Unless you’ve had your head buried in the health-and-fitness sand for the past 5 or so years, you’ve probably heard of ‘Bulletproof Coffee’: the drink where you heap up to 2-tablespoons each of butter and coconut oil, a grand total of 60g of saturated fat, into your coffee.

The ‘Bulletproof Coffee’ craze is a symptom of a confused nutrition public, where the populist message brewing since 2010 has been that

“we were wrong about saturated fats, it’s all sugar! Saturated fats are GOOD for you!”

But are they? Where is the genesis for this movement? And is it in fact correct for us to be stating that saturated fats are not an issue for cardiovascular disease?

This will be a 2-part article: Part 1 will look at saturated fats specifically, while Part 2 will focus on the current scapegoat, dietary sugar, and the role of free sugars in cardiometabolic disease.


The Genesis for the Movement

From a research standpoint, the tide turned with a systematic review and meta-analysis of 16 prospective cohort studies by Siri-Tarino et al. in 2010 [1], which concluded that saturated fats were not associated with an increase in cardiovascular disease [CVD] risk.

This paper was followed by a meta-analysis of the role of individual fatty acids in CVD risk from Chowdhury et al. in 2014 [2], which suggested that saturated fatty acids were “weakly” associated with CVD risk.

These studies have been championed by the low-carb high-fat paradigm, and are repeatedly cited as evidence that “we were wrong about saturated fat”. So why are these meta-analyses concluding a null association between saturated fat and CVD? There are several methodological flaws which need to be highlighted.


Critiquing the Meta-Analyses

The central issue in any meta-analysis is the quality of the individual primary studies included. Both meta-analyses have major limitations in this respect. In the Siri-Tarino et al. (2010) study, 16 prospective cohort studies looking at associations between saturated fat and CVD were analyzed. However, their analysis failed to compare “hard” outcomes, i.e. fatalities from CVD, to “soft” or total incidence CVD: a reanalysis of all 16 studies included in the meta-analysis revealed a significant 32% increase in risk of mortality from CVD, and of total incidence CVD, from saturated fat intake [3].

Another major issue with the Siri-Tarino et al. (2010) meta-analysis is that the authors included 7 studies which controlled for serum LDL and total cholesterol. These studies accounted for up to 50% of CVD events included in the meta-analysis, thus their overall conclusions, and their selective subgroup analysis of these studies, biased the results toward the null hypothesis, i.e. that there is no association between saturated fat and CVD [4]. Population studies can often minimize the relationship between saturated fats and CVD when looking at serum cholesterol parameters as endpoints, due to wide intra-individual variance in cholesterol levels that are independent of diet [5]. Mathematical modeling of errors in estimating both dietary exposure and the wide variance in endpoint measures (i.e. cholesterol) has shown that the greater the intra-individual variance, the more a true association between saturated fat and CVD would be attenuated [6].

The Chowdhury et al. (2014) meta-analysis looked at associations with both dietary and circulating levels of individual fatty acids. This is a major limitation, because even with the inclusion of the gold standard of evidence – the randomised controlled trial – the results of this meta-analysis reflect a major limitation of nutrition science: that the RCT was designed to test pharmaceuticals, i.e. single compounds with a targeted pharmacokinetic and pharmacodynamic profile. This design does not always adequately translate the effect of a single nutrient on a single endpoint into an accurate reflection of the effect of a complex food matrix of other nutrients, and an overall diet pattern, on multifactorial disease processes.

The Chowdhury et al. (2014) meta-analysis, as an analysis of individual fat subtypes and fatty acids, is a manifestation of this limitation: that the “inputs” were the foods being consumed at a population level, assessed by tools like food frequency questionnaires. Translating that into the effect of individual fatty acids on CVD risk can be misleading. For example, on the face of their analysis monounsaturated fats significantly increase CVD risk. Who knew that olive oil, nuts, and avocados are going to give you a heart attack? Of course, this is an absurd statement: we know from controlled, food-based intervention studies that these foods are associated with significant reductions in risk for CVD [7][8][9].

In another example, Chowdhury et al. (2014) included in their meta-analysis studies that specifically looked at consumption of dairy products and CVD risk, for which we have evidence supporting an inverse association between dairy products, dairy fats, and CVD [10][11]. These included studies would bias the results of the meta-analysis toward a non-significant association, which is ultimately what the authors concluded. The issue is that no food is exclusively one fat subtype, and the fatty acid compositions of the foods is relevant. With dairy, the short-chain saturated fatty acids are not associated with as deleterious an effect on serum cholesterol or CVD outcomes as other saturated fatty acids and foods themselves differ in effect: butter raises LDL-C more than cheese, despite having the same saturated fat content [12].

In addition, the monounsaturated fats in this meta-analysis (Chowdhury et al., 2014) came primarily from animal and processed foods, which may contain from 20-50% monounsaturated fatty acids. So by distilling foods into fatty acids, nutritional epidemiology can often come to conclusions inconsistent with controlled data (an exception is in relation to fish consumption translating to the cardioprotective and neuroprotective effect of marine omega-3 fatty acid intake).

These meta-analyses are a reflection of several issues in nutrition science that need to be considered:

  1. The use of “endpoints” to determine risk of disease from nutrients. Is the relationship between saturated fats and CVD primarily due to adverse effects on blood cholesterol levels?
  2. Focusing on one, single nutrient in the context of the complexity of diet. What is the effect of replacing one macronutrient with another for CVD risk?
  3. What dietary characteristics influence other markers of CVD risk – HDL cholesterol, triglyceride – aside from obesity?
  4. The need to look at diet patterns as a whole, and move from single-nutrient hypotheses.

Let’s clear up the saturated fat controversy by looking at these issues.


Saturated Fat, Blood Cholesterol and Cardiovascular Disease Endpoints

Advocates of this paradigm, particularly in the low-carb and/or “Paleo” crowds, often point to the ‘Seven Countries Study’ by a University of Minnesota researcher called Ancel Keys, which observed significant associations between high saturated fat intake and CVD across different populations. You’ll hear the oft used line, correlation is not causation. No, it’s not: but it is hypothesis-generating.

This highlights an issue with nutrition science that is never going to go away: the pharmaceutical model of evidence being applied to nutrition, the veneration of the randomised controlled trial as the only evidence worth its salt. Unfortunately, however, you can’t do an RCT feeding people at risk for CVD a high saturated fat diet using overt heart disease as an endpoint, because it’s unethical. But you can look at controlled feeding studies using risk factors as endpoints, like blood cholesterol levels.

Nutrition science needs more ‘big picture’ thinking than other health sciences, and that is how we look at saturated fat in CVD: epidemiology does matter, but it’s taken into account alongside mechanistic studies, and controlled studies assessing relevant endpoints.

So, is blood cholesterol a valid risk factor for CVD? Bottom line: Yes. There is a distinct “atherogenic lipoprotein phenotype” that is consistent and confirmed across multiple lines of evidence: high blood LDL-cholesterol, low HDL-levels, high circulating triglycerides [TGs], and a remodelling of LDL into small, dense lipoprotein subparticles [13].

In this respect, there can be little evidential dispute that saturated fats raise total blood and LDL cholesterol levels, with both parameters – but particularly LDL-C – strongly implicated in the pathogenesis of atherosclerosis and CVD [14][15][16][17]. Thus, if we take increased either total cholesterol and/or LDL-cholesterol as a primary endpoint, saturated fats are strongly implicated in CVD risk [14][15][16][17].

It is important to quantify that these risk factors do correlate to “hard” endpoints, i.e. cardiovascular disease fatalities: comparing “hard” vs. “soft” endpoints (incidence of CVD), the evidence disconcertingly suggests a significantly stronger association with saturated fat intake weighted by years of exposure and CVD fatalities, compared to total CVD [3]. It is important to note, however, that these issues relate to the effects of saturated fat on blood cholesterol levels, not dietary cholesterol on blood cholesterol levels, the latter of which we now have robust evidence that it is not an issue. Yes, eggs are absolutely to fine consume in your habitual diet (they are predominantly unsaturated in fatty acid composition), and dietary cholesterol is no longer considered a risk for CVD [18].

Thus on the basis of hard endpoints and on the basis of risk factors, like blood cholesterol, there is strong evidence to support a causative role of high saturated fat intake in cardiovascular disease. However, this approach – of looking at the relationship between a complex variable like diet and a multi-factorial disease process through single nutrient and outcome variables – has rightly been questioned for being too simplistic, having regard to the influence of other factors, in particular the balance of other macronutrients, and influence of diet on other markers of disease like triglycerides, HDL-C and lipoprotein subtypes [17][12][13].

JOIN NOW to continue reading...
All the science, none of the B.S. Sign up today. Monthly Gold membership is $12.99
Workout Builder

Choose from several training programs for different goals and difficulty level.

Video Q&A

Get YOUR questions answered every week by Layne himself.

Exclusive Content

Discover a plethora exclusive articles and videos on nutrition and training from some of the top experts in the world.

Webinar Replays

Layne hosted a series of webinars and live training sessions.

Amanda Bucci

"Not only does Layne talk the talk, he's walked the walk. I recommend listening to and reading his resources to any of my friends looking for science-backed fitness information to guide them on their path of knowledge expansion within the realm of fitness. He's one of the few people i've found to be a reliable, educational, no b.s. resource."

- Amanda Bucci (Bodybuilding.com Athlete)
Sign Up Now Members Login