The authors of a 2021 narrative review^[1] argued against the energy balance model of obesity. They instead argued for the carbohydrate-insulin model (CIM), which asserts that a high-glycemic diet increases fat gain and that this fat gain drives a positive energy balance, perpetuating a cycle of continued excess weight gain. Additionally, the authors of the narrative review stated that the EBM doesn’t consider “biological mechanisms that promote weight gain.”

The role of carbohydrates and insulin in the etiology of obesity continues to be debated. What does the evidence say?

The authors of the current narrative review responded to the aforementioned 2021 review, defending the EBM model. They described the EBM, CIM, and relevant data from studies assessing rodents, human genetics, epidemiology, and dietary and pharmacological interventions.

The Energy Balance Model of Obesity: The EBM states that the brain is the main organ responsible for body weight regulation, coordinating with the endocrine, metabolic, and nervous systems. The brain modulates food intake unconsciously, responding to environmental influences and the body’s energy needs. This occurs through short-term signals sent to the brain between and during meals (e.g., ghrelin, PYY, GLP-1) and long-term signals (e.g., leptin) released from adipose tissue, which tell the brain how much stored energy (i.e., body fat) is available. Thus, while daily energy intake and energy balance can be highly variable, regulation of energy balance is achieved over long durations.

In addition, macronutrient intake affects whole-body net oxidation rates of carbohydrate, fat, and protein, so that overall energy imbalances are primarily reflected as fat imbalances, regardless of the macronutrient composition of the diet. For example, if someone eats in an energy surplus, with the excess calories primarily derived from protein, the body will convert the excess protein to glucose, and fat oxidation (or “fat-burning”) will decrease because protein is being used as fuel, resulting in a positive fat balance.

The authors state that the characterization of the EMB in the aforementioned narrative review as a theoretical model that “essentially disregards knowledge about the biological influences on fat storage”^[1] is incorrect and is an oversimplification of the EBM.

The Carbohydrate-Insulin Model of Obesity: The CIM states that obesity results from high carbohydrate intake driving excess insulin secretion, causing adipose tissue to accumulate and “trap” fat, preventing nonadipose tissues from getting fuel. The authors of the current review noted that fat storage can occur in the absence of dietary carbohydrates or increases in insulin and that many factors beyond dietary carbohydrates determine insulin secretion.

The authors also noted that the CIM, as described in the recent narrative review, is different from its previous iterations in that “all obesogenic factors (e.g., amount of dietary protein, micronutrients, poor sleep, stress, physical inactivity, and environmental endocrine-disrupting chemicals) affect insulin secretion or adipocyte biology directly, with increased energy intake and decreased energy expenditure as necessary downstream consequences.” In this way, the authors state that the current CIM is an “oversimplified version of the EBM, with a focus on glycemic load as the main driver of excess energy intake.”

Evaluation of the EBM and CIM:

• Rodent studies: The authors noted that most standard laboratory rodent diets are high in carbohydrates and do not induce obesity, whereas diets with lower percent carbohydrates and higher percent fat often induce obesity. Additionally, standard laboratory rodent diets contain carbohydrates derived from corn starch, maltodextrin, and sucrose, all of which have a high glycemic index.

• Human genetics: The EBM implicates the brain as the primary organ responsible for obesity, whereas the CIM implicates adipose tissue. Other than rare mutations in the leptin gene, no genetic disorders primarily affecting fat calls or insulin have been reported to cause obesity. In contrast, genome-wide association and gene expression studies have determined that variations in adiposity between people are primarily due to differences in genes most highly expressed in the brain.

• Human epidemiological studies: Evidence does not suggest that carbohydrate intake is the main driver of the U.S. obesity epidemic. Epidemiological evidence finds that obesity risk is based on long-term adherence to various healthful dietary patterns, with variable carbohydrate contents.

• Human diet intervention studies: The CIM predicts that long-term weight loss occurs by reducing dietary carbohydrate and glycemic load, resulting in less hunger, lower food intake, and increased energy levels.^[1] However, diet intervention studies have found that low glycemic load diets do not result in greater weight loss than higher glycemic load diets. Additionally, long-term diet intervention studies comparing low-fat and low-carbohydrate diets controlling for protein intake do not report a difference in weight loss.

• Human pharmacological intervention studies: The CIM states that high insulin levels induce weight gain by inhibiting adipose tissue lipolysis, thereby trapping fat in fat cells.^[1] However, inhibiting adipose lipolysis with acipimox treatment does not affect energy intake, resting energy expenditure, or body composition, and GLP-1 receptor agonists (medications used to treat obesity) acutely increase insulin secretion.

The controversy surrounding carbohydrates, insulin, and obesity is unlikely to end anytime soon. As such, it’s worth revisiting prior issues of Study Summaries that have shed light on the topic:

The September 2020 issue of Study Summaries includes a review of a secondary analysis of another randomized controlled trial that analyzed the estimated caloric requirements of participants on either a low- or high-carb diet during a weight-maintenance period. The average estimated caloric requirements were about 245 kcal/day higher in the low-carb group than in the high-carb group. However, the interindividual variability was high, and some of the methodology used in the original trial has been hotly contested.

The December 2020 issue of Study Summaries includes a review of a 4-week crossover study comparing the effects of a plant-based low-carbohydrate diet (LCD: 10% carbs, 50% fat, 40% protein), an animal-based LCD (10% carbs, 60% fat, and 30% protein), and a low-fat diet (LFD: 61% carbs, 21% fat, and 18% protein) on body composition and blood markers. All three diets resulted in weight loss, but the plant-based LCD group lost more weight than the LFD.

The January 2021 issue of Study Summaries includes a review of a meta-analysis of 38 randomized controlled trials that investigated the effects of low-carb diets, low-fat diets, or both, on weight loss and blood lipids. Low-carb dieters lost about one additional kilogram of weight overall, and this difference was most notable when the diets ranged from 6 to 12 months. Triglycerides were reduced more by low-carb diets, whereas LDL-C, HDL-C, and total cholesterol were reduced more by low-fat diets.

The same January issue of Study Summaries includes a review of a systematic review of 8 randomized controlled trials comparing saturating fat intakes and measurements of body weight, blood glucose, cholesterol, and blood pressure among adults with a BMI of at least 25 who ate either a low-carb, high-fat diet (LCHF) or a low-fat diet (LF). Both diets resulted in significant weight reduction, improved blood glucose levels and inflammatory markers, and lowered blood pressure. The reduction in LDL cholesterol was stronger in the LF diet group, while increased HDL cholesterol and reduced triglycerides were observed in the VLCHF group.

The same January issue of Study Summaries includes a review of an 18-month study assessing the effects of a low-carbohydrate Mediterranean diet (LCMD) on hepatic fat content and visceral adipose tissue compared to a low-fat diet (LFD). The participants in the LCMD group had a greater reduction in hepatic fat content than the participants in the LFD group. After adjusting for changes in visceral adipose tissue, the LCMD group also experienced greater improvements in blood lipids, blood pressure, and cardiovascular risk score.

The March 2021 issue of Study Summaries includes a review of a crossover study in which participants followed a low-carb diet (LCD: 10% carbs, 75% fat) and a low-fat diet (LFD: 75% carbs, 10% fat) for two weeks each. The two diets led to similar weight loss (1–2 kg / 2.2–4.4 lb, over two weeks). However, the low-fat diet led to a greater reduction in fat mass, whereas the low-carb diet led to a greater reduction in fat-free mass.

The same March issue of Study Summaries includes a review of a meta-analysis of 18 randomized controlled trials that investigated the effects of a ketogenic diet compared to a low-fat control diet. The ketogenic diet reduced body weight, BMI, fat mass, fat-free mass, waist circumference and visceral fat, lean body mass, and body fat percentage, compared to a low-fat diet. Variability between studies was found for body weight, fat mass, and BMI. BMI was not reduced in studies that included only women.

The September 2021 issue of Study Summaries included a review of a meta-analysis of randomized controlled trials assessing the relative effectiveness of low-fat/high-carb (LFHC) and low-carb/high-fat (LCHF) diets on weight loss and cardiovascular risk factors. Compared to LFHC diets, LCHF diets resulted in a greater weight loss (−1.01 kg) and a greater increase in HDL-C (+7.7 mg/dL). However, LCHF diets resulted in a smaller decrease in total cholesterol and LDL-C (−24.4 and −22.8 mg/dL, respectively).

The December 2021 issue of Study Summaries includes a review of a 6-month, nonrandomized controlled trial in which participants consumed either a low-carb diet (LCD) or a low-fat diet (LFD). Compared to baseline, both diets improved triglycerides, HDL-C, diastolic blood pressure, fasting blood glucose, and waist circumference. The only significant difference between groups was for triglycerides, which showed greater reductions with the LFD.

The February 2022 issue of Study Summaries includes a review of a 6-month randomized controlled trial that randomized people with type 2 diabetes to follow a low-carb, high-protein diet (LCHP: 14% carbohydrates, 28% protein, 58% fat) or a low-fat diet (LFD: 53% carbohydrates, 17% protein, 30% fat). Body weight decreased in the LCHP group compared to the LF diet (−4.1 vs. −1.0 kg), and markers of glycemic control improved to a greater extent in the LCHP group compared to the LF diet.

The March 2022 issue of Study Summaries includes a review of a meta-analysis of 61 randomized controlled trials comparing the effects of low-carb weight-reducing (LCWR) diets with balanced-carbohydrate weight-reducing (BCWR) diets on body weight and cardiovascular risk factors among 6,925 participants with obesity. Participants without T2D experienced greater weight loss (−1.07 kg) in the low-carbohydrate diet group over 3 to 8.5 months, compared to the balanced-carbohydrate diet group. Similarly, there was greater weight loss (−0.93 kg) in the low-carbohydrate diet group over 1 to 2 years. Participants with T2D experienced greater weight loss (−1.26 kg) in the low-carbohydrate group over 3 to 6 months, compared to the balanced-carbohydrate diet group. However, at 1 to 2 years, there was no difference in weight loss between groups.