Learn more about Peter Attia’s digital program
There is no one-size-fits-all diet that can be applied to all individuals to reduce the prevalence of obesity. But the success of any diet – from vegetarianism to time-restricted feeding – ultimately relies on how it affects the balance between energy intake – the physiological processes that drive appetite – and energy expenditure – metabolic rate and activity level. And yet, despite decades of research, we still have no clear consensus on one question that is central to this equation: do different macronutrients vary in their effects on energy balance? We’ve addressed this topic from the energy intake perspective in a previous newsletter, but let’s now take a look at a study by Bikman et al. published in the European Journal of Clinical Nutrition that attempts to address the other half of the equation: can different macronutrients have differential effects on energy expenditure? [Disclosure: this study was partially funded by Nutrition Science Initiative at a time when I served as the president of that organization.]
The Carbohydrate-Insulin Model
As Dr. Stephan Guyenet has explained previously on the podcast, there exist various schools of thought regarding the underlying drivers of obesity. One theory is the carbohydrate-insulin model (CIM), which proposes that a high-carbohydrate diet, by increasing insulin signaling, leads to increased storage of calories as fat, which in turn reduces circulating fuels available for use by cells. This paradoxical low-energy state results in decreased energy expenditure and elevated drive to eat, creating a cycle that, if sustained, can eventually lead to obesity. However, mechanistic links between insulin, increased fat storage, and decreased energy expenditure remain poorly understood.
In their study, Bikman et al. sought to determine whether different fuels (carbohydrates vs. fats) might have a direct, calorie-independent impact on energy expenditure via effects on mitochondrial respiration rates. They hypothesized that carbohydrates would reduce adipose tissue mitochondrial activity and consequent energy output relative to dietary fat, tilting the balance of fuel partitioning toward fat deposition over fat oxidation.
About the Study
The 27 outpatient study participants (BMI ≥ 25) underwent a preliminary run-in phase in which the investigators provided them with meals at a level of 60% of their daily estimated calorie requirement in order to induce loss of 10-14% body weight. Participants were then stabilized at the reduced weights and randomized to one of three weight-maintenance diets: high-carbohydrate (carbohydrates accounted for 60% of total calories), moderate-carbohydrate (40%), or low-carbohydrate (20%). All groups received 20% of daily calories in the form of protein, and remaining calories were supplied by fat. Body weights were monitored throughout the study and total caloric intake was adjusted as necessary for each participant to ensure weight remained in the target range. Adipose tissue samples were obtained from participants post-weight loss stabilization (baseline measurement) and again after 10-15 weeks on their assigned diet. The researchers then assayed the tissue samples for various metrics of mitochondrial respiration capacity ex vivo.
To assess the effect of carbohydrate load on mitochondrial respiration, the investigators calculated the change from baseline (difference between the first and second biopsies) in activity at various stages of mitochondrial respiration (as determined by oxygen consumption using a specialized respirometer) and then compared the magnitude of the changes across the three groups. Compared to the low- and moderate-carbohydrate groups, the high-carbohydrate group had significantly lower maximal mitochondrial respiration rates, as might be expected given that high carbohydrate intake will modestly increase lactate production. This difference appeared to be caused by a large increase in maximal mitochondrial respiration relative to baseline among the low and moderate groups rather than by a decrease in the high group, as all groups demonstrated a positive change relative to baseline. In other words, some degree of elevation in mitochondrial respiration appears to have resulted from an unknown, systematic variable unrelated to carbohydrate intake, which may or may not have affected all groups equally. This upward shift across the board may indicate, for instance, that participants had not reached a metabolic equilibrium state following run-in weight loss.
No significant differences were observed between moderate- and low-carbohydrate diets. Ultimately, the investigators interpreted these data as suggesting that a high-carbohydrate diet impairs the ability of fat cells to generate energy via mitochondrial respiration. They propose an increase in insulin secretion as the possible mediator and a predisposition to obesity as a consequence of this phenomenon, suggesting these findings validate the carbohydrate-insulin model of obesity. But are these interpretations really justified?
The Gaps
At its surface, this study provides an intriguing insight into the impact of carbohydrate consumption on energy expenditure, but closer examination reveals a number of critical gaps which limit the conclusions we can make from these data.
For instance, mitochondrial respiration in adipose tissue represents only a small fraction of whole-body energy expenditure. Human tissues have varied rates of metabolic activity, and fat tissue certainly falls on the less active side of this spectrum, whereas skeletal muscles and internal organs are much more metabolically active. This study’s white adipose tissue sample hence may not be representative of whole-body energetics.
Additionally, because the energy derived from fat was not equivalent across diet groups, the varying proportion of fat in test diets can’t be excluded as possibly responsible for the observed changes. In other words, the elevations in mitochondrial respiration associated with lower levels of dietary carbohydrate might in fact have been caused by the corresponding higher levels of dietary fat.
The authors also failed to measure a critical pathway of energy metabolism which occurs outside of mitochondria: lactate production. Because carbohydrates – but not fats – can be metabolized anaerobically (via lactate production) or aerobically (via the Krebs cycle and mitochondrial respiration), we would expect an increase in dietary carbohydrate to correspond to an increase in lactate production. In other words, an increase in the proportion of carbohydrate in the diet means that more energy substrate is being diverted from mitochondrial respiration.
For mitochondrial assays, ex vivo samples were all treated with the same energy substrates, so lactate production due to an increase in carbohydrate fuel would not explain the between-group differences in respiratory capacity. However, this variable is important for understanding what the results might mean for whole-cell or whole-body energetics. The lactate pathway is far less energetically efficient than aerobic metabolism, meaning that it yields less energy in the form of ATP from a given calorie input. So if the investigators had measured lactate and found it to be elevated in the high-carbohydrate group, it might indicate that fat cells from those participants were burning more fuel overall, despite a reduction in mitochondrial respiration.
Not a Smoking Gun for the CIM
It’s also important to note the limitations of this study with regard to the evidence it provides for the CIM framework, as proposed by the investigators. The CIM posits that a high-carbohydrate diet increases insulin production, which alters metabolic fuel partitioning from fat oxidation toward fat deposition, causing obesity. But the data from this study didn’t actually show a reduction in energy expenditure associated with a high-carbohydrate diet. All diet groups showed elevated mitochondrial respiration relative to baseline – the high-carbohydrate group merely showed the smallest elevation. However, since the authors only report percent changes from baseline rather than absolute values for their measurements, we cannot know if perhaps any of the groups started at different levels.
Further, as the authors themselves admit, their investigation inexplicably left out any measures of insulin itself – the key link, according to the carbohydrate-insulin model, between dietary carbohydrates and obesity. While the investigators comment that rodent studies have shown that insulin can reduce adipocyte mitochondrial respiration, similar effects have not yet been characterized in humans. In fact, the extreme weight loss observed with use of GLP-1 agonists – which cause exaggerated postprandial insulin secretion – might suggest that insulin does not lead to increased fat deposition and reduced energy expenditure. (In other words, insulin spikes dramatically with use of GLP-1 agonists, yet this elevation is followed by appetite reduction and extreme weight loss – so how can we conclude that high insulin causes obesity?) Thus, we’re left with no clear evidence to support the investigators’ proposal of an increase in insulin secretion as a mechanism for the effects of a high-carbohydrate diet on adipose tissue mitochondrial respiration.
To evaluate mitochondrial respiration, the group studied isolated adipocyte mitochondria exclusively and used their results to hypothesize an obscenely simple pathway involving a mere three nodes: carbohydrate intake → insulin spike → fat deposition. But the biological pathways regulating energy balance are extremely complex and involve intricate (and not fully understood) cross-talk between many organ systems and cell types, and Bikman et al.’s ex vivo experimental design cannot tell us anything about how the many other players involved in energy regulation might have contributed during the diet intervention stage. Even if high-carbohydrate diets impair adipocytes’ maximal respiratory response, this does not necessarily imply an increase in fat deposition, let alone through the insulin-mediated pathway the researchers claim. For example, circulating carbohydrates, insulin levels, and fat stores are all detected by the brain, which in turn can modulate energy expenditure by signaling to peripheral organs.
The Bottom Line
While various models of obesity propose different mechanisms behind weight gain and adiposity, most acknowledge diet composition may play an important role. But debate rages on over precisely what that role might be and whether different macronutrients might impact energy intake or expenditure.
The paper published by Bikman et al. provides a clue in showing that a diet high in carbohydrates alters mitochondrial capacity for energy production in adipose tissue, but it offers no insights into the pathway(s) behind this effect and fails to provide key information which might have shed light on the implications for whole-body energetics. In short, this study provides no evidence whatsoever of insulin as a causal link or the impact of a diet low in carbohydrates and relatively high in fat on fat deposition. But regardless of the mechanism or its implications, these data do indicate that diet composition may have some effect on the mitochondrial capacity of fat cells. So while it tells us nothing about the validity of the CIM, this study does give us more reason to believe that how much we eat isn’t all that matters: what we eat also deserves our attention – and more thorough research.
For a list of all previous weekly emails, click here.
