The effects of canola and olive oils on lipid profile and fatty liver compared to sunflower oil in women with polycystic ovary syndrome: a randomized controlled trial


 BackgroundPolycystic ovarian syndrome (PCOS) is one of the most common endocrinopathies and metabolic disorders in women during their reproductive years and is often associated with dyslipidaemia and other cardiovascular diseases (CVD) risk factors. This study aimed to evaluate the effects of dietary intervention with canola and olive oils in comparison with sunflower oil on lipid profile and fatty liver severity among women with PCOS.MethodThis study was a 10 weeks intervention including 72 women with PCOS. Patients were randomly assigned to three groups to receive 25 g/day canola, olive, or sunflower oils for 10 weeks. The primary outcome was any changes in the lipid profile and the secondary outcome was a change in fatty liver severity.ResultAt the end of the study, 72 patients with a mean age of 29.31 were analyzed. Canola oil consumption led to a significant reduction in serum levels of triglycerides and TC/HDL (p = 0.021), LDL/HDL (p = 0.047), and TG/HDL (p = 0.001) ratios, but there was no significant reduction in lipid profile following olive oil consumption. Moreover, both the canola (p < 0.001) and olive oils (p = 0.005) could significantly reduce the fatty liver grade.ConclusionOverall, compared to the olive and sunflower oils, significant improvements in lipid profile and liver function were observed following consumption of canola oil in women with PCOS.Trial registrationIR.MUI.RESEARCH.REC.1397.315. Registered 30 JUNE 2019 - Retrospectively registered, https://www.irct.ir/trial/38684


Study design and patients' characteristics
This study was conducted as a randomized, double-blind, controlled clinical trial that conformed to the declaration of Helsinki and Good Clinical Practice Guidelines. The protocol of this study was reviewed and approved by Medical Ethics Committee at Isfahan University of Medical Sciences, Iran (ethics registration number: IR.MUI.RESEARCH. REC.1397.315) and was registered at Iranian Registry of Clinical Trials (approval code: 38684). Study participants included 90 patients with PCOS aged 18 to 45 years. Eligible patients were those with PCOS who referred to the Gynecology clinic in Isfahan. Participants were included in the study according to the Rotterdam criteria (34), in which two of the following features were con rmed: 1) Oligomenorrhea (interval between two menstrual periods more than 35 days) or amenorrhea (no vaginal bleeding for at least 6 months); 2) Clinical ndings of increased blood androgen levels (hirsutism scores greater than 7 or obvious acne), or increased blood testosterone levels (testosterone levels above 2 nmol) and 3) Polycystic ovaries in ultrasound scan (12 follicles measuring 2-9 mm in diameter, or ovarian volume >10 mL in at least one ovary). Patients were eligible if they were not menopause, did not consume any supplements with lipid-lowering properties, or omega 3 supplements in the last 3 months, or followed a special diet, BMI range between 25 to <40 and not having a severe weight loss history in the past six months. Participants were excluded if they had: pregnancy, adrenal hyperplasia, androgen-secreting tumors, Cushing syndrome, hyperprolactinemia, thyroid dysfunction, diabetes or other metabolic diseases, used lipid-lowering drugs, tamoxifen, raloxifene, oral or injectable corticosteroids (such as prednisone, prednisolone, dexamethasone, triamcinolone, hydrocortisone or betamethasone), any history of allergies, intolerance or harmful drug reactions to the canola and olive oils. Also, patients who were undergoing chemotherapy and smokers were excluded. All participants were given necessary explanations regarding the study protocol and written informed consent forms were obtained from all patients before commencing the study began.

Sample size
The sample size was calculated based on the standard formula for clinical trials, considering type 1 error (α) of .05 and type 2 error (β) of .10 (power = 90%) based on serum TG levels in Salar et al. study (35). Based on this information, 27 subjects were required to be included in each treatment group. Considering three probable dropouts in each group, the nal sample size was determined as 30 patients in each group.

Fatty acid compositions of olive oil and canola oil
The fatty acids (FA) composition of olive, canola, and sun ower oils (Oila, Iran) were evaluated in a reference food chemistry laboratory at the Isfahan University of Medical Sciences (Meyar Danesh Pars laboratory, Iran). The FA composition of the three types of oils was determined by high-performance gas chromatography (36). As shown in Table 1, olive oil contained 68.93% oleic acid as the main FA. While canola oil contained 59.62 % as oleic acid as its main FA. Sun ower oil contained 60.51% linoleic acid as its main FA. Also, the n-6/n-3 ratio was 18.4 for sun ower oil, 17.8 for olive oil, and 2.9 for canola oil.

Randomization and Intervention
Patients were randomly assigned into three equal groups: 1) received 25 g/day canola oil, 2) received 25 g/day olive oil, and 3) control group that consumed 25 g/day sun ower oil for 10 weeks. The random allocation was done by an investigator who was not directly involved in the trial. Participants were given the oils in similar bottles. Neither participants nor the researchers and the physician were aware of the type of oils consumed until the study was ended. Bottles containing the oils were given to the women for the rst ve weeks at the beginning of the study, and this was repeated in the next ve weeks. We also provide them a small container that contained exactly 25 grams to pour the oil they needed every day and used it until nighttime. All patients were advised to take a balanced diet with a macronutrient distribution of 45%-60% as carbohydrate, 30% to 35% as fat, and 15-18% as protein. The diet was described and each patient was given an exchange list to ensure the consistency of the diet during the 10-wk intervention. Subjects were advised to limit sh and nuts intake (do not eat walnuts and eat sh at most once a week), and to avoid taking omega-3 or axseed supplements. All patients were also asked not to change their physical activity patterns during the study. Participants were followed via short text messages and phone calls to detect any possible adverse effects. The compliance rate was determined by the number of empty bottles returned; those who consumed 85% of the oils or more were considered as an adherent to the study.

Anthropometric data and dietary intake
Participant's weight and body composition were evaluated by portable TANITA M780 (Tanita, Japan) and height was measured by Seca 763 scale (Hamburg, Germany). Weight was measured with light clothes without shoes in standing position. Height was measured in standing position, looking straight ahead, arms at sides, and shoulders relaxed with no shoes with a precision of 0.1 cm. BMI was then calculated by dividing the weight (kg) by the square of height (m2). A trained nutritionist evaluated the participants' dietary intake by a three days dietary recall questionnaire. Patients were asked to complete three-day food record questionnaires (including two consecutive days and a day-off) at the beginning and end of the study. Then, each food item was entered into a customized Nutritionist IV software (1997, First DataBank Inc., San Bruno, CA), and mean intake of energy, macro and micronutrients were calculated at the baseline and 10 weeks post-intervention.

Laboratory data
To assess the lipid pro le, blood samples were obtained from all participants after 12 hours of fasting. Blood samples were centrifuged at 3000 rpm for 10 minutes and separated serums were frozen at -80°C until further biochemical measurements. Serum concentrations of TG, HDL, LDL, and total cholesterol were measured using an enzymatic colorimetric method (Pars Azmoon, Tehran, Iran). Non-HDL cholesterol was calculated by subtracting HDL-C from TC and represented the LDL + IDL + VLDL cholesterol fractions (37). At the beginning and the end of the study, liver sonography was performed by a skilled radiologist for all patients using an ultrasound device (General Electric LOGIQ E9-using probe 3.5/5 MHz, USA). The radiologist was blind to the study process and study groups. A rating method was introduced to achieve a semi-quantitative measure of the extent of fat deposition in the liver. Based on parameters including liver echo-texture, the brightness of the liver, the contrast ratio of the liver-to-kidney, and blurred vessels, the degree of hepatic fatty in ltration was scored from grade 0 to grade 3.

Statistical analyses
To analyze the data, the SPSS software version 21 (IBM Corp. IBM SPSS Statistics for Windows, Armonk, NY) was used. The normality of quantitative data was evaluated by the Kolmogorov-Smirnov test. Data were presented as mean±standard deviation (SD) for quantitative variables and frequency (%) for categorical variables. To analyze the qualitative variables, we applied the chi-square test. Changes of the quantitative variables were compared pre-and postintervention using paired sample t-test. The participants' baseline characteristics were compared between the groups by independent samples t-test or Chi-Square test, whatever applicable. One-way analysis of variance (ANOVA) and LSD post-hoc tests were used to compare the groups in terms of quantitative variables. Also, an analysis of covariance (ANCOVA) was used to adjust the effect of confounding variables. The P-value<0.05 was considered as statistically signi cant.

Study baseline characteristics
In the present study, 72 (80%) participants completed the trial. Eighteen participants (six in each group) failed to follow the protocol (Fig. 1). The baseline characteristics of the participants are presented in Table 2. 72 patients with a mean age of 29.31 ± 6.52 were included in the nal analysis. There were no signi cant differences between the three groups in age, height, weight, disease duration, physical activity, marital status, taking OCP, and metformin in the last three months.
Compare the dietary intakes of the participants Table 3 summarizes the macronutrients and micronutrients intake of the study participants. Statistical analysis of energy, macronutrients, and dietary intake of PUFA, SFA, LA, and dietary ber showed no signi cant differences between the three groups at baseline and end of the study (P>0.05). However, at the end of the study, dietary intake of MUFA was signi cantly higher in the canola and olive oil groups (P<0.001). Also, dietary intake of ALA was signi cant only in the canola oil group (P=0.016).
Effect of canola and olive oil on the lipid pro le Table 4 compares the lipid pro le between canola, olive, and sun ower groups at the baseline and following 10 weeks of the study. At the beginning of the study, there were no signi cant differences were observed between the three groups in none of the lipid pro le variables. After 10 weeks intervention, there was a signi cant reduction in serum TG (P=0.012), Non-HDL (P=0.014), TC/HDL (P=0.007), LDL/HDL (P= 0.040), and TG/HDL (P=0.006) in the canola group, but not in the olive oil and sun ower groups. Also, we showed a marginal signi cant effect regarding TC (P=0.076) and LDL (P=0.055) in the canola oil at the end of the study. The results showed that canola oil supplementation compared to the olive oil and sun ower oil, caused a signi cant reduction in serum levels of TG (P=0.004), and TC/HDL (P= 0.021), LDL/HDL (P= 0.047), and TG/HDL (P=0.001) indexes. As shown in Table 4, the changes were not signi cant after adjusting the confounding factors.
Effect of canola and olive oil on the Fatty liver severity Within-and between-group changes in terms of fatty liver severity are summarized in Table 5. In within-group comparisons, fatty liver severity was signi cantly reduced in canola (P<0.001) and olive groups (P=0.005), while no changes in the control group were observed. Moreover, there were no signi cant differences between the canola and olive oil groups in terms of fatty liver severity reduction (P=0.404).

Discussion
In the current study, we investigated the effects of canola and olive oils for 10 weeks on lipid pro le and NAFLD severity among women with PCOS. We found canola oil has bene cial effects on some lipid pro le parameters and fatty liver severity. Also, olive oil consumption led to signi cant improvement in the fatty liver severity while had no effects on the lipid pro le. Previous studies have reported that changes in dietary fatty acid composition, especially replacement of SFA with PUFA and MUFA improve blood lipid levels even in those participants with low initial concentrations (1,2). In our study, participants in the canola oil group showed a signi cant decrease in serum TG and Non-HDL concentrations and TC/HDL, LDL/HDL, and TG/HDL ratios. Also, we showed a marginally signi cant effect regarding TC and LDL concentration in the canola oil. Similar to our ndings, Ghobadi et al. in a meta-analysis study showed that canola oil decreased LDL and TC compared to sun ower oil and saturated fat(3). Also, Sarkkinen et al. Performed a 6-month dietary intervention trial in hypercholesterolemic patients and found that LDL-C levels in the canola oil group were lowered (3.7%) from baseline levels (4). The effects of canola oil on TC and LDL may be dependent on the types of fatty acids that are replaced by the oil, study duration, and baseline levels of TC and LDL (5).
According to the previous ndings, the in uence of various oils on blood lipids is controversial. Some researchers reported that TG and very-low-density lipoproteins (VLDL) levels were increased after consuming olive oil compared with canola and sun ower oils (6, 7), while others have found different results (8,9). Both olive oil and canola oil contain high amounts of MUFA (10). A higher intake of MUFA can improve the entry of TG into the blood circulation that could speed up its clearance (11).In our trial, it could be the reason for the signi cant decrease in TG concentrations after consuming canola oil, however, this mechanism is not true about the effect of olive oil as it led to a slight increase in TG levels but it is not statistically signi cant. Canola oil is a good source of oleic acid, ALA, and phytochemicals (12). It is not well known the exact mechanisms of Canola oil on serum lipids, however, it is probably related to its fatty acid composition, which is high in MUFAs and PUFAs, particularly ALA (13). It has been shown that a higher amount of ALA intake can increase insulin secretion, improve insulin sensitivity, and lipoprotein lipase activity, leading to serum TG reduction (14). Moreover, the PUFA content of canola oil can downregulate the VLDL-c and apolipoprotein-B100 synthesis, which can lower serum TG concentrations (12,15). Previous studies have reported that a dietary pattern with a higher amount of PUFA can decrease serum LDL-c and TC, but not TG and HDL-c (16,17). In the present study, canola oil supplementation caused a marginally signi cant reduction in the LDL-c and TC concentrations (Table4). In line with our ndings, Salar et al. in a clinical trial showed that canola oil at the dose of 30 g/day decreased serum LDL-c more than sun ower oil, a poor source of ALA (18). Furthermore, in the Sacks et al.'s trial, they found a signi cant decrement in serum levels of LDL-c and TC/HDL ratio following the replacement of SFA by canola oil (19). We found that canola oil consumption led to a signi cant reduction in TC/HDL, LDL/HDL, and TG/HDL ratios, which are known as the main predictors of CVD (20). It has been reported that a higher intake of canola oil can reduce the risk of coronary heart disease and all-cause and cardiovascular mortality due to its PUFA content (13).
We were unable to nd any signi cant reduction in lipid pro le regarding the olive oil consumption. The results of some previous studies are in contradiction with our ndings. Jamal et al. in an RCT evaluated the effect of olive oil on lipid pro les and blood glucose levels in type 2 diabetic patients. They found that olive oil in a dosage of 30 mL/day caused a signi cant decrease in serum levels of TG, TC, and LDL-c (21). Venturini et al. in their clinical trial found that extra virgin olive oil (20 mL/day) in combination with sh oil (3 g/day) could decrease serum levels of TC and TC/HDL-c and LDL-c/HDL-c indexes (22). Also, it has been reported that extra virgin olive oil in a dosage of 4 g/day in mildly hypocholesterolemic participants was associated with favorable changes in plasma lipid pro le (23). In a recent meta-analysis, researchers reported that olive oil cause a signi cant reduction in serum levels of TC, LDL-c, and TG (24). The mechanisms by which olive oil can exert its bene cial antioxidative effect can be clari ed by polyphenol activity or through the cumulative protective effect of both its polyphenols and MUFA content (25).
The lack of bene cial effects of olive oil on blood lipids may arise for some reason. Olive oil contains more SFA than canola oil that can impair the lipid pro le. Besides, the n-6/n-3 ratio is higher in olive oil than that of canola oil (58). Several studies have illustrated that a higher intake of n-6 PUFA increases the risk of CVD and it is independently associated with elevated serum TC and TG (26).
In the present study, we showed that both olive and canola oils could decrease fatty liver severity, however, no signi cant difference was observed between them. In agreement with our nding, Nigam et al. in an interventional study including 93 males with NAFLD showed that consumption of olive oil and canola oil at the dose of 20 g/day led to a signi cant decline in fatty liver grade and other NAFLD risk factors (27). In another study, Kruse et al. compared the e cacy of canola oil and olive oil on hepatic steatosis in obese men. In their study, 27 obese men consumed 50 g/day of either canola or olive oils for 8 weeks and results showed that canola oil compared to olive oil caused a greater reduction in hepatic steatosis (28).
Dietary ingredients, especially the type and amount of fats, are important in the deposition of liver fat and are responsible for 15% of the liver fat content. Dietary fat can exacerbate hepatic steatosis in both direct and indirect pathways via in uencing the adipose tissues (29). The positive effects of olive oil on hepatic fat content can be clari ed by the faster oxidation of MUFAs than SFA in the postprandial phase (30). For instance, consumption of diets rich in MUFA (28% to total calories) for 8 weeks by patients with type 2 diabetes reduced liver fat by 30%, and this decline was associated with augment of postprandial β oxidation of fatty acids (31). Similarly, Errazuriz et al. in an RCT showed that high MUFA intake (28% Calorie, half as olive oil) in prediabetes patients for 12 weeks lowered hepatic fat and improved hepatic and total insulin sensitivity (32). Additionally, a high-MUFA diet boosts lipoprotein lipase activity more than a diet rich in SFA which leads to enhanced clearance of circulating triglyceride-rich lipoproteins (33). Also, a higher amount of MUFA and ALA in canola oil can exert bene cial effects against fatty liver by the improvement of insulin sensitivity, glucagon-like peptide-1 responses, and up-regulation of glucose transporter-2 expression in the liver of insulin-resistant participants (34). Besides, a higher intake of MUFA and ALA can increases lipid oxidation, inhibits hepatic triacylglycerol synthesis, and decreases insulin resistance (35).
In our study, we included only women. As mentioned, there is a sex difference in the conversion of ALA to EPA and DHA that is higher in women than in men. It has been reported that females had higher erythrocyte phospholipid EPA, lower adipose tissue EPA and higher plasma DHA content. Previous studies showed that there is an inverse relationship between the menopausal status of women and the age of female participants with the change in plasma EPA content after olive oil supplementation (36,37). A part of the inconsistency in our results compared to other studies could be due to these reasons.
We have not found any signi cant effect from the sun ower oil on lipid pro le and Fatty liver severity. Sun ower oil is a rich source of omega-6 fatty acids. Previous studies have shown that a high intake of foods rich in omega-6 fatty acids can exacerbate in ammation and liver damage. Due to the dietary pattern of the Iranian people and the cheapness of sun ower oil compared to olive or canola oil, most people, especially middle-income or low-income families, use this type of oil in their food preparation.
As a strength of our study, its design as a randomized clinical trial to assess three types of conventional oils from different sources can be mentioned. We encountered some limitations in our investigation. The main limitation of this study was the method of fatty liver evaluation. Ultrasound is not very accurate to detect mild cases of the fatty liver compared to the broscan method. Also, we couldn't measure the ALA, EPA, and DHA content of erythrocyte membranes as a good indicator of participants' adherence.

Conclusion
The results of this study showed that canola oil, as a good source of MUFA and ALA, can exert bene cial effects compared to the olive and sun ower oils in improving lipid pro le and fatty liver status among the women with PCOS. Further studies are needed to con rm these preliminary ndings and to determine the proper doses for achieving the best results.  NAFLD, Non-Alcoholic Fatty Liver Disease. P-values are resulted from ANOVA or Chi-Square or kruskal-wallis test.