Habitual and meal-speci c carbohydrate quality index and their relation to metabolic syndrome in a sample of Iranian adults

Maryam Majdi Tehran University of Medical Sciences School of Medicine https://orcid.org/0000-0002-9945-7785 Hossein Imani Tehran University of Medical Sciences Elham Bazshahi Tehran University of Medical Sciences Fatemeh Hosseini Tehran University of Medical Sciences Kurosh Djafarian Tehran University of Medical Sciences Azadeh Lesani Tehran University of Medical Sciences Zahra Akbarzade Tehran University of Medical Sciences Sakineh Shab-Bidar (  s_shabbidar@tums.ac.ir ) Tehran University of Medical Sciences


Dietary assessment, meal timing, and CQI calculation
Dietary data were gathered by using repeated but non-consecutive (Monday to Sunday) 24-hour dietary recall method. The 24-hour meal was structured and included breakfast. The rst 24-hour recall is obtained through interviews, the other two recalls are recorded by telephone during two repeated non-random days during the study, and information was recorded. Food and food groups were extracted through these questionnaires. Meals usually include breakfast, lunch, dinner, and snacks. If two or more meals were reported within 59 minutes, they were considered as one meal. Otherwise, mealtimes were coded with more energy content as a meal or as snacks. In some cases, lunch was not served, but dinner and supper were both reported (some people refer to the common concept of a conventional lunch in the United States, a half-day meal, dinner). If dinner and supper were more than an hour apart, and dinner had less energy than supper, dinner would be coded for lunch and supper for dinner. If more than one dinner was reported, the second dinner could not be recorded as lunch unless the rst dinner has already been recorded as lunch to maintain the timing of the meal (31) Finally, the total value of nutrients in all meals and snacks was calculated daily. CQI was calculated based on the energy-adjusted amount of total carbohydrate intake values calculated using the residual method (32) . CQI was de ned by summing up the following four criteria: 1) ratio of solid carbohydrates to avaiable carbohydrates, 2) dietary ber intake (g/day), 3) GI, and 4) ratio of whole grains to total grains (whole grains, re ned grains, and their products). Subjects were categorized into quintiles and take a value (ranging from 1 to 5) for each quintile according to each of these four criteria; however, the scoring of GI was reversed; thus, those in the fth quintile received one point, and those in the rst quintile received ve points. Finally, an overall CQI was computed by adding together all values of the four criteria (ranging from 4 to 20). It was also ranked into quintiles (25) . But Given that the Iranians are using whole grains in the diet is limited, this component is not calculated in this calculation, so the nal score ranges from 3 to 15.
GI values were obtained from international tables (33) , the glycemic index of Iranian foods (34) , and literature reviews. Glucose was used as the reference (GI for glucose= 100). The mean of the GI values was assigned if more than one eligible GI value was available for a speci c food item. The carbohydrate content of each food was determined using standard portion sizes from the United States Department of Agriculture food composition databases (35). All nuts and vegetables except starchy roots, considered as very low GI (ranged from 10 to 20). Solid carbohydrates were obtained by subtracting the amount of liquid carbohydrate (summing up sweetened beverages and fruit juice) from total carbohydrate intake.

Anthropometric assessment and biochemical tests
Weight was measured using a Seca weighing scale (Seca and Co. KG; 22 089 Hamburg, Germany; Model: 874 1321009; designed in Germany; made in China) with light clothing (without a coat and raincoat). A wall stadiometer board was used for the height of participants without shoes with a sensitivity of 0.1 cm height measurements. BMI was calculated as weight (in kilograms) divided by height (in meters squared). Normal weight was de ned as BMI ≤24.9 kg/m2, and the presence of overweight and obesity as BMI=25-29.99 kg/m2 and ≥30 kg/m2, respectively (36) . Waist circumference (WC) was measured according to the guiding protocol of WHO, at the midpoint between the lower border of the rib cage and the iliac crest, using a non-stretchable berglass measuring tape (37) and classi ed according to the International Diabetes Federation criteria (1) . The waist-to-height ratio (WHtR) was de ned as the WC divided by the measured height. WHtR ≥0.5 was adopted for overweight and abdominal obesity for the purpose of uniformity regarding age differences (38) . Blood pressure is measured by a digital barometer (BC 08, Beurer, Germany) after at least 10-15 minutes of rest and sitting. Blood pressure was measured twice for each person, and the average blood pressure was reported for each person. Of all participants, 10 ml of fasting blood was taken between 7-10 a.m. in the acid-washed test tubes without anticoagulant until after room temperature maintenance (RT Temperature (RT)) for 30 minutes. Minute blood clots and centrifuge at 1500 g for 20 minutes. The serums are poured into micro-clean tubes and stored in the 80 CC freezer until the test. Measurement of serum sugar and fat was performed using the enzymatic method, based on colorimetry, using commercial kits (Pars Azmoun, Iran) with the automatic machine (Selecta E, Vitalab, Netherland).
The tests were performed on people on the same day as the blood draw.

Sociodemographic and lifestyle variables
General information such as age, marital status, smoking status, living situation (alone or with someone), and disease status by asking participants with a general information questionnaire registered. International Physical Activity Questionnaire (IPAQ) is used to examine people's physical activity (39), which records three intensity levels of activity based on the metabolic equivalents (METs). METs were classi ed as low (< 600 MET-minutes/week), moderate (600-3,000 MET-minutes/week), and vigorous (> 3,000 MET-minutes/week).

Metabolic syndrome
MetS and its components were de ned using the following criterion(1). Individuals who have at least three or more of the following disorders were classi ed as having MetS: high waist circumference (≥88 for women and ≥102 for men); elevated triglyceride levels (≥150 mg/dl); low HDL-C levels (≤50 mg/dl for women and ≤40 mg/dl for men); high blood pressures (systolic blood pressure ≥ 130 mmHg and diastolic blood pressure ≥ 85 mmHg) or use of antihypertensive medication; and high fasting glucose levels (≥ 100 mg/dl) or use of hypoglycemic medication.

Statistical analysis
Energy-adjusted dietary CQI was used to classify participants into tertiles. According to the type of variables, the comparison of quantitative mean variables between the tertiles of subject characteristics and anthropometric measurement was performed using ANOVA (one-way variance analysis) and comparison of qualitative variables distribution between the tertiles with Chi-2 square test. Logistic regression was performed to investigate the relationship between CQI as an independent variable and MetS and its components as a dependent variable in an unadjusted and multivariable adjusted model. In this regard, age, sex, energy intake, physical activity, marital status, smoking status, educated status, underlying disease, and BMI were included as covariates in the modi ed regression model. All statistical analyses were done using IBM Statistical Package for Social Sciences (V.22; SPSS Inc.), and p<0.05 was considered statistically signi cant.

Results
The mean age of study participants with MetS was 46.1 ± 10, and the mean BMI with MetS was 29.2 ± 4.71. The prevalence of MetS among participants in the lowest and highest tertiles of CQI were 30.1 and 33.7, respectively (P = 0.6). The mean CQI in participants with MetS was 9.15 ± 2.83 (Supplementary Table 1). 30, 24, and 5 participants were removed from breakfast, lunch, and dinner, respectively, due to the lack of enough information for the analysis and lack of cooperation in their dietary intake reports. As a result, 820, 826, and 845 participants remained in the study for nal analysis at breakfast, lunch, and dinner meals.
General characteristics of study participants according to carbohydrate quality score based on habitual diet and meal is indicated in Table 1. Within lunch meal, those in the top tertiles of CQI were less likely to be a current smoker (p = 0.05). In habitual diet and all three meals, the distribution of participants in terms of other general characteristics across tertiles of CQI was not signi cantly different. In breakfast meal, participants in the top tertiles of CQI had lower TG and SBP compared with subjects in the lowest. Thus, we found a signi cant association between tertiles in participants for TG (P = 0.04). In addition, no signi cant statistical differences were found in other terms of laboratory characteristics across tertiles of CQI in all three meals (Table 2). Data are presented as mean ± standard deviation (SD). †One-way ANOVA test used for assessment variables.
* All values were adjusted for energy intake.
* P < 0.05 Abbreviations: FBG fasting blood glucose; TG triglyceride; HDL-C high density lipoprotein-cholesterol; SBP systolic blood pressure; DBP diastolic blood pressure mg milligram; dl deciliter The selected dietary intake of study participants across tertiles of CQI is shown in Supplementary Table 2 and Table 3. In habitual diet, we observed a signi cant association between tertiles in participants for linolenic acid, total sugar, glycemic index (P < 0.001 for all), and total ber (P = 0.002). Within breakfast meal, dietary intake of total sugar and glycemic index were signi cantly different across tertiles of CQI (P < 0.001 for all). In lunch meal, dietary intakes of total ber, total sugar, and glycemic index were signi cantly different across tertiles of CQI among participants (p = 0.001 for all). Moreover, the total energy intake of participants was signi cant across tertiles of CQI (P = 0.005). Within the dinner meal, participants in the top of tertiles of CQI had a higher intake of carbohydrate and a lower intake of SFA. Moreover, dietary intakes of total ber, total sugar, and glycemic index were signi cantly different across tertiles of CQI (p = 0.001 for all).

Discussion
The present study examined the relationship between Habitual and meal-speci c CQI and odds of MetS and its components in a sample of Iranian adults. Our ndings showed a non-signi cant association between CQI with MetS either before and after adjustment for potential confounders in all three meals and habitual diet. Also, we found no signi cant link between MetS components after adjustment for covariates. However, there was an increasing trend for elevated FBG in dinner meal and habitual diet.
Carbohydrates are a heterogeneous class of nutrients, and the consumption of re ned carbohydrates for enhancing the quality of received carbohydrates from a public health perspective has been suggested (40) . In this direction, the available studies on adults indicate that total carbohydrate intake or dietary carbohydrate proportion is not associated with the risk of obesity (41,42) . According to a report from a population with a balanced diet and lower carbohydrate intake, carbohydrate quality is more important factor to determine diet quality compared with the quality of fat (43) as far as, it has been proposed that a reduction in fat intake was compensated with an increased intake of re ned starches and sugars (44) . There is compelling evidence that carbohydrate quality has important effects on the progression and treatment of CVD, the MetS, T2D, and obesity (45) . Aspects of carbohydrate quality that may be important in these components include dietary ber, whole-grain, GI, and GL, and in particular, the intake of sugar-sweetened drinks. However, their properties are often highly inter-related, and it may be di cult to implicate one over another in any particular condition (45) . Therefore, in this context, this seems to be more important in Iranian populations because of higher intake of carbohydrate (46) .
To our knowledge, there has been no observational study examining the relationship between integrated carbohydrate quality in meals and metabolic disorders. Dietary approaches derive from meal timing are a hopeful plan for the modulation of circadian rhythms and clock-controlled metabolic functions in humans. Besides, studies have proposed that certain time windows are more suitable for the consumption of carbohydrate-rich or fat-rich food to maintain metabolic health. In a cross-over trial, Kessler and Pivovarova-Ramich (47) investigated that consumption of high carb meals in the evening has an unfavorable in uence on blood glucose level and glycemic control in individuals with impaired glucose metabolism. In agreement with this nding, other studies in humans suggested that a carbohydrate-rich diet at the beginning of the day could be safe against the development of diabetes and MetS (48,49) . In the modern lifestyle, the main meal is dinner (50) . Western main meals end with a sweet dessert. In addition, drinking beverages during the main meal was considered to be a part of modern lifestyle, as well as consuming special foods for breakfast that differ largely from the foods eaten at other meals (51) . On the other hand, skipping breakfast is very common in modern societies. In study of eight young men, in the rst condition, participants ate three main meals (breakfast, lunch, and dinner), while in the other condition, the same amount of energy was consumed at lunch and dinner times only. They found that skipping breakfast increased the average blood glucose concentration during the afternoon and sleep, subsequently resulting in an overall increased 24-hour average blood glucose concentration (52) . Another study assessed the glucose metabolism of healthy adults in two conditions of breakfast skipping and dinner skipping.
They showed, breakfast skipping resulted in higher glucose concentrations and insulin resistance after lunch (53) . In line with our ndings in habitual diets, a cross-sectional study conducted on Korean adults observed no signi cant associations between CQI and T2DM or MetS, although the quality of carbohydrate consumed is associated with the risk of obesity and high blood pressure (54) . Also, in another study, there was no association between GI and GL and MetS (55) .
Contrary to our ndings, a cross-sectional study from Ghana demonstrates that the diet with high CQI levels inversely related to general and abdominal obesity (56) . In another study identi ed that consuming re ned foods with high carbohydrate content was a positive association with a higher risk of abdominal obesity in Ghanaian University students (57) . A cohort study from Spain in university graduates reported an inverse association between dietary CQI and general obesity (58) . The result of a systematic review and meta-analysis study that were focusing on the association between carbohydrate quality and NCDs incidence and metabolic biomarkers showed that daily consumption of dietary ber was associated with a reduced risk of health-related consequence (59) .
These ndings are supported by cohort studies, which report a reduced risk of coronary heart disease incidence and mortality and incidence of diabetes.
However glycemic index of a meal can also be affected by its protein and fat content, in that dietary fat slows carbohydrate absorption from the gut, although this effect is probably minimal within the context of a mixed meal (60)(61)(62) .
High carbohydrate diets, which are common in developing nations, especially Asian countries, contain a high content of re ned sources (such as white rice and white bread), which are low in ber. These types of diets usually re ect poor food quality and mainly have high GI content, which can lead to negative metabolic consequences (63)(64)(65) . According to the National Food Consumption Survey, most of the calorie intake in Iranian people, which is about more than 60%, obtained from carbohydrates. In other words, the dietary intake of carbohydrates among Iranians is 450 g per day (rural areas: 413 g/d and urban areas: 518 g/d) (66) . Due to these data and consumption of whole grains in the Iranian diet is limited, so the quality of carbohydrates in this study is lower than other populations, which included whole grains.
For this reason, we may not have been to see an exact relationship between CQI with MetS and its components. The content of dietary ber and vitaminminerals in whole-grains is higher than re ned carbohydrates. The protective effects of these nutrients against the risk of chronic diseases are well-known (67,68) . Due to their physical structure and dietary ber content, whole grains are categorized as low GI foods. As mentioned earlier, most of the dietary carbohydrate intake in Iran is in the form of re ned grains, which usually contain a higher value than the GI and GL (34) . The use of whole grains is a useful way to increase the amount of ber in the diet and reduce the risk of non-communicable disease (NCDs). In addition, fruits and vegetables are important factors in ber intake in the diet. According to the above, our calculated CQI has low quality, but due to the consumption of medium to high ber content in this study, it can be concluded that the intake of fruits and vegetables and solid carbohydrates in the diet of individuals are higher. These substances contain fructose.
Fructose is found in large amounts in sucrose (table sugar), high fructose corn sweetener, natural fruit juices, and fruits. Fructose also led to a signi cant elevation in fasting glucose, insulin levels and decreased insulin sensitivity (45) . As our results, in this study showed a signi cant increase trend in FBG.
Although there were no statistically signi cant, the correlations between CQI and MetS and its components were positive. Based on the substantial role of carbohydrates in Iranian diets and their low-quality diets, it seems that focusing on the improve carbohydrate quality would be a practical and advantageous strategy to make better food choices among Iranians (46) .
In prospective studies, the consumption of liquid carbohydrates was associated with weight gain, whereas there was an inverse association between consumption of solid carbohydrates and high weight gain (69)(70)(71)(72) . Examination of the systematic evidence presented for the effect of long-term intervention with low GI and GL on fasting insulin level and proin ammatory markers showed that it could be effective in preventing obesity-related disease (73) .
High intakes of dietary ber and whole grains are more clearly associated with good health outcomes than measures of GI or GL. Although the glycemic index provides a measure of the glycemic potential of the carbohydrate content of foods, some low glycemic index foods might have other attributes that are not health-promoting. Foods with added fructose or sucrose, as well as mixed foods high in SFA and carbohydrate (i.e., confectionery products), may have a low GI (74) . ndings from a dose-response meta-analysis showed that diets identi ed by low dietary ber contribute to diverse NCDs and that administration of quantitative recommendations for dietary ber intake will be bene cial. While consumption in the range of 25-29 g/day is su cient, dose-responses data showed that amounts greater than 30 g/day have more bene ts (59) .
Given the effect of low GI in the development of obesity, there is evidence that low GI diets increase satiety by reducing voluntary food intake, thus reducing total energy intake, and it can be effective for body-weight maintenance and may prevent obesity (75)(76)(77) . In contrast, the intake of a high GI diet causes increases in hunger and subsequently leads to the increased intake of food, thus potentially affecting energy balance and body composition (78) . Fibercontaining foods should be chewed before passing through the stomach and into the small bowel, where they affect satiety, glucose and insulin responses, and lipid absorption. Whole foods that require chewing and retain much of their structure in the gut are more likely to cause a feeling of satiety, which in turn leads to weight loss and modulation of carbohydrate and lipid metabolism. In the large bowel, ber is almost completely broken down by the resident micro ora under a set of anaerobic reactions known as fermentation. The gut microbiota plays many important roles in human health (79) .
The present study has important strengths. To the best of the authors' knowledge, this study was the rst to investigate CQI in meals and its association with MetS and its components. We had a su cient sample size in this study that was done within 25 Health houses in the Tehran Metropolis. Despite these strengths, the study has some limitations. First, this study had a cross-sectional design, and the ndings do not establish causality between CQI and MetS; therefore, the results ought to be interpreted with caution. Second, using the questionnaire retrospectively may reduce information recall. There was also under-reporting and over-reporting of food items received. Third, even though the data were controlled for some potential confounders, the effects of eating behavior, menopausal status, and residual confounding cannot be discounted.

Conclusions
In conclusion, in the present study, CQI was not associated with MetS. Further investigations into the mechanisms underlying the role of carbohydrate quality in the development of metabolic disorders are warranted. Additionally to the quantity and quality of the food we eat, time to eat and the frequency of eating is also an important aspect of healthy eating habits. Optimization of mealtimes has signi cant bene ts for the prevention of chronic diseases and great promise for lifestyle interventions in the near future.

Declarations
Ethical approval and consent to participate