Why your pasta might not be the hunger villain after all: New research upends belief that carb-driven insulin spikes lead to overeating — and reveals what actually matters for appetite.
Study: Testing the carbohydrate-insulin model: Short-term metabolic responses to consumption of meals with varying glycemic index in healthy adults. Image Credit: YesPhotographers / Shutterstock
In a recent article published in the journal Cell Metabolism, researchers tested whether meals with different glycemic indices (GI) lead to distinct metabolic and hunger responses, thus affecting appetite and intake during subsequent meals. Their findings indicate that high GI alone may not drive hunger —and found no differences in perceived hunger across groups— countering conventional wisdom.
Background
The prevalence of obesity has increased sharply in recent decades, with bodies such as the World Health Organization raising the alarm about an ongoing global obesity epidemic. Uncovering the underlying causes of obesity is an important research focus as it increases the risk of several non-communicable diseases, including stroke, hypertension, cardiovascular disease, and diabetes mellitus.
One theory, the carbohydrate-insulin model (CIM), posits that eating high-GI foods can lead to weight gain since they cause spikes in blood sugar, triggering high insulin production and lowering glucagon levels. This promotes fat storage, leading to a subsequent drop in blood sugar below normal levels that the body perceives as starvation. It responds by slowing metabolism and increasing hunger, which may lead to the individual overeating during their next meal. Over time, this may create a cycle of fat storage and increased food consumption, leading to obesity.
The results of a previous study with adolescent participants supported the CIM but failed to separate the effects of carbohydrate consumption from those of other macronutrients. A more rigorous test of the CIM requires varying GI levels while keeping macronutrient composition levels constant, ensuring that results are attributable to differences in GI alone.
About the Study
The research team examined how meals with different GI levels affect metabolism in healthy adults. All the meals had the same macronutrient composition in terms of fats, proteins, and carbohydrates, varying only in GI. However, participants were unaware of the GI differences.
Researchers predicted that meals with higher GI would cause a sharper rise and fall in blood glucose, trigger greater insulin release, and subsequently lead to lower levels of circulating fuels such as fatty acids and glucose. This would increase hunger and result in higher food intake during the next meal.
Before the experiment, participants ate a baseline meal from a free-choice food table and underwent scans to measure their body fat; they were also fitted with continuous glucose monitors that measured glucose levels every five minutes. On the day of the experiment, they were asked to arrive after fasting for more than 10 hours. Their hunger levels (on a 100-point scale), height, and weight were measured, and they consumed a test meal (high, medium, or low GI).
Over the next five hours, they provided blood samples (to assess hunger-related hormones, lactate, ketones, fatty acids, and insulin) and information about their perceived hunger before they were given a test meal similar to the baseline meal, and food intake was measured.
The meals contained 20% fat, 20% protein, and 60% carbohydrates; the low-GI meal consisted of spaghetti pasta, the medium-GI meal consisted of buckwheat noodles, and the high-GI meal consisted of steamed bread.
Power analysis suggested that a minimum of 135 participants were needed for the study. The data was analyzed using correlation analysis, regression models, and analysis of variance (ANOVA) to compare the different GI groups and explore links between metabolic changes and food intake.
Findings
While 135 participants participated in screening for the study, only 130 participants enrolled, and 10 did not complete the study. The final sample comprised 40 participants in each GI group, with no significant differences in body mass index (BMI), height, weight, or age. There were also no significant differences in intake during the baseline or test meals.
Overall energy intake from baseline (pre-experimental day) to the experimental day increased by 71 kJ for the low GI group, 671 kJ for the medium GI group, and 617 kJ for the high GI group. Notably, the low-GI group showed a smaller increase in energy intake compared to the medium and high groups. However, absolute intake during the test meal did not differ significantly between groups. Sex was not correlated with energy intake. There were no significant changes in carbohydrate intake, and protein and fat intake were higher for the medium GI group, possibly reflecting subtle differences in meal composition despite efforts to standardize macronutrients.
In terms of hormonal and metabolic responses, glucose levels increased after all meals; while the low GI group exhibited the lowest initial rise, they also showed the highest levels after five hours. Overall, the high GI group had a significantly higher glucose response than the low GI group. Insulin levels increased significantly in the high GI group at several time points, and overall insulin response was also highest in this group. Other hormones did not vary with meal type.
Considering energy intake and insulin response, higher insulin levels before the test meal were linked to lower energy intake, a finding contrary to CIM predictions. The insulin-glucagon ratio was also negatively correlated with energy intake, but no relationship was found between changes in energy intake and other hormones or metabolites. Critically, there was no difference in hunger ratings between the three groups at any time point, contradicting a central tenet of the CIM.
Conclusions
While glucose and insulin levels followed the patterns the research team expected to see, the absence of hunger differences directly challenges the CIM’s proposed mechanism. Energy intake did increase relative to baseline with medium and high GI meals, but individual responses varied significantly. The negative correlation between insulin and subsequent intake also contrasts with CIM expectations, suggesting the model’s limitations in explaining real-world appetite regulation. As this experiment took place for a short period with partial blinding, further research is needed to examine the long-term effects of high GI diets.
Journal reference:
- Testing the carbohydrate-insulin model: Short-term metabolic responses to consumption of meals with varying glycemic index in healthy adults. Liu, Y., Mei, H., Xue, L., Cheng, C., Wu, Y., Chao, Z., Yu, Y., Gao, L., Zhang, H., Gao, X., Li, Q., Wang, L., Liu, J., Niu, C., Zhang, X., Hu, S., Speakman, J.R. Cell Metabolism (2025), DOI: 10.1016/j.cmet.2025.01.015, https://www.cell.com/cell-metabolism/fulltext/S1550-4131(25)00015-4