includes approximately 25% of population.
*Load factor less than 0.3 are omitted to simplify the table †Kaiser's Measure of Sampling Adequacy (MSA)= 0.690, Bartlett's test of sphericity = <0.001Table 2: Load factor of food groups for 2 major predominant dietary patterns †* Food grouping was performed according to Table 1 in order to be used in factor analysis to identify dietary patterns.Using factor analysis, two major dietary patterns in the study population were identified which are presented in Table 2. Healthy dietary pattern was high in vegetables, cabbages, tomatoes, yellow vegetables, fruits, fish, olives, beans.The second, Unhealthy food pattern was high in sugar, soft drinks, sweets and desserts, industrial juices, high-fat dairy products, butter, processed meat, eggs, refined grains and snacks.Altogether these two major dietary patterns form 15.7 % of total variance.12-14 h fasting venous blood samples were taken in order to measure serum lipids [45].Total cholesterol and triglycerides were measured by enzymatic methods using Pars Azmoon kits, Pars Azmoon Inc., Tehran, Iran.LDL-c and HDL-c levels were assayed according to lab kit instructions, in qualified lab certified ISO 9002.Lipid profiles of the patients were divided into two groups of normal or abnormal according to the Standards of Medical Care in Diabetes (Diabetes Care) of the Diabetes Association of America, 2014 [46].

Physical activity
Daily physical activity evaluated in terms of metabolic equivalent × hours per day (MET.h.day) by a physical activity questionnaire which has been developed in Europe in previous studies and its validity has been confirmed [47].Reliability and validity of this questionnaire have been confirmed by Kelishadi, et al. study in Iran [48].

Statistical Analysis
General characteristics of participants in quartiles of dietary patterns were compared using one-way analysis of variance (ANOVA) for quantitative variables and chi-square test for qualitative variables.Binary logistic regression analysis in multiple models was used to test dietary patterns association with dyslipidemia, with adjustment for different variables such as age, sex, family history of diabetes, duration of diabetes, BMI, physical activity, use of lipid-lowering drugs and energy intake.IBM SPSS Statistics version 21 was used for all steps of the analysis.All p values were based on two-sided tests and compared with significance level of 0.05.

Result
Table 3 shows some general characteristics of participants in this study based on the presence or absence of lipid disorders.The interesting results are, those with normal triglyceride levels had a lower daily energy (p<0.05) and also lower lipid lowering drugs intake (p<0.05)compared with those with hyper triglyceridemia.Individuals with normal LDL-c levels have higher incidence of CVD histories (p<0.001).However higher level of LDL-c was observed in people with family history of diabetes.Patients with normal HDL-c concentration have lower percent of alcohol consumption (p<0.05) and incidence of CVD histories (p<0.001).

Discussion
According to the study results, after cofounder adjustment, the most adherences to healthy food pattern decreases the risk of having high LDL by 55%.Also the chance of increased HDL-c in the highest adherence quartile of the pattern is 2.4 times higher than those in the lowest quartile of the group.
Our investigating on general characteristics of subjects in dietary patterns adherence quartiles shows those in the highest quartile of adherence healthy dietary pattern have higher energy intake and physical activity levels, compared to the lowest quartile (p=0.01).In contrast, highest quartile for unhealthy dietary pattern adherence had higher BMI, physical activity and energy intake compared to the lowest quartile (in all; p=0.01).They also had a lower mean age (p=0.01).Besides higher intake of alcohol was observed compared to lowest quartile (p=0.024).
The odds ratio of dyslipidemia in quartile of dietary patterns scores with adjustment for confounding variables has been shown in the two models (Table 4).In model 1 (without adjusting for variables) healthy dietary pattern adherence was significantly associated with low levels of LDL-c and high HDL-c levels.After adjustment for confounding factors, the association between the healthy dietary pattern and LDL-c and HDL-c level still remained significant.Comparing to the first quartile of healthy food patterns, individuals with highest score had lower odds ratio for having high LDL-c (OR=0.45,95% CI=0.28-7.2).Also having high HDL-c in the highest adherence to this pattern was 2.4 times higher than those in the lowest adherence of this group (OR=2.4,95% CI =1.1-5.1).Adherence to Unhealthy dietary pattern no significant relationship was observed in two models (without and after adjusting for variables).
As Asian population has different dietary styles and habits and in spite of the fact most information about the association between food patterns and risk of dyslipidemia is mainly derived from studies investigated Western, European or U.S. populations.It is not clear in what extent these findings are applicable in populations around the world [37].Moreover in this case, no specific study has been done on diabetic populations.However several studies evaluated dietary pattern role on the metabolic syndrome prevalence.As the diabetic patients are told and learnt to obey specific diet, we investigated their dietary patterns role on risk of hyperlipidemia.
*Total cholesterol: >200mg/dl, Triglyceride: >150mg/dl, LDL-c: >100mg/dl, HDL-c: ≤50mg/dl for females, ≤40mg/dl for males †p value based on comparison of the first quartile (baseline) and the fourth quartile ‡Unadjusted (crude) §Adjusted for age, gender, family history of diabetes, duration of diabetes, body mass index, physical activity, energy intake and lipidlowering drugs intake Two major food patterns, Healthy and Unhealthy dietary patterns were identified in this cross sectional survey.Dietary patterns obtained in this study are similar to the study on Tehran North-East women [49].Other similar studies in Iran, presented three major dietary patterns [50,51] which Western and Healthy dietary pattern of these studies were resemble to our Unhealthy and Healthy patterns respectively.Although dietary patterns among countries and cultures are different and not comparable, Unhealthy and Healthy food patterns in our study have resemble to those studied patterns in the world [14,15,[31][32][33][34].
Healthy pattern obtained in this study is consistent with the findings of Ismail-Zadeh and colleagues [39] on female teachers in Tehran in which those in the highest quintile of the healthy dietary pattern (high intake of fruits, vegetables, tomatoes, poultry, beans and green leafy vegetables, tea, juice and whole grain) had lower odds for dyslipidemia.Healthy dietary pattern in Finnish adults in Mikkila, et al. [33] showed on longitudinal analysis, LDL-c concentration was inversely correlated with score of Healthy dietary pattern (high intake of vegetables, beans and nuts, tea, rye, cheese and other dairy products).In Asian population several studies have shown lower likelihood of having low HDL-c and/or lower risk of having high TG [27][28][29].The protective effect of these dietary patterns can be explained by healthy components (such as vegetables, fruits, fish and whole grains).However Denova-Gutierrez [34] examined major dietary patterns relation to risk of metabolic syndrome.Prudent dietary pattern in this study, consistent with our Healthy pattern, no significant association was observed between this dietary pattern and metabolic syndrome components.Similar to this finding Akter in Japanese working population [35] and Naja in Lebanese adults [36] found no association between dietary patterns and the chance of having high LDL-c or low HDL-c.These confliction may suggest other environmental and clinical factors besides dietary factors in controlling lipid levels.
Besides Traditional dietary pattern of Ismail-Zadeh study (high intake of refined grains, potatoes, tea, whole grains, hydrogenated fats, beans and broth) were significantly associated with higher odds of dyslipidemia.Amini, et al. [30] studied patients with glucose intolerance expressed Western dietary pattern (high intake of sweets, butter, soft drinks, mayonnaise, sugar, cakes, meat , hydrogenated fats and eggs) is associated with a high chance of increasing triacylglycerol.
"Sweets" dietary patterns in Noel, et al. [31], Traditional dietary pattern in Mikkila [33], processed food products pattern in Olinto, et al. [34] also observed that pattern of were positively associated with high LDL-c and/or low HDL-c concentrations.These patterns were similar to our Unhealthy pattern in high consumption of snacks, sweets and sugar, fast foods, animal fat rich foods.Studies on Asian population, Song shows [28] traditional pattern (high consumption of white rice and kimchi) an increased risk of having low HDL-c compared with the other two patterns.The rice oriented pattern of Song [40] was positively associated with hypertriglyceridemia in men and LDL-c in both men and women.This pattern was characterized by a high intake of white rice but low intake of vegetables, fruits, meat, and dairy products.These patterns are somehow comparable with our Unhealthy pattern in high consumption of refined grains.However Cho, et al. [27] study in Korean women shows higher HDL-cholesterol levels in the participants with the highest Western dietary pattern score.Western pattern was identified by high intake of fast foods, animal rich fat foods, fried foods, grilled meat and seafoods, sweet foods, resemble to Unhealthy pattern in current study.Taking all the studies, consumption of refined grains, sweets, sugar, butter and animal fat rich foods can increase the risk of having dyslipidemia.However we found no relation between unhealthy dietary patterns with lipid disorders.
The Cross-sectional nature of the study does not allow cause and effect conclusions.Findings obtained should be confirmed in prospective studies.On the other hand, the assessing dietary intake using a food frequency questionnaire, it is likely to have measurement error such as difficulties in recalling consumed items, under-reporting or over-reporting about some or total number of food items.In the current study, dietary patterns were assessed only upon dietary intake.However, some researchers rely on FFQ as dietary assessment instrument, nutritional behavior such as eating habits and timing and frequency of meals and snacks should be noticed [52].Another limitation, there is no gold standard to determine the number of factors in factor analysis, and final decision to determine the number of factors to be realized is up to investigator.However the reliability and validity of this method has been shown in several studies [17,53].Most of our participated individuals were over weighted (83.3%) and also 37.9% have CAD history, so make the judgment somehow blurred.Moreover high number of our participant were under lipid-lowering treatments, despite this factor was adjusted in analysis; we could not identify the intensity of treatment since it is a critical factor that influence lipid profile.
Overall, the findings suggest that a dietary pattern rich in vegetables, fruits, fish, beans and olives is associated with a lower risk of developing high LDL-c levels and lower chance of having low HDL-c in diabetic patients.Yet the results obtained should be confirmed in prospective and interventional studies in healthy diabetic population with normal weight.

Table 3 :
Characteristics of participants in terms of blood lipid profile

Table 4 :
Odds ratios and 95% confidence interval of dyslipidemia according to score quartiles of dietary patterns in the study population