Fats and oils – a scoping review for Nordic Nutrition Recommendations 2023

This scoping review for the Nordic Nutrition Recommendations 2023 summarizes the available evidence on fats and oils from a food level perspective. A literature search for systematic reviews (SRs) and meta-analyses was conducted in PubMed. There are few SRs and meta-analyses available that investigate the association between fats and oils (food level) and health outcomes; the majority report associations at the nutrient level (fatty acid classes). All identified SRs and meta-analyses were of low methodological quality, thus the findings and conclusions presented within this scoping review should be interpreted cautiously. Based on this limited evidence, the following results were indicated: the intake of olive oil may be associated with reduced risk of cardiovascular disease (CVD), type 2 diabetes (T2D), and total mortality in prospective cohort studies. The intake of butter was not associated with the risk of CVD but may be related to slightly lower risk of T2D and higher risk of total mortality in prospective cohort studies. For cancer, the evidence is sparse and primarily based on case-control studies. The intake of olive oil may be associated with reduced risk of cancer, whereas the intake of butter may be associated with increased risk of certain cancer types. Butter increases LDL-cholesterol when compared to virtually all other fats and oils. Palm oil may increase LDL-cholesterol when compared to oils rich in MUFA or PUFA but may not have any effect on glucose or insulin. Coconut oil may increase LDL-cholesterol when compared to other plant oils but may decrease LDL-cholesterol when compared to animal fats rich in SFA. Canola/rapeseed oil may decrease LDL-cholesterol compared to olive oil, sunflower oil and sources of SFA and may also reduce body weight compared to other oils. Olive oil may decrease some inflammation markers but may not have a differential effect on LDL-cholesterol compared to other fats and oils. The effect on risk markers likely differs depending on the type/version of oil, for example, due to the presence of polyphenols, phytosterols and other minor components. Taken together, based on the available evidence, oils rich in unsaturated fat (e.g. olive oil, canola oil) are to be preferred over oils and fats rich in saturated fat (e.g. butter, tropical oils).


D
ietary sources of fat are usually, and conveniently, defined by their fatty acid compositions as saturated (SFA), monounsaturated (MUFA) or polyunsaturated (PUFA) fat rich sources (illustrated in Fig. 1) as the degree of saturation is known to have a meaningful impact on important risk markers (e.g.blood lipids).

Popular scientific summary
• Fats and oils vary mainly in fatty acid composition but also in their content of polyphenols and other bioactive components.• Among the Nordic and Baltic countries, the use of fats and oils appears to be highest in Denmark and Finland.• The intake of olive oil is associated with reduced risk of cardiovascular disease, type 2 diabetes, some types of cancer, and total mortality.• Butter, palm oil, and coconut oil increase LDL cholesterol levels when compared to plant oils high in unsaturated fatty acids.• Choosing vegetable oils rich in unsaturated fatty acids, such as olive and rapeseed oil, over tropical plant oils and animal fats high in saturated fatty acids seems most beneficial for health.Fredrik Rosqvist and Sari Niinistö Although the degree of saturation is likely to be the primary mediator in terms of health effects of dietary fats and oils, it is not the only factor.Vegetable oils may also contain different levels of other bioactive components, such as polyphenols, antioxidants, vitamins and phytosterols, potentially mediating parts of the overall health effect.These bioactive compounds may be more or less preserved depending on the degree of processing (1).Olive oil is a well-known example, where the level of polyphenols may vary widely due to agronomic factors (e.g.ripeness of the olives and cultivation conditions), extraction technology, and mixing of different fractions (2).In general, extra-virgin olive oil has the highest content of polyphenols whereas levels may be very low in refined olive oils.Similarly, the natural content of phytosterols (known to have beneficial effects on blood cholesterol levels) varies markedly between different fats and oils, being markedly higher in, for example, rapeseed and corn oil than in, for example, sunflower, soybean and olive oil (1).However, and similar to polyphenols on olive oil, the levels of phytosterols in oils will be affected by the cultivation conditions, extraction method, and degree of refinement (1).Thus, the overall health effect of various sources of fat may differ, even if the fatty acid composition is similar, due to complex interactions between nutrients, non-nutrients and physical structure (known as food matrix-effects).Furthermore, for some fats and oils, multiple fractions may be obtained.An illustrative example of this is palm oil, consisting of roughly equal parts of saturated and unsaturated fatty acids and is viewed as a source of saturated fat in the diet.However, multiple different fractions can be obtained from palm oil.These fractions have different physicochemical properties and differ markedly in their fatty acid composition (3).For example, compared to 'palm oil', 'palm olein' has a lower content of palmitic acid (and total SFA) and is semi-solid in room temperature, whereas both 'super olein' and 'top olein' are liquid in room temperature due to even lower contents of palmitic acid (and total SFA) (3).In contrast to 'palm oil', 'palm stearin' has a markedly higher content of palmitic acid (and total SFA) and is solid at room temperature (3).
In extreme contrast to 'palm oil', 'palm olein' and 'palm stearin', palm kernel fractions (e.g.palm kernel oil and -olein) have a very low content of palmitic acid and is dominated primarily by lauric acid (~45%) but also myristic acid (~15%) (3).Finally, red palm oil (crude/unprocessed) has a high content of carotenoids, antioxidants, vitamin E, and phytosterols but has limited utility and is seldom used.Thus, in order to draw conclusions about the health effects of palm oil compared to other oils, it is important to know which fraction(s) that have been used.Palm oil is a so-called 'tropical oil', which means that it is a vegetable oil rich in saturated fat derived from the tropical zone.These vegetable oils (also including e.g.coconut oil) are distinctly different from the so-called 'non-tropical oils', which are instead characterized by their low content of saturated fat.The aim of this scoping review is to describe the totality of evidence for the role of fats and oils for health-related outcomes as a basis for setting and updating food-based dietary guidelines in the Nordic Nutrition Recommendations (NNR) 2023 (Box 1).This scoping review primarily considers results at a food level perspective (i.e.studies reporting on specific food sources of fat, e.g.olive oil).Results from studies reporting at a nutrient level perspective (e.g.saturated fat) are considered in an accompanying paper (4).

Methods
This review follows the protocol developed within the NNR2023 project (5).The sources of evidence used in this review follow the eligibility criteria described previously (6).No de novo NNR2023 systematic review (SR) was performed for this paper, and no previously published qualified SRs reporting at a food level perspective were available (5,7).However, one qualified SR primarily reporting at a nutrient level could be used as supporting evidence for the association between butter and cardiovascular disease (CVD) (8) The search was repeated without filtering for 'Humans' which resulted in an additional three papers of relevance.The quality of the included studies (Table 1) was evaluated using the tool AMSTAR 2, modified for NNR (5).The certainty of evidence was not assessed.Excluded studies are presented in Table 2.The topics discussed in this scoping review are, to a large degree, defined by the topics covered by the SRs.In case of multiple and overlapping SRs, we chose the most recent and most updated SRs over the older, or used both if they were complementary.

Diet intake in Nordic and Baltic countries
It is challenging to directly compare the intake of fats and oils between the Nordic and Baltic countries as food groupings and assessment methods are not harmonized.Food group definitions, and thus level of information, differ between countries.Furthermore, data was collected at different time points (from 2010 to 2020) using different instruments (2 × 24 h recall, 4-day web-based food record, and 7-day food record), and participation rate ranged between 33 and 90%.Finally, contribution of alcohol to total energy intake was not included in Latvia and Lithuania.However, based on available data (9) it appears that the intake of fats and oils is higher in Denmark (men: 47 g/day, women: 35 g/day) and Finland (men: 53 g/day, women: 38 g/day) than in Norway (men: 39 g/day, women: 24 g/day) and Iceland (men: ~20 g/day, women: ~15 g/day).Data for Sweden only includes 'spreads on bread' (being 13 g/day for men, 10 g/day for women) and is thus difficult to compare to the other countries.In the Baltic countries, intake appears higher in Estonia (men: ~26 g/day, women: 19 g/day) than in Latvia (men: 13 g/day, women: 11 g/day) and Lithuania (men:12 g/day, women: 9 g/day).However, there is a large standard deviation/range in all countries.

Olive oil
In a meta-analysis from 2014 (10) based on seven cohort studies (population ranged from n ~3,300 to n ~41,000 with follow-up durations ranging from 3.7 to 11.3 years), the intake of olive oil (top vs. bottom third) was associated with a 28% reduction in the risk for combined • This paper is one of many scoping reviews commissioned as part of the Nordic Nutrition Recommendations 2023 (NNR2023) project (5) • The papers are included in the extended NNR2023 report but, for transparency, these scoping reviews are also published in Food & Nutrition Research • The scoping reviews have been peer reviewed by independent experts in the research field according to the standard procedures of the journal • The scoping reviews have also been subjected to public consultations (see report to be published by the NNR2023 project) • The NNR2023 committee has served as the editorial board • While these papers are a main fundament, the NNR2023 committee has the sole responsibility for setting dietary reference values in the NNR2023 project Another meta-analysis from 2014 based on the same data observed similar results for both stroke and CHD when exposure was expressed as 'per 25 g/day increment'.The beneficial effect of olive oil on CVD indicated by meta-analyses of observational studies is supported by the ~5 year-long PREDIMED trial, where the risk of CVD (composite of myocardial infarction, stroke and death from cardiovascular causes) decreased by ~30% in the group receiving extra-virgin olive oil in the context of a Mediterranean diet compared to the control group (receiving advice on a low-fat diet).

Palm oil
In a SR from 2018 (13), one single case-control study (n = 2,111 cases) from Costa Rica was identified comparing the intake of palm oil with soybean oil and 'other oils' regarding the risk of myocardial infarction.The results indicated that palm oil was associated with increased I 2 = 22%) when the duration of follow-up ranged from 4 to 22 years.Further, in a dose-response meta-analysis, 10 g increase of olive oil daily was associated with 9% lower risk (RR 0.91 [0.87-0.95]).In addition, a non-linear association was observed so that olive oil intake showed stronger protective association up to the 13 g daily intake (14).However, the quality of meta-evidence of these studies was graded as low (14).Moreover, a SR from 2015 (15) reported that olive oil was found to be protective from T2D in two randomized controlled Similarly, high olive oil consumption was related to lower risk for developing a gastrointestinal cancer, that is, colorectal, esophageal, gastric, and pancreatic cancer in overall meta-analysis that combined two cohort and 13 case-control studies (RR 0.77 [0.66-0.89],I 2 = 41%) as well as among 13 case-control studies (RR 0.72 [0.61-0.85],I 2 = 39%), whereas a protective association was not observed in a meta-analysis of cohort studies (RR 0.97 [0.75-1.24],I 2 = 21%) (18).Further, high intake of olive oil was protectively associated with upper aero-digestive cancer (laryngeal, nasopharyngeal, and oral/pharyngeal) among six case-control studies (RR 0.74 [0.60-0.91],I 2 = 33%) and urinary tract cancer (prostate and bladder) in a meta-analysis including six case-control studies (RR 0.46 [0.29-0.72],I 2 = 73%).
Furthermore, in a meta-analysis from 2022 (18) based on three cohort and 11 case-control studies, high versus low consumption of olive oil showed 33% lower risk of breast cancer (RR 0.67 [0.52-0.86],I 2 = 83%).However, in a subgroup analysis by study design, this protective association was observed only among 11 case-control studies (RR 0.63 [0.45-0.87],I 2 = 80%), while there was no association within the three cohort studies (RR 0.67 [0.29-1.56],I 2 = 78%).In another meta-analysis from 2015 (19) based on five cohorts and 11 case-control studies (11,161 cases in over 150,000 women from Europe, the US and Iran), the highest versus lowest category of intake of vegetable oils, including olive, safflower seed, peanut, soya, corn, or mixed oil, was not associated with the risk of breast cancer (OR 0.88 [0.77-1.01]).However, there was substantial heterogeneity among studies (I 2 = 74%).Neither dose-response meta-analysis (n = 12 studies with 6,253 cases in 73,842 women) did show any association between 10 g increment of daily vegetable oil intake and the risk of breast cancer (OR 0.98 [0.95-1.01])(19).Subgroup analysis showed no indication that menopause or hormone receptor status may modify the association between vegetable oils and the risk of breast cancer (19).

Other health outcomes
A meta-analysis from 2014 (25) investigated the association between dairy foods and risk of Parkinson's disease.Based on three effect sizes from two prospective cohorts (one from Finland, one from USA), the intake of butter (highest vs. lowest category) was not associated with the risk of Parkinson's disease (RR 0.76 [0.51-1.13],I 2 = 0.0%).
In a meta-analysis from 2021 (27) based on one prospective cohort and two case-control studies, higher butter consumption was associated with a 27% increased risk of endometriosis in the females compared to lower intake (1.27 [1.03-1.55],I 2 = 0%).Studies included 1,173 endometriosis cases from the US and Europe.
In the PREDIMED trial, supplementation with extra-virgin olive oil in the context of a Mediterranean diet was associated with a non-significant 10% reduction in risk for total mortality (HR 0.90 [0.69-1.18]).

Butter
In a meta-analysis from 2016 based on two large studies consisting of European populations (nine country-specific cohorts, ~10 years of follow-up, ~28,000 events), the intake of butter was associated with a 1% increased risk for all-cause mortality per 14 g/day increment (RR 1.01 [1.00-1.03])without heterogeneity (I 2 = 0%).A previous meta-analysis from 2013 (29) based on partly the same data (one overlapping study) observed no association between intake of butter and all-cause mortality (RR 0.96 [0.85-1.08]per each additional serving/week), with considerable heterogeneity (I 2 = 78%).

Butter
Butter increases LDL-cholesterol when compared to basically all other fats/oil, as demonstrated in a SR from 2021 (30) and a network meta-analysis from 2018 (31).The difference in LDL-cholesterol (−0.25 to −0.42 mmol/L per 10 E% isocaloric exchange) is largest when butter is compared to oils rich in unsaturated fats (safflower oil, sunflower oil, rapeseed oil, flaxseed oil, corn oil, olive oil, and soybean oil), but butter increases LDL-cholesterol also in comparison to other sources of SFA (coconut oil, palm oil, and beef fat).Similar effects were observed for total cholesterol.Fewer differences are observed for triglycerides and high-density lipoprotein [HDL]-cholesterol.However, sunflower oil, soybean oil, and palm oil were observed to decrease triglycerides compared to butter (−0.04 to −0.06 mmol/L per 10 E% isocaloric exchange), and coconut oil was observed to increase HDL-cholesterol compared to butter (0.04 mmol/L per 10 E% isocaloric exchange).

Palm oil
The most recent meta-analysis (32) included studies using palm oil or palm olein, but not studies using palm stearin, palm kernel oil or red palm oil.Compared to MUFA-rich diets, palm oil increased LDL-cholesterol (0.24 mmol/L [0.06-0.42])when based on n = 16 cross-over studies including a total of n = 365 participants, but decreased LDL-cholesterol (−0.28 mmol/L [−0.53 to −0.03]) when based on one parallel study including 60 participants.Compared to PUFA-rich diets, palm oil increased LDLcholesterol (0.26 mmol/L [0.06-0.45])when based on five cross-over studies including a total of 114 participants but not when based on three parallel studies including a total of 152 participants (0.54 mmol/L [−0.45 to 1.52]).Similar results, overall, were observed for total cholesterol.Palm oil increased HDL-cholesterol compared to PUFA-rich diets, both when based on five cross-over studies (0.08 mmol/L [0.01-0.15])and when based on three parallel studies (0.08 mmol/L [0.00-0.16]).Compared to MUFArich diets, there was no difference in HDL-cholesterol when based on 16 cross-over studies (0.03 mmol/L [−0.01 to 0.07]), but palm oil increased HDL-cholesterol when based on one parallel study (0.18 mmol/L [0.09-0.27]).Finally, palm oil had no differential effect on apoA1 or apoB when compared to either MUFA-rich diets (n = 3 cross-over studies) or PUFA-rich diets (n = 2 cross-over studies and n = 1 parallel study).
A previous meta-analysis (3) included only studies using palm olein and found that palm olein decreased LDLcholesterol when compared to other SFA-rich oils/fats (coconut, lard) (−0.50 mmol/L [−0.70 to −0.30]), but there was no difference when compared to MUFA-rich oils (olive oil, peanut oil, canola oil, high-oleic sunflower oil), PUFArich oils (soybean oil) or 'all other oils'.Results were similar for total cholesterol whereas no differential effect was found for triglycerides.Palm olein increased HDL-cholesterol when compared to MUFA-rich diets (0.04 mmol/L [0.00 to 0.07]) and decreased HDL-cholesterol when compared to other SFA-rich oils (−0.06 mmol/L [−0.11 to −0.00]), whereas no differential effects were found when compared to PUFA-rich oils or 'all other oils'.

Olive oil
In a meta-analysis from 2019 (35) based on 27 randomized trials, olive oil had no differential effect on LDLcholesterol compared to oils rich in n-3 PUFA (12 studies), n-6 PUFA (10 studies) or SFA (3 studies).When compared to 'all other plant oils combined', olive oil was found to increase LDL-cholesterol (n = 24 studies), but stratified analyses showed that this effect was only observed in studies lasting a maximum of 30 days (15 studies) and not in studies with longer duration (10 studies).Olive oil increased total cholesterol when compared to oils rich in n-3 PUFA (11 studies), n-6 PUFA (12 studies) and 'all other plant oils combined' (26 studies) but not when compared to SFA (3 studies).Olive oil increased both HDLcholesterol and triglycerides when compared to oils rich in n-3 PUFA and 'all other plant oils combined' but not when compared to oils rich in n-6 PUFA or SFA.Finally, olive oil had no differential effect on apoA1 or apoB when compared to 'all other plant oils combined' (10 studies).
Another two meta-analyses from 2019 compared the effects of different types of olive oil (36,37).Olive oil with a high content of polyphenols decreased both LDLcholesterol and oxidized LDL-cholesterol compared with olive oil with a low content of polyphenols or 'refined' olive oil.Furthermore, olive oil with a high content of polyphenols increased HDL-cholesterol when compared with olive oil with low content of polyphenols but not when compared to 'refined' olive oil.No differential effect was observed for diastolic blood pressure, but olive oil polyphenols may decrease systolic blood pressure.

Canola oil
A meta-analysis from 2020 (38) observed that canola oil decreased LDL-cholesterol when compared to olive oil (−0.17

Flaxseed oil
Based on a network meta-analysis (a technique for comparing three or more interventions simultaneously) from 2018 (31), the effect of flaxseed oil on LDL-cholesterol is similar to that of most other fats/oils (safflower oil, sunflower oil, rapeseed oil, hempseed oil, corn oil, olive oil, soybean oil, palm oil, coconut oil, beef fat, and lard), but lowers LDL-cholesterol in comparison to butter (−0.37 mmol/L per 10 E% isocaloric exchange).Effects were similar for total cholesterol.Flaxseed oil had no differential effects on HDL-cholesterol or triglycerides compared to any other fats/oils.However, all comparisons are based on only 1-2 studies.Another meta-analysis from 2016 (39) indicates that flaxseed oil may lower diastolic (−4.1 mmHg [−6.8 to −1.4]), but not systolic (−4.6 mmHg [−11.9 to 2.6]), blood pressure; however, the type of comparison was not specified.Finally, a SR and meta-analysis from 2021, based on nine RCTs, found no differential effect of flaxseed oil on total cholesterol, LDL-cholesterol, HDL-cholesterol or triglycerides compared to other oils (sunflower oil, safflower oil, corn oil) in patients with dyslipidemia-related diseases (40).

High-oleic vegetable oil
A SR from 2015 (41) investigated the effects of oils rich in oleic acid (primarily high-oleic sunflower oil, but also safflower, olive, and canola oil) on cardiovascular risk factors when compared to other fats and oils (primarily palm oil, but also butter, cocoa butter, lard, and margarines with high content of trans fat).Based on 23 comparisons from 17 crossover interventions (8 studies in hyper-and nine studies in normocholesterolemic subjects), higholeic oils reduced LDL-cholesterol, apoB, and total cholesterol when compared to oils/fats rich in saturated fat in the majority of comparisons, whereas no differential effect was observed for triglycerides, HDL-cholesterol or apoA1.When compared to trans-fat-containing oils/ fats, high-oleic oils decreased total-and LDL-cholesterol, apoB and triglycerides in all or almost all comparisons, whereas HDL-cholesterol and apoA1 were increased in the majority of comparisons.High-oleic oils had no differential effect on any lipoprotein marker compared to oils rich in PUFA.

Inflammation markers
The largest available meta-analysis from 2015 (42) (n = 15 RCTs), including altogether 541 participants in the olive oil groups and 731 in control groups, showed that olive oil compared to control oil or diet reduced levels of c-reactive protein (CRP) (mean difference 0.64 mg/L [−0.96, −0.31],I 2 = 66%).Also, olive oil decreased concentration of interleukin 6 (IL-6) (mean difference −0.29 pg/mL ([−0.7,−0.02], seven RCTs with a total of 416 subjects in olive oil and 441 in control groups, I 2 = 62%)) as well as improved endothelial function (mean difference 0.76% ([0.27-1.124];eight RCTs with 335 subjects in olive oil and 516 in control groups, I 2 = 26%)).The study duration varied from four to 208 weeks.However, a considerable heterogeneity among studies indicates that results of this meta-analysis should be interpreted with caution.
In a SR from 2019 (47) results from eight RCTs that compared the effect of palm oil on glucose metabolism in interventions that lasted from 3 to 7 weeks and the number of participants in intervention groups ranged from 15 to 100 participants were presented.Palm oil compared to other vegetable oils was not seen to affect fasting plasma glucose or insulin levels (47).

Body weight and other anthropometric measures
In a meta-analysis from 2019 (48) based on 22 RCT studies including 1,078 participants, canola oil reduced body weight 0.3 kg (−0.52, −0.08) compared to many other types of vegetable oils, fish oil or control diet in interventions of varying durations from three to 28 weeks.These studies were done in several countries around the world including Finland and Sweden and there was no heterogeneity between the studies (I 2 = 0%).However, canola oil was not observed to affect BMI or other anthropometric measures.

Summary of evidence
Associations with diseases and mortality

•
Consumption of olive oil was associated with lower risk of CVD, T2D, and total mortality in prospective cohort studies.• Consumption of butter was not associated with risk of CVD but it was related to slightly lower risk of T2D and higher risk of total mortality in prospective cohort studies.• Consumption of olive oil was associated with reduced risk of some cancer types whereas intake of butter was associated with higher risk of certain cancer types.However, the evidence was scarce and mainly based on case-control studies.
Effects on risk markers and body weight Non-tropical vegetable oils are rich in unsaturated fatty acids, and should be used as their dietary source.Particularly, the essential fatty acids, n−6 linoleic acid (LA) and n−3 alpha-linolenic acid (ALA), must be obtained from the diet since the human body is not capable of synthesizing them.Several vegetable oils are good sources of LA, whereas only a few oils are rich in ALA (Table 3).Good sources of ALA are canola/rapeseed, linseed, hempseed, soybean and walnut oils.For example, the recommended daily amount of both n-6 PUFA and n-3 PUFA can be obtained by 2-3 tablespoon of canola/ rapeseed oil.In a Nordic context, rapeseed oil may thus be considered as the primary source of added fat due to its nutritional profile as well as being locally produced.However, olive oil may entail additional health benefits beyond fatty acid composition (e.g.polyphenols) (36,37) and may also be used frequently.Unfortunately, very few studies are available that specifically investigates health effects of rapeseed oil.However, based on individual RCTs, rapeseed oil decreases liver fat content during isocaloric conditions compared to olive oil (53), and a diet rich

Examples of recommended quantities
Daily minimum intake* with an average expenditure of 2,000 kcal 5.6 g (2.5 E%) 1.1 g (0.5 E%) Intake with 2,000 kcal during pregnancy and lactation 8.9 g (4.0 E%) 2.2 g (1 E%) 1 For margarine and butter mixes, an average of LA and ALA is calculated.
Values are from Finnish, Swedish, and US national food databases.
*Because minimum requirements of cis-PUFA for adults are not known, the estimates are based on threshold intake data from children.The recommendation for essential fatty acids, that is, LA and ALA, is 3E% of which at least 0.5E% should be ALA.Recommended energy percentages were calculated as kcal of 2,000 kcal.Then, recommended intake of LA and ALA as grams were calculated based on that 1 g fat yields 9 kcal energy.For example, calculation for LA: 2.5% of 2,000 kcal = 50 kcal; thus, 50 kcal / 9 kcal = 5.6 g.
in canola oil was found to decrease abdominal fat even when compared to diets with higher content of PUFA (blends of corn oil, safflower oil and flaxseed oil) (54).Furthermore, meals rich in rapeseed oil increased 24-h fat oxidation compared to meals rich in palm oil (55), and diets rich in rapeseed oil markedly improve blood lipid profile compared to butter (56,57).Rapeseed oil has a higher content of phytosterols compared to many other oils (1), perhaps explaining the LDL-cholesterol reducing effect of rapeseed oil compared to both olive oil and sunflower oil demonstrated in a recent meta-analysis (38).Finally, a rapeseed oil-based margarine was used in the Lyon Diet Heart Study, a secondary prevention trial testing an ALA-enriched Mediterranean diet in patients having survived a first acute myocardial infarction (58,59).A striking protective effect was observed on the risk of recurrence (50-70% reduction) after 4 years of follow-up.Although these positive results cannot be ascribed specifically to rapeseed oil, it further supports the role of rapeseed oil as part of a cardioprotective diet.Vegetable oils are 100% fat and therefore include a high energy content.Although adding healthy plant oils to the diet will improve the overall fat composition of the diet, it will also increase the energy content and density of the diet.Thus, from the perspective of weight management, it is advisable to use healthy plant oils in moderate amounts.Processing of oils, such as extraction method, cold-pressing, heating and refining, has no effect on fatty acid composition of oils but they impact on how other bioactive compounds are preserved (2).Refining, however, removes taste, unwanted compounds, and reduces oxidation products that will induce further oxidation if not removed.Further, cooking temperature, light exposure during storage as well as storage temperature and time influence bioactive compounds in oils since they are sensitive to heating and light.Light and heat are also factors that increase oxidation.In general, oils are suitable for frying and cooking (despite their relatively high content of unsaturated fatty acids) ( 60), but they should not give smoke at high temperatures when cooked.The temperature during normal pan frying is typically 140-175°C, and negative effects of frying vegetable oils high in unsaturated fatty acids are marginal even at 200 °C for extended time periods (60).

Data gaps for future research
Further systematic investigations would be desirable of different consumption levels of vegetable oils in relation to disease outcomes, mortality, blood lipids as well as overweight and obesity among both adults and children.Current evidence indicates that the type/version (e.g.high or low in polyphenols) of oil may affect risk markers which suggests importance of degree of processing and advantageous health effects of bioactive components.
Further studies are needed to clarify the effect of these minor components on health.Margarines and butter mixes are commonly used products in the Nordic countries; however, there is scarce evidence about their overall health effects compared to other sources of dietary fat.Margarines may differ a lot regarding both fat content and fatty acid composition (i.e. which oils/fats that have been used), and more specific studies are needed.As the oils/fats that are used in margarine production are typically highly refined, the levels of bioactive compounds (e.g.polyphenols) may potentially be lower in margarines compared to less refined versions of the parent oils/fats.On the other hand, many beneficial components could be added in margarines, for example, antioxidants, vitamin D or plant sterols and stanols.Furthermore, oils/fats used in margarine production are often interesterified (i.e.fatty acids are rearranged within the triglycerides) to achieve desirable physical structure/functionality.Although no negative effects of interesterification have been indicated in short-term human studies, the evidence is scarce and more studies are warranted.For similar reasons (physical structure), margarines typically contain a variable fraction of tropical oils (e.g.palm, coconut) influencing both the fatty acid and sustainability profile.However, vegetable oil-based margarines usually contain less saturated and trans-fat (levels of trans-fat are very low to absent) than butter or butter mixes, and could therefore be considered a healthier choice as the fatty acid composition is likely the primary factor determining the overall health effect.Margarines with a higher fat content (compared to lower) are generally better sources of essential fatty acids (Table 3).

Limitations
The main limitation is that there were no recent qualified SRs available reporting at a food level perspective, and neither was a de novo NNR2023 SR performed for this topic.Overall, there were few SRs available, and many included SRs had low or critically low confidence, thus the summarized evidence should be interpreted with caution.Furthermore, only few fats/oils (olive oil, butter) have been investigated in relation to disease outcomes in SRs; for example, there was no SR concerning canola/ rapeseed oil and disease outcomes.Thus, evidence of olive oil was emphasized in relation to disease outcomes.Furthermore, there were hardly any SRs that included margarines, butter mixes or shortenings.All studies that filled inclusion criteria were done in adults, and therefore evidence of health effects of fat and oils among vulnerable groups, that is, children, adolescents, pregnant and lactating women could not be taken into consideration.There were no SRs available for several diseases, for example, autoimmune diseases, asthma, and allergies.Most of the evidence on disease outcomes are based on observational data, and although statistical adjustment for potential confounders are typically performed, residual confounding likely remains.Data on dietary intake are most often self-reported, and various degrees of misreporting, and thus misclassification, can be expected.Another general limitation that can be discussed is the statistical modelling of the dietary data, where specific substitution models are generally not available, which may hamper the overall interpretation and advice on practical implementation.Evidence was identified in scientific literature only since 2011 and the search was done in one database.Therefore, there might be some relevant older studies which did not meet inclusion criteria.Furthermore, as the search only included SRs and meta-analyses, there may be individual studies of high quality (not included in any SR) that has not been considered.

Fig. 1 .
Fig.1.Compositional differences between various fats and oils.Values are taken from the Swedish, Finnish and US food databases which showed the amount of different types of fatty acids as grams per 100 g fat or oil which equals to percentages.

Table 1 .
(Continued) n = 3 for SFA-rich oils, n = 12 for n-3-rich oils, n = 10 for n-6-rich oils, n = 24 for all other plant oils combined Olive oil increase LDLcholesterol compared to all other plant oils combined, no difference compared to SFA-rich-, n-3-rich-or n-6-rich oils Low confidence Schwingshackl (2019) Type of olive oil Refined olive oil, low polyphenol olive oil n = 9 for olive oil, n = 11 for sunflower oil, n = 10 for SFA, n = 35 for all other oils combined

Table 1 .
(Continued) (page number not for citation purpose)A scoping review for Nordic Nutrition Recommendations 2023

Table 2 .
List of excluded studies

Table 2 .
List of excluded studies

Table 3 .
Fats and oils as dietary sources of essential fatty acids1