Background
Hypertension, as a worldwide health problem, is a main contributing factor to cardiovascular disease (CVD) and premature death, affecting about 1 billion adults globally [
1,
2]. In contradiction of trends reported in the USA and northern Europe [
3], evidence indicates that CVD mortality is elevating in Iran; hypertension is prevalent in 24% of women and 22% of men in this population [
4]. It is anticipated that the burden of hypertension on communities rises continuously along with population ageing. Evidence has demonstrated that age, genetics, ethnicity, gender, and socioeconomic status are risk factors for hypertension [
5]; nevertheless, this disease is modifiable [
6,
7] and well-controlled blood pressure (BP) can enhance quality of life, improve prognosis, and prevent its clinical complications [
8,
9].
The changes in lifestyle, such as physical activity, weight reduction and dietary modifications are the main approaches for the prevention and management of hypertension [
10]. Excessive consumption of dietary sodium, saturated fatty acids, cholesterol, and alcohol are shown to be linked to an elevated risk of hypertension, while intake of vegetables/fruits and foods rich in magnesium, calcium, potassium, and unsaturated fatty acids is reported to reduce blood pressure [
5,
11]. Whereas previous dietary recommendations for the prevention of hypertension focused mainly on single micronutrients or food items, a more effective approach for this purpose is recommendations that cover the entire quality of diet using dietary pattern, considering potential interactions of food items/nutrients [
12‐
16]. The overall quality of diet could be assessed using the a posteriori and the a priori methods, in which the a posteriori approaches derive dietary patterns exploratory according the intake of foods reported by the studied population, but, the a priori approaches evaluate the adherence of participants to a predefined healthy dietary index [
17]. Dietary approach to stop hypertension (DASH) dietary pattern, featured by low intake of salt, saturated fat, red meat, and high intake of low-fat dairy products, fish, nuts, fruits, whole grains, vegetables, magnesium, potassium, calcium, and fiber, is the most effective dietary pattern to reduce BP [
18]. Moreover, evidence shows that higher diet quality assessed by other a priori dietary indices, including the Mediterranean dietary score (MDS), diet quality index-international (DQI-I), healthy eating index-2015 (HEI-2015), and dietary diversity score (DDS) are protective against the risk of CVD [
19‐
22]. Nevertheless, the relation of these indices to hypertension is not well-established.
The Fasa Cohort Study [
23] provided this opportunity to evaluate diet–disease associations from an epidemiological standpoint. The present analysis was conducted to explore the relation of MDS, HEI-2015, DQI-I, and DDS to the risk of hypertension in an Iranian population.
Results
A total of 10,112 individuals (45.14% male), with mean age of 48.63 ± 9.57 years, participated in this study. The prevalence of hypertension among the study population was 28.26% (21.59% in males and 33.74% in females).
It was observed that in men and women with hypertension compared with those with normal blood pressure, the means of age, weight, waist to hip ratio (WHR), BMI, SBP, and DBP were significantly higher, but their education, height, and physical activity were significantly lower. Also, the frequency distribution of individuals in terms of history of diabetes, alcohol consumption, active smoking, taking supplements, and history of ischemic heart disease were significantly higher in men with hypetension compared with men with normal BP. For women with hypertension compared with those with normal blood pressure, the frequency distribution of individuals in terms of history of diabetes, alcohol consumption, active smoking, taking supplements, and history of ischemic heart disease were significantly lower in women with hypetension compared with women with normal BP. Also, in the groups with high blood pressure, compared to the groups with normal blood pressure, the mean score of DDS, MDS, DQI-I, and HEI-2015 was lower (Table
1).
Table 1
Distribution of baseline variables and mean score of dietary indices in men and women with and without hypertension
Age (year)a | 48.63 ± 9.57 | 53.81 ± 9.97 | 47.19 ± 9.00 | ≤0.001 | 53.62 ± 9.16 | 46.11 ± 8.64 | ≤0.001 |
Education (year)a | 4.66 ± 3.88 | 4.99 ± 4.23 | 6.05 ± 4.09 | ≤0.001 | 2.51 ± 2.90 | 4.32 ± 3.42 | ≤0.001 |
Height (cm)a | 161.67 ± 9.65 | 168.16 ± 6.44 | 169.19 ± 7.09 | ≤0.001 | 155.17 ± 6.60 | 155.92 ± 7.29 | ≤0.001 |
Wight (kg)a | 67.02 ± 13.32 | 74.68 ± 13.64 | 67.72 ± 13.37 | ≤0.001 | 67.44 ± 12.86 | 64.04 ± 12.45 | ≤0.001 |
WHRa | 0.93. ±0.06 | 0.94 ± 0.06 | 0.9 ± 0.06 | ≤0.001 | 0.96 ± 0.06 | 0.93 ± 0.06 | ≤0.001 |
BMI(kg/m2)a | 25.64 ± 4.85 | 26.36 ± 4.34 | 23.59 ± 4.26 | ≤0.001 | 27.94 ± 4.94 | 26.28 ± 4.73 | ≤0.001 |
Physical activity(MET)a | 41.47 ± 11.34 | 42.81 ± 13.67 | 45.84 ± 14.42 | ≤0.001 | 37.72 ± 6.49 | 38.77 ± 6.81 | ≤0.001 |
DBP (mmHg)a | 74.65 ± 11.99 | 87.66 ± 11.81 | 70.69 ± 8.85 | ≤0.001 | 83.93 ± 12.37 | 70.28 ± 8.98 | ≤0.001 |
SBP (mmHg)a | 111.36 ± 18.50 | 131.16 ± 18.29 | 104.83 ± 12.30 | ≤0.001 | 127.98 ± 20.12 | 103.94 ± 12.50 | ≤0.001 |
Marital statusb | | | | 0.06 | | | 0.001 |
Non-married | 1118 | 96 (2.7) | 16 (1.6) | | 620 (16.9) | 386 (20.6) | |
Married | 8994 | 3483 (97.3) | 970 (98.4) | | 3055 (83.1) | 1486 (79.4) | |
History of diabetesb | 1244 | 185 (5.2) | 169 (17.1) | ≤0.001 | 400 (10.9) | 490 (26.2) | ≤0.001 |
History of ischemic heart disease b | 1098 | 202 (5.6) | 200 (20.3) | ≤0.001 | 247 (6.7) | 449 (24.0) | ≤0.001 |
Alcohol drinkerb | 210 | 179 (5.0) | 31 (3.1) | 0.01 | 0 (0.0) | 0 (0.0) | – |
Active smokingb | 2735 | 2025 (56.6) | 432 (43.8) | ≤0.001 | 133 (3.6) | 145 (7.7) | ≤0.001 |
Supplement useb | 1647 | 382 (10.7) | 74 (7.5) | 0.003 | 871 (23.7) | 320 (17.1) | ≤0.001 |
Obesity status b | | | | ≤0.001 | | | ≤0.001 |
Underweight BMI<18.4) | 576 | 396 (11.1) | 28 (2.9) | | 128 (3.5) | 24 (1.3) | |
Normal weight (BMI = 18.5–24.9) | 4119 | 1890 (53.0) | 355 (36.2) | | 1371 (37.4) | 503 (27.0) | |
overweight (BMI = 25–29.9) | 3613 | 1033 (29.0) | 417 (42.5) | | 1413 (38.5) | 750 (40.2) | |
obese (BMI ≥ 30) | 5474 | 247 (6.9) | 181 (18.5) | | 757 (20.6) | 589 (31.6) | |
Family history of diabetesb | 4498 | 1397 (39.0) | 430 (43.6) | 0.009 | 1754 (47.7) | 917 (49.0) | 0.37 |
Family history of hypertensionb | 6394 | 2008 (56.1) | 651 (66.0) | ≤0.001 | 2351 (64.0) | 1384 (73.9) | ≤0.001 |
Family history of ischemic heart diseaseb | 5369 | 1699 (47.5) | 456 (46.2) | 0.49 | 2127 (57.9) | 1087 (58.1) | 0.89 |
DQI-I scorea | 54.77 ± 11.66 | 54.62 ± 11.40 | 58.33 ± 11.75 | ≤0.001 | 53.95 ± 11.67 | 54.79 ± 11.76 | 0.01 |
HEI-2015 scorea | 50.00 ± 11.41 | 48.77 ± 10.85 | 51.99 ± 11.54 | ≤0.001 | 49.72 ± 11.57 | 51.83 ± 11.72 | ≤0.001 |
DDS scorea | 4.79 ± 1.96 | 5.20 ± 1.90 | 5.26 ± 1.95 | 0.40 | 4.23 ± 1.91 | 4.51 ± 1.88 | ≤0.001 |
MDSa | 4.50 ± 1.52 | 4.29 ± 1.50 | 4.44 ± 1.50 | 0.008 | 4.58 ± 1.50 | 4.74 ± 1.56 | ≤0.001 |
The characteristics of study participants among different quartiles of dietary indices stratified by hypertension status are reported in supplemental Tables
1,
2,
3 and
4. Both subjects with and without hypertension in the top quartile of MDS, compared with those in the first quartile, were more likely to have higher height, physical activity, be male, and be active smoker, while, they were less educated, and had a fewer prevalence for diabetes and dietary supplements use (
P < 0.05). In contrast to hypertensive subjects, normotensive individuals in the top quartile of MDS had a lower age, family history of diabetes and hypertension, and a healthier profile for obesity measures (weight, WC, BMI, and HC) (
P < 0.05) (supplemental Table
1). Both normotensive and hypertensive individuals with greater adherence to HEI-2015 were younger, had lower WC and obesity rates, diabetes, and ischemic heart disease, in comparison to those from lower levels (P < 0.05). In hypertensive patients, there were more subjects with a family history of hypertension among those with the highest scores of HEI-2015 than those with the lowest scores (
P = 0.04). In contradiction of hypertensive participants, normotensive people in the highest HEI-2015 quartile had lower weight, HC, and BMI, while were more likely to have a family history of diabetes as well as to be physically active, male, and smoker, than those in lower quartiles (
P < 0.05) (supplemental Table
2). For DDS, it was identified that, irrespective to hypertension status, with increase in adherence to DDS, the levels of education, height, obesity measures (weight, WC, BMI, and HC), physical activity, and the frequency of males, smoker and married participants increased, but, age and DBP decreased (
P < 0.05). For subjects without hypertension, higher score for DDS was significantly related to a lower frequency of alcohol drinking and supplement use (P < 0.05). In hypertensive patients, there were less subjects with a family history of ischemic heart disease among those with the highest scores of DDS than those with the lowest scores (
P ≤ 0/001) (supplemental Table
3). Moreover, regardless of hypertension status, increase in DQI-I was significantly related to a lower education level and a reduction in diabetes prevalence, obesity rate and obesity-related parameters (weight, WC, BMI, and HC). It was also observed that in normotensive people, higher score for DQI-I was significantly related to a lower systolic and diastolic BP and a decreased likelihood to be married, but the rate of smoking increased (
P < 0.05). Hypertensive patients with higher adherence to the DQI-I had significantly lower physical activity level, compared with those with a low adherence (supplemental Table
4).
Multivariable-adjusted odds ratio and 95% CI for hypertension across quartiles of dietary indices for the whole population, males, and females are presented in Table
2. In the analysis of the whole population, after controlling for potential covariates including daily energy intake, age, gender, physical activity, smoking, family history of hypertension, BMI, and the level of education, a significant negative association was identified between MDS (OR: 0.86, 95%CI = 0.75–0.99; P-trend = 0.03) and HEI-2015 (OR: 0.79, 95%CI = 0.68–0.90; P-trend ≤0.001) with the risk of hypertension in the highest quartile compared with the lowest quartile. In addition, in the stratified analysis by gender, HEI-2015 remained as a significant preventive dietary approach against hypertension in both males (OR: 0.80, 95%CI = 0.64–0.99; P-trend = 0.05) and females (OR: 0.78, 95%CI = 0.66–0.94; P-trend = 0.04). The relation of MDS to hypertension was disappeared in the fully adjusted model for males and females. However, it remained significant in the model adjusted for daily energy intake and age (males: OR: 0.72, 95%CI = 0.58–0.89; P-trend = 0.008, females: OR: 0.80, 95%CI = 0.68–0.95; P-trend = 0.01). No significant relationship was detected between DQI-I and DDS with the odds of hypertension in the overall analysis and the stratified analysis by gender.
Table 2
Logistic regression analysis for the relationship between diet quality indices and the risk of hypertension in the whole population (N = 10,112), males (N = 4565), and females (N = 5547)
| MDS | Quartile 1 | 1 | | 1 | | 1 | |
Quartile 2 | 0.94 (0.83–1.06) | 0.35 | 0.97 (0.85–1.11) | 0.73 | 1.03 (0.90–1.18) | 0.63 |
Quartile 3 | 0.80 (0.71–0.91) | 0.001 | 0.81 (0.71–0.93) | 0.003 | 0.88 (0.77–1.01) | 0.08 |
Quartile 4 | 0.79 (0.70–0.89) | ≤0.001 | 0.76 (0.67–0.87) | ≤0.001 | 0.86 (0.75–0.99) | 0.03 |
P -trend | | ≤0.001 | | ≤0.001 | | 0.03 |
HEI-2015 | Quartile 1 | 1 | | 1 | | 1 | |
Quartile 2 | 0.53 (0.47–0.60) | ≤0.001 | 0.83 (0.73–0.95) | 0.007 | 0.91 (0.79–1.04) | 0.17 |
Quartile 3 | 0.62 (0.55–0.69) | ≤0.001 | 0.72 (0.63–0.81) | ≤0.001 | 0.80 (0.70–0.91) | 0.001 |
Quartile 4 | 0.75 (0.66–0.85) | ≤0.001 | 0.66 (0.58–0.75) | ≤0.001 | 0.79 (0.68–0.90) | 0.001 |
P -trend | | ≤0.001 | | ≤0.001 | | ≤0.001 |
DDS | Quartile 1 | 1 | | 1 | | 1 | |
Quartile 2 | 0.91 (0.81–1.03) | 0.15 | 1.01 (0.89–1.15) | 0.78 | 0.97 (0.84–1.11) | 0.65 |
Quartile 3 | 0.95 (0.84–1.08) | 0.49 | 1.08 (0.94–1.23) | 0.25 | 0.95 (0.82–1.09) | 0.48 |
Quartile 4 | 0.86 (0.75–0.99) | 0.04 | 1.07 (0.92–1.24) | 0.37 | 0.93 (0.79–1.09) | 0.39 |
P -trend | | 0.19 | | 0.65 | | 0.84 |
DQI-I | Quartile 1 | 1 | | 1 | | 1 | |
Quartile 2 | 0.83 (0.62–1.13) | 0.25 | 0.87 (0.64–1.18) | 0.39 | 0.86 (0.63–1.18) | 0.37 |
Quartile 3 | 0.80 (0.57–1.11) | 0.19 | 0.83 (0.60–1.17) | 0.29 | 0.87 (0.62–1.24) | 0.46 |
Quartile 4 | 0.76 (0.53–1.09) | 0.14 | 0.83 (0.57–1.19) | 0.31 | 0.90 (0.61–1.31) | 0.58 |
P -trend | | 0.41 | | 0.67 | | 0.81 |
Males | MDS | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.92 (0.75–1.13) | 0.47 | 0.94 (0.76–1.17) | 0.61 | 0.99 (0.79–1.24) | 0.97 |
| Quartile 3 | 0.82 (0.67–1.01) | 0.06 | 0.80 (0.64–0.99) | 0.04 | 0.88 (0.71–1.11) | 0.30 |
| Quartile 4 | 0.77 (0.63–0.93) | 0.01 | 0.72 (0.58–0.89) | 0.002 | 0.88 (0.70–1.09) | 0.25 |
| P -trend | | 0.04 | | 0.008 | | 0.50 |
HEI-2015 | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.71 (0.58–0.87) | 0.001 | 0.78 (0.63–0.96) | 0.01 | 0.89 (0.72–1.11) | 0.33 |
| Quartile 3 | 0.52 (0.42–0.63) | ≤0.001 | 0.60 (0.49–0.74) | 0.001 | 0.75 (0.60–0.93) | 0.009 |
| Quartile 4 | 0.48 (0.39–0.59) | ≤0.001 | 0.59 (0.48–0.73) | 0.001 | 0.80 (0.64–0.99) | 0.05 |
| P -trend | | ≤0.001 | | ≤0.001 | | 0.05 |
DDS | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 1.02 (0.82–1.28) | 0.82 | 1.07 (0.87–1.39) | 0.39 | 0.99 (0.77–1.27) | 0.96 |
| Quartile 3 | 1.19 (0.95–1.49) | 0.11 | 1.26 (0.99–1.60) | 0.06 | 1.05 (0.82–1.34) | 0.69 |
| Quartile 4 | 1.08 (0.86–1.37) | 0.47 | 1.25 (0.98–1.59) | 0.06 | 1.00 (0.77–1.30) | 0.97 |
| P -trend | | 0.35 | | 0.16 | | 0.95 |
DQI-I | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.72 (0.42–1.21) | 0.22 | 0.81 (0.47–1.40) | 0.46 | 0.81 (0.45–1.46) | 0.49 |
| Quartile 3 | 0.50 (0.27–0.92) | 0.02 | 0.54 (0.29–1.01) | 0.06 | 0.68 (0.34–1.35) | 0.27 |
| Quartile 4 | 0.69 (0.34–1.39) | 0.30 | 0.81 (0.39–1.68) | 0.58 | 0.96 (0.42–2.20) | 0.94 |
| P -trend | | 0.16 | | 0.30 | | 0.70 |
Females | MDS | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.98 (0.84–1.14) | 0.81 | 0.99 (0.84–1.17) | 0.98 | 1.05 (0.89–1.25) | 0.51 |
| Quartile 3 | 0.82 (0.70–0.96) | 0.01 | 0.83 (0.70–0.98) | 0.03 | 0.89 (0.75–1.07) | 0.22 |
| Quartile 4 | 0.86 (0.73–1.01) | 0.06 | 0.80 (0.68–0.95) | 0.01 | 0.87 (0.73–1.04) | 0.14 |
| P -trend | | 0.04 | | 0.01 | | 0.13 |
HEI-2015 | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.79 (0.68–0.92) | ≤0.001 | 0.87 (0.74–1.03) | 0.11 | 0.92 (0.75–1.09) | 0.35 |
| Quartile 3 | 0.72 (0.62–0.84) | ≤0.001 | 0.80 (0.67–0.94) | 0.008 | 0.84 (0.71–1.008) | 0.06 |
| Quartile 4 | 0.60 (0.51–0.70) | ≤0.001 | 0.71 (0.60–0.83) | 0.001 | 0.78 (0.66–0.94) | 0.008 |
| P -trend | | 0.004 | | 0.001 | | 0.04 |
DDS | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.91 (0.79–1.06) | 0.24 | 1.04 (0.89–1.22) | 0.58 | 0.99 (0.84–1.18) | 0.99 |
| Quartile 3 | 0.88 (0.75–1.03) | 0.13 | 1.05 (0.88–1.24) | 0.56 | 0.93 (0.78–1.12) | 0.47 |
| Quartile 4 | 0.79 (0.66–0.96) | 0.01 | 1.08 (0.88–1.32) | 0.45 | 0.92 (0.74–1.14) | 0.45 |
| P -trend | | 0.11 | | 0.88 | | 0.80 |
DQI-I | Quartile 1 | 1 | | 1 | | 1 | |
| Quartile 2 | 0.89 (0.61–1.28) | 0.54 | 0.91 (0.62–1.31) | 0.61 | 0.89 (0.60–1.30) | 0.56 |
| Quartile 3 | 0.94 (0.63–1.40) | 0.78 | 0.97 (0.64–1.45) | 0.88 | 1.00 (0.66–1.52) | 0.97 |
| Quartile 4 | 0.77 (0.50–1.18) | 0.24 | 0.82 (0.53–1.26) | 0.37 | 0.86 (0.55–1.33) | 0.50 |
| P -trend | | 0.69 | | 0.81 | | 0.84 |
Discussion
In the present study, we assessed the relation of four a priori defined diet quality indices, including MDS, DQI-I, HEI- 2015, and DDS to the risk of hypertension in a large population of Iranian adults. The findings revealed that people in the highest quartile of MDS and HEI-2015, compared with individuals in the lowest quartile, had significantly lower odds for having hypertension, but, no such relationship was observed for DQI-I and DDS. Furthermore, higher adherence to MDS, DQI-I, and HEI-2015 was associated with a significant decrease in obesity-related parameters including weight, WC, BMI, and HC. In contrast, greater score for DDS was related to an increase in obesity rates.
In the studied population of the present study, which included middle-aged/elderly people in Iran (aged from 35 to 70), hypertension was more prevalent among females than men. In line with our study, in a study in India, hypertension was more prevalent among Elderly females than men [
30]. It is worth noting that the mean BMI (26.84 ± 4.86 vs. 24.19 ± 4.42) and WHR (94.88 ± 6.19 vs. 91.54 ± 6.33) were significantly higher in women compared with men in our study. Higher BMI and WHR have been strongly related to increased BP [
31]; this might, in part, justify higher prevalence of hypertension observed in females in this study.
It has been proposed that high-quality diets, such as the MDS and HEI, may prevent development of hypertension in healthy people and could further reduces blood pressure in hypertensive people who use antihypertensive drugs [
9,
32]. A recent meta-analysis found that intervention with the Mediterranean diet for 1 year decreased both SBP and DBP [
33]. Supporting our results, the study by Shim et al. [
9] reported that a high adherence to the Korean version of HEI is favorably related to BP control in men with physician-diagnosed hypertension. Moreover, the cross sectional study of Saraf-Bank et al. [
34] revealed that a greater adherence to HEI-2010 is negatively liked to high BP among Iranian adult women. Nevertheless, the study by Daneshzad et al. [
35] found no association between HEI and mean systolic and diastolic BP. Both MDS and HEI-2015 have many similarities and emphasize on high intake of fish, fruits, vegetables, monounsaturated fat, legumes, whole grains, and low consumption of saturated fatty acids. They also differ in some components such that the MDS scores dairies negatively but HEI-2015 scored diaries positively [
26,
27]. Furthermore, HEI-2015 recommends lowers intakes for added sugars, refined grains, and sodium [
26], which its association with hypertension is established [
36]. These diet are rich in fiber and have a low energy density and low glycemic load, all of which are p[protective against hypertension [
37]. The favorable impacts of vegetables/fruits on hypertension has been also demonstrated [
38]. This relationship could also be attributed to potential healthy constituents such as phytochemicals, fibers, vitamins, magnesium, and potassium and antioxidants in vegetables and fruits [
39], which have been independently related to a reduction in BP [
40‐
42]. The study by Shah et al. [
43] proposed lower inflammation, lower endothelial oxidative stress, and greater endothelial functioning as potential novel underlying mechanisms through which components of MDS might affects vascular changes related to improved cardiovascular risk.
There are very limited evidence regarding the relation of DDS and DQI-I to BP. Some previous studies reported lower systolic or diastolic BP in higher categorizes of DDS and DQI-I [
21,
44]. A cross-sectional study of 82 adults living in Saba Island in the eastern Caribbean Basin identified a significant link between a diet with poor diversity and the increased risk of hypertension [
44]. In contrast, a cross-sectional study on 230 Iranian women with type 2 diabetes identified no significant difference in SBD and DBP across categorizes of DQI-I [
35]. In this study we found that, compared with individuals in the lowest quartile, people in the top quartile of the DDS had significantly lower mean of DBP, and for DQI-I, had a lower mean of both systolic and diastolic BP; nevertheless, multivariable adjusted logistic regression analysis found no significant association between DDS, DQI-I and hypertension. It should be considered that the results of logistic regression analysis is a more reliable as large sample size of the study makes small differences in means as significant; hence this study concluded a null association between DDS, DQI-I and risk of hypertension. The heterogeneity in the results of this study campared with the previous studies might be due to differences in sample size, method used to assess dietary intake, study design, health status of participants, different level of adjustment for confounders, and method of data analysis.
In line with our study, some studies [
4,
21], but not all [
45], reported that higher DDS is related to higher prevalence of obesity. This association is expectable as consumption of a more varied diet is linked to higher nutritional adequacy, greater intake of macronutrients and also energy [
46], proposing that intake of a large number of foods may result in a higher calorie intake, and thus, obesity [
21]. It is notable that the MDS, DQI-I, and HEI-2015 were related to a healthier profile for obesity measures, which has also been observed in some previous investigations [
47‐
49]. These negative relationships with obesity parameters were expected, as these dietary patterns are rich in fiber and have low energy density and low glycemic load, all of which have a protective effect against obesity [
50,
51]. However, the association of these dietary patterns on health outcomes is derived from the interaction of all components and could not be attributed to a single food item. The prospective study by Funtikova et al. [
52] on 2181 participants revealed that, during a 10-year follow-up, a higher DQI score at baseline related to a decrease in WC. Another study on Chinese diabetic patients showed that the HEI, but not DQI-I, score had a negative relationship with obesity [
53]. In a cross-sectional study recruiting 1062 Mexican women, with age between 35 to 69 years, MED was linked with lower WC whereas other a priori dietary patterns (DQI –I and HEI) were not predictors of anthropometric measures [
54]. Moreover, unlike our findings, adherence to HEI, MDS, and DQI-I could not predict BMI and WC in Tehran Lipid and Glucose Study after 6.7 years of follow-up [
55]. The capability of dietary indices to predict obesity measures relies on how well these indices correlate with dietary energy intake as the chief cause of obesity. However, these indices do not assign negative scores to excessive energy intakes; this limitation makes it difficult to reach a consistent conclusion across studies. The discrepancies regarding the relation of diet quality indices to obesity-related parameters may also be derived from differences in genetic background, interaction of gene-environmental factors, and difference in adjustment for potential confounders and disease pattern in various studies.
The strengths of the present study are its large population representative sample derived from a recent nationwide study and consideration of the potential covariates in the analysis. Moreover, another positive point of this study is face to face interview for data collection of FFQ, which makes the data more reliable. Also, conducting the anthropometric measurements instead of self-reporting can be considered as a strength. However, there are several limitations in this study that should be considered when inferring the findings. First, since our study had a cross-sectional design, causal association could not be inferred; although, cross-sectional studies provide valuable data on diet–disease associations. Future longitudinal studies can provide stronger evidence in this regard. Second, FFQ is susceptible to recall bias, such that subjects might under- or overestimate their food intake resulting in misclassification of food intake. As a structured dietary assessment method, the FFQ is less precise than methods of daily intake such as 24-h recalls and food records. Although 24-h food recall is used mostly in calculating HEI-2015, there are several well-designed studies that have used FFQ for calculating HEI-2015 [
56]. Because of severe attenuation, both FFQ and multiple 24-h recalls cannot be recommended as an instrument for evaluating relations between absolute intake of diet and disease. These instruments could be used for detecting diet-disease association when considering the relative dietary intake of participants. Finally, however it was attempted to control for known covariates, the probability of residual confounding could not be omitted in our results.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.