The Relationship between Dietary Patterns and Metabolic Phenotypes: A Cross-Sectional Study in a Representative Sample of Iranian Industrial Employees

Abstract

Background and Aim:

There is limited evidence to support the relationship between dietary patterns and metabolic phenotypes. Therefore, this study aimed to assess the association of dietary patterns with metabolic phenotypes among a large sample of Iranian industrial employees.

Methods:

This cross-sectional study was conducted among 3,063 employees of Esfahan Steel Company, Iran. Using exploratory factor analysis, major dietary patterns were obtained from a validated short form of food frequency questionnaire. The metabolic phenotypes were defined according to Adult Treatment Panel III guidelines. The independent-sample t-test, one-way analysis of variance, χ ² test, and multivariable logistic regression were applied to analyze data.

Results:

Three major dietary patterns were identified by factor analysis: the Western dietary pattern, the healthy dietary pattern, and the traditional dietary pattern. After controlling for potential confounders, subjects in the highest tertile of Western dietary pattern score had a higher odds ratio (OR) for metabolically healthy obese (MHO; OR 1.58, 95% confidence interval [CI]: 1.29–1.94), metabolically unhealthy normal weight (OR 1.93, 95% CI 1.08–3.45), and metabolically unhealthy obese (MUHO) phenotypes (OR 2.87, 95% CI 2.05–4.03) than those in the lowest tertile. Also, higher adherence to traditional dietary pattern was positively associated with a higher risk of MHO (OR 1.91, 95% CI 1.56–2.34) and MUHO phenotypes (OR 2.33, 95% CI 1.69–3.22) in the final model.

Conclusion:

There were significant associations between dietary patterns and metabolic phenotypes, suggesting the necessity of nutritional interventions in industrial employees to improve metabolic phenotype, health outcomes, and, therefore, job productivity in the workforce population.

Introduction

Metabolic disorders, including dyslipidemia, hypertension, and insulin resistance put people at risk for chronic diseases, such as cardiovascular disease, diabetes, and metabolic syndrome.^1,2 Although such disorders are more prevalent among obese people,³ there are some people with obesity who are metabolically healthy, while some individuals with normal weight are metabolically unhealthy.^4
–6 According to their weight and metabolic status, individuals are divided into four groups: metabolically healthy normal weight (MHN), metabolically healthy obese (MHO), metabolically unhealthy normal weight (MUHN), and metabolically unhealthy obese (MUHO).^7,8

Previous studies have investigated the effects of sociodemographic and behavioral factors^9,10 and dietary intake on metabolic phenotypes.^11,12 Epidemiological studies have often focused on the effects of individual foods or nutrients on metabolic health.^2,13 The effects of foods and nutrients associated with a healthy diet, including fruits, vegetables, whole grains, vitamin D, and Omega 3, on the metabolic indicators have been recently addressed in a review study.¹⁴ Given the synergic effects of food and nutrients and their interactions when eating together, as well as undiscovered components in foods, it seems therefore that the assessment of dietary patterns is an effective and comprehensive tool for understanding the impact of nutrition on metabolic health.^15,16 Furthermore, the dietary pattern methods can provide more detailed information about individuals’ dietary habits. Thus, they may be more informative on the nutritional etiology of metabolic conditions.^15,16

The statistical analysis of dietary patterns can be distinguished into a priori and a posteriori methods.^15,17 A priori approach assigns the dietary scores and indices (i.e., Mediterranean score, glycemic index) based on the current nutritional knowledge of dietary factors affecting health. Conversely, a posteriori approach defines the dietary patterns (i.e., Western or unhealthy patterns) based on dietary data obtained from the study population.^15,17

Despite a growing body of evidence on the association between dietary patterns and health, to date, the relationship between dietary patterns and metabolic phenotypes has been rarely investigated, especially in developing countries and among special populations at high risk of chronic diseases. Therefore, the aim of this study was to investigate the association between the a posteriori dietary patterns and the different metabolic phenotypes among industrial workers in Iran.

Methods

Study design and participants

This cross-sectional study was conducted on 3,500 full-time employees of Iran Esfahan Steel Company (ESCO) in the Epidemiological Survey of Chronic Diseases on Manufacturing Employees (ESCOME).¹⁸ Considering 0.1, 0.05, and 0.01 as the values for the prevalence of psychological disorders, type one error rate, and sampling error rate, respectively, the sample size was estimated to be 3,500. Using a multistage cluster sampling method and based on the size of company departments, the participants were selected among 16,000 people from seven main departments of the company. All workers with at least 1 year of work experience and those who were willing to participate in the study were eligible to enroll in this study. Participants who did not answer more than 10% of the questions were excluded from the analysis (n = 437). A total of 3,063 participants returned the complete questionnaires (response rate: 0.87) and were included in the analysis.

The study protocol was approved by the Medical Research Ethics Committee of the Isfahan University of Medical Sciences (Project number: 87115), and written informed consent was obtained from each participant.

Dietary assessment

Dietary intake was evaluated by a 48-item (including 38 food items/groups along with 10 dietary habits) self-administered food frequency questionnaire (FFQ) with the assessed validity and reliability in the Iranian population. The reproducibility of the FFQ was evaluated using the intraclass correlation coefficient, which ranged from 0.47 to 0.69 across various food groups.¹⁹ Participants were asked to report the frequency of consumption of 38 food items during the past year on a daily, weekly, or monthly basis. The consumed frequency of each food item was converted to weekly consumption (“times” per week). Also, food items that participants rarely or never consumed were considered “zero.”

Anthropometric assessment

Anthropometric indices were measured by trained health care staffs according to the standard protocols. Weight and height were measured with subjects wearing minimal clothing and no shoes. Weight was measured in kilograms (kg) and height was measured to the nearest 0.1 cm using an unstretched tape. Waist circumference (WC) was measured in midway between the lowest rib and the iliac crest, and hip circumference (HC) was measured at the point with the maximum size of the buttocks. Body mass index (BMI) was calculated as the ratio of weight in kilograms divided by height in meters squared (m²).

Assessment of biochemical variables and blood pressure

The overnight fasting blood specimens were collected for biochemical analysis. In this regard, a commercial kit (Pars Azmoon, Iran) was used to quantify the concentrations of total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglyceride (TG). The level of fasting blood glucose (FBS) was also measured using the Biosystems kit (France).

In ESCO’s laboratory occupational medicine center, all experiments were conducted in biological duplicates. Blood pressure was recorded by professional personnel utilizing a standard mercury sphygmomanometer. The average of two readings was calculated after measuring each arm twice in a sitting position following a 5-min delay.

Assessment of other variables

The sociodemographic characteristics of the subjects were collected using a self-administered questionnaire, which included age (based on years), sex, marital status (single or married), number of household members, educational levels (0–5, 6–12, and >12 years), smoking habits [cigarette smoking (Yes/No)], shift work (daily/rotational), having an extra career (Yes/No), and the total sleep time (hr).

Physical activity [metabolic equivalent (MET) (hr/week)] was assessed by the Short Form Persian version of the International Physical Activity Questionnaire,²⁰ in which participants were asked to report the time spent walking, and physical activity with vigorous and moderate intensity over the past week (numbers of days and the average time per day). The Persian version of the 23—item effort–reward imbalance questionnaire (ERI-Q) was also employed to measure the ERI among individuals.²¹

Definition of metabolic/obesity phenotypes

Obesity/overweight and normal weight were described as BMI ≥ 25 kg/m² and BMI < 25 kg/m², respectively. The metabolic phenotypes were defined based on the following components as recommended by Adult Treatment Panel III (ATP III): abdominal adiposity (WC > 88 cm in women and > 102 cm in men); high serum triacylglycerol levels (≥150 mg/dL); low serum HDL cholesterol (<40 mg/dL in men and <50 mg/dL in women); abnormal glucose homeostasis (fasting plasma glucose ≥110 mg/dL) and elevated blood pressure (>130/85 mmHg). If a participant met three or more of these criteria, he/she was considered as having the unhealthy metabolic phenotype.²²

Statistical analysis

We employed exploratory factor analysis via the principal component extraction technique in order to extract dietary patterns on the basis of the 38 food items/groups, and for interpretability purposes, using the Varimax orthogonal transformation technique, the factors were rotated. The scree plot and eigenvalues >1 were utilized in order to decide the suitable number of factors.²³ In accordance with the earlier literature and loaded food items/groups on the individual derived factors, the labeling of obtained factors (i.e., dietary patterns) was conducted. Through summation of intakes of food items/groups, which were weighted via their factor loadings, the factor score was determined for every pattern,²³ and a factor score was assigned to every subject for each recognized pattern. The categories of subjects were determined according to tertiles of dietary patterns scores. This enabled us to evaluate the association of different levels of intake from each dietary pattern with metabolic phenotypes in our study participants. Quantitative data were expressed as mean ± SD, while the categorical data were presented as frequency (percentage). To evaluate significant differences in the subjects’ characteristics across the tertile of dietary patterns scores and metabolic/obesity phenotype, χ ² and one-way analysis of variance tests were carried out for categorical and continuous variables, respectively.

The relationship between dietary patterns and metabolic phenotypes was evaluated using the multinomial logistic regression in different models. The association between dietary patterns and metabolic phenotypes was first obtained in the crude model odds ratio (OR), then in model I, the adjustment was done for demographic variables of age, sex, marital status, and educational levels, and then further adjustment was made for the lifestyle-related confounders including sleep duration, physical activity [MET (hr/week)], household size, and ERI ratio (as an indication of job stress) in model II. In all multivariable models, the first tertile of dietary patterns scores was considered as a reference category. Also, the MHN phenotype was considered as a reference category of the multinomial response variable in all analyses. The Mantel–Haenszel extension χ ² test was used to assess the overall trend of ORs across increasing tertiles of dietary patterns scores. We used SPSS software (version 16; SPSS Inc., Chicago, IL, USA) for all statistical analyses.

Results

The descriptive characteristics of the study participants across metabolic status are shown in Table 1. The MHO metabolic phenotype was the most prevalent (45%), followed by the MHN and MUHO phenotypes (41% and 11%, respectively). The MUHN phenotype was identified in 3% of the participants.

Table 1.

General Characteristics of Study Participants According to Metabolic Phenotypes

Variables	Metabolically healthy normal weight	Metabolically healthy obese	Metabolically unhealthy normal weight	Metabolically unhealthy obese	P value*
Variables	(n = 1270)	(n = 1366)	(n = 80)	(n = 345)	P value*
Age (years)	34.96 ± 7.14	37.18 ± 7.01	39.09 ± 1.17	41.00 ± 6.90	0.00
Gender
Male (%)	1135 (89.4)	1263 (92.5)	74 (92.5)	330 (95.7)	0.001
BMI (kg/m²)	22.26 ± 1.97	27.80 ± 2.35	22.62 ± 2.05	29.77 ± 3.30	<0.001
Sleep duration (hr)	7.17 ± 1.18	7.08 ± 1.17	7.24 ± 1.12	7.00 ± 1.11	0.04
ER ratio	0.65 ± 0.21	0.67 ± 0.23	0.64 ± 0.21	0.68 ± 0.05	0.04
Household size	3.55 ± 1.17	3.61 ± 1.04	3.61 ± 1.08	3.83 ± 1.02	0.001
Physical activity (MET-hr/wk)	7.13 ± 3.51	7.42 ± 3.70	7.39 ± 3.42	7.68 ± 3.92	0.06
Education level (%)
0–5 years	109 (8.6)	96 (7.0)	12 (15.0)	38 (11.0)	0.01
6–12 years	770 (60.6)	864 (63.3)	50 (62.5)	222 (64.3)
>12 years	391 (30.8)	406 (29.7)	18 (22.5)	85 (24.6)
Shifting work (%)
Daily	579 (45.6)	615 (45.0)	39 (48.8)	146 (42.3)	0.65
Shift	691 (54.4)	751 (55.0)	41 (51.2)	199 (57.7)
Second job—yes (%)	107 (8.4)	136 (10.0)	8 (10.0)	34 (9.9)	0.57
Marital status
Married (%)	1075 (84.6)	1270 (93.0)	76 (95.0)	336 (97.4)	<0.001
Income (Rials) (%)
<5,000,000	232 (18.3)	202 (14.8)	6 (7.5)	41 (11.9)	<0.001
5,000,000–8,000,000	684 (53.9)	684 (50.1)	52 (65.0)	165 (47.8)
8,000,000–12,000,000	244 (19.2)	316 (23.1)	15 (18.8)	84 (24.3)
>12,000,000	106 (8.3)	158 (11.6)	7 (8.8)	55 (15.9)
No answer	4 (0.3)	6 (0.4)	0 (0.0)	0 (0.0)
Smoking—Yes (%)	363 (28.6)	393 (28.8)	20 (25.0)	115 (33.3)	0.27

Values in the table are mean ± SD for continuous and percentage for categorical variables; P-values resulted from one-way ANOVA for continuous and χ ² test for categorical variables.

P-values obtained from χ ² test and ANOVA for categorical and continuous variables, respectively.

ANOVA, analysis of variance; BMI, body mass index; ER ratio, effort—reward imbalance ratio; MET-hr/wk, metabolic equivalent-hr/wk.

For both metabolically healthy and unhealthy groups, the means of age, ERI ratio, and household size in nonobese subjects were significantly lower than those in obese ones (P < 0.05). On the contrary, sleep duration in non-obese subjects was significantly higher than that in obese ones in both metabolically healthy and unhealthy groups (P = 0.04). Obese subjects in both metabolically healthy and unhealthy groups were more likely to be male, highly educated, married, and had higher incomes (P < 0.01).

Biochemical and anthropometric measures of the participants in different categories of metabolic phenotypes are shown in Table 2. Among different phenotypes, MHN subjects had the lowest mean FBS, cholesterol (CH), TG, BMI, WC, waist to hip ratio (WHR), systolic blood pressure (SBP), and diastolic blood pressure (DBP) in comparison to other metabolic phenotypes (P < 0.001 for all). Obese groups, either metabolically healthy or unhealthy had higher levels of FBS, CH, TG, BMI, WC, HC, WHR, and DBP (P < 0.001). In contrast, HDL levels were higher in nonobese subjects in both healthy and unhealthy groups (P < 0.001). Metabolically unhealthy subjects had higher levels of FBS, CH, TG, HDL, LDL, BMI, WC, WHR, and SBP/DBP than metabolically healthy ones (P < 0.001).

Table 2.

Anthropometric and Clinical Traits of Study Participants According to Metabolic Phenotypes

	Metabolically healthy normal weight	Metabolically healthy obese	Metabolically unhealthy normal weight	Metabolically unhealthy obese	P value*
FBS (mg/dL)	88.19 ± 13.15	89.93 ± 14.02	98.07 ± 24.13	105.54 ± 33.91	<0.001
CH (mg/dL)	172.30 ± 31.34	183.52 ± 32.89	203.58 ± 38.95	204.28 ± 37.54	<0.001
TG (mg/dL)	118.77 ± 70.97	153.58 ± 92.50	175.18 ± 108.14	221.55 ± 138.25	<0.001
HDL (mg/dL)	42.88 ± 8.48	41.12 ± 7.24	47.76 ± 8.39	44.01 ± 8.05	<0.001
LDL (mg/dL)	87.01 ± 19.92	93.30 ± 20.33	105.52 ± 26.82	101.83 ± 22.55	<0.001
BMI (kg/m²)	22.26 ± 1.97	27.80 ± 2.35	22.62 ± 2.05	29.77 ± 3.30	<0.001
WC (cm)	81.12 ± 6.80	92.96 ± 7.07	82.65 ± 6.71	100.38 ± 7.90	<0.001
HC (cm)	95.59 ± 5.30	103.31 ± 5.27	95.08 ± 7.07	106.44 ± 6.41	<0.001
WHR (cm)	0.85 ± 0.06	0.90 ± 0.06	0.87 ± 0.11	0.94 ± 0.05	<0.001
SBP (mmHg)	11.30 ± 0.10	11.70 ± 0.93	12.92 ± 1.07	12.78 ± 1.11	<0.001
DBP (mmHg)	7.46 ± 0.78	7.82 ± 0.75	8.47 ± 0.54	8.56 ± 0.75	<0.001

Values are presented in mean ± SD.

P-values resulted from χ ² test and analysis of variance.

FBS, Fasting blood sugar; CH, cholesterol; TG, triglyceride; HDL, High density lipoprotein; LDL, Low density lipoprotein; BMI, Body mass index; WC, Waist circumference; HC, Hip circumference; WHR, Waist to hip ratio; SBS, systolic blood pressure; DBP, diastolic blood pressure.

Table 3 shows the factor-loading matrix of the extracted dietary patterns. Using factor analysis, three major dietary patterns were excluded based on 38 food items/groups: Western dietary pattern composed of butter, cream, sheep's head and trotters, processed meat, carbonated drinks, commercial fruit juices, jams, cake, cookie and sweets, biscuits, fresh fruit juice, seeds, egg, pasta, canned food, fast foods, and mayonnaise; the healthy dietary pattern composed of fruit, fresh vegetables, low-fat dairy products, nuts, beans, and garlic; and the traditional dietary pattern composed of high-fat dairy products, refined bread, rice, potato, poultry, red meat, and fried foods. Three patterns explained 7.8%, 6.3%, and 6.1% of the whole variance in 38 original food items, respectively (total variance of 20.2%). Factor loadings <0.2 were not reported.

Table 3.

Factor Loadings for Food Items in Three Extracted Factors as Dietary Patterns

Food items/groups	Western	Healthy	Traditional
Butter	0.38
Cream	0.44
Organ meat	0.44
Sheep's head and trotters	0.31
Proceed meat	0.53
Pasta	0.35
Carbonated drinks	0.53
Jams	0.40
Cakes, cookies and sweets	0.38
Seeds	0.36
Industrial fruit juices	0.47
Egg	0.31
Canned Foods	0.34
Fast food	0.55
Restaurant foods	0.11
Mayonnaise	0.40
Non-hydrogenated vegetable oil		0.24
Cheese		0.20
Low-fat milk		0.35
Low-fat yogurt		0.46
Yogurt drink		0.44
Fish		0.37
Fruits		0.43
Fresh fruit juice		0.24
Vegetables and salads		0.50
Nuts		0.43
Beans		0.40
Soy protein		0.28
Garlic		0.37
Hydrogenated vegetable oils			0.26
High-fat milk			0.39
High-fat yogurt			0.51
Red meat			0.50
Poultry			0.39
Bread			0.42
Rice			0.46
Potato			0.42
Fried foods			0.38
% of Variance explained	7.8%	6.3	6.1

The highest factor loading across the factors was reported for each food item. Moreover, assignment of food items to each factor was done based on highest factor loading.

Table 4 summarizes the distribution of sociodemographic, anthropometric, and clinical characteristics of the study participants across tertiles of the extracted dietary patterns. People with the highest adherence to Western dietary pattern were significantly more likely to be single, younger, cigarette smokers, and not have second jobs (P < 0.05). These subjects also had higher mean values of BMI, WC, and WHR anthropometric measurements (P < 0.05). Subjects who had higher levels of healthy dietary pattern consumption were predominantly male, older, married, rotational-shift workers, and without second jobs (P < 0.05). They also were significantly more likely to have an educational level between 6 and 12 years, a larger household size, and a lower WHR value (P < 0.05). Our results also indicated that participants at the highest tertile of traditional dietary pattern adherence were significantly more likely to be male, smokers, rotational-shift workers, have a larger household size, and in the middle-income level (P < 0.05). The values of BMI, WC, and WHR in subjects in the upper tertile were significantly lower than those who were in the lower tertile. The prevalence of hyperlipidemia was significantly higher in subjects in the upper tertile of the Western and traditional dietary patterns (P < 0.05), while its prevalence was significantly lower in subjects in the upper tertile of the healthy dietary pattern (P < 0.05). Moreover, the prevalence of hypertension was significantly higher in subjects who were in the upper tertile of the Western dietary pattern (P < 0.05).

Table 4.

Distribution Characteristics of Study Participants Across Tertiles of Dietary Pattern Scores

Variables	Western dietary pattern				Healthy dietary pattern				Traditional dietary pattern
Variables	Tertile 1	Tertile 2	Tertile 3	P value	Tertile 1	Tertile 2	Tertile 3	P value	Tertile1	Tertile 2	Tertile 3	P value*
Age (years)	38.7 ± 7.4	36.8 ± 7.0	34.7 ± 6.9	<0.001	36.3 ± 7.47	36.5 ± 7.2	37.2 ± 7.2	0.027	36.9 ± 7.2	36.8 ± 7.4	36.5 ± 7.2	0.458
Gender
Male	90.6	91.6	92.5	0.31	90	91.4	93.3	0.027	89.6	92.3	92.8
Female	9.4	8.4	7.5		10	8.6	6.7		10.4	7.7	7.2
Marital status
Married	92.7	92.6	85.1	<0.001	88.3	90.6	91.5	0.049	89.2	90	91.2
Single	7.3	7.4	14.9		11.7	9.4	8.5		10.8	10	8.8
Household size	3.7 ± 1.1	3.59 ± 1.1	3.6 ± 1.1	0.052	3.5 ± 1.1	3.6 ± 1.1	3.7 ± 1.1	0.010	3.5 ± 1.0	3.6 ± 1.1	3.7 ± 1.1	0.008
Education level
0–5 (years)	9.5	9	6.6	0.144	9.5	6.9	8.7	0.004	7.6	8.6	8.9
6–12 (years)	62.1	61.2	63.4		58.4	62.9	65.4		60.9	61	64.8
>12 (years)	28.4	29.8	30		32.1	30.2	25.9		31.5	30.4	26.3
Ever smoked cigarettes (Yes)	23.8	29.5	33.9	<0.001	28.5	28.8	29.9	0.765	28.7	26.6	31.9	0.031
Sleep duration, hr	7.10 ± 1.13	7.15 ± 1.17	7.10 ± 1.18	0.62	7.07 ± 1.18	7.11 ± 1.10	7.17 ± 1.20	0.20	7.02 ± 1.16	7.13 ± 1.15	7.20 ± 1.17	0.002
Physical activity (METs, hour/week)	7.2 ± 3.7	7.4 ± 3.5	7.4 ± 3.7	0.609	7.2 ± 3.5	7.3 ± 3.7	7.5 ± 3.8	0.098	7.3 ± 3.7	7.3 ± 3.5	7.5 ± 3.7	0.453
Effort–reward imbalance ratio	0.7 ± 0.2	0.7 ± 0.2	0.7 ± 0.2	0.932	0.7 ± 0.2	0.7 ± 0.2	0.7 ± 0.2	0.157	0.7 ± 0.2	0.7 ± 0.2	0.7 ± 0.2	0.492
Work shift
Daily	45	46.3	43.6	0.490	51.3	43	40.6	<0.001	49.3	43.8	41.8
Rotating	55	53.7	56.4		48.7	57	59.4		50.7	56.2	58.2
Having a second job
Yes	11.1	8.8	7.9	0.041	7.4	11	9.4	0.020	8	9.4	10.4
No	88.9	91.2	92.1		92.6	89	90.6	92	90.6	89.6
Hypertension (yes)	7.8	11.6	16.8	<0.001	12	12.8	11.4	0.633	32.3	11.4	13.1	0.457
Hyperlipidemia (yes)	38.9	46.2	48.0	0.001	47.1	44.4	41.5	0.035	44.5	40.8	47.3	0.048
BMI (kg/m²)	24.8 ± 3.6	25.6 ± 3.8	26.3 ± 3.8	<0.001	25.4 ± 3.7	25.8 ± 3.8	25.5 ± 3.7	0.108	26.1 ± 3.7	25.6 ± 3.9	24.9 ± 3.7	<0.001
Waist (cm)	88.7 ± 9.5	88.7 ± 9.8	90.4 ± 9.9	<0.001	88.7 ± 9.7	88.1 ± 9.7	88.0 ± 10.0	0.059	90.1 ± 9.5	88.6 ± 9.9	86.9 ± 9.8	<0.001
Waist–hip ratio	0.9 ± 0.1	0.9 ± 0.1	0.9 ± 0.7	<0.001	0.9 ± 0.1	0.9 ± 0.1	0.9 ± 0.1	0.047	0.9 ± 0.6	0.9 ± 0.1	0.9 ± 0.1	0.001

Values are mean ± SD or number.

P-values resulted from χ ² test and analysis of variance for categorical and continuous variables, respectively.

Table 5 presents the OR along with 95% confidence interval (CI) for OR of the association between three identified dietary patterns and metabolic phenotypes considering the first tertile of consumption as the reference category. Subjects in the third tertile of Western pattern consumption were more likely to present MUHO phenotype than MHN in both crude (OR = 3.67, 95% CI 2.68–5.03) and fully adjusted models (OR = 2.87, 95% CI 2.05–4.03). High adherence to Western pattern was also associated with higher odds of MHO (OR = 1.58, 95% CI 1.29–1.94) and MUHN (OR = 1.93, 95% CI 1.08–3.45) phenotypes in fully adjusted models.

Table 5.

Crude and Multivariable Adjusted ORs (95% CI for OR) of the Association between Dietary Patterns and Metabolic Phenotypes

Variables	Western dietary pattern			Healthy dietary pattern			Traditional dietary pattern
Model	T1	T2	T3	T1	T2	T3	T1	T2	T3
Crude
MHN	1	1	1	1	1	1	1	1	1
MHO	1	1.45 (1.19–1.73)	1.73 (1.43–2.10)	1	0.99 (0.82–1.20)	0.81 (0.67–0.98)	1	1.27 (1.05–1.53)	1.76 (1.45–2.13)
MUHN	1	1.18 (0.64–2.18)	2.54 (1.47–4.39)	1	1.21 (0.69–2.11)	1.00 (0.57–1.77)	1	1.05 (0.62–1.79)	1.02 (0.58–1.81)
MUHO	1	1.90 (1.36–2.64)	3.67 (2.68–5.03)	1	1.00 (0.75–1.35)	0.82 (0.61–1.10)	1	1.26 (0.93–1.71)	2.05 (1.53–2.76)
Model I
MHN	1	1	1	1	1	1	1	1	1
MHO	1	1.30 (1.08–1.58)	1.51 (1.24–1.84)	1	1.02 (0.84–1.23)	0.85 (0.71–1.04)	1	1.27 (1.05–1.54)	1.82 (1.49–2.21)
MUHN	1	1.01 (0.55–1.88)	1.99 (1.14–3.47)	1	1.28 (0.73–2.25)	1.09 (0.62–1.92)	1	1.07 (0.63–1.83)	1.08 (0.61–1.91)
MUHO	1	1.53 (1.09–2.16)	2.58 (1.86–3.58)	1	1.08 (0.80–1.46)	0.91 (0.67–1.24)	1	1.25 (0.91–1.72)	2.16 (1.59–2.94)
Model II
MHN	1	1	1	1	1	1	1	1	1
MHO	1	1.27 (1.05–1.55)	1.58 (1.29–1.94)	1	1.01 (0.83–1.24)	0.89 (0.73–1.08)	1	1.35 (1.11–1.64)	1.91 (1.56–2.34)
MUHN	1	1.03 (0.55–1.98)	1.93 (1.08–3.45)	1	1.20 (0.68–2.11)	0.99 (0.54–1.74)	1	1.06 (0.61–1.83)	1.01 (0.56–1.84)
MUHO	1	1.44 (1.01–2.07)	2.87 (2.05–4.03)	1	1.10 (0.81–1.51)	0.98 (0.71–1.35)	1	1.31 (0.94–1.83)	2.33 (1.69–3.22)

Model I, adjusted for demographic variables of age, sex, marital status, and education levels.

Model II, further adjusted for lifestyle variables of sleep duration, physical activity [MET (hr/week)], household size and effort–reward imbalance ratio.

Multivariable logistic regression model was used to estimate ORs with 95% CIs for OR for the association of the metabolic phenotypes and dietary patterns. All comparisons were made compared to the first tertile as reference category in each pattern. Among metabolic phenotypes, the MHN phenotype was used as a reference in all analyses.

P-values for trend were obtained from Mantel–Haenszel χ ² test. T1, T2, and T3 are first to third tertiles of dietary patterns, respectively.

CI, confidence interval; MHN, metabolically healthy normal weight; MHO, metabolically healthy obese; MUHN, metabolically unhealthy normal weight; MUHO, metabolically unhealthy obese; OR, odds ratio.

The traditional pattern was also associated with MUHO phenotype in a positive way, with a 105% increase in the risk of this phenotype occurrence in the third tertile of consumption (OR = 2.05, 95% CI 1.53–2.76) in the crude model and 133% increase in the fully adjusted model (OR = 2.33, 95% CI 1.69–3.22). This pattern was also associated with the MHO phenotype in both crude (OR = 1.82, 95% CI 1.49–2.21) and adjusted models (OR = 1.76, 95% CI 1.45–2.13). However, we did not observe a straightforward trend in the association of traditional dietary pattern with MUHN.

According to healthy dietary pattern, a significant association was only observed between higher adherence to healthy dietary pattern and 19% decreased odds of MHO phenotype in terms of the crude model (OR = 0.81, 95% CI 0.67–0.98). No significant association was found between healthy dietary pattern and other studied metabolic phenotypes among the study population.

Discussion

To the best of our knowledge, this is the first study with a large sample size evaluating the association between dietary patterns and metabolic phenotypes among Iranian industrial employees using a comprehensive and advanced statistical method. Recently, socioeconomic transition coupled with an increased prevalence of obesity, in addition to poor metabolic status, may be attributed to several chronic conditions.^24,25 It is, therefore, important to modify metabolic health, particularly through diet and lifestyle changes. Due to the fact that foods and nutrients are eaten together, a nutritional analysis focusing on separate food or nutrients makes it difficult to assess the impact of dietary intakes on health outcomes.²⁶ There is a biological interaction among foods and nutrients; it is thus more informative to consider the effects of whole dietary patterns on nutrition-related diseases.²⁷

As mentioned in the Results section, three major dietary patterns were observed in this population: Western, healthy, and traditional. Although, the Western and traditional dietary patterns both shared high-calorie food items, the Western dietary pattern contained more fast food, canned food, and processed meat, and the traditional dietary pattern contained more red meat and nonprocessed food items. According to the results, the prevalence of MHN, MUHN, MHO, and MUHO was 41%, 3%, 45%, and 11%, respectively, indicating that more than half of the obese subjects were metabolically healthy in the present study. Similar to our results, in the study conducted on Brazilian adults,²⁸ the MHN phenotype was the most prevalent (44%). However, MUHO and MHO phenotypes presented similar prevalence of 23% and MUHN was identified in 10% of this population. In contrast, in another study of Iranian adults,²⁹ 40% of the subjects presented MHO phenotype, followed by 33% MHN, 23% MUHO, and 4% MUHN phenotype. The discrepancy between metabolic phenotypes in different populations may be due to the different definitions of metabolic phenotypes and the variations in dietary and lifestyle habits.^3,9 The results indicated a positive association between the identified Western and traditional dietary patterns with obesity phenotypes (MHO and MUHO). However, there was no significant association between healthy pattern and any of the evaluated phenotypes. All associations were independent of other lifestyle factors. Our results indicated a significant association between some sociodemographic indicators and adherence to various dietary patterns. Thus, sociodemographic variables should be considered in lifestyle intervention programs to better outcomes achievement.

Our results showed that the Western dietary pattern, which was mainly composed of high consumption of meat, sweets, and saturated fats, was significantly associated with the increased odds of all metabolic phenotypes with respect to metabolic status and obesity in ESCO employees. This result is in line with previous studies indicating the association between higher adherence to Western pattern and metabolic abnormalities.^30
–32 A recent meta-analysis indicated that higher adherence to the Western dietary pattern is associated with a 19% increased risk of an unhealthy metabolic profile.³³ The negative impacts of the Western dietary pattern on metabolic status could be attributed to the higher intakes of proinflammatory foods and the lower intake of beneficial foods and nutrients contained in this pattern. Higher intake of refined grains³⁴ and saturated fat³⁵ in this dietary pattern trigger proinflammatory responses, and inflammation is the leading cause of metabolic abnormalities.³⁶ Also the main contents of Western dietary pattern are high-energy foods which can promote weight gain followed by endothelial dysfunction and metabolic disturbances.³⁷

Our results also indicated a significant positive association between traditional dietary pattern and unhealthy metabolic phenotypes in both normal weight and obese subjects, mediated by higher intake of high-fat dairy products, rice, bread, and hydrogenated fats among participants with higher adherence to this dietary pattern. Our study demonstrated a positive association between Iranian traditional pattern and metabolic status and weight status in the participants, but the two previous studies conducted in Iran did not show any association between the traditional dietary pattern and metabolic status.^30,38 This discrepancy may be attributed to the complex nature of this dietary pattern containing both healthy (whole grains and poultry) and unhealthy (hydrogenated fats, refined grains, and potatoes) foods. Although the unhealthy food items of this pattern have unfavorable effects on the metabolic status, the healthy items have protective effects on the metabolic markers.^30,35 Thus, the food items selected from traditional pattern is paramount to health outcome improvements, especially in at-risk populations.

In this study, there was a decreasing trend in odds of unhealthy metabolic phenotypes along with higher adherence to healthy dietary pattern, but it was not statistically significant. Despite the proven effects of healthy dietary patterns on healthy metabolic profile, our results were in agreement with some previous studies, suggesting no association between healthy dietary pattern and metabolic syndrome.^39,40 In the present study, there was no significant difference between most of the metabolic parameters with respect to the tertiles of healthy dietary pattern. This initial lack of differences may interfere with the true impacts of healthy dietary pattern on metabolic status in the study population. Also, it is likely that the consumption of fruits and vegetables was lower than the recommendations in this population, leading to the distortion of results.

Several limitations should be considered in the interpretation of the current study’s results. The lack of data on portion sizes in the 38-item FFQ resulted in the lack of adjustment for energy and other nutrients confounders in the association between dietary patterns and metabolic outcomes. However, using a simplified FFQ without portion size questions in this large population might decrease the respondents’ burden and increase data completion. Also, our study population mainly consisted of males, and, therefore, the limited number of females may not allow us to identify true associations, and generalize the findings to the whole population particularly women workers. Finally, the cross-sectional nature of this study was another limitation of our study. However, a large sample of participants with relatively high job stress degrees could be considered as the main strength of this study. Also, the adjustment of a wide variety of sociodemographic, lifestyle, and job-related confounders was another strength of this study.

In conclusion, unhealthier dietary patterns with higher consumption of refined and processed foods were associated with an increased risk of having an unhealthy metabolic and weight profile. Accordingly, future nutritional strategies targeting metabolic criteria and obesity are recommended to improve metabolic phenotype and related health outcomes among Iranian industrial workers.

Footnotes

Authors' Contributions

A.F. conceptualized the idea and supervised the current secondary study under the framework of ESCOME study. A.F. analyzed the data. S.A.T contributed in preparation of the current study’s proposal. S.G.H. and S.J.N. prepared the first draft of the manuscript and did the revisions during review process. H.R conceptualized the idea of the main study (ESCOME). H.R and N.S. supervised ESCOME study. All authors read and approved the final version of article.

Author Disclosure Statement

No conflicting financial interests exist.

Funding Information

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

Grundy

, Brewer

Jr. , Cleeman

, et al. Definition of metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation, 2004; 109(3):433–438; doi: 10.1161/01.CIR.0000111245.75752.C6

Kimokoti

, Judd

, Shikany

, et al. Food intake does not differ between obese women who are metabolically healthy or abnormal. J Nutr, 2014; 144(12):2018–2026; doi: 10.3945/jn.114.198341

Pajunen

, Kotronen

, Korpi-Hyövälti

, et al. Metabolically healthy and unhealthy obesity phenotypes in the general population: The FIN-D2D Survey. BMC Public Health, 2011; 11:754; doi: 10.1186/1471-2458-11-754

Blüher

. The distinction of metabolically ‘healthy’ from ‘unhealthy’ obese individuals. Curr Opin Lipidol, 2010; 21(1):38–43; doi: 10.1097/MOL.0b013e3283346ccc

Hinnouho

, Czernichow

, Dugravot

, et al. Metabolically healthy obesity and risk of mortality: Does the definition of metabolic health matter? Diabetes Care, 2013; 36(8):2294–2300; doi: 10.2337/dc12-1654

Sims

. Are there persons who are obese, but metabolically healthy? Metabolism, 2001; 50(12):1499–1504; doi: 10.1053/meta.2001.27213

Phillips

. Metabolically healthy obesity across the life course: Epidemiology, determinants, and implications. Ann N Y Acad Sci, 2017; 1391(1):85–100; doi: 10.1111/nyas.13230

Wildman

, Muntner

, Reynolds

, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: Prevalence and correlates of 2 phenotypes among the US population (NHANES 1999-2004). Arch Intern Med, 2008; 168(15):1617–1624; doi: 10.1001/archinte.168.15.1617

Goday

, Calvo

, Vázquez

, et al. Prevalence and clinical characteristics of metabolically healthy obese individuals and other obese/non-obese metabolic phenotypes in a working population: Results from the Icaria study. BMC Public Health, 2016; 16:248; doi: 10.1186/s12889-016-2921-4

10.

Xia

, Dong

, Gong

, et al. Association between indices of body composition and abnormal metabolic phenotype in normal-weight Chinese adults. Int. J. Environ. Res, 2017; 14(4); doi: 10.3390/ijerph14040391

11.

Fung

, Rimm

, Spiegelman

, et al. Association between dietary patterns and plasma biomarkers of obesity and cardiovascular disease risk. Am J Clin Nutr, 2001; 73(1):61–67; doi: 10.1093/ajcn/73.1.61

12.

Neumann

, Martins

, Marcopito

, et al. Dietary patterns associated with risk factors for cardiovascular disease in a Brazilian city. Rev Panam Salud Publica, 2007; 22(5):329–339; doi: 10.1590/s1020-49892007001000006

13.

Shin

, Kim

, Kang

, et al. Effect of fruits and vegetables on metabolic syndrome: A systematic review and meta-analysis of randomized controlled trials. Int J Food Sci Nutr, 2015; 66(4):416–425; doi: 10.3109/09637486.2015.1025716

14.

Australia Y.B. Aust N Z J Stat. Canberra, Australia. 2008; 161.

15.

. Dietary pattern analysis: A new direction in nutritional epidemiology. Curr Opin Lipidol, 2002; 13(1):3–9; doi: 10.1097/00041433-200202000-00002

16.

Moeller

, Reedy

, Millen

, et al. Dietary patterns: Challenges and opportunities in dietary patterns research an Experimental Biology Workshop, April 1, 2006. J Am Diet Assoc, 2007; 107(7):1233–1239; doi: 10.1016/j.jada.2007.03.014

17.

Jacques

, Tucker

. Are dietary patterns useful for understanding the role of diet in chronic disease? Am J Clin Nutr, 2001; 73(1):1–2; doi: 10.1093/ajcn/73.1.1

18.

Roohafza

, Feizi

, Tirani

, et al. Epidemiological survey of chronic diseases on manufacturing employees (ESCOME): Design and methodology of a large cross-sectional study. Clin. Epidemiology Glob. Health, 2020; 8(1):275–279; doi: 10.1016/j.cegh.2019.08.012

19.

Mohammadifard

, Sajjadi

, Maghroun

, et al. Validation of a simplified food frequency questionnaire for the assessment of dietary habits in Iranian adults: Isfahan Healthy Heart Program, Iran. ARYA Atheroscler, 2015; 11(2):139.

20.

Klishadi

, Khosravi

, Famouri

, et al. Assessment of physical activity of adolescents in Isfahan. J Shahrekord Univ Med Sci, 2001; 3(2) http://eprints.skums.ac.ir/id/eprint/5577

21.

Yadegarfar

, Alinia

, Hassannezhad

, et al. Validation and localization of Farsi version of effort–reward imbalance questionnaire to measure job stress among employees of Isfahan Polyacryle Corporation. Iran. J. Epidemiology, 2012; 8(3):73–83.

22.

Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA, 2001; 285(19):2486–2497; doi: 10.1001/jama.285.19.2486

23.

Kim

J-O

, Mueller

. Factor analysis: Statistical methods and practical issues. Vol 14: sage; 1978.

24.

Ghassemi

, Harrison

, Mohammad

. An accelerated nutrition transition in Iran. Public Health Nutr, 2002; 5(1a):149–155; doi: 10.1079/PHN2001287

25.

Galal

. Nutrition-related health patterns in the Middle East. Asia Pac. J. Clin. Nutr, 2003; 12(3).

26.

Kant

. Dietary patterns and health outcomes. J Am Diet Assoc, 2004; 104(4):615–635; doi: 10.1016/j.jada.2004.01.010

27.

Millen

, Quatromoni

, Pencina

, et al. Unique dietary patterns and chronic disease risk profiles of adult men: The Framingham nutrition studies. J Am Diet Assoc, 2005; 105(11):1723–1734; doi: 10.1016/j.jada.2005.08.007

28.

Pereira

, Juvanhol

, Silva

, et al. Dietary patterns and metabolic phenotypes in Brazilian adults: A population-based cross-sectional study—ERRATUM. Public Health Nutr, 2019; 22(18):3377–3383; doi: 10.1017/S1368980019002593

29.

Nikniaz

, Abbasalizad Farhangi

, Tabrizi

, et al. Association of major dietary patterns and different metabolic phenotypes: A population-based study of northwestern Iran. BMC Endocr Disord, 2019; 19(1):131; doi: 10.1186/s12902-019-0455-3

30.

Esmaillzadeh

, Kimiagar

, Mehrabi

, et al. Dietary patterns, insulin resistance, and prevalence of the metabolic syndrome in women. Am J Clin Nutr, 2007; 85(3):910–918; doi: 10.1093/ajcn/85.3.910

31.

Lutsey

, Steffen

, Stevens

. Dietary intake and the development of the metabolic syndrome: The atherosclerosis risk in communities study. Circulation, 2008; 117(6):754–761; doi: 10.1161/CIRCULATIONAHA.107.716159

32.

Arisawa

, Uemura

, Yamaguchi

, et al. Associations of dietary patterns with metabolic syndrome and insulin resistance: A cross–sectional study in a Japanese population. J Med Invest, 2014; 61(3-4):333–344; doi: 10.2152/jmi.61.333

33.

Fabiani

, Naldini

, Chiavarini

. Dietary patterns and metabolic syndrome in adult subjects: A systematic review and meta-analysis. Nutrients, 2019; 11(9):2056; doi: 10.3390/nu11092056

34.

Esmaillzadeh

, Mirmiran

, Azizi

. Whole-grain consumption and the metabolic syndrome: A favorable association in Tehranian adults. Eur J Clin Nutr, 2005; 59(3):353–362; doi: 10.1038/sj.ejcn.1602080

35.

Freire

, Cardoso

, Gimeno

, et al. Dietary fat is associated with metabolic syndrome in Japanese Brazilians. Diabetes Care, 2005; 28(7):1779–1785; doi: 10.2337/diacare.28.7.1779

36.

Deng

, Shivappa

, Tang

, et al. Association between diet-related inflammation, all-cause, all-cancer, and cardiovascular disease mortality, with special focus on prediabetics: Findings from NHANES III. Eur J Nutr, 2017; 56(3):1085–1093; doi: 10.1007/s00394-016-1158-4

37.

Rodrigues

PRM

, Pereira

, Cunha

, et al. Fatores associados a padrões alimentares em adolescentes: Um estudo de base escolar em Cuiabá, Mato Grosso. Rev Bras Epidemiol, 2012; 15(3):662–674; doi: 10.1590/S1415-790X2012000300019

38.

Falahi

, Anbari

, Ebrahimzadeh

, et al. Dietary patterns associated with metabolic syndrome: The Khorramabad Study. J. Food Secur, 2017; 2(3):201–212.

39.

Akter

, Nanri

, Pham

, et al. Dietary patterns and metabolic syndrome in a Japanese working population. Nutr Metab (Lond), 2013; 10(1):30; doi: 10.1186/1743-7075-10-30

40.

Shokrzadeh

, Aslani

, Rahimi—Foroushani

, et al. Interaction between a variant of vitamin D receptor gene and a posteriori dietary patterns on metabolic syndrome and its components: A nutrigenetic observational study. NFS, 2018; 48(5):780–794; doi: 10.1108/NFS-11-2017-0254