In total, 126 healthy overweight and obese subjects were assessed for participation eligibility through advertisements in the local newspaper and at the Shahid Motahari Medical Teaching Center (one of the main teaching hospitals of Urmia University of Medical Sciences). A total of 27 subjects met the inclusion criteria, defined as follows: body mass index (BMI; in kg/m²) of 27–40 and aged 20–50 years. The exclusion criteria were as follows: weight loss of ≥ 10% of body weight within the 6 months and ≥ 5% of body weight within the 1 month before enrollment in the study; subjects were changed their routine physical activity during the 12-week intervention; pregnancy; lactation; menopause; history of acute disease; acute or chronic inflammatory diseases; cardiovascular disease; renal disease; liver disease; endocrine disease; abnormal thyroid hormone concentrations; cancer or chemotherapy/radiotherapy; use of drugs potentially affecting glucose and lipid metabolism; use of antihypertensive drugs; taking antibiotics, anti-obesity drugs, and medications that could affect energy expenditure, prebiotics, probiotics, within 2 weeks before the intervention or during the intervention; if they had a fiber intake of > 30 g/day; taking antioxidant; vitamin and/or mineral supplements; smoking; and inability to give informed consent.

This randomized, double-blind, placebo-controlled, 2 × 2 crossover trial was conducted from September 2013 to December 2013. Written informed consent was obtained from all participants before enrollment. The study protocol was approved by the Urmia University of Medical Science Ethics Committee and was registered at ClinicalTrials.gov as NCT01992783.

At baseline, subjects were randomly assigned to receive either resistant starch supplements or the identical-appearing placebo (maltodextrin). Follow-up assessments were performed every 4 weeks at weeks 4, 8, and 12 after randomization. Randomization lists were computer-generated by a statistician. Subjects, investigators, statistician, and laboratory staff were blind to the intervention assignment until the end of the study.

In this study, we used the formula suggested for crossover trials sample size calculation. With α = 0.05, β = 5% (power = 95%), S = 0.82, and Δ = 0.46 (based on a study conducted by Mutlu-Türkoğlu et al. [16] and mean change in malondialdehyde), the sample size needed was estimated to be 21 participants. Considering a drop-out rate of 30%, the final sample size needed was estimated to be 27 participants.

The participants in the first period were given 13.5 g Hi-Maize 260 containing up to 60% resistant starch (Ingredion Inc., Manchester, United Kingdom), while the participants in second period were given an identical appearing placebo (13.5 g maltodextrin) daily for 4 weeks, separated by a 4-week washout period. The appearance of the placebo was indistinguishable in shape, color, size, packaging, smell, and taste from the resistant starch supplement. Following the washout, the groups were crossed over to receive the opposite intervention for 4 weeks. The participants were instructed to consistently take the supplements with the main meals. At the onset of the study, subjects were requested to maintain their regular diet and levels of physical activity throughout the trial period. Adherence was assessed by supplement counts confirmed at each visit. Compliance with the supplements was monitored every week by a telephone call. None of the subjects completing the study had any serious adverse events such as gastrointestinal, central nervous system, skin symptoms, and hypoglycemia [1718].

Biochemical testing was performed at baseline and at weeks 4, 8, and 12 to determine serum lipid profiles, glycemic variables, lipid peroxidation marker, and antioxidant status. Blood (10 mL) was drawn at the Urmia University of Medical Sciences reference laboratory after 12 hours of fasting and was centrifuged at 1,465 ×g at room temperature for 10 minutes into 4 aliquot microtubes. Microtubes of each serum sample were frozen immediately at −80°C until analysis. Serum concentrations of total cholesterol, triacylglycerols (TGs), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and fasting blood sugar (FBS) were measured by using routine enzymatic assays with commercial kits (Pars Azmoon, Tehran, Iran) with the use of an autoanalyzer (Hitachi 902 Auto analyzer; HiTAShi Ltd., Tokyo, Japan). Serum concentrations of insulin were quantified using an immunoradiometric assay kit (DIAsource INS-IRMA Kit; DIAsource, Louvain-la-Neuve, Belgium). The measurement of superoxide dismutase (SOD) was based on the principle that xanthine reacts with xanthine oxidase to generate superoxide radicals, which react with 2-(4-iodophenyl)-3-(4-nitrophenol)-5-phenyltetrazolium chloride (INT) to form a red formazan dye which is read at 505 nm [19]. The procedure was done based on spectrophotometric method (Perkin-Elmer, D-7770 Uberlingen, Germany) using a Biorex kit (Biorex, Antrim, UK). Total antioxidant status (TAS) was measured by using the determination ferric reducing ability power (FRAP) [20]. In this method ferric, present in FRAP reagent, is reduced to the ferrous form at low pH and the absorbance of resultant color is read at 593 nm using a spectrophotometer (Perkin-Elmer, D-7770 Uberlingen, Germany). Malondialdehyde (MDA) concentration was measured using Thiobarbituric Acid Reactive Substances method [21]. The method is based on the MDA reaction with thiobarbituric acid at temperature of 90°C–100°C for 30 minutes and pH = 2–3 to obtain a pink pigment that absorbs at 535 nm (Perkin-Elmer, D-7770 Uberlingen, Germany). All biochemical measurements were performed in the same laboratory with the use of standard laboratory methods.

Homeostatic model assessment of insulin resistance (HOMA-IR) and quantitative insulin sensitivity check index (QUICKI) were also calculated based on suggested formulas [22].

Anthropometric measures were taken in the standing position, with participants wearing light clothing and footwear at baseline and at weeks 4, 8, and 12. Height and weight were measured with an accuracy of 0.1 cm and 0.1 kg, respectively. BMI was calculated by dividing weight in kilogram by the square of height in meters. Waist circumference was measured at the middle point between the lower rib margin and the iliac crest. Blood pressure was measured after 10 minutes of rest at baseline and at weeks 4, 8, and 12. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured with participants in the sitting position with the use of a validated mercury sphygmomanometer.

A 3-day (1 weekend day and 2 nonconsecutive weekdays) food records were performed at baseline and at weeks 2, 6, and 10. A nutritionist individually taught subjects how to measure their food portions and how to complete the food records. Dietary records were analyzed using a nutrition analysis program (Nutritionist IV, version 3.5.2; First DataBank Division, San Bruno, CA, USA) modified for Iranian foods.

The primary outcome measures were changes in serum lipid profiles, glycemic variables, lipid peroxidation marker, and antioxidant status. Secondary outcome measures were changes in blood pressure and anthropometric variables.

To determine the effect of resistant starch on serum lipid profiles, glycemic variables, lipid peroxidation marker, antioxidant status, blood pressure, and anthropometric variables, we applied a general linear model analysis of variance for repeated measurements. Also, we computed the median value of studied variables and divided the subjects into 2 groups as higher than or lower than median groups to show the effect of resistant starch on that variable in subjects with high or low level of variable. Multiple comparisons between the intervention and the placebo were followed by the Bonferroni post hoc test. In this method, in case the Mauchly's test of sphericity assumption was violated, the Epsilon and Huynh-Feldt test results are reported. Data were analyzed with the use of SPSS version 16 (SPSS, Inc., Chicago, IL, USA). For all analyses, p < 0.05 was considered significant and all data are shown as means ± standard deviations if not indicated otherwise.

As shown in Table 3, we found no significant differences in serum concentrations of total cholesterol, TGs, HDL-C, LDL-C, FBS, insulin, HOMA-IR, QUICKI, SOD activity, and MDA in the 3 phases of baseline, after intervention with resistant starch and after placebo. However, changes in TAS (p = 0.04) and serum concentrations of insulin in 52.4% participants with insulin levels above 16 µIU/mL at the baseline (p = 0.04) were significantly different in the three phases (Figure 2). In addition, the mean of serum HDL-C after the intervention was significantly higher than after baseline value (p = 0.04) (Table 3).

There were no significant differences in SBP and DBP, weight, BMI, or waist circumference in the three phases of baseline, after intervention with resistant starch, and after placebo (Table 3).