Comparison of Canola and Soybean Oils on Serum Lipid and Glucose Proles and Anthropometric Parameters in Overweight and Obese Type 2 Diabetes Mellitus Patients: A Randomized Clinical Trial

No earlier human study compared inuences of canola and soybean oils on patients with type 2 diabetes (T2D). Current study aimed to investigate effects of canola and soya oils on blood and anthropometric parameters in overweight and obese Iranian diabetic (II) patients. A total of sixty-six T2D subjects were randomly allocated to three groups. Canola oil (CO; n 23, received 30g canola oil); Soya oil (SO; n 19, received 30g soya oil) and control group (n 24, their usual intake of dietary oils) for 8 weeks. Lipid and glycemic proles as well as anthropometric indicators were evaluated before and after the intervention. Repeated-measures ANOVA was used to evaluate time×group interactions for the outcome variables followed by a t test (signicance level, p < 0.05). (WC; −4.1 P=0·031) (−3.1 and kg; P=0·048) canola high density lipoprotein in T2D patients. Changes were more considerable in those consumed soybean oil. Canola oil decreased central obesity indices (waist circumference and weight) in T2D patients. Further studies are needed to shed light on this issue. T2D subjects. Consumption of canola oil decreased central obesity indices (WC and weight) in T2D subjects. Given the high prevalence of diabetes and obesity, there is a growing need for further studies about effects of canola and soybean oil instead of other edible oils on patients with T2D and other chronic diseases.


Introduction
Today, global prevalence of type 2 diabetes is increasing continuously, along with the increased prevalence of overweight and obesity [1]. Fat accumulation in the visceral area is a major risk factor of diabetes mellitus [2,3] Chronic diseases, such as obesity and type 2 diabetes mellitus (T2D), are the main health concerns in the current century. They have put a great burden to health systems all over the world [4,5]. The global prevalence of T2D in 2000 was about 150 million people, which is estimated to increase to 300 million in 2025 [6]. In addition, more than four million Iranian adults are suffering from T2D, which has been increased by 35% over the past seven years [7].
Diabetes is a chronic disease that requires continual medical and self -care training. Nutrition therapy is an integral part of diabetes management, playing an essential role in the treatment of the disease [8]. It seems that the control of cholesterol, triglycerides and glucose in diabetic patients plays an effective role in the prevention and treatment of atherosclerosis [9]. High consumption of saturated fatty acids and carbohydrates causes several diabetes complications [8].
The majority of cooking oils used in Iran are hydrogenated and semi-hydrogenated oils [10]. Researches show that substitution of this type of oil with liquid vegetable oils signi cantly increases serum high density lipoprotein (HDL-) and decreases serum low density lipoprotein (LDL) [11]. Therefore, more consumption of polyunsaturated fatty acids (PUFA) (having more than a single double bond) instead of saturated fatty acids (SFA) is commonly recommended in T2D patients [12]. In addition, higher intake of omega-3 fatty acids has been associated to better serum lipid pro le and decreased cardiovascular risks [13]. Moreover, studies suggest that adherence to a diet containing high amounts of MUFA (Mono Unsaturated Fatty Acids), having one double bond, is a good choice for patients with T2D [14].
Canola oil (rapeseed oil) contains the lowest amount of saturated fatty acids (6.7% of the oil's total fatty acids), in comparison to other common types of cooking oils. In addition, it holds 18.7% PUFA (n-6), 65.3% MUFA, and almost 11% alpha-linolenic acid (essential n-3 fatty acids), which is the highest rate in comparison to other types of cooking oils [15,16]. Soybean oil is primarily comprised of PUFA (55 to 58%), 12 to 15% saturated fat, and 22 to 30% MUFA (oleic acid) [17]. Although, rare clinical trials have been done on the effects of these two oils on metabolic pro les and anthropometric measures in patients with type 2 diabetes mellitus, ndings are controversial. A systematic review and meta-analysis in 2019 showed that canola oil consumption had a modest effect only on body weight, no on other obesity indices [18]. Moreover, most previous studies about the effects of canola and soybean oil on lipid and glucose pro les were done on animal models. Therefore, more human studies are needed. Furthermore, these two types of edible oils are commonly considered to be healthy, however, no comparison has been done between them to nd which is healthier among T2D patients. Therefore, current randomized clinical trials aimed to compare effects of canola and soybean oils on lipid and glucose pro les and anthropometric measures among a group of Iranian patients with type 2 diabetes mellitus.

Study design
This was an 8-week parallel double-blind randomized clinical trial (RCT). This RCT was conducted between mid-July 2012 to mid-January 2013 in one hospital in Zahedan, Iran. All study participants received informations regarding the study design and objectives and then they signed a written consent. This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving patients were approved scienti cally and ethically by the Zahedan University of Medical Sciences (ZUMS) (identi cation: IR.ZAUMS.REC.1393.2160). The trial registration number at irct.ir is IRCT2012062510110N1. Figure 1 shows the Consolidated Standards of Reporting Trials (CONSORT) 2010 diagram of the ow of participants from enrollment to analysis; the completed CONSORT checklist is provided in Table S1.

Study participants
Participants were recruited from the Ali-Asqar Diabetes Clinic in Zahedan, Iran. All patients with type 2 diabetes mellitus who met following criteria and was agreed to participate in the current trial, were recruited: (a) were overweight or obese (25≤BMI≤39.9), (b) aged 30-65 years old, (c) and had a mean Fasting Blood Sugar (FBS) of ≥126mg/dl. Subjects with the following criteria were not included: (a) patients with hyperglycemia because of other diseases rather than type 2 diabetes mellitus, (b) patients with liver/ kidney/thyroid diseases, (c) patients who received insulin therapy, (d) patients who received cholesterollowering agents or beta blockers, and (e) patients who drank alcohol. We also excluded patients with active and intense infectious diseases during blood collection as well as patients who were hospitalized over the plan implementation.

Study protocol
Before intervention, all patients entered a run-in period for 2 weeks through which they instructed to have a weight maintenance diet according to the American Diabetes Association guidelines [19]. All patients were also asked to not use canola and soya oil for two weeks prior to the study. Instead, they were allowed to use only corn oil for cooking.
Then, patients were divided into 3 groups through using computer-generated random numbers: 1) CO (receivers of 30 g canola oil; n 23); 2) SO (receivers of 30 g soya oil; n 19); and 3) control group who continued their usual diet; n 24. Randomization was done by a third investigator out the current study. Therefore, all participants and study directors were blinded. The two rst groups were asked to: a) replace their usual oil with canola and soya oil and to not use any other types of solid or liquid oils; b) prepare and cook their foods separately from the other family members; and c) use only 30 grams of the allowed oil in their daily diet. The control group was asked to continue its usual diet. Bottles of canola or soybean oils were given to the participants for 4 weeks. The intervention period was 8 weeks. We assessed dietary intakes of participants by employing three 24 h recalls questionnaires (including a weekend day) in the beginning of study, which was repeated in the end of the intervention period. Data on demographic, medical history, physical activity, and duration of diabetes were gathered using a general questionnaire by the face-to-face interview. After 4 weeks, all subjects were visited to evaluate their compliance and to receive new bottles of the oils for another 4 week (for the canola and soybean oil groups). Compliance was evaluated by counting the empty bottles.

Anthropometrics measurements
Anthropometric measurements were done at the study beginning and end of the 8th week. Weight was measured with light clothing and without shoes using a digital scale (Seca 808) to the nearest of 0·1 kg. Height was measured without shoes using a stadiometer (Seca) to the nearest of 0·1 cm. Waist circumferences (WC) and hip circumferences (HC) were measured by a tape to the nearest of 0·1 cm. BMI was calculated using the equation BMI (weight (kg)/height 2 (m 2 )). WHpR (waist hip ratio) was calculated using the equation WHpR (waist (cm)/hip (cm)). Mid arm circumference (MAC) was measured midway between olecranon and acromion [20]. Calf circumference (CC) was measured on the left leg in a sitting position, with the knee and ankle at a right angle and feet resting on the oor, at the point of greatest circumference [20]. We used non-elastic tape to measure MAC and CC.

Biochemical measurements
At the study beginning and the end of trials, participants were invited to attend the diabetes clinic laboratory while they were fasting for 12-14 h. Then, 10 cc blood was collected from each patient. Blood samples were divided into tubes without the anticoagulant (EDTA). FBS was measured using enzymatic method by using commercial kits (Pars Azmoon) and an auto-analyzer system (Selectra E; Vitalab). Glycosylated hemoglobin (HbA1c) was measured using colorimetric method after an initial chromatographic separation (BioSystems). TG (Triglycerides), TC (Total Cholesterol), LDL (Low density lipoprotein) and HDL (High density lipoprotein) were measured by the enzymatic method (Pars Azmoon kit). ESR (Erythrocyte Sedimentation Rate) was measured by the Westergren method, and CBC (Cell Blood Counter) was measured by a cell counter. In addition, CRP (C-reactive protein) was measure using a qualitative method (OMEGA). Finally, serum creatinine (Cr) and urea were assayed by the JAFFE and UV-test, respectively (Pars Azmoon).

Statistical analysis
The sample size was de ned according to type I error of α = 0.05 and type II error of β = 80%. Initially, participants were classi ed into three groups: control, canola oil and soya oil. Normal distribution of data was checked using the Kolmogrov-Smirnov. Comparison of quantitative variables between the CO, SO and control groups was performed using Independent Samples T tests (for normally distributed variables) or Mann-Whitney U tests (for non-normally distributed variables) and analysis of covariance (ANCOVA), that adjusted for energy intake. Repeated-measures ANOVA was used to evaluate time×group interactions for the outcome variables, with time and group as factors. In case of signi cant time-group interaction, between-group comparison of changes at week 8 was carried out using ANOVA followed by Tukey's post hoc analysis with polynomial contrast analysis for trend when indicated. When time effect was signi cant, the within-group comparison of values was performed by paired sample t tests. All statistical analyses were carried out using the Statistical Package for Social Sciences (SPSS) version 18. P values <0·05 were considered as statistically signi cant.

Results
Study population characteristics Figure 1 shows the owchart of study procedure. Study participants were 31 males and 35 females with an average age of 46.59 years (SD = 7.8). The three groups were not signi cantly different in terms of mean age (p=0.18), mean years lasted from rst diagnosis of the disease (p=0.32), and gender (p=0.23) ( Table 1).

Dietary intake of study population according to groups
Dietary intakes of participants has been compared before and after the intervention between the three groups ( Table 2). No signi cant differences were seen between three groups in terms of energy intake, carbohydrates, proteins or fats (p 0.05). The results of the comparison showed that mean linoleic and linolenic were higher in CO and SO compared with control group, even after adjusting for energy intake (p 0.05).

Comparison of biochemical parameters in study groups
At baseline, no signi cant differences were seen between the three groups in terms of serum concentrations of the study main outcomes (Table 3). However, a signi cant effect of time was observed in serum FBS (p< 0.001), TG (p< 0.001), TC (p< 0.001) and LDL (p=0.028) concentrations from week 0 to week 8 (Table  3). A signi cant reduction of TG, TC, and LDL was observed in the CO group (p = 0.01 for TG, p = 0.01 for TC, p = 0.04 for LDL) and in the SO group (p = 0.002 for TG, p = 0.001 for TC, p = 0.01 for LDL) after 8 weeks, which was not signi cant in the control group (p = 0.6 for TG, p = 0.7 for TC, p = 0.3 for LDL). HDL increased in CO and SO groups (p = 0.001, p = 0.001, respectively), with no signi cant changes in control group (p =0.3) ( Table 3). Serum Cr level signi cantly increased only in the CO group, and no signi cant changes were observed in other groups (SO and control) (p = 0.063 and p = 0.06, respectively).
Signi cant time × group interaction was observed in the study groups in terms of serum concentrations of TC (p=0.007), LDL (p=0.013), HDL (p=0.038) and FBS (p< 0.001) ( Table 3). Tukey's post hoc test showed signi cant greater reduction of serum concentration of TC in SO and CO groups than the control group after the intervention (p= 0.009 and p= 0.034 respectively). Moreover, reduction in TC concentration in SO group was signi cantly greater than CO group (p= 0.049).
LDL levels signi cantly decreased in SO (p= 0.01) and CO (p= 0.04) groups. Tukey's post hoc test showed signi cant differences between SO and control groups (p = 0.01), CO and control (p = 0.023) groups, and reduction in LDL concentration in SO group was signi cantly greater than CO group (p = 0.051).
Signi cant differences in HDL concentrations were found between CO and control group (p = 0.039) and also between SO and control group (p = 0.034); however, no signi cant differences were observed between CO and SO groups (p = 0.97). Moreover, signi cant differences were seen between CO and control (p < 0.001) groups and SO and control (p = 0.001) groups in terms of FBS, while no signi cant difference was observed between CO and SO groups (p = 0.47) ( Table 3).

Comparison of anthropometric parameters in study groups
Effects of dietary interventions on anthropometric measures have been presented in Table 4. No signi cant differences in anthropometric measures were seen between three groups at the study beginning. The effect of time on body weight, HC, WC, and WHpR from week 0 to week 8 was signi cant in CO group (p= 0.02, 0.02, 0.0001, and 0.03, respectively), with signi cant reduction of weight, WC and WHpR and signi cant increase in HC. However, there was no signi cant effect of time on these variable in SO (p = 0.48, 0.09, 0.19, and 0.82, respectively) and control (p = 0.44, 0.72, 0.1, and 0.18, respectively) group (Table 4).
A signi cant time × group interaction was observed in the study groups in terms of weight and WC (Table 4). Tukey's post hoc test showed that after 8 weeks, CO group had signi cantly greater weight loss than the control and SO groups (p = 0.001 and p = 0.049), respectively. However, no signi cant differences were observed between SO and control (p = 0.54) groups (Table 4).
Moreover, a signi cant difference was observed between the CO and SO groups (p = 0.041) and the CO and control (p < 0.001) groups in terms of WC, while no signi cant difference was observed between SO and control groups (p = 0.63) ( Table 4).

Discussion
Current study showed that consumption of 30 g/day of both canola or soybean oil in comparison to control signi cantly decreased fasting blood glucose and increased proinsulin to insulin ratio in overweight and obese patients with T2D. Consumption of these two oils also signi cantly reduced serum levels of TC and LDL and increased HDL concentrations, which was more considerable among those who consumed soybean oil. Body weight and waist circumference signi cantly decreased only in canola oil group. No signi cant changes were seen in serum levels of TG and HbA1C, and serum levels of in ammatory factors (ESR and CRP), urea and Cr, and in white blood cell (WBC), red blood cell (RBC), and platelet (PLT) levels.
Consumption of either soybean or canola oil resulted in signi cant reduction of FBS. Soybean and canola oil consumption also was associated to increased proinsulin to insulin ratio. Increased proinsulin to insulin ratio is a common manifest of T2D associated to insulin resistance or impaired conversion of proinsulin to insulin [21]. Signi cant improvement in blood glucose without changes in HbA1C and signi cant elevation of proinsulin to insulin ratio shows that consumption of these two oils had no signi cant effect on insulin resistance, as the major complication of T2D mellitus. Consumption of non-fried soya, but not its oil, resulted in signi cant reduction of insulin resistance in Japanese adults in a cross-sectional study [22]. Consumption of canola oil in patients with T2D reduced insulin resistance and serum levels of insulin in a randomized clinical trial [23]. It should be noted that most studies in this area have been done on animal models and further human studies are needed to reach a clear conclusion. Our study also found signi cant reduction of anthropometric measures in T2D patients following consumption of canola oil. Due to signi cant bene cial changes in anthropometric measures, it might be suggested that consumption of soybean or canola oil for a longer time in uences more considerably insulin resistance in patients with T2D.
This study also showed that consumption of 30 g/day of soybean and canola oil has bene cial effects on lipid pro les, including TC, LDL, and HDL in patients with type 2 diabetes mellitus (T2D). These effects were more pronounced in the patients who consumed soybean oil that those at CO group. Earlier studies reported inconsistence ndings for the association of soybean and canola oil with serum lipid pro les. Gulesserian et al. examined seventeen 4-19year-old children and adolescents with baseline TC of 233±35 mg/dl who received canola oil for 5 months (15 gr/day in the rst two months and 22 gr/day in the next three months). Serum levels of TG, TC and LDL-c decreased following canola oil consumption in that study, with no signi cant changes in serum HDL-c concentrations Small sample size and short duration of intervention were among the main limitations of the current study. In addition, because participants were at an age range of 45-49, the results cannot be generalized to young and elderly populations. It is necessary to done similar studies on other individuals and groups in order to investigate effects of canola and soybean oils on biochemical and anthropometric indices in patients with diabetes mellitus.

Conclusion
In conclusion, results of our study showed that daily intake of canola and soybean oil for 8 weeks improved serum FBS, TC, LDL and HDL in T2D subjects.    Study ow diagram (CONSORT 2010). BP, blood pressure.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.