Health Effects of Various Edible Vegetable Oil: An Umbrella Review

Vegetable oils, derived from diverse sources such as seeds, nuts, and some fruits, play a significant role in dietary health. However, the current evidence on the health effects of different types of vegetable oil consumption remains controversial. This umbrella review aims to synthesize evidence from systematic reviews and meta-analyses to assess the health outcomes associated with various vegetable oils. A comprehensive literature search was performed up to 31 July, 2023 on 12 databases for studies examining the association of different vegetable oils with health outcomes in adults. Data was extracted independently by 2 authors, with evidence strength assessed using the grading of recommendations, assessment, development, and evaluation criteria. A total of 48 studies, including 206 meta-analyses, were included. Moderate to very low certainty evidence showed that monounsaturated and polyunsaturated fatty acids such as canola oil, virgin olive oil, and rice bran oil are beneficial in reducing serum total cholesterol and low-density lipoprotein (LDL) concentrations. Conversely, low to very low certainty evidence suggests that oils high in saturated fats, such as coconut oil and palm oil, increase total cholesterol and LDL concentrations but also raise high-density lipoprotein concentrations. Very low certainty evidence showed the consumption of olive oil, sesame oil, and coconut oil could improve blood sugar control. Low certainty evidence showed olive oil consumption reduced risk of breast, digestive, and other cancers. Moderate to very low certainty evidence suggested that canola oil and sesame oil consumption reduced body weight. The consumption of vegetable oil appears to offer different health benefits, with summary estimates indicating beneficial effects on reducing lipid concentrations, especially with monounsaturated and polyunsaturated rich oils when consumed in recommended amounts. Future research should focus on long-term studies and comprehensive dietary assessments to better understand the health impacts of vegetable oils, providing a basis for informed dietary recommendations. This study was registered at PROSPERO as CRD42021239210.


Introduction
Vegetable oils are a heterogenous group of oils that are extracted from plant seeds (e.g., flaxseed oil, canola oil), nuts (e.g., peanut oil), flesh of fruits (e.g., palm oil, olive oil) as well as bran (e.g., rice bran oil) [1].In the past decades, studies on dietary fats and oils have gained interest.The diet-heart hypothesis suggests that the deposition of cholesterol in the arterial wall can be reduced by the modification of dietary fats, wherein saturated fats are replaced with vegetable oils rich in unsaturated fats [2,3].This subsequently slows the progression of atherosclerosis and coronary artery disease, and the survival rate can be improved.Dietary guidelines recommend the substitution of saturated fats with unsaturated fats, including vegetable oils, to improve health outcomes.This is consistent with the beneficial effects observed in cardiometabolic health and longevity found in the prospective cohort studies using canola oil and olive oil [4,5].Nevertheless, there has been growing controversy on the consumption of vegetable oils on health outcomes, including the incidence of cardiovascular diseases (CVDs), all-cause deaths, metabolic syndrome, and nonfatal cardiovascular events.Some randomized controlled trials (RCTs) revealed that the benefits of vegetable oils in lowering serum cholesterol concentrations did not translate into better clinical outcomes in terms of disease risks and survival [2,3].
Studies have shown that the benefits or harms of vegetable oils are highly dependent on the fatty acid profile in terms of the type and fraction of fatty acids present [6].For instance, oils predominantly composed of monounsaturated oleic acids, such as olive oil, canola oil, peanut oil, and rice bran oil, have been linked with anti-inflammation and lipid-lowering properties [7][8][9].In addition, research has documented the role of oils containing PUFAs such as the n-3 α-linolenic acid (ALA) and n-6 linoleic acid (LA) including flaxseed oil (~50% ALA), traditional sunflower oil (~60-70% LA), soybean oil (~51% LA), sesame oil (~43% LA) and peanut oil (~30% LA) [1].Both ALA and LA are important as they provide essential fatty acids that cannot be synthesized by the human body, along with reported cardioprotective effects [10].On the contrary, palm oil contains a balanced ratio of SFA and unsaturated fatty acids (~50% palmitic and stearic acid, along with ~40% oleic acid) [11].The palmitic acid content contributes to the heat stability of palm oil, yet the high degree of saturation has been associated with increased LDL cholesterol [12].Similarly, coconut oil mainly comprises SFAs, particularly the predominant medium-length saturates, lauric acid.Although the use of coconut oil has garnered interest for its purported benefits related to lauric acid, these remained inconclusive given that lauric acid may not biologically act as medium-chain triglycerides, along with the presence of other long-chain SFAs in coconut oil [13,14].
Considering the association between nutritional habits and overall health, there is now a suggestion to shift from focusing on the effects of individual fatty acids to edible cooking oils as a whole concerning health outcomes [6,7].Several reviews have studied the effects of various edible oils on health.For instance, a significant reduction in the risk of CVD events and weight gain was reported on olive oil through observational studies and RCTs [8].Another systematic review found some improvement in fasting blood glucose and insulin sensitivity with the consumption of flaxseed oil.
To date, there is no comprehensive overview of studies comparing the various health effects of edible vegetable oils, suggesting that this aspect needs to be revisited.We aimed to summarize the existing evidence comparing the health effects of vegetable oils on health outcomes, including blood pressure, blood lipid concentrations, and glucose concentrations, to serve as a reference for future researchers in this area.

Methods
The study was registered in PROSPERO (CRD42021239210).The systematic literature search adhered to the guidelines outlined in the PRISMA.

Literature search and selection criteria
A systematic literature search was performed up to 31 July, 2023 using the following databases: PubMed, Embase, Psy-cINFO, Database of Abstracts of Review of Effectiveness, Health Technology Assessment databases and reports, National Health Service (NHS) Economic Evaluation Database, HealthSTAR, BIOSIS, Science Citation Index, Cochrane Central Register of controlled trials, CINAHL Plus and Allied and Complementary Medicine Database.The search focused on systematic reviews and meta-analyses investigating the health effects of vegetable oil consumption, with no language restriction.The search terms are described in Supplemental Text 1.This was supplemented with a review of the reference lists of eligible reviews and metaanalyses.
Studies were included if they met the following criteria: 1) systematic reviews or meta-analyses of studies in adults that 2) investigated the association of different vegetable oils (e.g., soybean oil, palm oil, peanut oil, and sunflower oil) on 3) health outcomes, (e.g., lipid parameters, weight).Studies were excluded if they were 1) primary studies, 2) vegetable oils that are used in the form of dietary supplements, or 3) the individual sterol components of the vegetable oils (e.g., β-sitosterol, campesterol, and brassicasterol).Publications that reported only on the exposure to plasma concentrations or biomarkers without the dietary intake of vegetable oils were further excluded.

Data extraction
Two authors independently extracted the data.When the study included a meta-analysis, the effects of each comparison on health outcomes, along with their 95% confidence intervals (CIs) and quality scores, were also extracted.All extracted data were double checked and verified by the senior author (SWHL), and any disagreement was resolved through consensus.If multiple published meta-analyses were identified on the same association, the primary studies from each meta-analysis were extracted for each exposure to avoid the inclusion of duplicate studies.

Assessment of methodological quality
The assessment of the methodological quality of each included published meta-analysis was conducted independently by 2 reviewers using the validated AMSTAR-2 (A measurement tool to assess systematic review) tool [9].Subsequently, the assessments were examined by another 2 reviewers.In cases of disagreement and failed consensus, further consultation was obtained from a third reviewer (SWHL).

Assessment of summary effects
All outcomes were synthesized and narratively described, with the findings presented in a tabular format.The review characteristics and findings were presented in the summary table.The meta-analysis and corresponding 95% CI were recalculated using the DerSimonian and Laird random effects model for each outcome.This was performed using summary effects of the primary studies that compared 2 different oils reported in the published meta-analyses.If 2 or more reviews assessed the same outcome and consumption of edible oil, data from the latest study were included.However, both sets of findings were compared.The quality of evidence provided in the meta-analyses was evaluated using the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) criteria.The evidence was rated from very low, low, moderate, and high certainty in the evidence based on the 5 recommended domains [10].
Heterogeneity was examined through Cochran's Q test and I 2 statistic with the identification of heterogeneity based on a P value of <0.1 or I 2 50%.Funnel plots were visually inspected to assess publication bias and small study effects.STATA version 16.0 (StataCorp LLC) was used for all analyses.

Literature search
The literature search identified 4166 articles, of which 3175 articles were screened.Eighty articles were reviewed, with 48 articles included in the final umbrella review.Reasons for exclusion included conference abstracts (n ¼ 3), noncooking oils (n ¼ 7), examined effects of diet (n ¼ 9), not systematic review (n ¼ 4), and they examined active ingredients of oils (n ¼ 10) (Figure 1 and Supplemental Text 2).

Methodological quality
AMSTAR-2 tool analysis found that only 1 included review was evaluated as high in confidence level.Seven (14.6%) reviews were moderate confidence, 30 (62.5%) were low confidence, and 10 (20.8%) were rated as critically low confidence (Supplemental Figures 1 and 2).These reviews were rated low or critically low confidence as they had the following critical flaw(s) with 35 reviews lacking a list of excluded studies with justification for their exclusion; 17 reviews did not address the risk of bias in the interpretation of the results, and 14 reviews did not mention the establishment of the review methods before the reviews were conducted (Supplemental Figures 1 and 2).

Summary of findings
The main results, together with the effect sizes for each of the reported outcomes, were summarized in      ) and high strength of evidence (n ¼ 2; 4.3%), and were primarily related to lipid outcomes and anthropometric measurements (Supplemental Tables 1-10).
Overall, some beneficial effects on lipid profiles were reported with the consumption of canola, virgin olive, and rice bran oil, whereas the consumption of sesame and flaxseed oil was found to improve blood pressure.Meanwhile, the consumption of both sesame and olive oils was reported to improve glycaemic profiles, whereas sesame oil consumption could reduce body weight and BMI (in kg/m 2 ), and the consumption of olive oil reduced some cancer risk.The full description of these effects is described in detail below.
In terms of cholesterol ratio, the effects of vegetable oils on LDL: HDL, TC: HDL, and TG: HDL were heterogeneous.However, canola oil was found to reduce apolipoprotein B concentrations [47], whereas coconut oil increased apolipoprotein A1 concentrations [24].Olive oil [11], rice bran oil [53], and palm oil [43] were found to have no significant effects on apolipoprotein concentrations.

Glycaemic outcomes
Reports on the effect of vegetable oil consumption on blood sugar controls were scarce.Hemoglobin A1c was found to decrease significantly with coconut oil [24], olive oil [14], and sesame oil [52], whereas fasting blood glucose concentrations were reduced with olive oil [14] and sesame oil [52].No significant effects were found on fasting blood glucose concentrations or HOMA-IR with canola oil [47], peanut oil [58], coconut oil [24], and palm oil [42].

Anthropometric indices
The use of canola oil and sesame oil was reported to reduce body weight between 0.35 kg and 0.30 kg; however, peanut oil consumption has been shown to increase weight (0.90 kg; 95% CI: 0.40, 1.40).No significant effects were found on waist circumference, body fat, waist: hip ratio, android: gynoid fat ratio, hip circumference, or lean mass for all other vegetable oils.

Cancer risk
Only 1 study investigated the effect of olive oil consumption on the risk of developing cancers [16].The report indicated that olive oil consumption was associated with lower odds of developing cancer, including breast cancer and digestive cancer [16].

Discussion
To the best of our knowledge, this umbrella review represents the first effort to synthesize the evidence from existing systematic reviews and meta-analyses on the effects of vegetable oils on health outcomes.The present review supports the growing body of evidence suggesting vegetable oils rich in MUFAs and PUFAs, such as canola oil and rice bran oil, have desirable effects in reducing TC and LDL concentrations.MUFAs and PUFAs are antiatherogenic and anti-inflammatory and demonstrate beneficial effects on cholesterol concentrations, consistent with epidemiological and clinical data [59].Notably, the current review showed that only virgin olive oil [11] enriched with higher concentrations of polyphenols are commonly associated with additional health benefits, including antioxidant effects, anti-atherosclerotic potential, and anti-inflammatory properties that lower lipid parameters.
In contrast, we found that oils relatively rich in saturated fats, such as coconut and palm oil, tend to increase TC and LDL.On the contrary, coconut oil has gained popularity in recent years for its potential health benefits attributed to the presence of medium-chain TGs that are absorbed intact and directly in the liver [60].However, our present review found limited evidence supporting these claims.On the contrary, it has been reported that palm oil can adversely affect blood lipid profiles due to its higher saturated fat content when compared to other vegetable oils, which contain lower concentrations of saturated fats [42].Such effects may be associated with CVD risks, but this needs to be taken in light of the beneficial effects of increasing HDL concentrations.
Palm oil is a versatile product that can be fractionated into palm olein, palm stearin, and palm mid-fractions.These palm oil fractions have very different physicochemical and nutritional properties as well as functions.Therefore, it is imperative for this study to differentiate between these products and avoid categorizing them under the common term "palm oil."Our review found that when focusing on palm olein, the liquid fraction of palm oil commonly used in cooking in many Southeast Asian countries, it has minimal impact on lipid parameters when compared to other vegetable oils rich in unsaturated fatty acids, such as olive oil, canola oil, soybean oil, and high-oleic sunflower oil.This observation may be attributed to the positional distribution of fatty acids on the TG backbone present in palm olein [57,61].Additionally, palm olein exhibits a higher degree of unsaturation, with elevated concentrations of oleic acid (39-45%) and LA (10-13%) [61].Similar to other vegetable oils rich in unsaturated fatty acids, the stereospecific number-2 position of the TG backbone in palm olein predominantly contains MUFA and PUFA, which are known to influence lipid profiles, a topic that has been discussed [62,63].
There is now a growing interest in the consumption of virgin oils as functional foods compared to the refined form of the oils [64].This interest stems from the refining processes involved, which may reduce the amount of antioxidants and polyphenols in the oils, compounds known for their health-protective effects [20].For instance, virgin olive oil, which contains high concentrations of polyphenols, can reduce oxidative stress that is often associated with chronic diseases [55].Although our current review aimed to explore the benefits of different oils and their extraction methods, we only identified 2 studies that have examined this aspect, which reported the health effects of virgin olive oil consumption compared to olive oil [55,56].Therefore, future research should delve deeper into this topic and address any existing controversies.
The health benefits of various vegetable oils were also reflected in other outcomes to an extent, including blood pressure, hemoglobin A1c, weight and BMI.Of note, the consumption of flaxseed oil demonstrated a slight reduction of 4.10 mmHg in diastolic blood pressure, which is nearing the clinical relevance associated with lower disease risks [65].The distinctive feature of flaxseed oil is the amount (>50%) of (n-3) ALA that can be converted to EPA and DHA [1].Along with EPA and DHA, ALA was also suggested to have cardioprotective potential, thereby improving blood pressure [32].
Our current review complements the recent review by Teasdale et al. [66], which reported the profile of nutrients and bioactive contents of edible oils and fats to better inform dietary recommendations, albeit with some differences.Unlike our review, they attempted to delineate the impact of various bioactive contents in each oil.For instance, the review found that most studies to date have reported some beneficial effects on the cardiovascular effects of these edible oils.These beneficial effects were generally attributed to the use of biophenols and flavonoids, which can be found in oils such as extra virgin olive oils, coconut oils, hemp seed, and avocado oil.Emerging evidence also suggests some beneficial effects with the use of tocopherols and squalene, which are often rich in oils such as palm oil, soybean oil, canola oil, and sunflower oil.

Strengths and limitations
Our study demonstrated several strengths.The study offers a comprehensive systematic overview of the evidence from all published systematic reviews and meta-analyses on the health benefits of various vegetable oils.The methodological quality of reviews was assessed, and the quality of evidence was graded using validated tools.However, this study has certain limitations which cannot be neglected.First, the primary studies included in each meta-analysis were not assessed directly.Therefore, the results could have been influenced by primary studies not included in the published meta-analyses or additional studies published after the reviews.Recalculation for 54 of the 206 meta-analyses was impossible as no estimates were reported.We also could not determine the populations examined for meal compositions, including the types of fatty acids (SFA, MUFA, and PUFA) intake and their quantities, which are important for interpreting results.For example, studies that used standardized feeding methodology might have provided a more accurate representation of the effects of vegetable oils on lipid and glycemic outcomes rather than cross-sectional studies that confounding variables might influence.
Additionally, the populations examined in all studies were heterogeneous and might have included those with hyperlipidemia; therefore, they might not reflect a true direct comparison of the effects.As such, to derive recommendations, further investigations are warranted.Our analysis included the largest number of primary studies for each outcome or most recent publication and oil type.Therefore, we might not have chosen meta-analyses with the highest quality of evidence.Nevertheless, most updated reviews often included the same primary studies as prior reviews.Finally, due to the limited number of primary studies, most of the outcomes reported in our umbrella review were only of low-quality evidence.As such, including a meta-analysis with even fewer primary studies would not change the quality of evidence compared to those in our umbrella review.

Implications for practice
This review has identified several significant potential implications for the public, clinicians, and policymakers.Firstly, limited evidence shows that consumption of vegetable oils rich in MUFA and PUFA instead of SFAs can reduce coronary events, including mortality, despite various mechanisms and evidence supporting MUFA-and PUFA-rich vegetable oils benefits in lowering lipid concentrations in adults with or without comorbidities [47,53].Secondly, this review also noted that the different types of vegetable oils offer different health benefits due to their unique fatty acid profiles.It is crucial to emphasize that the reported health benefits of these vegetable oils consumed are based on the amounts recommended in dietary guidelines and not by overconsumption/increasing the total amounts of calories consumed daily, as this may lead to weight gain and possibly obesity.
Finally, it is necessary to implement research investigating the effects of the long-term safety and efficacy of various vegetable oils, given that most of the reviews included in this study only examined RCTs, which are often conducted over a short period of 6 mo.Studies should also explore the question of internal consistencies, such as comparing virgin coconut oil with coconut oil and the influence of mediating factors such as body fats.
In conclusion, numerous studies have examined the effects of various vegetable oils on health outcomes.This review suggests that different vegetable oils offer different health benefits, which provide potential primary preventive effects of diseases.Given the challenges in evaluating the impacts of various vegetable oils independent of other dietary practices and the trial durations studied, upcoming research should prioritize obtaining comprehensive dietary data and concentrate on long-term clinical outcomes such as cardiovascular events and mortality.

Table 2
. Based on the random effects model, with 206 analyses performed, 47 (22.8%)FIGURE 1. PRISMA flow chart showing the selection of studies.PRISMA, preferred reporting items for systematic reviews and meta-analyses.P.T. Voon et al.Advances in Nutrition 15 (2024) 100276

TABLE 1
Characteristics of included studies in the current review

TABLE 2
Study summary of the included reviews (continued on next page) were found statistically significant.These results were supported mainly by a very low strength of evidence (n ¼ 26; 55.3%), followed by low strength of evidence (n ¼ 11; 23.4%), moderate strength of evidence (n ¼ 8; 17.0%