Association of the major dietary patterns with the risk of prediabetes among population from south-west of Iran: findings from Hoveizeh Cohort study

doi:10.21203/rs.3.rs-2064059/v1

Download PDF

Research Article

Association of the major dietary patterns with the risk of prediabetes among population from south-west of Iran: findings from Hoveizeh Cohort study

https://doi.org/10.21203/rs.3.rs-2064059/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Introduction: Type 2 diabetes (T2DM) has become a growing concern in recent years due to its great risks for public health. Nowadays dietary pattern has drawn considerable attention regarding the relationship between diet and diseases. In the present study we aimed at identifying major dietary patterns and their relationship to prediabetes.

Methods: In this case control study 5377 individuals (1335 people with prediabetes and 4042 controls) were recruited from Hoveizeh cohort study part of the Persian cohort study. A semi quantitative food frequency questionnaire was used to collect participants’ dietary intake, and major dietary patterns were identified by factor analysis. Adherence to the patterns was divided in to quartiles and the relationship between dietary patterns and prediabetes was then assessed using logistic regression analysis. The correlation between patterns and components of lipid profile was also assessed by logistic regression.

Results: four major patterns were identified: 1) fruits and vegetables 2) traditional 3) Sweets and snacks 4) prudent. None of which was associated with prediabetes after adjusting for confounding factors. But fruits and vegetables pattern was associated with low HDL.c (fourth VS first quartile OR: 1.539, 95%CI: 1.094–2.164). Also prudent pattern was negatively associated with high LDL.c (fourth VS. first quartile OR: 0.769, 95%CI: 0.622–0.952)

Conclusion: although No significant association was found between identified dietary patterns and the risk of prediabetes, an association was found between fruits and vegetables pattern and low HDL.c in people with prediabetes due to high consumption of refined grains. There was also a negative association between prudent pattern and high LDL.c because of high amounts of olive oil and vegetable oil and low intake of hydrogenated oil and animal fat.

prediabetes

dietary pattern

factor analysis

Iran

Hoveizeh

Persian cohort

Diabetes is a chronic, metabolic disease known by elevated levels of blood glucose which can lead to serious damage to the heart, blood vessels, eyes, kidneys and nerves. The most common is Type 2 Diabetes Mellitus (T2DM), usually in adults, which happens when the body becomes resistant to insulin or does not produce enough insulin [1]. It is diagnosed if one or more of the following criteria are met: Fasting blood sugar (FBS ≥ 126 mg/dl) or 2-h plasma glucose concentration > 200 mg/dl after a 75-g glucose load on the oral glucose tolerance test (OGTT)[2]

Prediabetes is defined as impaired fasting glucose (IFG), impaired glucose tolerance (IGT) or both. [3] It is characterized by an elevated blood glucose level which is above normal but not high enough to be defined as diabetes [2]. IFG is Known by 110 < FBS < 125 mg/dl. IGT is diagnosed if both of these criteria are met: (FBS < 126 mg/dl) and 2-h plasma glucose concentration > 140 and < 200 mg/dl. [2]. The American Diabetes Association (ADA) recommends diagnosing ‘prediabetes’ with HbA1c values between 5.7–6.4% and impaired fasting glucose when the fasting plasma glucose is between 100–125mg/dl.

T2DM is very prevalent around the word and is recognized as the seventh leading cause of death in 2016[4]. According to International Diabetes Federation 463 million people (%9.3) had diabetes in the world in 2019 which is predicted to reach 578.4 million (%10.2) in 2030 and 700.2 million (%10.9) in 2045 if the current trend continues. In Middle East and North Africa (MENA) the rate of diabetes was 54.8 million (%12.8) in 2019 and it is anticipated to reach 76 million (%14.2) and 107.6 million (%15.7) in 2030 and 2045 respectively [2]. In Iran about 4.5 million adults had diabetes in 2011. It is estimated that by the year 2030, this number will rise to 9.2 million [5]. A combination of factors can lead to diabetes. Examples of which are sedentary life style, smoking, genetics and unhealthy diet [6]. Regarding diet, a lot of studies suggest that some foods and nutrients are protective and some have adverse effects on the T2DM risk [7]. For instance consuming foods with high amounts of saturated fatty acids, refined grains and red meat, has been said to increase the risk of diabetes [8]. The World Health Organization (WHO) and the Food and Agriculture Organization (FAO) suggest limiting saturated fat and eating enough fiber also WHO recommends lowering the free sugar intake to prevent T2DM [9]. How food groups and nutrients affect the onset of diabetes and its complications, have been evaluated broadly before [10]. However people consume a variety of foods together as a meal rather than separately [7, 11] so the nutrients definitely have interactions with each other so analyzing the impact of one food or nutrient cannot determine these synergistic or antagonistic effects [7]. As a result determination of people’s dietary patterns seems more practical [6] and can provide us with new insights and more useful solutions for diabetes prevention[9].

T2DM has an upward trend in Iran and other parts of the world so it is crucial that diabetes-related dietary patterns be recognized and changed in order to prevent pre-diabetes and T2DM. As dietary patterns are extremely varied in different parts of the world and even in different regions of a country, the existing data is not completely applicable. There are a lot of studies about the relation of dietary pattern and diabetes. But the studies about dietary pattern and pre-diabetes are scarce. To the best of our knowledge there are not any studies about the association of dietary patterns and pre-diabetes in Khuzestan, Iran. So we aimed our work at identifying the major dietary patterns and studying their relation with the risk of prediabetes in adult participants of Hoveizeh Cohort in Hoveizeh city, Khuzestan. Iran.

Study population

This case-control study was conducted based on the data from Hoveizeh Cohort which is part of the Prospective Epidemiological Research Studies in Iran (the PERSIAN Cohort Study) [12], including 180 000 Iranian adults [13].

Hoveizeh Cohort is a prospective study designed to assess the NCDs in southwest of Iran which recruited all the Iranian residents of Hoveizeh who were between 35 and 70 years old from May 2016 to August 2018 except for blind, deaf, mute, mentally ill or physically disabled people who could not pay a visit to Cohort center or those unwilling to participate after being asked three times in different time periods. All of these people filled four questionnaires: two clinical, one general and one FFQ, Which were all valid and reliable [13].

In our study we used the data from Hoveizeh Cohort phase 1 such as anthropometric indexes, blood variables and dietary intake. We included people who were between 35 and 60 and excluded those who were pregnant, had cancer, kidney diseases, liver disorders, hypothyroidism, and hyperthyroidism, took blood sugar-lowering medicine, were on a special diet in the past year, consumed less than 800 or more than 4200 Calories or if 20 percent of questionnaire information was missing. In the end 5377 participants remained who were considered as our case group (100 < FBS < 126 mg/dl, 1335 people) or control group (FBS < 100 mg/dl, 4042 people).The study was approved by ethics committee of Ahvaz Jundishapur University of Medical Sciences. The ethical number is IR.AJUMS.REC.1399.724.

Questionnaires:

General questionnaire included questions about 1) personal information, 2) socio-economic status (whether or not they own a house/a car, the number of people they live with, the number of rooms in the house, home appliances, how often they travel ) 3) work experience 4) the condition of living place (the material which was used in the building, the kind of fuel they use, whether or not they have access to clean water, contact with animals) 5) sleep (the amount of sleep they get, whether or not they use sleeping pills) 6) physical activity. To calculate the physical activity of study subjects, the metabolic equivalent of task (MET) was computed to show the intensity of physical activities. The daily physical activity questionnaire was used to calculate MET for all participants' activities during 24 hours a day. In the end each person was given a number depending on the amount of their physical activity [13].

Clinical questionnaire consisted of questions about non communicable diseases, family history of the illnesses, oral health, women’s fertility status, the kind of medication and supplements people use, alcohol consumption and smoking habits.

Participants’ dietary intake was assessed using a 130-item food frequency questionnaire which was valid and reliable [13]. During the interview trained nutritionists explained the serving sizes of food items to the participants and showed them pictures of real sizes of these items if necessary. Consumption frequency of each food item was asked and recorded on a daily, weekly, monthly or yearly basis in the past year. Then the reported amounts of consumed food was changed in to grams. After that, the energy and nutrient content of the food items were determined using US Department of Agriculture’s (USDA) Food Composition Table (FCT) in Nutrition 4. Iranian FCT was used for items not included in USDA FCT.

Anthropometric measures

These measures were assessed when people were recruited as subjects of Hoveizeh Cohort. Weight was measured using Seca 755 mechanical column scale without shoes or heavy clothes. Height measurements were carried out using Seca 206 stadiometer, with shoulders at normal position and no shoes on. Body mass index (BMI) was calculated as weight (kg) divided by height squared (m2). Waist circumference was measured using a non-elastic tape between the lowest rib and iliac crest, in a standing position and without any pressure on the body. Hip circumstance was also measured with the same instrument and in the above mentioned position. The widest part of the hips was measured with the accuracy of 0.1 centimeter (the same as waist).Blood pressure was measured twice from the left and right arm after resting for 5 minutes and with a 10-minute interval in the sitting position. The mean of the two numbers was considered as the subject’s blood pressure.

Biochemical assessment

Blood samples were collected between 7 to 9 a.m. after fasting overnight for 8 to 12 hours. 25 milliliters of blood, including one clot tube (7 ml) and two k2-EDTA tube (9 mL) (KANG JIAN, China) by Vacutainers (Greiner Bio-One International GmbH, Kremsmunster, Austria) were taken from all participants. Centrifuged (Sigma, Germany) at 3000 rpm for 15 minutes to separate serum. Then, the required serum levels were measured by BT 1500 autoanalyser (Biotecnica Instruments, Italy). Serum total cholesterol, HDL-cholesterol, triglycerides, and Fasting Blood Sugar (FBS) were measured. LDL-cholesterol calculated using Friedewald formula. According to American Diabetes Association (ADA) FBS between 100 and 125 was considered to be prediabetes.

Statistical analysis

In order to extract the patterns, factor analysis was performed. In this study principal component analysis with varimax rotation was used. Two statistical tests were performed in order to determine the sufficiency of sampling: Bartlett’s test of sphericity (which should be significant), and Kaiser-Myer-Olkin (KMO) (which ranges from 0 to 1 and values ≥ 0.6 shows that a good factor analysis can be performed). In the present study adequacy of the sampling was approved by KMO values > 0.6 and P ≤ 0.05 for Bartlett’s test sphericity test. In order to compare continuous and categorical variables between case and control, independent t-test and chi-square test were used respectively. Quantitative variables were shown as mean ± standard deviation, and qualitative variables as percentage. The number of extracted factors was determined using the scree plot and eigenvalue

> 1.2. This way four patterns were extracted and in each pattern items with loadings of ≥ 0.3 were reported. The positive and negative sign showed a direct or inverse relationship between food groups and dietary pattern. The patterns were named based on the loaded food groups. Afterwards dietary patterns were divided into quartiles. The first quartile showed the lowest adherence and the last one indicated the highest. For each pattern the first quartile was considered as the reference. To find the correlation between patterns and prediabetes, binary logistic regression was performed and the odd ratios (ORs) and %95 confidence intervals (CIs) were determined in different models. First the crude model was obtained. The second model was adjusted for age, gender, education, marital status and wealth. And model three was further adjusted for energy and BMI. To find the correlation between patterns and abnormal lipid profile (LDL.c > 100, HDL.c < 40, TC > 200, TG > 150), in people with prediabetes, also binary logistic regression was used in the same models.

Table 1 and 2 depict the demographic characteristics of the participants. Table 1 shows the quantitative variables and Table 2 shows the qualitative ones. Patients had significantly higher weight, BMI, WC, WHR, SBP, DBP, LDL.c, TG, cholesterol (p < 0.001), less MET (p < 0.001), consumed less energy (p = 0.001), were significantly older (p < 0.001) and less educated (p = 0.002).

Table 1 characteristics of the participants (quantitative variables)

variables	Prediabetes(n=1335)	Control(n=4042)	P-value
Age	47.55±7.03	45.10±6.75	< 0.001
Weight.Kg	80.46±15.2	76.15±14.75	< 0.001
Height.cm	163.81±8.88	164.83±9.08	< 0.001
BMI. kg/m2	30±5.35	28.03±5.03	< 0.001
Energy.kcal	2803.19±706.04	2876.01±682.31	0.001
Waist circumference	101.71±11.83	96.94±11.66	< 0.001
Hip circumference	106.11±9.99	103.54±9.56	< 0.001
WHR	0.95±0.06	0.93±0.06	< 0.001
Systolic blood pressure. mmHg	113.9±17.63	109.26±16.33	< 0.001
Diastolic blood pressure. mmHg	72.25±11.33	69.88±10.75	< 0.001
LDL	109.96±32.96	104.88±30.98	< 0.001
HDL	50.05±12.01	50.55±12.01	0.188
TG	174.3±89.78	146.43±93.54	< 0.001
Cholesterol	194.28±40.35	184.73±36.77	< 0.001
MET	37.15±5.51	37.81±5.49	< 0.001

P-value < 0.05 was considered significant

P-value based on the t-test

Table 2 characteristics of the participants (qualitative variables)

variables	Prediabetes(n=1335)	Control(n=4042)	P-value
Gender			0.008
Male	467(35%)	1577(39%)
female	868(65%)	2465(61%)
education			0.002
Illiterate	843(63.1%)	2340(57.9%)
Primary school	202(15.1%)	725(17.9%)
Secondary school	88(6.6%)	315(7.8%)
High school degree	111(8.3%)	307(7.6%)
university	91(6.8%)	355(8.8%)
Marital status			< 0.001
Single	62(4.6%)	189(4.7%)
Married	1150(86.1%)	3628(89.8%)
Widowed	93(7%)	154(3.8%)
divorced	30(2.2%)	71(1.8%)
Wealth			0.362
Very poor	248(18.6%)	815(20.2%)
Poor	274(20.5%)	882(21.8%)
Moderate	290(21.7%)	832(20.6%)
Rich	282(21.1%)	781(19.3%)
Very rich	241(18.1%)	732(18.1%)

P-value < 0.05 was considered significant

P-value based on the chi-squared test

After using Principal Component Analysis, four dietary patterns were extracted based on the eigenvalue >1.2 and scree plot examination. They were named Based on the loaded food groups. The first pattern (fruits and vegetables) was identified by high loadings of tomatoes, garlic and onions, fruits, green vegetables, refined grain, white meat and vegetable oil. The second pattern named “traditional” Was characterized by high intake of green vegetables, legumes, other vegetables, dairy products, whole grain, organ and processed meat and red meat. The third pattern labeled “sweets and snacks” was identified by high intake of white meat, red meat, sweets, cake and cookies, nuts, pickled vegetables, chips and cheese puffs. The last pattern named “prudent” Was identified by high consumption of vegetable oil and olive oil, and low intake of refined grain, hydrogenated oil and animal fat. Factor loadings of the food groups are shown in Table 3. The variance explained by these four patterns were 9.98, 9.469, 8.901 and 7.756 respectively. The cumulative variance was 36.106. The association between pre diabetes and the quartiles of dietary patterns is depicted in table 4. Model 1 was crude. Model 2 was adjusted for age, gender, education, marital status, and wealth. Model 3 was further adjusted for BMI and energy. In model 1 an elevated risk of prediabetes was shown in people in the second quartile OR=1.258, 95%CI=1.017-1.557, third OR=1.287, 95%CI=1.04-1.593 and fourth quartile OR=1.371, 95%CI=1.11-1.693 of prudent Pattern. And a significant dose response relationship was also detected (P trend=0.014). But after adjusting for confounding factors in model 3 it was not significant any more. There was no significant association between fruits and vegetables, traditional and sweets and snacks dietary pattern and prediabetes. In Table 5 and 6 the association between components of lipid profile across quartiles of dietary pattern is demonstrated. Model 1 was crude, model 2 was adjusted for age, gender, education, marital status and economical status. Model 3 was further adjusted for energy consumption and BMI. All of the quartiles of prudent pattern had a negative association with the risk of elevated LDL.c. Q2: OR=0.808, CI=0.668-0.976. Q3: OR=0.806, CI=0.66-0.983. Q4: OR=0.769, CI=0.622-0.952 (P-trend=0.009). Also all of the quartiles of fruits and vegetables pattern were associated with the risk of reduced HDL.c. (Q2: OR=1.472, CI=1.107-1.959. Q3: OR=1.404, CI=1.034-1.905. Q4: OR=1.539 CI=1.094-2.164). But the trend was not significant (P-trend=0.076).

Table 3 Rotated factor loading matrix for the four dietary patterns


Food groups	Fruits& vegetables	traditional	Sweets& snacks	prudent
tomato	0.695	-	-	-
Garlic/onion	0.575	-	-	-
fruits	0.559	-	-	-
green vegetables	0.515	0.5	-	-
Refined grain	0.446	-	-	-0.337
White meat	0.4	-	0.33	-
legumes	-	0.605	-	-
Other vegetables	-	0.587	-	-
Dairy products	-	0.48	-	-
Whole grain	-	0.465	-	-
Organ meat/processed meat	-	0.411	-	-
Red meat	-	0.317	0.306	-
sweets	-	-	0.589	-
Cake/cookies	-	-	0.559	-
nuts	-	-	0.524	-
Pickled vegetables	-	-	0.519	-
Chips/cheese puffs	-	-	0.476	-
Hydrogenated oil/animal fat	-	-	-	-0.729
Vegetable oil	0.344	-	-	0.646
0live oil	-	-	-	0.434
Tea/coffee	-	-	-	-

Values less than |0.3| were excluded Bartlett’s Test of Sphericity: 8142.864, sig <0.0001; Kaiser-Meyer-Olkin test= 0.741

Table 4 Odds ratio and 95% confidence intervals for the association between dietary patterns and prediabetes

	Q1	Q2	Q3	Q4	P-trend
Fruits and vegetables pattern
Model 1	1	0.932(0.756-1.149)	1.044(0.85-1.283)	0.923(0.749-1.138)	0.671
Model 2	1	0.974(0.9787-1.206)	1.108(0.897-1.367)	1.005(0.81-1.248)	0.682
Model 3	1	0.991(0.789-1.245)	1.113(0.872-1.42)	1.011(0.762-1.342)	0.645
traditional pattern
Model 1	1	1.04(0.846-1.277)	0.92(0.746-1.135)	0.977(0.793-1.205)	0.622
Model 2	1	1.044(0.847-1.287)	0.93(0.751-1.152)	1(0.806-1.24)	0.741
Model 3	1	1.063(0.86-1.314)	0.95(0.764-1.182(	1.025(0.819-1.281)	0.987
Sweets and snacks pattern
Model 1	1	1.043(0.849-1.281)	0.909(0.738-1.121)	0.974(0.791-1.199)	0.195
Model 2	1	1.142(0.925-1.409)	1.058(0.852-1.314)	1.2(0.963-1.494)	0.441
Model 3	1	1.179(0.951-1.463)	1.077(0.859-1.35)	1.239(0.978-1.57)	0.297
prudent pattern
Model 1	1	1.258(1.017-1.557)	1.287(1.04-1.593)	1.371(1.11-1.693)	0.014
Model 2	1	1.255(1.01-1.559)	1.267(1.014-1.582)	1.301(1.034-1.637)	0.08
Model 3	1	1.183(0.947-1.479)	1.153(0.912-1.456)	1.183(0.922-1.518)	0.514

Model 1: crude

Model 2: adjusted for age, gender, marital status, education and economical status

Model 3: further adjusted for energy and BMI

Logistic regression analysis was used

Table 5 Odds ratio and 95% confidence intervals for the association between dietary patterns and components of lipid profile for “fruits and vegetables “and “traditional” pattern

Traditional pattern					Fruits and vegetables pattern
P-trend	Q4	Q3	Q2	Q1	P-trend	Q4	Q3	Q2	Q1
										High cholesterol
0.269	1.09(0.9-1.32)	1.06(0.88-1.29)	1.07(0.88-1.3)	1	0.657	0.97(0.8-1.17)	0.86(0.71-1.05)	0.87(0.72-1.05)	1	Model 1
0.264	1.1(0.9-1.34)	1.07 (0.88-1.31)	1.07 (0.88-1.3)	1	0.971	1(0.82-1.22)	0.88(0.72-1.07)	0.88 (0.72-1.07)	1	Model 2
0.175	1.12(0.91-1.37)	1.08 (0.89-1.32)	1.08 (0.88-1.31)	1	0.663	1.03(0.8-1.33)	0.9(0.72-1.13)	0.89(0.73-1.10)	1	Model 3
										High LDL
0.766	1(0.84-1.21)	1.08(0.9-1.29)	1.06(0.88-1.27)	1	0.803	0.97 (0.81-1.17)	0.85(0.71-1.03)	0.89(0.74-1.07)	1	Model 1
0.791	1.01(0.83-1.21)	1.08(0.89-1.3)	1.05(0.87-1.26)	1	0.702	1.02 (0.85-1.24)	0.89(0.73-1.07)	0.91(0.76-1.10)	1	Model 2
0.62	1.02(0.84-1.23)	1.09(0.9-1.31)	1.06(0.88-1.27)	1	0.597	1.02(0.8-1.3)	0.88(0.71-1.09)	0.92 (0.75-1.12)	1	Model 3
										High TG
0.004	1.25(1.04-1.51)	1.08 (0.9-1.3)	1.01(0.83-1.21)	1	0.036	1.23 (1.03-1.49)	1.06(0.88-1.28)	1.08 (0.90-1.31)	1	Model 1
0.033	1.18(0.97-1.43)	1.04 (0.86-1.27)	0.98(0.81-1.19)	1	0.71	1.06(0.87-1.28)	0.96(0.79-1.16)	1(0.82-1.21)	1	Model 2
0.097	1.15(0.94-1.4)	1.04 (0.86-1.27)	0.99(0.81-1.2)	1	0.126	0.87 (0.68-1.12)	0.84(0.68-1.05)	0.93 (0.76-1.14)	1	Model 3
										Low HDL
0.12	0.92(0.72-1.18)	0.93(0.73-1.18)	1.03(0.82-1.31)	1	<0.001	2.31(1.78-2.991)	1.81(1.39-2.36)	1.82 (1.4-2.37)	1	Model 1
0.008	0.81(0.63-1.05)	0.87(0.67-1.11)	0.99(0.77-1.26)	1	0.002	1.64 (1.25-2.14)	1.46(1.11-1.93)	1.51 (1.15-1.98)	1	Model 2
0.008	0.8(0.62-1.04)	0.86(0.67-1.11)	0.98(0.77-1.26)	1	0.076	1.53 (1.09-2.16)	1.4 (1.03-1.9)	1.47 (1.10-1.95)	1	Model 3

Model 1: crude

Model 2: adjusted for age, gender, marital status, education and economical status

Model 3: further adjusted for energy and BMI

Logistic regression analysis was used

Table 6 Odds ratio and 95% confidence intervals for the association between dietary patterns and components of lipid profile for “sweets and snacks “and “prudent” pattern

Prudent pattern					Sweets and snacks pattern
P-trend	Q4	Q3	Q2	Q1	P-trend	Q4	Q3	Q2	Q1
										High cholesterol
0.113	0.86(0.71-1.05)	0.99(0.82-1.2)	1.034(0.854-1.251)	1	0.004	0.77(0.64-0.94)	0.85(0.7-1.03)	0.86(0.71-1.04)	1	Model 1
0.089	0.84 (0.68-1.04)	0.98(0.8-1.2	1.024(0.843-1.243)	1	0.321	0.91 (0.74-1.11)	0.96(0.79-1.17)	0.92(0.76-1.12)	1	Model 2
0.031	0.8 (0.63-1)	0.94(0.76-1.16)	0.995(0.816-1.213)	1	0.578	0.93 (0.75-1.15)	0.97(0.79-1.19)	0.93(0.76-1.13)	1	Model 3
										High LDL
0.195	0.88 (0.73-1.05)	0.88(0.73-1.05)	0.85(0.71-1.02)	1	0.036	0.82 (0.68-0.98)	0.86(0.71-1.03)	0.9(0.75-1.09)	1	Model 1
0.03	0.8(0.66-0.98)	0.84(0.69-1.02)	0.83 (0.69-1)	1	0.641	0.93 (0.77-1.13)	0.94(0.78-1.14)	0.96(0.79-1.15)	1	Model 2
0.009	0.76(0.62-0.95)	0.8(0.66-0.98)	0.8(0.66-0.97)	1	0.856	0.94 (0.77-1.15)	0.95 (0.78-1.15)	0.97(0.8-1.17)	1	Model 3
										High TG
0.804	1(0.83-1.20)	1.14(0.95-1.37)	1.12(0.93-1.35)	1	0.884	1.01(0.84-1.22)	1.09(0.9-1.31)	1.02(0.85-1.23)	1	Model 1
0.307	1.07(0.88-1.32)	1.18(0.97-1.43)	1.1(0.91-1.34)	1	0.528	0.97(0.79-1.18)	1.07(0.88-1.29)	0.99(0.82-1.2)	1	Model 2
0.298	1.08(0.86-1.34)	1.15(0.94-1.42)	1.08(0.89-1.32)	1	0.162	0.9(0.73-1.12)	1.03(0.84-1.26)	0.98(0.81-1.2)	1	Model 3
										Low HDL
0.757	1(0.78-1.28)	1.16(0.91-1.48)	1.14(0.89-1.45)	1	<0.001	1.6(1.29-2.12)	1.45(1.13-1.87)	1.35(1.05-1.74)	1	Model 1
0.838	1.04(0.79-1.37)	1.14(0.88-1.48)	1.1(0.85-1.41)	1	0.272	1.2(0.92-1.57)	1.18(0.9-1.53)	1.14(0.88-1.49)	1	Model 2
0.789	1.02(0.76-1.37)	1.11 (0.85-1.45)	1.07(0.83-1.38)	1	0.428	1.17 (0.89-1.55)	1.16(0.89-1.52)	1.14(0.87-1.49)	1	Model 3

Model 1: crude

Model 2: adjusted for age, gender, marital status, education and economical status

Model 3: further adjusted for energy and BMI

Logistic regression analysis was used

In this case-control study the association between dietary pattern and prediabetes was examined. Four major dietary patterns were identified: fruits and vegetables, traditional, sweets and snacks, and prudent. Our findings showed that none of the patterns were associated with the risk of prediabetes. As for lipid profile in people with prediabetes, fruits and vegetables pattern was associated with low HDL.c and prudent pattern had a negative association with high LDL.c. after the previously mentioned adjustments.

In the current study, fruits and vegetables pattern was not associated with the risk of prediabetes. This pattern was characterized by high intake of tomatoes, garlic and onions, green vegetables, fruits, refined grain, white meat and vegetable oil. Fruits and vegetables pattern is comparable to a pattern in a study performed by Koloverou et al. With high loadings of fruit, vegetables, legumes, bread, rusk and pasta which reduced the risk of diabetes among people aged 45–55 years, but not in other age groups. After adjusting for carbohydrate intake, the statistical significance was lost[14]. According to Lin et al. a pattern with high intake of vegetables, vegetables with oil or dressing, fruits, seafood, legumes, soy products, rice and flour products was inversely associated with TC and LDL.c but directly associated with hyperglycemia which may be because of high consumption of simple carbohydrates like fruits, rice, noodles and bread that affected glucose metabolism [11]. The result of a meta-analysis showed that high intake of vegetable, fruit, whole grain and nuts reduces the risk of prediabetes [15]. Some systematic reviews studying the Mediterranean diet, showed that it can lead to glycemic control, reduced HbA1C, insulin resistance and oxidative stress compared to low fat and low carbohydrate diet [16, 17]. In the present study refined grain had the highest loading in fruits and vegetables pattern, so it can be concluded that it is not as healthy as Mediterranean diet with high intake of olive oil and whole grain or other patterns with high loadings of whole grain. Also, in the current study, a direct association between fruits and vegetables pattern and increased risk of low HDL.c was observed. The results of a review study showed that low carbohydrate, glycemic index and glycemic load diet was beneficially associated with HDL.c levels [18]. In the present study fruits and vegetables pattern had high loadings of refined grain which may be the reason for its association to low HDL.c levels.

In traditional pattern green vegetables, other vegetables, legumes, dairy products, whole grain, red, organ and processed meat were highly loaded. No association was found between this pattern and prediabetes or lipid profile components. In one study an increased risk of diabetes was observed in relation to processed meat consumption in women [22]. Jiang et al observed that Iron from red meat is related to increased risk of diabetes in men. But the reason for this correlation may be because of other components of red meat like saturated fats and cholesterol as there was no association between meat from other sources and prediabetes [23]. The food groups in traditional pattern consist of both healthy (vegetables, legumes and whole grain) and unhealthy (red, organ and processed meat) components. Dairy products were considered high fat, as in this region people tend to consume their local products which are mostly high in fat. In other studies low fat dairy was loaded in healthy patterns and high fat dairy in unhealthy ones. So not having an association with prediabetes was expected in traditional pattern.

Sweets and snacks pattern was characterized by high loadings of white and red meat, sweets, cake and cookies, nuts, pickled vegetables, chips and cheese puffs. No association was observed between this pattern and prediabetes or lipid profile. It is similar to “western” or “unhealthy” pattern in other studies. For instance SSMM pattern identified by consumption of sweets, solid fat, red meat, mayonnaise and refined grain was directly associated with prediabetes [8]. And western pattern with high consumption of refined grain, butter, sweets, hydrogenated oil, red and processed meat was positively correlated with the risk of metabolic syndrome and insulin resistance [20]. It also contains food group with high amount of salt. According to Radzeviciene et al. there was an association between adding salt to prepared meal and type 2 diabetes [24]. In another study by Feng et al. consuming high amounts of salt was associated with increased intake of sweetened beverages in children and young adults [25]. Another loaded food group in this pattern were nuts. As Muly et al reported in a systematic review, consuming tree nuts can reduce FBG as it can lower the glycemic load [27]. Another study conducted by Hernandez et al. had similar results. One of the reasons is the fiber content of nuts. Fiber can increase insulin sensitivity and reduce plasma insulin levels, as a result carbohydrate absorption slows down promoting better glycemic control. The second reason is that fiber is resistant to digestion so it will produce short chain fatty acids in colon which results in a reduction in hepatic glucose production and the secretion of GIP and GLP1 is stimulated. It triggers insulin secretion and promotes β cell proliferation and maintenance of blood glucose levels [28]. Nuts have high content of PUFA and MUFA which can decrease insulin resistance. It is believed that higher intake of unsaturated fatty acids affects cell membrane and facilitates glucose receptor movement to the cell surface increasing insulin sensitivity. Nuts are high in calcium, magnesium and potassium which are said to have advantageous effect on glycemic control [28]. Nut consumption can also improve lipid profile parameters [29].

This pattern has high loadings of cakes, sweets and pickled vegetables but then again it has high amounts of nuts. Despite high consumption of sweets and salty snacks, it seems that eating nuts as healthy snacks can prevent this pattern from increasing the risk of prediabetes.

Prudent pattern was identified by high intake of vegetable oil and olive oil, and low consumption of refined grain, hydrogenated oil and animal fat. No association was found between this pattern and prediabetes, but it reduced the risk of high LDL.c in some studies high intake of olive oil was associated with reduced risk of T2DM, improved glucose metabolism, and decreased insulin resistance [30, 31]. Olive oil is a good source of vitamin E, polyphenols especially Flavonoids which are anti-inflammatory and antiatherogenic and reduce oxidative stress and improve endothelial function, lipid profile, insulin sensitivity and glycemic control [30]. Polyphenols can also affect glucose metabolism through reducing carbohydrate absorption, decreasing glucose release from liver, and stimulating glucose uptake by cells [31]. According to Violi et al. olive oil reduced postprandial glucose which was the result of an increase in GLP1 and GIP and a reduction in Dipeptidyl peptidase4 (DPP4) [32]. In a systematic review Schwab et al. concluded there is a probable association between olive oil intake and insulin sensitivity and fasting insulin levels, but no correlation was found with FBS. The results showed no association between total fat and diabetes, but some limited evidence revealed that linoleic acid (LA) is inversely associated with T2DM. But some other n-6 PUFAs like GLA, DGLA, DTA and n-6 DPA were correlated with higher incidence of T2DM. Plant-derived n-3 PUFA alpha linolenic acid (ALA), was oppositely associated with T2DM but it was not true about marine-derived n-3 PUFAs [33]. The results were in line with another study performed by Forouhi et al. The possible reason is that ALA can induce insulin secretion through direct action on G protein receptors expressed in β cells, and increase insulin sensitivity through Insulin-like growth factor1 (IGF1) pathways. So it is important to consider individual PUFA and MUFAs rather than the whole circulating fatty acids [34]. In the present study unlike other studies there was no significant correlation between prudent pattern and prediabetes. It may be explained through dissimilarities between this pattern and extracted patterns in previously mentioned studies. Moreover, we did not know about the kinds of vegetable oil used by people. Since the results of using different kinds of vegetable oil were diverse, it is recommended to include it in FFQ.

According to Violi et al. consuming olive oil resulted in a reduction in LDL.c levels it may be because of an interference between extra virgin olive oil and LDL.c absorption or a direct activity on LDL.c biosynthesis [32]. The results of Schwab’s study showed replacing SFA with PUFA and MUFA can lead to a reduction in total cholesterol and LDL.c [33]. According to Hernaez et al. olive oil polyphenols can reduce LDL.c concentration in men which may be the result of an improvement in systematic oxidative status or because of an elevation in the gene expression of lipoprotein lipase (LPL) [35]. In our study prudent pattern was significantly associated with reduced risk of high LDL.c which was reasonable and in line with most of the studies considering high loadings of vegetable and olive oil and negative loadings of hydrogenated and animal fat. This study has some limitations. First, like other case-control studies our study cannot determine a cause effect relationship. Second, factor analysis was used which can result in different patterns based on the researchers’ opinions. Third some behaviors like the number of the meals, the way of preparing them and time are also important but were not included. And finally recall bias probably exists since data collection is highly dependent on people’s memory. However the present study was the only one that have inspected the relationship between dietary patterns and prediabetes in Khuzestan Iran. The sample size was also very large. In conclusion the identified patterns were not associated with prediabetes, but fruits and vegetables pattern was associated with low HDL.c and prudent pattern had an inverse association with high LDL.c levels. Future studies are required to confirm these results in other populations.

T2DM: Type 2 diabetes mellitus; FBS: Fasting blood sugar; OGTT: oral glucose tolerance test; IFG: impaired fasting glucose; IGT: impaired glucose tolerance; HbA1c: hemoglobin A1C; HDL.c: high density lipoprotein cholesterol; LDL.c: low density lipoprotein cholesterol; NCD: non communicable diseases; FFQ: food frequency questionnaire; MET: metabolic equivalent of task; BMI: body mass index; WC: waist circumference; HC: hip circumference; WHR: waist to hip ratio; SBP: systolic blood pressure; DBP: diastolic blood pressure; TG: triglyceride; GIP: gastric inhibitory polypeptide; GLP1: glucagon-like peptide-1; PUFA: poly unsaturated fatty acid; MUFA: mono unsaturated fatty acid; DPP4: dipeptidyl peptidase4 ; LA: linoleic acid; ALA: alpha linolenic acid; GLA: gamma linolenic acid; DGLA: dihomo gamma linolenic acid; DTA: Docosatetraenoic acid; DPA: Docosapentaenoic acid.

Acknowledgments

This study was a part of a Master of Science thesis in Nutrition Sciences by Pegah Sajadi, supported by Jundishapur University of Medical Science, Ahvaz, Iran (grant No.NRC-9908). The Iranian Ministry of Health and Medical Education has contributed to the funding used in the PERSIAN Cohort through Grant No.700/534.

Authors’ contributions

A.M designed the original intervention. A.H and B.Ch collected the data. A.M and P.S analyzed and interpreted the data and were the major contributors in writing the manuscript. All authors reviewed and confirmed the final manuscript.

Funding

The study was supported by Jundishapur University of Medical Sciences, Ahvaz, Iran (grant No. NRC-9908).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

The study was approved by ethics committee of Ahvaz Jundishapur University of Medical Sciences. The ethical number is IR.AJUMS.REC.1399.724.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

diabetes [https://www.who.int/health-topics/diabetes#tab=tab_1]
Rhys Williams (Chair) SCDC, Reem Almutairi, Pablo Aschner Montoya, Abdul, Basit DB, Stéphane Besançon, Christian Bommer, Wenche Borgnakke, Edward Boyko, Dominic, Bright JC, Gisela Dahlquist, Hema Divakar, Alireza Esteghamati, Nita Forouhi, Laercio Franco,, Edward Gregg LG, Mohamed Hassanein, Calvin Ke, Dinky Levitt, Lee-Ling Lim, Ayesha, Motala GO, Katherine Ogurstova, David Owens, Chris Patterson, Meda Pavkov, Jonathan, Pearson-Stuttard AR, Wolfgang Rathmann, Musarrat Riaz, Gojka Roglic, Jonathan, Shaw DS, Alan Sinclair, Eugene Sobngwi, Gyula Soltesz, Rebecca Thomas, Nigel Unwin,, Heather Ward SW, Xilin Yang, Lili Yuen, Ping Zhang: IDF DIABETES ATLAS. vol. 9, 9th edition: international diabetes federation; 2019.
Agarwal A, Hegde A, Yadav C, Ahmad A, Manjrekar PA, Srikantiah RM: Assessment of oxidative stress and inflammation in prediabetes—A hospital based cross-sectional study. Diabetes & Metabolic Syndrome: Clinical Research & Reviews 2016, 10:S123-S126.
Sciberras J, Camilleri LM, Cuschieri S: The burden of type 2 diabetes pre-and during the COVID-19 pandemic–a review. Journal of Diabetes & Metabolic Disorders 2020, 19:1357-1365.
Esteghamati A, Larijani B, Aghajani MH, Ghaemi F, Kermanchi J, Shahrami A, Saadat M, Esfahani EN, Ganji M, Noshad S: Diabetes in Iran: prospective analysis from first nationwide diabetes report of National Program for Prevention and Control of Diabetes (NPPCD-2016). Scientific reports 2017, 7:1-10.
Shu L, Shen X-M, Li C, Zhang X-Y, Zheng P-F: Dietary patterns are associated with type 2 diabetes mellitus among middle-aged adults in Zhejiang Province, China. Nutrition journal 2017, 16:1-9.
Lee KW, Woo HD, Cho MJ, Park JK, Kim SS: Identification of dietary patterns associated with incidence of hyperglycemia in middle-aged and older Korean adults. Nutrients 2019, 11:1801.
Bagheri F, Siassi F, Koohdani F, Mahaki B, Qorbani M, Yavari P, Shaibu OM, Sotoudeh G: Healthy and unhealthy dietary patterns are related to pre-diabetes: a case–control study. British journal of nutrition 2016, 116:874-881.
Song S, Lee JE: Dietary patterns related to triglyceride and high-density lipoprotein cholesterol and the incidence of type 2 diabetes in Korean men and women. Nutrients 2019, 11:8.
Zaroudi M, YAZDANI CJ, Mehrabi S, Ghorbani E, Norouzkhani J, Shirashiani H, Nikzad B, Seiedpour M, Izadi M, MIRZAEI MR: Dietary patterns are associated with risk of diabetes type 2: a population-based case-control study. 2016.
Lin L-Y, Hsu C-Y, Lee H-A, Wang W-H, Kurniawan AL, Chao JC-J: Dietary patterns in relation to components of dyslipidemia and fasting plasma glucose in adults with dyslipidemia and elevated fasting plasma glucose in Taiwan. Nutrients 2019, 11:845.
Poustchi H, Eghtesad S, Kamangar F, Etemadi A, Keshtkar A-A, Hekmatdoost A, Mohammadi Z, Mahmoudi Z, Shayanrad A, Roozafzai F: Prospective epidemiological research studies in Iran (the PERSIAN Cohort Study): rationale, objectives, and design. American journal of epidemiology 2018, 187:647-655.
Cheraghian B, Hashemi SJ, Hosseini SA, Poustchi H, Rahimi Z, Sarvandian S, Malehi AS, Alipour M, Eghtesad S, Fatahiasl J: Cohort profile: The Hoveyzeh Cohort Study (HCS): A prospective population-based study on non-communicable diseases in an Arab community of Southwest Iran. Medical journal of the Islamic Republic of Iran 2020, 34:141.
Koloverou E, Panagiotakos D, Pitsavos C, Chrysohoou C, Georgousopoulou E, Grekas A, Christou A, Chatzigeorgiou M, Skoumas I, Tousoulis D: Adherence to Mediterranean diet and 10‐year incidence (2002–2012) of diabetes: correlations with inflammatory and oxidative stress biomarkers in the ATTICA cohort study. Diabetes/metabolism research and reviews 2016, 32:73-81.
Maghsoudi Z, Ghiasvand R, Salehi-Abargouei A: Empirically derived dietary patterns and incident type 2 diabetes mellitus: a systematic review and meta-analysis on prospective observational studies. Public health nutrition 2016, 19:230-241.
Sleiman D, Al-Badri MR, Azar ST: Effect of mediterranean diet in diabetes control and cardiovascular risk modification: a systematic review. Frontiers in public health 2015, 3:69.
Martín-Peláez S, Fito M, Castaner O: Mediterranean diet effects on type 2 diabetes prevention, disease progression, and related mechanisms. A review. Nutrients 2020, 12:2236.
Yanai H, Katsuyama H, Hamasaki H, Abe S, Tada N, Sako A: Effects of carbohydrate and dietary fiber intake, glycemic index and glycemic load on HDL metabolism in Asian populations. Journal of clinical medicine research 2014, 6:321.
Zhang J, Li L, Song P, Wang C, Man Q, Meng L, Cai J, Kurilich A: Randomized controlled trial of oatmeal consumption versus noodle consumption on blood lipids of urban Chinese adults with hypercholesterolemia. Nutrition Journal 2012, 11:1-8.
Esmaillzadeh A, Kimiagar M, Mehrabi Y, Azadbakht L, Hu FB, Willett WC: Dietary patterns, insulin resistance, and prevalence of the metabolic syndrome in women. The American journal of clinical nutrition 2007, 85:910-918.
Schulze M, Manson J, Willett W, Hu F: Processed meat intake and incidence of type 2 diabetes in younger and middle-aged women. Diabetologia 2003, 46:1465-1473.
Song Y, Manson JE, Buring JE, Liu S: A prospective study of red meat consumption and type 2 diabetes in middle-aged and elderly women: the women’s health study. Diabetes care 2004, 27:2108-2115.
Jiang R, Ma J, Ascherio A, Stampfer MJ, Willett WC, Hu FB: Dietary iron intake and blood donations in relation to risk of type 2 diabetes in men: a prospective cohort study. The American journal of clinical nutrition 2004, 79:70-75.
Radzeviciene L, Ostrauskas R: Adding salt to meals as a risk factor of type 2 diabetes mellitus: a case–control study. Nutrients 2017, 9:67.
He FJ, Marrero NM, MacGregor GA: Salt intake is related to soft drink consumption in children and adolescents: a link to obesity? Hypertension 2008, 51:629-634.
Lin Y, Chattopadhyay K, Yang X, Li J-L, Chen Y-S, Zhou Y, Li L: Association between dietary salt and plasma glucose, insulin and hemoglobin A1C levels among type 2 diabetes patients in Eastern China. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy 2021, 14:4811.
Muley A, Fernandez R, Ellwood L, Muley P, Shah M: Effect of tree nuts on glycemic outcomes in adults with type 2 diabetes mellitus: A systematic review. JBI Evidence Synthesis 2021, 19:966-1002.
Hernández-Alonso P, Camacho-Barcia L, Bulló M, Salas-Salvadó J: Nuts and dried fruits: an update of their beneficial effects on type 2 diabetes. Nutrients 2017, 9:673.
Stróżyk AK, Pachocka L: The role of nuts consumption in the primary and secondary prevention of type 2 diabetes. Clinical Diabetology 2017, 6:26-33.
Guasch-Ferré M, Hruby A, Salas-Salvadó J, Martínez-González MA, Sun Q, Willett WC, Hu FB: Olive oil consumption and risk of type 2 diabetes in US women. The American journal of clinical nutrition 2015, 102:479-486.
Schwingshackl L, Lampousi A, Portillo M, Romaguera D, Hoffmann G, Boeing H: Olive oil in the prevention and management of type 2 diabetes mellitus: a systematic review and meta-analysis of cohort studies and intervention trials. Nutrition & diabetes 2017, 7:e262-e262.
Violi F, Loffredo L, Pignatelli P, Angelico F, Bartimoccia S, Nocella C, Cangemi R, Petruccioli A, Monticolo R, Pastori D: Extra virgin olive oil use is associated with improved post-prandial blood glucose and LDL cholesterol in healthy subjects. Nutrition & diabetes 2015, 5:e172-e172.
Schwab U, Lauritzen L, Tholstrup T, Haldorsson TI, Riserus U, Uusitupa M, Becker W: Effect of the amount and type of dietary fat on cardiometabolic risk factors and risk of developing type 2 diabetes, cardiovascular diseases, and cancer: a systematic review. Food & nutrition research 2014, 58:25145.
Forouhi NG, Imamura F, Sharp SJ, Koulman A, Schulze MB, Zheng J, Ye Z, Sluijs I, Guevara M, Huerta JM: Association of plasma phospholipid n-3 and n-6 polyunsaturated fatty acids with type 2 diabetes: the EPIC-InterAct case-cohort study. PLoS medicine 2016, 13:e1002094.
Hernáez Á, Remaley AT, Farras M, Fernández-Castillejo S, Subirana I, Schröder H, Fernández-Mampel M, Munoz-Aguayo D, Sampson M, Solà R: Olive oil polyphenols decrease LDL concentrations and LDL atherogenicity in men in a randomized controlled trial. The Journal of nutrition 2015, 145:1692-1697.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Association of the major dietary patterns with the risk of prediabetes among population from south-west of Iran: findings from Hoveizeh Cohort study

Status:

Version 1

Abstract

Introduction

Materials And Methods

Study population

Questionnaires:

Anthropometric measures

Biochemical assessment

Statistical analysis

Results

Discussion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1