A large network meta-analysis of 54 RCTs compared the effects of 13 different oils (safflower, sunflower, canola, hempseed, flaxseed, corn, olive, soybean, palm, coconut) and solid fats (beef fat, lard, butter) on blood lipids in generally healthy participants, and in participants at CVD risk [21••]. For low-density lipoprotein cholesterol (LDL-C) and triacylglycerols, no clear differences between coconut oil and unsaturated rich oils were observed (Table 3). However, a 10% isocaloric exchange of butter with coconut oil reduced LDL-C by −0.23 mmol/L (95% CI: −0.40 to −0.07), total cholesterol by −0.18 mmol/L (95% CI: −0.34 to −0.02), and improved HDL-C by 0.04 mmol/L (95% CI: 0.01 to 0.08). On the contrary, several unsaturated rich oils such as safflower, sunflower, canola, and corn oil reduced total cholesterol compared to coconut oil [21••]. However, in this network meta-analysis, the authors did not distinguish between the effects of virgin vs. refined coconut oil.

Another comprehensive meta-analysis included 16 RCTs and found that higher coconut oil intake increased plasma LDL-C (0.27 mmol/L, 95% CI: 0.08 to 0.46) but also HDL-C (0.10 mmol/L, 95% CI: 0.06 to 0.15), whereas no effect was detected for triacylglycerols, body weight, body fat, and markers of glycaemia and inflammation in comparison with non-tropical vegetable oils [22••]. The heterogeneity between the RCTs included in this meta-analysis is high and most of the RCTs did not report on the types of coconut oil used. The information was only available for six RCTs; two RCTs used organic extra virgin coconut oil, two used refined, bleached, and deodorized oil, one used fractionated coconut oil, and one RCT used filtered coconut oil obtained by pressing dehydrated coconut pulp. Due to the limited information, the authors were unable to conduct stratified analysis by the types of coconut oil used.

The largest RCT so far in general healthy participants on coconut oil was conducted in the UK in 2017 [23•]. A total of 96 participants were randomized to one of the three interventions (50g/day virgin coconut oil, extra virgin olive oil or unsalted butter) for 4 weeks. LDL-C concentrations were increased on butter compared to coconut oil (0.42 mmol/L, 95% CI: 0.19 to 0.65) with no differences in the change of LDL-C in coconut oil compared to olive oil. Coconut oil also increased HDL-C compared to butter (0.18 mmol/L, 95% CI: 0.06 to 0.30) or olive oil (0.16 mmol/L, 95% CI: 0.03 to 0.28). There were no differences in changes in weight, body mass index, central adiposity, fasting blood glucose, and systolic or diastolic blood pressure among any of the three intervention groups.

The most recent RCT was not included yet in systematic reviews because it was published recently in 2021. It included 48 participants with metabolic syndrome aged 20–50 years [24•]. Compared to a control group, consumption of 30g/day virgin coconut oil did not result in differences in anthropometric outcomes and blood pressure. However, virgin coconut oil improved HDL-C (0.19 mmol/L, 95% CI: 0.12 to 0.26) and triacylglycerols (−0.61 mmol/L, 95% CI: −0.91 to −0.30). On the contrary, detrimental effects were observed for LDL-C (increase by 0.57 mmol/L [95% CI: 0.35 to 0.79]), total cholesterol (increase by 0.95 mmol/L 95% CI: 0.65 to 1.26), and asymmetric dimethyl arginine (increase by 2.12 μg/L 95% CI: 0.78 to 3.45). In line with these findings, Vogel and colleagues [25] showed some improvements of extra virgin coconut oil on fasting glucose and HDL-C compared to soybean oil, but the detrimental effects on LDL-C were not confirmed. In another recent RCT, Maki and colleagues [26] showed that the consumption of 4 tablespoons per day (54 g/day) of corn oil reduced non-HDL-cholesterol compared to coconut oil. Similar to the findings described above, when virgin coconut oil (30 ml/day) was compared to an equal amount of safflower oil, an increase in total cholesterol, LDL-C, and HDL-C was observed, while no differences were detected for anthropometric outcomes [27].

Neumann and colleagues [28] investigated the impact of different meal fatty acid compositions on postprandial lipemia in metabolically healthy adults and adults at risk of CVD. Although coconut oil provoked a weaker postprandial lipemic response than cocoa butter, butter, and lard in two RCTs, another RCT did not find any differences when comparing the AUC 0–3 h of postprandial triacylglycerol concentrations after meals enriched with coconut oil, tallow, and milk fat.

The largest RCT on secondary prevention so far was conducted in India between 2009 and 2014, and included 200 patients with coronary artery disease who were randomly assigned to a coconut or sunflower oil group (15% of daily energy intake) [29]. Blood lipid profile at 3 months after 1 or 2 years did not show differences in both of the groups. Of the 200 patients, two in each intervention group underwent revascularization, which can be considered a cardiovascular outcome.

Since the evidence for the effects of coconut oil on cardiovascular risk factors is limited, the impact of individual fat classes is of importance in understanding the role of coconut oil in the etiology of CVD.

A well-conducted meta-regression analysis by Mensink [30••] showed that when carbohydrates were isocalorically replaced (by 1% of energy intake) with lauric acid (the predominant fatty acid in coconut oil), total cholesterol slightly increased by 0.029 mmol/L (95% CI: 0.014 to 0.045), LDL-C by 0.017 mmol/L (95% CI: 0.003 to 0.031), and HDL-C by 0.019 mmol/L (95% CI: 0.016 to 0.023), while triacylglycerols decreased by −0.015 mmol/L (95% CI: −0.023 to −0.007). When carbohydrates were isocalorically replaced (by 1% of energy intake) with myristic acid, total cholesterol increased by 0.060 mmol/L (95% CI: 0.042 to 0.077), LDL-C by 0.044 mmol/L (95% CI: 0.028 to 0.060), and HDL-C by 0.021 mmol/L (95% CI: 0.017 to 0.025), and triacylglycerols reduced by −0.011 mmol/L (95% CI: −0.020 to −0.002). In addition, the isocaloric replacement of carbohydrates (1% of energy intake) with palmitic acid increased total cholesterol by 0.041 mmol/L (95% CI: 0.030 to 0.052), LDL-C by 0.036 mmol/L (95% CI: 0.026 to 0.046), and HDL-C by 0.010 mmol/L (95% CI: 0.007 to 0.013), and reduced triacylglycerols by −0.011 mmol/L (95% CI: −0.017 to −0.006). Moreover, isoenergetic substitution of palmitic acid with oleic acid lowers total cholesterol, LDL-C, and apoB concentrations [31]. In a systematic review on RCTs that focused on interventions to reduce intake of SFA, no effect was observed for systolic and diastolic blood pressure [32].

Regarding glycemic parameters, an isocaloric substitution (5% of total energy) of SFA with polyunsaturated and monounsaturated fatty acids improved biomarkers of glycemic control such as fasting glucose, HbA1c, C-peptide, and HOMA-index based on a systematic review of feeding RCTs [33]. However, there was insufficient information available to classify the subtypes of fatty acids, so these findings must be considered primarily relevant to the effects of total dietary SFA (predominantly palmitic acid).

Although no data are available for the association between coconut oil and the incidence of CVD, evidence from RCTs and prospective observational studies is available regarding the dominant fatty acid type in coconut oil, SFA. In systematic reviews of prospective observational studies, a higher total SFA intake was not associated with risk of CVD [34,35,36,37]. However, this finding needs to be interpreted with caution, since the meta-analyses were mainly based on comparisons of high versus low intakes, and are therefore less informative than findings from substitution analyses. In the pooling project including eleven American and European prospective cohort studies, for a 5% lower energy intake from SFAs and a concomitant higher energy intake from polyunsaturated fatty acids, there was an inverse association with risk of coronary events [38]. Findings from RCTs focusing on coronary events confirmed these beneficial effects of replacing SFA with polyunsaturated fatty acids [39].