Really-low-fat vs somewhat-lower-carb - a nuanced analysis


    With so many low-carb trial results rolling in each year, you might think that it’s case closed: everything there is to know is known. But there are still a few key pieces missing, and one of those pieces has just been released in the form of a six-day feeding study. Why only six days? That can’t tell us anything, right? Read on to see how revealing this study actually was, as well as what it can’t show (and likely wasn’t designed to show).

    Low-carbohydrate diets have become even more popular in the past few years, bolstered by the so-called “carbohydrate-insulin hypothesis of obesity”. This hypothesis suggests that carbohydrates are the main culprit of weight gain. Things get complicated here, because there are both practical factors (e.g. going low-carb means limiting your food options, which typically makes snacking on junk food more difficult) and physiological factors at play. For the latter category, advocates claim that you can harness the power of decreased insulin levels (from carb restriction) and lose more fat due to factors such as elevated free fatty acid release from fat cells and increased fat oxidation.

    Some low-carb advocates believe that carbohydrates are uniquely fattening, due to the effects of insulin. Despite a plethora of studies on low carb effects, there are still important areas left to research.

    Lo and behold, both meta-analyses and long-term trials often show low carb diets to be as good or better than other diets for weight loss. However, participants usually self-report their food intake, and the longer the trial, the more likely life will get in the way. It’s a catch-22 of sorts: you want long trials to make sure the diet is sustainable, but the longer the trial the more likely there will be unwanted variations in diet.

    Despite all the low-carb trials of recent years, there was still a lack of highly-controlled studies that solely altered carbs and fat in participants living in (aka stuck in) a metabolic ward. To address this issue, Kevin Hall’s team at the National Institutes of Health designed a short-term study to isolate the different effects of a restricted-fat diet versus restricted-carb diet on body weight, energy expenditure, and fat balance.

    Low-carb performs fairly well for weight loss in trials over the course of months. A recent short-term study controlled several extra variables to isolate the effects of carb reduction on fat loss.

    The Study


    The participants were 19 obese volunteers (ten males and nine females) with no apparent disease.

    All participants were required to reside within a metabolic unit, where they …

    • First received a baseline weight maintenance diet for five days before being randomized to either ….
    • The Restricted Fat group or Restricted Carbohydrate group, where caloric intake was reduced by 30% from baseline for six days, by either fat or carb reduction.

    The carb levels ended up being 352 grams for Restricted Fat versus 140 for Restricted Carb, and the fat levels 17 versus 108. In other words, (moderately lower carb than typical diets) versus (oh my goodness I can count my fat gram intake on my fingers and toes!).

    This trial wasn’t designed to explore a real-life 100-gram-and-under low carb diet and especially not a ketogenic diet. Rather, it was a mechanistic study designed so that they could reduce energy substantially and equally from fat or carbs, but without changing more than one macronutrient. If they lowered carbs much more in the Restricted Carb group (like under 100 grams), they’d then have to go into negative fat intake for the Restricted Fat group. And negative fat intake is impossible (*except for in quantum parallel universes). One more note: all participants kept dietary protein constant and exercised on a treadmill for an hour a day.

    19 study participants spent six days in a metabolic unit, eating either a moderately low-carb diet (140 grams) or a really darn low-fat diet (17 grams). The trial used these specific levels in order to achieve isocaloric diets with large carb/fat reductions, without having to alter more than one macronutrient.

    After the completion of this first phase, subjects went home for a two to four week washout period where they resumed their normal eating habits. Participants then returned to the study center to undertake the same protocol except with switched intervention groups. So those formerly in the restricted carbohydrate group would now be in the restricted fat group, and vice-versa.

    This crossover design was one of the many ways in which this trial was more stringent than most previous studies (other reasons are shown in the above graphic), since crossing over eliminates much of the variability that normal randomized trials have. Randomized trials may be considered the gold standard, but this trial was a mix between gold, platinum, and titanium. Very strong, very valuable. They even had participants wear accelerometers on their hips to measure physical activity.

    The crossover design was one of many rigorous aspects of the study, along with a relatively large sample size for a metabolic ward study, highly controlled variables, and multiple accurate measurement techniques.

    Why not use just DXA for measuring body composition?

    The primary method used to measure body composition was to calculate the difference between dietary fat intake and fat oxidation, as measured in a metabolic chamber by indirect calorimetry, which estimates the heat released by a person based on the amount of O2 they consumed and CO2 they produced over a specific period of time. While indirect calorimetry isn’t perfect, it’s accurate and reliable enough to be the standard method used to measure energy expenditure in these types of studies.


    Body weight and a dual-energy X-ray absorptiometry (DXA) scanner were also used, but the former can’t calculate body fat, and the latter isn’t sensitive enough to detect the minute difference in body fat loss that usually occurs between eucaloric diet interventions of different macronutrients.

    DXA is quite accurate for changes over the long term, but indirect calorimetry is needed for a study of this nature.

    Results, limitations, and other considerations

    As expected, the researchers found that the Restricted Carb diet resulted in a decrease in daily insulin secretion (by 22%) and a sustained increase in fat oxidation, whereas the Restricted Fat diet resulted in no significant change of either. Despite this, by the end of the six-day period, the Restricted Fat diet resulted in greater fat loss than did the Restricted Carb group (463g vs. 245g).

    The Restricted Carb dieters also had lower energy expenditure, to the tune of 98 fewer kcal/d compared to only 50 fewer in the Restricted Fat group. This isn’t enough calories to account for the difference in fat loss though. So why exactly did the Restricted Fat group lose that much more fat than the Restricted Carb group? The paper doesn’t get into this much, but gives some hints:

    "Model simulations suggest that the differences in fat loss were due to transient differences in carbohydrate balance along with persistent differences in energy and fat balance. The model also implicated small persistent changes in protein balance resulting from the fact that dietary carbohydrates preserve nitrogen balance to a greater degree than fat”

    … so their mathematical model points to a few possible minor factors, including a possible small benefit from dietary carbs benefiting protein balance. The use of this complex model to extend the results out further is interesting, as it partly compensates for having a six-day-only study (which is normal in the world of metabolic ward studies), but on the other hand the model isn’t something that is easily understandable by people other than the study authors. Maybe it’s really accurate, maybe it’s not.

    “Very low carbohydrate diets were predicted to result in fat losses comparable to low fat diets. Indeed, the model simulations suggest that isocaloric reduced-energy diets over a wide range of carbohydrate and fat content would lead to only small differences in body fat and energy expenditure over extended durations.”

    … ah, so if the researchers were able to reduce carbs to a much lower level (which they couldn’t, due to the study design factors described earlier), the diets would have actually led to similar weight loss. That makes those “New Study Shows Low-Carb Failure!” headlines sound a bit silly. If you take a really-darn-low-fat diet like the Restricted Fat diet, and compare it to a very-low-carb diet, you’re comparing two extreme diets and are more likely to get some metabolic advantage. Our bodies are typically accustomed to a somewhat balanced mix of fuel, and extreme macronutrient diets can probably game the system a bit for a modicum of extra fat loss.

    More importantly, the authors put the results really important context: the differences in body fat loss between a wide range of different carb intakes are predicted to be very small (although the Restricted Fat diet was predicted to sustain its slight advantage over the course of months). This study wasn’t meant to demonstrate that low(ish) carbs are bad or low-fat is good, it was simply testing the hypothesis that carb reduction provides some secret sauce for fat loss in highly controlled conditions.

    Surprisingly, the paper doesn’t mention the term “glycogen” even once. The study participants did an hour of incline treadmill a day, and since they had an average BMI around 36, that could mean a decent amount of glycogen burn with prolonged activity at a high body weight. So if liver and muscle glycogen happened to be relatively more depleted in the Restricted Carb group (since they replenished glycogen less by fewer dietary carbs), that might mean less fat loss in the short run, which may not apply as much to the long run when glycogen is in a steady-state.

    Although the Restricted Fat group lost a bit more fat, fat loss over time was predicted to be similar over a range of carb intakes based on a mathematical model of metabolism. The main application of the study may be that despite a reduction in insulin, there was no extra fat loss advantage for the Restricted Fat diet … which more so argues against the “Carbohydrate-Insulin Theory of Obesity” rather than denying the efficacy of low-carb diets.

    As always, there are a few more limitations to consider.

    Due to the sample population chosen, the results of this study only apply to obese adults who are otherwise healthy. Further limiting the generalizability of the results, the tightly controlled study design does not accurately represent the free-living world, as most of us do not have strict external controls on our food choices.

    And to repeat a very important point: this study was not meant to inform long-run dietary choices. In the long-run, the choice between restricting fat or restricting carbs to achieve a caloric deficit may come down to one thing: diet adherence.

    While preference for certain foods may dictate which diet is easier to adhere to, this isn’t always the case. For instance, it seems that insulin-resistant individuals have an easier time adhering to a low-carbohydrate diet. Nowadays, new dieters often pair low-carb with higher protein, the latter of which can boost weight loss. And since there are plenty of high-sugar but low-fat junk foods (see Mike and Ike, et al.) but not so many high-fat but low-carb junk foods, low carb intakes can sometimes mean an easier time staying away from junk food when compared to low fat diets.

    What about the six day trial duration? Does that mean the results are less valid? Well, it depends on the question you want answered. There aren’t any six-month-long metabolic chamber studies because they would both be ludicrously expensive and turn into studies of hospital patients rather than free-living people. So the researchers chose to shed light onto this question:

    “Could the metabolic and endocrine adaptations to carbohydrate restriction result in augmented body fat loss compared to an equal calorie reduction of dietary fat?”

    Quite clearly, the answer was no in this study. Many in the low-carb blogosphere have argued that six days was too short for fat-adaptation. Maybe, but the paper also said:

    “Net fat oxidation increased substantially during the RC (restricted carbohydrate) diet and reached a plateau after several days, whereas the RF (restricted fat) diet appeared to have little effect.”

    So restricting fat didn’t change metabolism much, while restricting carbs increased fat oxidation at first but not after a few days. It’s possible that other physiological mechanisms (e.g. mitochondria-related factors) may take longer to adapt (especially in very low carb / keto diets), but this wasn’t a study of keto diets, it was a study testing whether carb restriction leads to extra fat loss compared with fat restriction.

    And while not without their limitations, free-living studies have generally shown that the low-carb groups tend to lose a bit more fat mass by the six-month mark (even when controlling for energy intake), but weight loss at the end of the trial tends to be similar. A big part of that is likely the increased protein that is typically coupled with lower carb. Those are things that media reports won’t mention when covering the current study -- context matters, and the totality of research suggests that media harping on each low-carb trial in isolation is dumb.


    As usual, don’t bother with media headlines -- this study is NOT a blow to low-carb dieting, which can be quite effective due to factors such as typically higher protein and more limited junk food options. Rather, this study shows that a low-carb diet isn’t necessary for fat loss and that lowering carbs and insulin doesn’t provide a magical metabolic advantage.

    It bears repeating: if you even try to apply this study to the real world of dieting choices, you will be frowned upon strongly. Even the lead author writes:

    "Translation of our results to real-world weight-loss diets for treatment of obesity is limited since the experimental design and model simulations relied on strict control of food intake, which is unrealistic in free-living individuals."

    This study was strictly meant to fill in a gap in the knowledge base of diet physiology. Got it?

    If you need a broad and simple takeaway from this study, here is one: weight loss does not rely on certain carb levels or manipulation of insulin, it relies on eating less. Don’t be scared that eating carbs will cause insulin to trap fat inside your fat cells.