Randomised trial of coconut oil, olive oil or butter on blood lipids and other cardiovascular risk factors in healthy men and women
Kay-Tee Khaw, Stephen J Sharp, Leila Finikarides, Islam Afzal, Marleen Lentjes, Robert Luben, Nita G Forouhi (2018) BMJ Open

https://bmjopen.bmj.com/content/8/3/e020167.full

Background
Coconut oil, an oil high in saturated fat (approximately 90%), has gained popularity as a dietary ingredient in the UK, promoted on television cookery programs, in recipe books and on social media with claims of health benefits frequently hitting the headlines (Lockyer & Stanner, 2016). Two recent reviews of intervention studies, including randomised controlled trials, on the effect of coconut oil on cardiovascular risk markers provide no suggestion from the weight of evidence that consumption of coconut oil rather than unsaturated plant oils would benefit cardiovascular health. Rather, these reviews suggested that, compared to unsaturated oils (such as safflower, soybean and olive oil), consumption of coconut oil raises total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C), but also increases high-density lipoprotein cholesterol (HDL-C) (Lockyer & Stanner, 2016; Eyres et al., 2016). These effects are consistent with expectations of the blood lipid effects of a saturated fat based on laboratory studies. However, both reviews highlighted the limited number of intervention studies to date and that those available often had small numbers of subjects and were conducted in populations not largely representative of the UK (e.g. Sri Lankans, Pacific Islanders, Malaysians).


Findings from a new randomised study have recently been published in the BMJ, adding to the limited research base behind a product that, on the basis of current evidence, might have been given a disproportionate health halo by popular media articles.


Study objective
A randomised study was carried out aiming to investigate the effect of daily consumption of 50 g of either extra virgin coconut oil, extra virgin olive oil or butter for one month on blood lipids and other metabolic measures in free-living UK subjects.


Study design
In this trial, middle-aged men and women living in the general community were randomly allocated to consume 50 g extra virgin coconut oil or 50 g butter or 50 g extra virgin olive oil for 4 weeks. Subjects who did not have a known medical history of heart disease, stroke, cancer or diabetes and without gall bladder or bowel problems, not taking cholesterol-lowering medication e.g. statins and living in Cambridgeshire, were recruited through BBC advertising between May and June 2017. Study fats were provided along with measuring cups and suggestions of how to include the fats in the diet including recipes. Information on the saturated fatty acid composition of the test oils used can be found in the table below.

 

   Extra virgin coconut oil  Extra virgin olive oil  Butter
 Saturated fat content 94% 19% 66%
 Main saturated fatty acids Lauric acid (C12:0): 48%

Myristic acid (C14.0): 19%

Palmitic acid (C16:0): 9%

Palmitic acid (C16:0): 15%

Stearic acid (C18:0): 3%

Palmitic acid: (C16:0) 28%

Stearic acid: (C18:0) 12%

Myristic acid: (C14:0) 11%

 

Outcomes
The primary outcome was change in LDL-C.
Secondary outcomes were change in TC, HDL-C, triglycerides, TC/HDL-C ratio, non-HDL-C, fasting blood glucose, C-reactive protein, weight, body mass index (BMI), body fat percentage, waist circumference, systolic blood pressure and diastolic blood pressure.


Results
Ninety six subjects were randomised to consume 50 g of either extra virgin olive oil (n=33 randomised, n=30 completed), extra virgin coconut oil (n=30 randomised, n=28 completed) or unsalted butter (n=33 randomised, n=33 completed) daily for 4 weeks. Two thirds of the subjects were women, mean age was 60 years and mean BMI for the 3 groups ranged from 24.8 (standard deviation 3.5) to 25.5 (standard deviation 4.5) kg/m2. Dietary intake (energy, fat, protein, carbohydrate and alcohol), as assessed by a 24 hour dietary assessment questionnaire, was similar across the intervention groups at baseline. Plasma lipids were raised at baseline in all groups (TC: 5.9 mmol/L, 5.9 mmol/L, 6.0 mmol/L; LDL-C: 3.5 mmol/L, 3.5 mmol/L and 3.7 mmol/L in the coconut oil, butter and olive oil groups respectively).


Changes in outcomes
In summary, in pairwise comparisons of changes in plasma lipids at 4 weeks:

  • Coconut oil did not significantly raise LDL-C concentrations compared with olive oil, while butter significantly raised LDL-C concentrations compared with both coconut oil and olive oil.
  • Coconut oil significantly raised HDL-C concentrations compared with both butter and olive oil.
  • Butter also significantly raised TC:HDL-C ratio and non-HDL-C compared to coconut oil and olive oil but there were no differences between coconut oil and olive oil for changes in TC:HDL-C and non-HDL-C.
  • The surprising finding from this study was that there was no significant difference between coconut oil and olive oil for LDL-C and TC measures, despite their differing fatty acid profiles.


There were no significant changes in body weight, BMI, central adiposity, fasting blood glucose, systolic or diastolic blood pressure in any of the groups over the 4 week period.

 

   Coconut oil vs. olive oil  Butter vs. coconut oil  Butter vs. olive oil
 LDL-C

-0.04 mmol/L

(-0.27 – 0.19) p=0.74

+0.42 mmol/L*

(0.19-0.65) p <0.0001

+0.38 mmol/L*

(0.16-0.60) p<0.0001
    Primary outcome - LDL-C
 TC

+0.19 mmol/L

(-0.08-0.46)

 +0.19 mmol/L

(-0.08 – 0.45)

+0.38 mmol/L*

(0.11 - 0.64)

 TC:HDL-C ratio

-0.14 mmol/L

(-0.33-0.05)

 +0.36 mmol/L*

(0.18-0.54)

 +0.22 mmol/L*

(0.04-0.40)

 Non-HDL-C

+0.002 mmol/L

(-0.23-0.24)

+0.39 mmol/L*

(0.16-0.62)

+0.39 mmol/L*

(0.16-0.62)
 HDL-C

+0.16 mmol/L*

(0.03-0.28)

 -0.18 mmol/L*

(-0.30—0.06)

 -0.02 mmol/L

(-0.14-0.09)

    Secondary blood cholesterol outcome measures: TC, HDL-C, triglycerides; TC:HDL-C ratio, non-HDL-C

 *, statistically significant


Self-reported compliance (measured by specific questionnaire) was high, with over 80% of subjects reporting more than 75% compliance in all groups, though average quantities of the test fats consumed were not reported in the paper. A 24 hour dietary assessment at 4 weeks indicated that fat intake increased in all groups but intakes of carbohydrate, protein and alcohol intake was similar. Most subjects reported no change in physical activity and supplementary data shows that excluding those that reported increasing physical activity had little effect on the results.

   

Coconut oil

Mean (SD)
 

Butter

Mean (SD)
 

Olive oil

Mean (SD)
 

Energy MJ

(change from baseline)
 

9.6 (3.2)

+ 0.3 (2.9)
 

8.6 (2.4)

+ 0.5 (2.0)
 

9.6 (3.1)

- 0.4 (2.8)
 

Fat (total) g/d

(change from baseline)
 

127 (47)

+ 29 (43)
 

94 (37)

+ 14 (36)
 

138 (38)

+ 28 (40)
 

Protein g/d

(change from baseline)
 

71 (26)

-7 (33)
 

77 (20)

+3 (30)
 

78 (31)

-12 (26)
 

CHO g/d

(change from baseline)
 

215 (84)

-31 (74)
 

214 (64)

+ 4 (69)
 

197 (101)

-55 (81)
 

Alcohol g/d

(change from baseline)
 

9 (15)

-8 (22)
 

13 (15)

-5 (23)
 

8 (18)

-11 (27)
 Weight Change from baseline (kg)  0.27 (0.77)  0.04 (1.00)  -0.04 (0.84)


Author conclusions
Two different dietary fats (butter and coconut oil) which are predominantly saturated fats, appear to have different effects on blood lipids compared with olive oil, a predominantly monounsaturated fat, with coconut oil more comparable to olive oil with respect to its effect on LDL-C.


Limitations

The authors considered a number of limitations:

  • It was a short-term trial of 4 weeks intervention, so we are unable to know what would have happened if the intervention had continued for a longer period. Studies of longer duration in this area would be useful.
  • In addition the study did not include a run-in period. According to EFSA, ‘blood lipids tend to stabilise after about 4 weeks in response to fixed nutritional interventions. However, the time needed to reach such stabilisation may depend on the study characteristics (e.g. appropriate run-in period) and the nature of the intervention. Evidence on the sustainability of the effect with continuous consumption of the food/constituent over longer periods of time (e.g. 8 weeks) should be provided.’ (EFSA, 2018).
  • There was no attempt to control other aspects of usual diet, in particular, total energy intake.
  • The study intervention gave the option of either consuming the test fat in addition to their normal diet or using it to replace fats and oils habitually consumed. The results therefore cannot be taken to reflect what would happen when the only change to a diet is the substitution of one fat with another (e.g. replacing butter with coconut oil or replacing butter with olive oil).
  • Individuals may have changed their behaviours in different ways to accommodate the additional fat. For example, participants may have modified other aspects of their diet for instance, increasing foods such as bread and potatoes, curries, salads or baked goods in order to incorporate the test fats into the diet or consciously reducing intake of other foods, though detailed dietary intake data was not reported in the paper.
  • The results presented in the paper detail an increase of total fat intake of 14 – 29 g (saturated fat intake was not reported), compared to the 50 g intervention, and changes in weight were small. This implies some compensatory behaviour - it may be that other foods containing fat were avoided or restricted in compensation, for example, cakes and biscuits or other foods that may be high in refined carbohydrates.
     

Other study limitations:

  • Subjects were not blinded to the intervention.
  • Compliance data relied on self-reported questionnaires.
  • Intervention dietary assessment was assessed on the basis of a single 24 hour recall.
  • Subjects were fasted for a minimum of four hours; rather than the suggested overnight (8-12 hour) duration, although the authors report that a majority of subjects were fasted overnight. 


Relevance

  • This interesting study highlights some of the complexities in research on saturated fat which may be important in their impact on blood lipids and health outcomes including different profiles of individual fatty acids, as well as the foods in which they are consumed or dietary patterns. Studies carried out in free-living subjects rather than more controlled conditions are liable to compensatory effects (i.e. subjects making changes to their diet or lifestyle).
  • The saturated fatty acid profiles of different dietary fats vary substantially. For example coconut oil is around 48% lauric acid (12:0) compared with butter which is about 40% palmitic (16:0) and stearic (18:0) acids, leading to suggestions that coconut oil may not have the same health effects as other foods high in saturated fat. However, a 2016 systematic review and regression analysis carried out by Mensink on behalf of the WHO which included data from 84 studies demonstrated that there was a significant increase in total cholesterol and LDL cholesterol when carbohydrates were replaced with lauric acid, myristic acid or palmitic acid.
  • Currently published reviews (Lockyer & Stanner, 2016; Eyres et al., 2016) have concluded that whilst trials investigating an association between coconut oil consumption and blood lipids or cardiovascular risk are mostly poor quality, trials have typically reported that coconut oil consumption raises LDL-C in comparison to unsaturated oil.

Whilst the findings of this particular study are interesting, the authors themselves agree that these findings do not alter current dietary recommendations to reduce saturated fat intake in general. The weight of evidence from other studies would still suggest that we should be reducing saturated fat and replacing with unsaturated fat. Indeed, national and international dietary guidance which is based on the totality of scientific evidence recommends reducing saturated fat in the diet and replacing this with small amounts of unsaturated fat in order to decrease CVD risk. In the UK, dietary advice is to reduce intake of all saturated fatty acids to less than 10% of total dietary energy (including alcohol), and this has recently been supported by a draft Scientific Advisory Committee on Nutrition review (SACN, 2018). However, it is increasingly recognised that dietary advice should focus on foods and overall dietary patterns, rather than single nutrients. Healthy dietary patterns typically include plenty of fruit and vegetables, wholegrains, some lower-fat dairy foods like milk and yogurt, seafood, nuts, seeds, pulses and lower intakes of fatty/processed meat, refined grains, sugar-sweetened foods and beverages, as well as choosing lower salt and lower saturated fat options.

Coconut oil can be included in the diet, but as it is high in saturated fats, should only be included less often and in small amounts and as part of a healthy, balanced diet.

 

References

EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies) (2018) Guidance for the scientific requirements for health claims related to antioxidants, oxidative damage and cardiovascular health (Revision 1). EFSA Journal 16(1):5136, 21 pp.

Eyres, L., Eyres, M. F., Chisholm, A., & Brown, R. C. (2016). Coconut oil consumption and cardiovascular risk factors in humans. Nutrition Reviews, 74(4), 267-280.

Lockyer, S., & Stanner, S. (2016). Coconut oil–a nutty idea?. Nutrition Bulletin, 41(1), 42-54.

Mensink, R. P., & World Health Organization. (2016). Effects of saturated fatty acids on serum lipids and lipoproteins: a systematic review and regression analysis. http://apps.who.int/iris/bitstream/handle/10665/246104/9789241565349-eng.pdf?sequence=1 

Scientific Advisory Committee on Nutrition (2018) Draft report: Saturated fats and health. https://www.gov.uk/government/consultations/saturated-fats-and-health-draft-sacn-report

Yuan, C., Spiegelman D., Rimm, E.B. et al. (2018) Relative Validity of Nutrient Intakes Assessed by Questionnaire, 24-Hour Recalls, and Diet Records as Compared With Urinary Recovery and Plasma Concentration Biomarkers: Findings for Women, American Journal of Epidemiology, 187(5), 1051–1063.