This study, named FatFunc (#FatFunc) was a neat RCT that sought to examine differences between two weight loss diets: high fat low carb (HFLC) - 10% carbs, 17% protein, 73% fat - and a low-ish fat diet (LF) - 53% carb, 17% protein, 30% fat.
Is it the fat from the cream that's bad? Is it the sugar? Is it that there are 400 calories per half cup?
The study was in 44 men - obese, without diabetes, no severe diseases, or those who had more than two drinks a week. The primary outcome registered on clinicaltrials.gov was to look at changes in adipose tissue.
Men were instructed on how to follow their diets and anthropomorphic measurements and biomarkers were taken at baseline, week 8, and week 12. Diet adherence was measured through calculation of respiratory quotient (which can represent percent of calories from fat) and food diaries.
Interestingly, the authors thought that to really test the effects of macronutrients (specifically saturated fat vs. carbohydrate), participants should get the majority of their calories from the same food items, though in differing amounts. That way - it seems, differences seen can’t be attributed to effects from specific foods common in one diet not seen in the other. It’s worth noting that both groups were instructed to take a multivitamin, so differences in outcomes due to micronutrient intake were minimized.
There are a few key details to the protocol which make this study unique:
1. All participants were instructed to eat the same number of calories per day.
Unlike some trials that pit low-carb vs low-fat where either or both arms are ad-libitum, this study restricted calories equally in both arms. The authors did this “Because our purpose was to compare responses to well-defined diets sharply contrasting in carbohydrate and fat within a normal caloric range, and not explicitly examining the effects of energy restriction.”
2. All men were told to avoid trans-fats but ALSO to restrict plant oils with high amounts of omega-6 fatty acids.
This caused a huge increase in saturated fat intake in the LCHF group. Average intake of saturated fat in the LCHF group shot up to 80.6 g (+/- 9.4) per day. MUFA intake went up to 54 g (+/- 4.3). Compared to 30.5 g (+/-4.4) SFA and 20.0 g (+/- 3.1) MUFA in the LF group that’s a pretty significant difference. If saturated fat were beneficial, this would be one way to test it.
3. Men on the LF diet were instructed to drink fruit juice.
In the authors’ words: “To meet the energy requirement from carbohydrate on the LFHC [low-fat] diet, while avoiding large portion sizes of vegetables and rice, we chose to include juice as a regular carbohydrate source in this diet.”
It isn’t totally clear to us why more vegetables and rice weren’t recommended.
Study authors created a special cookbook for each diet. These were designed to minimize differences between the foods eaten in both groups.
This: Both groups lost weight and experienced significantly improved biomarkers. All the men saw an average decrease in BMI of -3.6 kg/m2. They both saw an average decrease in 6.3% of body fat. There were no significant differences between groups of body composition measurements. This is pretty interesting in light of claims made about the unique metabolic benefits of consuming a low carbohydrate diet for weight loss.
What happened to changes in plasma lipids and sugars?
Cholesterol and triglyceride levels were converted from mmol/L to mg/dL using this and glucose was converted thusly.
The interesting component of this trial is really the fact that there weren’t any major differences between the two diet groups in terms of anthropometric markers. However, the media picked up on the increases in HDL cholesterol after consuming the very high saturated fat diet and somehow interpreted this data as novel and as meaning that we are all wrong about saturated fats.
Firstly, It’s worth noting that both groups of healthy, obese men had OK levels of HDL at baseline (>= 40 mg/dL). Not ideal (>60 mg/dL) but not at major risk (<40 mg/dL). Secondly, while observational evidence indicates an inverse association between HDL levels and risk, it’s not so clear that this relationship is causal and that raising HDL will positively impact outcomes.
Evidence from human genetics, where individuals have genetically high HDL, and drug interventions which raise HDL levels, have not indicated a benefit to high HDL levels. Conversely, even slight increases in LDL due to genetic variants that increase circulating LDL, lead to increased CVD risk over a lifetime. Researchers are most interested now on markers of HDL activity, and the current clinical evidence doesn’t look so hot for saturated fat. It’s unclear why headlines are just now suddenly making a big deal about the effect of saturated fats on HDL.
Considering the stronger evidence for LDL as a precursor for heart disease, the increase in total LDL from ~140 to ~150 is concerning. Especially since the LF group saw the same weight benefits and a reduction from borderline high cholesterol ~140 to “near desirable” ~112. That’s a sizable reduction in LDL cholesterol that the low carbohydrate group, which normally would have reduced LDL in the context of weight loss, likely missed out on. (note: it’s difficult to compare these two groups directly because we don’t have a control arm in the study who were randomized but didn’t receive an intervention)
Furthermore, by holding PUFAs still in both study groups, the RCT really examined the difference between MUFAs and SFAs vs. carbohydrates. Current dietary recommendations emphasize replacing SFAs with PUFAs, so headlines stating that this study affected recommendations regarding SFAs are misguided.
The authors of FATFUNC say that they kept PUFAs low because “Results of previous studies that replaced PUFAs with SFAs or vice versa might be partly ascribed to changes in PUFA intake. Controlled studies have shown that PUFAs modulate metabolic function, inflammatory response, and risk of CVD.”
Namely - it might be PUFA causing “beneficial” effects that makes saturated fat look bad.
However, in this case, a diet high in carbohydrates focusing on unprocessed foods and low in added sugar (even including some juice in the LF group) produces more beneficial changes in LDL, a more reliable biomarker for future risk.
Ultimately the FATFUNC study showed only changes in biomarkers, not clinical endpoints. But the changes in biomarkers that it did show were consistent with evidence demonstrated in previous RCTs as well as observational studies.
High sat fat intake increased HDL and LDL.
Lower fat diet lowered LDL.
Energy reduced diets reduced weight and improved markers of glycemic control across both groups.
In a “whole foods” weight loss diet low in added sugars, refined carbohydrates, and highly processed foods, maintaining a diet lower in saturated fat leads to better cholesterol outcomes.
The only way this study can be construed to say that saturated fat is good for you, is to suggest that raising HDL has a greater beneficial effect than lowering LDL. And at this point there is not a lot of evidence to suggest that is the case.
The focus of this trial should be on what it was designed to look at, the primary outcome of changes in adiposity. It adds onto a growing list of trials which look more critically at claims about the supposed metabolic superiority of a low carbohydrate diet, especially in that it gave more defined dietary recommendations compared to most free-living low carb vs low fat trials. The fact that the media took off with a secondary outcome of the trial is possibly a side-effect of the #fakenews environment we live in, and likely fueled by an overly enthusiastic press-release.
Did you read the first installment of this two-part series? Check it out now!