Obesity which is increasing globally is a major risk factor for a wide range of chronic diseases1. The World Health Organization (WHO) has defined overweight and obesity as abnormal or excessive fat accumulation, a body mass index (BMI) ≥ 25 kg/m2 and ≥ 30 kg/m2, respectively2. According to the latest report by the WHO, over 1.9 billion adults were overweight, and of these, 650 million were obese in 20163. Also, in 2016, the prevalence of overweight and obesity was 60.9% and 25.5% in Iran, respectively4,5. The results of several studies showed a higher prevalence of obesity in women. Furthermore, females with a higher BMI are at increased risk for breast cancer, atherosclerotic cardiovascular disease, hypertension, dyslipidemia, type 2 diabetes (T2D), and endocrine disorders6,7,8.

Obesity is commonly defined using BMI, while the evidence shows that this indicator is not a strong predictor of medical risks. Given the complicated function of adipose tissue, the distribution of lipids in different anatomic regions is more important for predicting diseases9. LAP and VAI, novel insulin resistance biomarkers are measured through anthropometric indices and metabolic parameters. LAP is calculated from waist circumference (WC) and fasting concentration of TGs, and VAI is calculated using the combination of BMI, WC, TGs, and high-density cholesterol (HDL)10,11. A systematic review and meta-analysis showed that, LAP is an inexpensive method to evaluate the risk of all-cause mortality, and hypertension. Also, it is an accurate indicator for diagnosing and evaluating diabetes, which can perform better than anthropometric indicators in this field12. Furthermore, another systematic review study reported a strong association between diabetes risk and LAP13.

The evidence has shown that lifestyle changes with diet modification are necessary to prevent obesity and its health outcomes14,15. Given foods and nutrients are consumed together, the dietary pattern approach enables researchers to examine the whole diet16. DDRRs was created by Rhee et al. to indicate a higher consumption of coffee, nuts, cereal fibre, and a high ratio of polyunsaturated fats (PUFA)/saturated fats (SFA), and a lower intake of high glycemic index (GI) foods, sugar-sweetened beverages (SSB), red and processed meats, and trans fatty acids17. While DDRRs includes lower GI foods and higher cereal fibre intake, which are components of a healthy diet and reduce the incidence of overweight and obesity, no previous study has examined the association between DDRRs with overweight and obesity in Iranian adults18.

Given the increasing prevalence of overweight and obesity and the importance of body composition as a key factor for predicting chronic conditions, this study for the first time assessed associations between DDRRs and LAP, VAI, and SMM in overweight and obese Iranian women.

Materials and methods

Study population

This cross-sectional study used the multistage random sampling method and included 291 women aged 18–48 years old from 20 Tehran Health Centers in 2018. Indeed, 20 health centers were randomly selected from all health centers of the Tehran University of medical sciences (Fig. 1). The women who referred to Tehran health centers, if met the inclusion criteria, were randomly recruited to enter the study. The inclusion criteria were: consent to participate in the study, general health and not having a history of chronic disease mentioned in the exclusion criteria, and having BMI between ≥ 25 and ≤ 40. Exclusion criteria were regular use of oral contraceptives, medicines, and supplements including weight loss supplements or medication for blood lipids, blood sugar, and blood pressure reduction, diagnosed with diabetes mellitus, hypertension (HTN), impaired renal function, cardiovascular diseases (CVDs), and impaired liver function, smoking, alcohol consumption, pregnancy, lactation period, menopause, and the history of weight loss in recent years. Furthermore, participants who did not answer more than 70 questions of the semi-quantitative FFQ and reported daily energy intakes over 4200 kcal/day or lower than 800 kcal/day were excluded19. The protocol of this study was approved by the ethics committee of the Tehran University of Medical Sciences (IR.TUMS.VCR.REC.1395.1597). All methods were performed in accordance with the relevant guidelines and regulations, and all participants were fully informed about the study protocols and signed an informed consent form before participating. The sample size was computed according to the following formula: where β = 0.95 and α = 0.05, then, with 95% confidence and 95% power, and r = 0.37.

$${\text{n}} = [({\text{Z1 }}{-} \, \alpha + {\text{Z 1}} - \beta ) \times \surd {1} - {\text{r}}^{{2}} ]/{\text{r}})^{{2}} + {2})$$
Figure 1
figure 1

Flow chart of subjects’ enrolment.

Sociodemographic characteristics

A demographic questionnaire was used to collect information on the medical history and current use of medications and supplement history, smoking habits, age, education, occupation, and marital status. The participant's level of physical activity was assessed using a validated international physical activity questionnaire (IPAQ)20. According to the IPAQ scoring criteria, physical activity was categorized into three levels: low (< 600 MET-min/week), moderate (≥ 600, < 3000MET-min/week), and high (≥ 3000 MET-min/week)20.

Dietary intake assessment

A 147-item semi-quantitative FFQ was used to assess the usual dietary intake of participants. The validity and reliability of the FFQ have been previously demonstrated21. This questionnaire was completed by a trained dietitian. Participants reported the frequency of intake of a given serving of each food item over the last year on a daily, weekly, monthly, or yearly basis. Portion sizes of the food groups were converted to grams using household measurements, and individual’s dietary intake data were analyzed using the Nutritionist IV software22.

Calculation of dietary diabetes risk reduction score

The DDRRs comprises eight components including a higher intake of cereal fibre, nuts, coffee, and PUFA to SFA ratio (P:S) and lower intake of red or processed meats, SSBs, trans fatty acids, and high GI foods17. To calculate the DDRRs, individuals were classified into quartiles according to their intake. For cereal fibre, nuts, coffee, and P:S ratio, the score range was between 1 and 4 assigned to the lowest and the highest intake, respectively. On the contrary, for red or processed meats, SSBs, trans fats, and high GI foods, scores between 1 and 4 were assigned to the highest and the lowest intake, respectively. The score of every component was summed up to calculate the total DDRRs score. The total DDRRs ranged between 8 (the lowest adherence) and 32 (the highest adherence)23. The DDRR score was categorized into tertiles. As a result, the score < 18 was the lowest, ≥ 18 to < 21 was the median, and ≥ 21 was the highest.

Body composition

The body composition was measured using a multi-frequency BIA (InBody720, South Korea, the reliability of our BIA test–retest in our laboratory is r = 0.98) after 12 h of overnight fasting and according to the manufacturer's protocol precautions24. Participants were asked to remove extra clothes, including coat, sweater, shoes, and metal utensils/jewelry, such as rings, watches, and also avoid unusual physical activity for 72 h prior to the assessment. The body composition indicators including BMI, fat mass (FM), fat-free mass (FFM), BF%, visceral fat (VF), waist-to-hip ratio (WHR), bone mineral content (BMC), SMM, skeletal lean mass (SLM), FMI, lean trunk (LT), intracellular water (ICW) and extracellular water (ECW) were also measured25.

Anthropometric indices

Anthropometric indices including weight, height, WC, and hip circumference (HC) were measured for each participant by a trained dietitian. Weight was measured using BIA, and height was measured with an accuracy of 0.1 cm using a Seca scale 206 while participants were in a standing position without shoes. WC was measured in the narrowest area of the waist and on bare skin without any pressure on the body, at the end of the natural exhalation, using a non-elastic tape with an accuracy of 0.5 cm. Using a strapless tape on the most prominent part that was marked, we measured the HC with an accuracy of 0.5 cm. To measure the arm circumference (AC), it was kept in a contracted position in line with the body and the elbow was bent 90° upwards, then its most prominent part was measured using a caliper. WHtR was calculated as WC (cm) divided by height (cm). All measurements were taken in  morning before breakfast and were performed by one person to reduce the measurement errors.

LAP and VAI equations

VAI was calculated using sex-specific formulas, where both TGs and HDL levels are expressed in mmol/L10.

$$\begin{aligned} & {\text{Males}}: \, \left( {{\text{WC}}/{39}.{68} + \left[ {{1}.{88} \times {\text{BMI}}} \right]} \right) \times \left( {{\text{TGs}}/{1}.0{3}} \right) \times \left( {\left[ {{1}.{31}/{\text{HDL}}} \right]} \right). \\ & {\text{Females}}: \, \left( {{\text{WC}}/{36}.{58} + \left[ {{1}.{89} \times {\text{BMI}}} \right]} \right) \times \left( {{\text{TGs}}/0.{81}} \right) \times \left( {{1}.{52}/{\text{HDL}}} \right). \\ \end{aligned}$$

LAP was calculated as (WC/65) × TG in men, and (WC/58) × TG in women26.

Blood sampling

Participants in this study were referred to the Nutrition and Biochemistry Laboratory of the School of Nutritional and Dietetics at Tehran University of medical sciences. After fasting for 10–12 h, 12 cm3 of venous blood samples were taken. Blood samples were collected in two tubes (one tube contained EDTA anticoagulant while another tube lacked this substance). The blood was centrifuged for 15 min at 3000 rpm, and the remaining blood was washed three times with 0.9% NaCl solution. Following serum separation, it was kept at − 80 °C for laboratory assessments.

Blood pressure assessment and laboratory measurements

Before the blood pressure measurement, participants were asked about their intake of coffee and tea, as well as recent physical activity. Blood pressure was measured using a standard mercury sphygmomanometer, with appropriate cuffs, after 15 min of resting. A mean of two measurements was calculated for each individual27. The serum fasting glucose concentration was measured using an enzymatic colourimetric method with the glucose oxidase technique. The insulin level was assessed using the enzyme-linked immunosorbent assay (ELISA) kit (Human insulin ELISA kit, DRG Pharmaceuticals, GmbH, Germany). Serum TG level was measured using the glycerol-3-phosphate oxidase phenol 4-amino antipyrine peroxidase (GPO-PAP) method. ALT and AST were measured based on the standard protocols. Total cholesterol (CHOL) levels were assessed based on the enzymatic endpoint method. Low-density lipoprotein-cholesterol (LDL-C) and HDLC were measured using direct enzymatic clearance. All evaluations were performed using Pars Azmoon laboratory kits (Test Pars Inc, Tehran, Iran).

HOMA and ISQUICKI calculations

Insulin resistance was measured using HOMA. The HOMA was calculated according to the following equation: HOMA = [Fasting Plasma Glucose (mmol/L) × Fasting Plasma Insulin (mIU/L)]/22.528. Insulin sensitivity quantitative insulin sensitivity check index (ISQUICKI) was assessed based on the equation: ISQUICKI = 1/[log (fasting insulin) + log (fasting glucose)29.

Statistical analysis

Statistical analysis was performed using the IBM SPSS software version 25.0 (SPSS, Chicago, IL, USA) and P-value < 0.05 was considered statistically significant and 0.05, 0.06, and 0.07 were considered marginally significant. Continuous and categorical variables were reported as means and standard deviations (SD), and number and percentage, respectively. The Kolmogorov–Smirnov test was used to determine the normal distribution of independent continuous variables (P > 0.05). A one-way analysis of variance (ANOVA) test was used to analyze continuous variables and a Chi-square test was used to compare qualitative variables according to tertiles of DDRRS. The analysis of covariance (ANCOVA) test was used to adjust the analysis for confounders and covariates including age, BMI, physical activity, and energy intake. Post-hoc (Bonferroni) analyses were performed to analyse the mean differences in continuous variables across tertiles of DDRRs. Linear regression analysis was used to examine associations between DDRRs and LAP, VAI, SMM, and other body composition components in the crude and adjusted models. The analysis was adjusted for potential confounders including age, energy intake, and physical activity in the first model and further for marital status and economic status in the second model. Findings were reported as Beta (β), standard error (SE), and 95% confidence intervals (CIs).

Ethics approval and consent to participate and consent for publication

Ethics approval for the study protocol was confirmed by The Human Ethics Committee of Tehran University of Medical Sciences (Ethics Number: IR.TUMS.VCR.REC.1398.142). All participants signed a written informed consent that was approved by the Ethics committee.


General characteristics of the study population

The characteristics of participants are presented in Table 1. The mean (SD) of age, BMI, FFM, VAI, and LAP of participants were 36.67 (9.10) years, 31.26 (4.29) kg/m2, 46.52 (5.71) kg, 2.46 (2.28) and 54.05 (41.72), respectively. The majority of participants were married (72.4%) and employed (99.5%).

Table 1 General characteristics according to tertiles of DDRRS in overweight and obese women (n = 291).

General characteristics across DDRRs tertiles

The general characteristics of participants over DDRRs tertiles are shown in Table 1. In the crude model, there was a significant mean difference in age (P = 0.003), physical activity (P = 0.008), TG (P = 0.077), AST (P = 0.033), ALT (P = 0.042), and insulin (P = 0.040) over DDRRs tertiles. After adjustment for potential confounders including age, energy intake, physical activity, and BMI, the mean difference remained significant for all variables (P < 0.05). Furthermore, HOMA-IR (P = 0.063), marriage status (P = 0.009), and economic status (P = 0.061) was significantly associated with DDRRs after controlling for confounding variables (age, energy intake, physical activity, and BMI).

Dietary intake across tertiles of DDRRs

Table 2 represents the intake of nutrients and food groups across tertiles of DDRRs. There was no significant mean difference in macronutrients, including carbohydrate, protein, fat (P > 0.05) over DDRRs tertiles. A significant lower intake of SFA (P < 0.001) across tertiles of DDRRs was observed after adjustment for energy intake.

Table 2 Intake of macronutrients, micro-nutrients, and food groups according to tertiles of DDRRS in overweight and obese women (n = 291).

As shown in Table 2, after controlling for energy intake, there was a significant difference in the mean of potassium (P < 0.001), Β-carotene (P = 0.005), iron (P < 0.001), vitamin B6 (P = 0.049), folate (P < 0.001), biotin (P = 0.007), phosphor (P = 0.068), copper (P < 0.001), manganese (P < 0.001), chromium (P = 0.066), total fibre (P < 0.001), and caffeine (P = 0.027) over tertiles of DDRRs.

After adjustment for energy intake, participants with the highest tertile of DDRR score had a higher intake of whole grain, vegetables, nuts, legumes, tea and coffee (P < 0.001) and a lower intake of SSB (P < 0.001), compared to those in the lowest tertile.

Anthropometric indices, VAI, and LAP across DDRRs tertiles

The association between anthropometric indices including BFM, FFM, VFA, SMM, VAI, and LAP over tertiles of DDRRs was presented in Table 3. While no significant mean difference in the crude model was observed, after controlling for confounders including age, energy intake, physical activity, marriage, and economic status, significant mean differences for VAI (P = 0.016) and LAP (P = 0.041) across tertiles of DDRRs were found. The results from Bonferroni posthoc test showed that the mean of VAI and LAP was higher in the first tertile compared to the second tertile.

Table 3 Primary outcomes including VAI, LAP, and muscle-mass across DDRRS tertiles in overweight and obese women (n = 291).

Associations between DDRRs and VAI, LAP, SMM, and anthropometric indices

The association between DDRRs and anthropometric indices is shown in Table 4. In the crude model, a significant positive association between DDRRs and VAI in tertile 2 (β: 0.96, 95% CI: 0.08, 1.83, P = 0.031) and a marginal inverse association between DDRRs and SLM in tertile 3 (β: − 1.49, 95% CI: − 2.99, 0.01, P = 0.053) was found. However, the significant association disappeared after adjustment for confounders (age, energy intake, physical activity, marital status, and economic status) in model 2. There was no significant association between DDRRs and LAP, trunk fat, BFM, FFM, SMM, BF%, WHR, VFA, VFL, FFMI, and FMI in the crude model (P > 0.05). However, after controlling for potential confounders in model 2, a negative association was found between DDRRs and trunk fat (P-value = 0.024), BFM (P-value = 0.023), BF% (P-value = 0.045), VFA (P-value = 0.0.26), and FMI (P-value = 0.045). There was no significant association between DDRRs and LAP, FFM, SMM, WHR, and VFL (P > 0.05). Furthermore, VAI (Ptrend = 0.052) and LAP (Ptrend = 0.069), TF (kg) (Ptrend = 0.27), TF% (%) (Ptrend = 0.074), BFM (Ptrend = 0.026), WHR (Ptrend = 0.066), VFA (Ptrend = 0.026), VFL (Ptrend = 0.064), FMI (Ptrend = 0.048) decreased with increasing tertiles of DDRRs (Table 4).

Table 4 Association between DDRRS and VAI, LAP, SMM, and anthropometric variables in overweight and obese women (n = 291).


According to our knowledge, this study is the first study investigated associations between DDRRs and LAP, VAL and SMM in overweight and obese women. According to our findings, there is an inverse and significant association between DDRRs and components of glycemic profiles (insulin, HOMA-IR), lipid profiles (TG), liver function enzymes (ALT, AST), and body composition indices (TF, BFM, FMI, BF%, VFA). Furthermore, body composition indices including VAI, LAP, TF, BFM, WHR, VFA, VFL, and FMI decreased significantly over DDRRs tertiles. However, no significant association was observed between VAI, LAP, and SMM and DDRRs.

The findings of this study showed a significant inverse association between DDRRs and BFM. In accordance with the results of our study, Perry et al. revealed that higher adherence to the DASH-style diet is associated with lower body fat in obese older American adults. The DASH diet was characterized by a higher intake of nuts, whole grains, fruits, vegetables, and legumes and a lower intake of carbonated beverages and red meat that is comparable to the components of DDRRs in this study30.

Our findings showed that the higher DDRRs is associated with a lower level of lipid profiles (serum triglycerides (TGs)), insulin profiles (insulin level and homeostasis model assessment-insulin resistance (HOMA_IR)), liver enzymes (aspartate aminotransferase (AST) and alanine transaminase (ALT)). In line with our findings, previous studies reported that higher adherence to the DASH diet was associated with improved lipid profiles, reduced TG and liver enzymes, and improved glycemic profiles, reduced serum insulin levels and HOMA-IR score31. The existing evidence showed that the Mediterranean diet characterized by a higher intake of healthy food groups including whole grains, MUFA, plant proteins, seafood, fruits, and vegetables, significantly reduced the BFM, which was consistent with the results of our study32,33. Furthermore, in agreement with the findings of this study, the evidence showed that the Mediterranean dietary pattern reduced weight, BMI, WC, fasting insulin levels, HOMA-IR, fatty liver indexes, TG, fasting plasma glucose, AST, and ALT34,35. In addition, in the direction confirming the results of our study, previous studies showed that participants in the lower tertiles compared to those in the higher tertiles of DDRRs, had higher HOMA-IR, triglycerides, and alanine transaminase as well as greater adiposity levels that could be due to higher intake of refined grains, sugary drinks, and saturated and trans-fat and lower intake of whole grains and PUFA36,37.

The higher intake of coffee, nuts, fibre, and PUFAs as components of DDRRs has been individually associated with lower BFM, lipid profiles, glycemic profiles, and liver enzymes. A recent study has reported that daily coffee consumption was inversely associated with BMI, BF%38, VFA39, total abdominal fat39, insulin and insulin resistance40, and levels of ALT and AST41. These associations could be explained through various mechanisms. Coffee comprises various components with pharmacologic effects, including caffeine and chlorogenic acid (CGA)38. Previous evidence revealed that CGA consumption increased postprandial energy expenditure and fat utilization in healthy participants and showed a suppressing effect on the accumulation of body fat39,42,43. There is also a possible explanation that antioxidants in coffee could improve insulin sensitivity and inhibit the induction of liver enzymes40,44,45. Furthermore, caffeine, an important chemical component of coffee, can reduce the risk of Type-2 diabetes and serum triglyceride levels46,47. However, the existing evidence regarding the effect of coffee is mixed. In a study conducted with a larger sample size of both genders in Greek adults, regular coffee consumption was negatively associated with VAL and LAP levels48. A systematic review suggests that adding nuts to habitual diets tends to lower body weight, FM and improve insulin sensitivity46,47. This effect might be explained by the fact that nuts comprise magnesium, linolenic acid, l-arginine, antioxidants, and MUFA may function against inflammation and insulin resistance47. Also, nuts are high-fibre, protein, and low-glycemic food groups, that cause weight loss through increasing satiety49. However, the evidence of the effect of nuts is inconsistent. A recent meta-analysis of randomized controlled trials demonstrated a diet with a higher intake of nuts had no significant impact on adiposity-related measurements compared to the control group50.

This study revealed that higher DDRRs is associated with a lower level of lipid profiles, insulin profiles and liver enzymes. A possible explanation may be that high fibre intake which is one of the components of DDRRs reduces body fat distribution51, lipid profiles51,52, fasting insulin, HOMA-IR score53, and liver function54. Furthermore, the low energy density of insoluble dietary fibre can improve postprandial satiety, lead to weight loss, and improves liver enzymes55,56. On the other hand, soluble fibre can reduce insulin resistance and inflammation53.

Opposite to our findings, which showed no significant association between DDRRs and VAL and LAP, Mazidi et al. reported that higher fibre intake in a healthy dietary pattern is associated with lower levels of VAL and LAP. The conflict results may be due to including a large number of participants from both genders in this study compared to our study, which included only women57.

As mentioned, this study showed that higher DDRRs is associated with lower lipid profiles, insulin profiles, and liver enzymes, which may be related to PUFAs as one of the components of DDRRs. Recent studies demonstrated that a high ratio of PUFA/SFA is associated with lower body fatness58, insulin resistance59, lipid metabolism60, and hepatic enzyme parameters61. Furthermore, it has been suggested that n-3 PUFAs may activate a metabolic change in adipocytes including increased β-oxidation, lipogenesis suppression in abdominal fat62, and inducing apoptosis in the adipose tissue (AT)63. Also, n-3 PUFAs activate the peroxisome proliferator-activated receptor (PPAR) alpha, which in turn stimulates fatty acid oxidation64, and PPAR gamma increases insulin sensitivity65, inhibits hepatic lipogenesis, and reduces hepatic reactive oxygen species66. While a randomized controlled trial study in 2021 showed that omega-3 (n-3 PUFAs) supplementation improved LAP and VAI levels, this study found no significant association which might be due to the fact that our study design was cross-sectional, while their study was a randomized controlled trial on diabetic patients with nonalcoholic fatty liver disease (NAFLD)67. Finally, it is likely that the anti-inflammatory, anti-atherogenic, decreasing visceral adiposity and improving dyslipidemia and hyperinsulinemia effects of DDRRs is due to its components, including antioxidants, vitamins and minerals, phenolic compounds, and unsaturated fatty acid17,23.

The current study has several limitations that should be considered in interpreting the results. Firstly, due to the cross-sectional design, causality cannot be conferred. As a result, further prospective observational studies and randomized clinical trials are needed to confirm the effect of DDRRs on LAP, VAL, and SMM. Secondly, using FFQs can result in under or over-reporting dietary intake. Thirdly, this study included only women; thus, it is impossible to generalize the results to the whole population. Lastly, using the categorical confounders might result in residual confounding. This study also has several strengths. This study is the first to show the link between DDRRs and LAP, VAL and SMM in adult women. This study included a large sample size and the analysis was controlled for various potential confounders.


The findings of this study showed an inverse association between DDRRs and the percentage of BF, VFA, FMI, BFM, TF, serum TG and insulin level, HOMA_IR, AST, and ALT in overweight and obese women. While a higher adherence to DDRRs tertiles was negatively associated with lower VAL and LAP, DDRRs had no significant association with VAL, LAP, and SMM. Further prospective or interventional research is needed to confirm whether the association represents a cause-effect relationship.