Participants whose ages were between 20 and 65 were voluntarily recruited. Informed consent was obtained from all participants for being included in the study. Flow diagram was presented in Figure 1. A total of 104 subjects were examined for inclusion criteria, having at least one MetS risk factor. According to the 2005 International Diabetes Foundation, the risk factors for MetS include the following: waist circumferences 90 cm or higher for men and 80 cm or higher for women, TG levels 150 mg/dL or higher, HDL-cholesterol less than 40 mg/dL for men and 50 mg/dL for women, blood pressure 130/85 mmHg or higher, and fasting glucose 100 mg/dL or higher [12]. MetS is the condition which meet more than three risk factors [12]. Height, body weight and body fat composition were measured using a bioimpedance analyzer (Inbody 370, Biospace, Seoul, Korea). Blood pressure was checked on the left arm using an automatic blood pressure monitor (FT-500R PLUS, Jawon Medical, Gyeongsan, Korea) at Yonsei University health service center. The waist circumference was measured with a measuring tape around the mid-section between the margin of the last rib and the iliac crest by trained personnel. Exclusion criteria included the following: 1) heavy alcohol/drug consumption; and 2) pregnancy, breastfeeding, or intention to become pregnant during the time of the study. Finally, 90 subjects were selected for a randomized, double-blind clinical trial and divided into two groups, RC+F and RC+FP. During the six weeks, 10 subjects dropped out and 20 subjects were excluded for the analysis in order to have comparable initial characteristics of the subjects between RC+F and RC+FP groups. The number of subjects for each component of risk factor in RC+F and RC+FP groups is presented in Supplemental Table 1. Only those who finished 6 weeks' intervention were included for the analysis. Blood was collected at the beginning and the end of the study, respectively. Serum and plasma were separated by centrifugation at 3,000 rpm for 10 minutes. Aliquots were then frozen at -80℃ until analysis. All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. The protocol was approved by the Institutional Review Board of Yonsei University (trial registration number, 1040917-201311-HRBR-121-03).

Table 1 showed the composition of food ingredients used to manufacture the rice cakes. Psyllium (0.51% w/w), fructooligosaccharide (10.11% w/w), and resistant starch (5.28% w/w) were used for both RC+F and RC+FP. Gynura procumbens Merr. (2.72% w/w) and Artemisia annua (2.90% w/w) were only added to RC+FP. The RC+F and RC+FP were manufactured in three forms: songpyeon, seolgidduk, and chaldduk. RC+F and RC+FP were sent to the study subjects via an express delivery in vacuum packed and frozen condition. The daily portion of RC+F or RC+FP was individually packaged, which was 100 g and approximately 200-250 kcal. In every week, RC+F and RC+FP were to be consumed in the form of songpyeon three times, seolgidduk twice and chaldduk twice. The average daily intake of rice cake was equivalent to a total of 42.9 g of songpyeon, 28.6 g of seolgidduk, and 28.6 g of chaldduk. The manufacturing process of seolgidduk, chaldduk, and songpyeon was described in the supplemental materials and methods.

Protein, fat, dietary fiber, and moisture contents were analyzed by the methods according to the food standard code: the Kjedahl method for protein, the chloroform-methanol extraction method for fat, the total dietary fiber content method for dietary fiber, and the drying method by heating under normal pressure for moisture content [13]. Energy (kcal), sodium, and carbohydrate contents were analyzed using a Standard food composition table [14].

Study subjects were educated on how to record the food records and how to estimate measurements using a ruler, a spoon size and a small paper cup which was commonly found in public spaces for automated coffee machine. A food record booklet and a serving size booklet were provided prior to the participation at the beginning of the study. During the six weeks of the study period, food records at least for two days including one week day and one weekend were asked to be recorded every week. Food records were collected in photographed images taken by a cellular phone and were sent via a cell phone or email to us every week. Total at least 12 to 18 days' food records per each subject were analyzed for dietary intake with the use of the Computer Aided Nutritional Analysis Program (CAN pro4.0, the Korean Nutrition Society, Seoul, Korea). Dietary between intake in the two groups was presented in Supplemental Table 2. Even if there was a modest increased tendency of dietary fiber intake in RC+FP compared to RC+F, there were no significant differences in total calorie intake and dietary intake of protein, carbohydrate, lipid and dietary fibers between the two groups (Supplemental Table 2).

Plasma levels of fasting glucose, Total-cholesterol (Total-C), HDL-cholesterol (HDL-C), and TG were analyzed using an automated blood chemistry analyzer (DRI-CHEM 4000i, Fujifilm, Tokyo, Japan). LDL and very low density lipoprotein (VLDL)-cholesterol levels were calculated using the Friedewald formula. LDL–C = Total-C – HDL-C – TG/5 and VLDL–C = TG/5. The atherogenic index (AI) was calculated using Haglund's assay [15]. AI = (Total-C – HDL-C) / HDL–C. homeostasis model assessment-insulin resistance (HOMA–IR) was calculated, where HOMA-IR = fasting insulin (mU/L) × fasting blood glucose (mg/dL)/405 [16].

Insulin (Human insulin ELISA kit, Sunredbio, Shanghai, China), malondialdehyde (MDA) (Lipid peroxidation assay kit, Biovision, Milpitas, CA, USA), and 8-hydroxy deoxyguanosine (8-OHdG) Quantitation (Human 8-OHdG ELISA kit, Mybiosource, California, CA, USA) were analyzed using commercial kits according to the manufacturers' instructions. Blood urea nitrogen (BUN), creatinine, aspartate transaminase (AST) and alanine transaminase (ALT) were analyzed using Modular analytics (PE, ROCHE, Germany).

Statistical analysis was performed using SPSS ver. 21.0 (SPSS Inc., IL, USA). All data was presented as mean ± standard deviation (SD). p < 0.05 was considered to be statistically significant. Data which didn't follow normalization was log transformed. Independent t-test was used to compare the general characteristics between RC+F and RC+FP groups. A comparison of the net change values between the two groups was performed by analysis of covariance with adjustment for baseline values.

Total 90 subjects were enrolled at the beginning of the study. But 10 people dropped out for personal reasons (5 in RC+F and 5 in RC+FP) and 10 subjects per each group were additionally excluded in order to have statistically similar sex ratios between the two groups. Finally 30 subjects per each group successfully completed the study (Figure 1).

Both RC+F and RC+FP had the mixture of fructooligosaccharide, resistant starch, and psyllium (Table 1); however, Artemisia annua and Gynura procumbens were added only to RC+FP, resulting in higher total polyphenol contents compared to RC+F (Table 1 and Figure 2). The average daily portion of RC+F or RC+FP contained 243.6 kcal and 222.4 kcal, 49.4 g and 43.7 g of carbohydrate, 5.3 g and 5.0 g of protein, 2.5 g and 2.6 g of fat, 6.5 g and 7.0 g of fibers, and 39.4 g and 89.2 g of total polyphenols, respectively.

The mean age for a study population was 35.0 ± 11.1 years. 61.7 % of the subjects were male and 38.3% were female. The comparison of general characteristics and mean values of MetS risk factors in RC+F and RC+FP groups were presented in Table 2. The average ages in RC+FP group (n = 30) was significantly higher than those in RC+F group (n = 30). However, the average values of BMI, waist circumference, TG, HDL-C, fasting blood glucose, and systolic, and diastolic blood pressures were not significantly different between the two groups.

Compared to the initial values, body weight, BMI and waist circumferences were not significantly changed in both groups after 6 weeks (Figure 3). However, blood pressures in both systolic and diastolic blood pressures were significantly decreased only in RC+FP group (Figure 3). These data indicated that RC+FP might have a blood pressure lowering effect.

Only in RC+FP group, serum levels of FBG and HOMA-IR were decreased at the end of the study compared to the beginning (Figure 4A and 4B). Serum levels of TG, Total-C, LDL-C, MDA and 8-OHdG, and AI were not significantly changed after 6 weeks compared to the initials in both groups (Figure 4C-H). Consumptions of RC+F and RC+FP did not induce the changes in BUN, AST, and ALT (data not shown). Our findings indicated that RC+FP appeared to improve FBG levels and HOMA-IR.

After 6 weeks, the proportions of the subjects with at least one MetS risk factor were decreased to 70.0% in RC+F and 60.0% in RC+FP groups (Figure 5). The number of subjects with MetS risk factor of two or higher was also decreased from 76.6% to 43.3% in RC+F group and from 76.7 % to 26.6% in RC+FP group. Subsequently, the percentage of subjects without a MetS risk factor increased in both groups (30% in RC+F and 40% in RC+FP) (Fig. 5A and 5B). Next, we examined whether those who had a specific MetS risk factor had improvement in values of their corresponding risk factor after 6 weeks' intervention. RC+FP group had decreased the levels of FBG and systolic and diastolic blood pressures whereas RC+F group had reduced systolic blood pressure (Table 3). These results suggested that RC+FP was more effective in reducing the risk of MetS compared to RC+F.