Type 2 diabetes and progressive metabolic deterioration

Over the years, type 2 diabetes has been described primarily through static values, such as fasting glycemia and glycated hemoglobin (HbA1c). For a long time, these parameters have been the pillars of diagnosis and disease monitoring, but they only capture a single moment or an average of glucose exposure over several months. It is somewhat like evaluating the traffic of an entire city by looking at a single photograph taken at an intersection: the image may be accurate at that precise moment, but it does not tell you how intense the traffic was, how much it varied throughout the day, or if sudden traffic jams occurred. Similarly, two people can have the same fasting glycemia or the same HbA1c while having completely different glycemic profiles. One may maintain relatively stable values throughout the day, while the other alternates between continuous postprandial spikes and rapid drops in glycemia. Today, we know that these very oscillations represent one of the most important aspects of the pathophysiology of diabetes, as they directly reflect the body's ability to maintain glucose homeostasis (Acciaroli et al., 2018) [11].

The introduction of Continuous Glucose Monitoring (CGM) systems has profoundly changed this perspective. Unlike traditional point-in-time measurements, CGM records thousands of glycemic values during normal daily life, allowing us to observe not only the average glycemic value but also its stability, the frequency of excursions, the duration of postprandial spikes, and the body's ability to quickly return glucose to physiological limits. This dynamic assessment has made it possible to understand that the deterioration of glucose metabolism does not appear suddenly when glycemia exceeds a certain threshold, but develops through a long process characterized by a progressive loss of glycemic stability, often many years before a diabetes diagnosis (Cichosz et al., 2025) [12].

One of the most important findings to emerge from recent studies is the demonstration that glycemic variability increases gradually along the entire metabolic continuum. By analyzing over eight hundred participants from five different clinical studies, it was observed that mean glucose increases progressively, moving from approximately 99 mg/dL in normoglycemic subjects to over 180 mg/dL in patients with poorly controlled diabetes. In parallel, the time spent above 140 mg/dL, the Mean Amplitude of Glycemic Excursions (MAGE), and numerous other variability indices increase, while the body's ability to quickly correct small glycemic oscillations progressively decreases. Even the complexity of the glycemic signal, an expression of an efficient regulatory system, reduces progressively as dysglycemia advances, indicating a loss of physiological balance that precedes the onset of overt diabetes (Cichosz et al., 2025) [12].

These observations explain why prediabetes (the alteration of glycemic regulation that precedes type 2 diabetes) can no longer be considered a simple phase characterized by a slight elevation in mean glycemia. Continuous monitoring shows that profound alterations in glucose regulation appear as early as this initial phase. In subjects carefully matched for age, sex, and body mass index, CGM has shown that people with prediabetes spend less time within the optimal glycemic range, have higher average values, and, above all, develop significantly greater oscillations compared to metabolically healthy subjects. Parameters such as the coefficient of variation (CV) and MAGE are already significantly increased, while the response to the glucose tolerance test shows higher glycemic exposure and slower normalization after the load. These results indicate that glycemic instability is one of the first signs of the loss of metabolic homeostasis and often precedes the stable increase in fasting glycemia or HbA1c (Rizos et al., 2024) [13].

This new vision also challenges one of the traditional principles of diabetes diagnosis: the idea that a single fasting glycemia measurement can faithfully represent a person's metabolic state. A large study conducted on over eight thousand non-diabetic adults has shown that fasting glycemia exhibits significant intra-individual variability, with daily oscillations exceeding 7 mg/dL. A variability of this magnitude is sufficient to cause the same individual, on different days, to shift from normoglycemia to the prediabetes or even diabetes range. It is therefore not surprising that about 40% of subjects initially classified as normoglycemic presented at least one subsequent measurement compatible with prediabetes, and that 3% even exceeded the diagnostic threshold for diabetes. Fasting glycemia thus appears as the final result of numerous physiological factors, including hepatic glucose production, basal insulin secretion, sleep, physical activity, dinner composition, the dawn phenomenon, and stress response, making it clear that a single value cannot describe the entire complexity of glucose metabolism (Shilo et al., 2024) [14].

Metrics derived from continuous monitoring, on the other hand, allow us to describe this complexity through numerous mathematical indices. Some measure how much glycemic values vary and deviate from their mean value during a given period, such as standard deviation and the coefficient of variation; others evaluate the magnitude of major glycemic excursions, such as MAGE; still others estimate the time spent above or within certain glycemic ranges or quantify the risk of hypoglycemia and hyperglycemia. Although there are dozens of different parameters, numerous studies show that only some of them are truly informative. Among these, the coefficient of variation emerges as one of the most robust indicators, while MAGE is likely the most sensitive metric for describing the loss of physiological control of glycemia, as it simultaneously integrates the magnitude of glycemic spikes and subsequent reductions (Acciaroli et al., 2018) [11]; (Zahalka et al., 2024) [15].

The importance of glycemic variability, however, is not limited to a simple description of metabolism. Glycemic oscillations seem to anticipate the development of overt diabetes by many years. The prospective AEGIS study, which followed nearly five hundred non-diabetic subjects for six years, demonstrated that those who later developed diabetes already exhibited significantly higher glycemic oscillations at the time of enrollment compared to participants who remained normoglycemic. Among all the parameters analyzed, the standard deviation of glycemia proved to be the best predictor of the future onset of the disease, reaching a very high discriminatory capacity. In other words, the loss of glycemic stability precedes the onset of diabetes and can be detected years before the disease becomes clinically evident (Rodríguez García et al., 2025) [16].

This predictive capacity likely stems from the fact that glycemic oscillations are not just a statistical phenomenon, but reflect well-defined biological processes. Repeated increases and reductions in glycemia promote the production of reactive oxygen species, increase oxidative stress, activate inflammatory pathways, impair endothelial function, and subject pancreatic β-cells to continuous functional overload. Glycemic variability thus becomes not only an early indicator of the disease, but also one of the possible mechanisms through which it progresses over time (Rodríguez García et al., 2025) [16].

A central role is played by glycemic responses after meals. While fasting glycemia primarily reflects hepatic glucose production and basal insulin secretion, the postprandial response simultaneously involves numerous physiological mechanisms: insulin secretion, peripheral insulin sensitivity, incretin activity (intestinal hormones that help the pancreas regulate glycemia), gastric emptying rate, intestinal absorption, and hepatic metabolism. A prospective study with nearly ten years of follow-up demonstrated that the area under the glycemic curve after a normal mixed meal is a powerful independent predictor of the future development of diabetes, maintaining its prognostic value even after correction for insulin sensitivity, insulin secretion, and body composition. This suggests that the body's ability to handle a real meal constitutes a much more complete measure of metabolic health than traditional diagnostic tests based exclusively on fasting glycemia or oral glucose tolerance (Mitchell et al., 2024) [17].

Continuous monitoring has also highlighted another aspect of extraordinary importance: different individuals can respond in profoundly different ways to the same food. By analyzing subjects who were carefully characterized from a metabolic standpoint, it was observed that some develop the greatest glycemic increase after consuming rice, others after bread, potatoes, or pasta. These differences do not depend solely on the food, but reflect specific metabolic phenotypes characterized by different levels of insulin resistance, β-cell functionality, lipid metabolism, gut microbiota composition, and distinct profiles. Even the effectiveness of simple nutritional strategies, such as the intake of fiber, protein, or fat before carbohydrates, varies considerably depending on the individual metabolic state, paving the way for true precision nutrition (Wu et al., 2025) [18].

Taken together, these findings are profoundly changing the way we interpret type 2 diabetes. The disease does not begin when glycemia exceeds a certain diagnostic threshold, but much earlier, when the body gradually loses the ability to keep glucose stable throughout the day. Glycemic variability therefore represents one of the first manifestations of metabolic deterioration, often invisible to traditional laboratory tests but clearly detectable through continuous monitoring. Although large-scale prospective studies are still needed to define shared diagnostic criteria, CGM is establishing itself as a tool capable not only of monitoring overt diabetes, but also of identifying at-risk subjects early, personalizing lifestyle interventions, and, potentially, modifying the natural history of the disease (Zahalka et al., 2025) [19].

The glycemic index (GI) and glycemic load (GL) are now two of the most important concepts for understanding the relationship between nutrition and blood sugar control. For many years, the focus was almost exclusively on the quantity of carbohydrates consumed in the diet, based on the assumption that all carbohydrates exerted substantially similar metabolic effects. However, evidence accumulated over the last few decades has profoundly changed this view, demonstrating that foods containing the same amount of carbohydrates can lead to completely different glycemic responses. This depends on the speed at which glucose is digested, absorbed, and released into the bloodstream—a process influenced by the food's structure, the presence of fiber, the degree of refinement, industrial processing, and the overall composition of the meal. It follows that it is not just how many carbohydrates are consumed that matters, but also how quickly they can change blood sugar levels after a meal (Augustin et al., 2015) [20].

To describe these aspects, two complementary indicators have been developed. The glycemic index measures the speed at which a carbohydrate-containing food increases blood sugar compared to a reference food, generally pure glucose. The glycemic load, on the other hand, integrates the glycemic index with the actual amount of carbohydrates present in the portion consumed, providing a more realistic estimate of the meal's metabolic impact. A food may indeed have a high glycemic index but contain few available carbohydrates and therefore result in a modest glycemic load; conversely, a food with a medium glycemic index, when consumed in large quantities, can produce a significantly higher glycemic exposure.

For this reason, the international consensus of the International Carbohydrate Quality Consortium has recognized the glycemic index and glycemic load as scientifically valid, reproducible, and complementary tools for evaluating carbohydrate quality, emphasizing that they must be interpreted alongside other elements that characterize a healthy diet, such as fiber content, fat quality, the presence of whole grains, and overall nutritional density (Augustin et al., 2015) [20].

The reason for this focus lies in the physiological consequences of the postprandial glycemic response. After a meal characterized by a high glycemic index or glycemic load, glucose rapidly reaches the bloodstream, causing a marked increase in blood sugar and a consequently intense insulin secretion. If this phenomenon repeats daily, the body is subjected to continuous glycemic and insulin spikes that promote the onset of oxidative stress, low-grade chronic inflammation, endothelial dysfunction, and the progressive deterioration of pancreatic β-cell function. Conversely, low glycemic index foods release glucose more slowly, producing more gradual glycemic curves, a less pronounced insulin response, and greater metabolic stability during the hours following the meal (Augustin et al., 2015) [20].

This stability now represents one of the main goals of preventive nutrition. Thanks to continuous glucose monitoring, it has been possible to demonstrate that the overall quality of nutrition directly influences daily glycemic variability. The analysis of over six hundred participants in the Framingham Heart Study highlighted that a higher intake of carbohydrates, especially when replacing proteins, is associated with an increase in average blood sugar and glycemic oscillations.

Conversely, a diet characterized by a high fiber content, a favorable ratio between carbohydrates and fiber, and better overall nutritional quality is systematically associated with a reduction in average blood sugar, the coefficient of variation, MAGE, the time spent above 140 mg/dL, and other main indices of glycemic variability. In subjects with prediabetes, consuming at least one gram of fiber for every nine grams of carbohydrates reduces the time spent in hyperglycemia by 7 to 10%, demonstrating how carbohydrate quality exerts measurable effects even in the very early stages of metabolic deterioration (Bakhshi et al., 2025) [21].

These results explain why the focus today is progressively shifting from the individual food to the meal as a whole. Traditional glycemic index tables are constructed by evaluating foods in isolation, whereas in daily life, they are almost always consumed together with other nutrients. An important experimental study demonstrated that, with the same amount of carbohydrates, the addition of high amounts of protein reduces the area under the glycemic curve by about 25% and significantly lowers the actually measured glycemic index and glycemic load.

This effect derives mainly from the slowing of gastric emptying mediated by intestinal hormones such as GLP-1, GIP, and cholecystokinin, as well as from the greater insulin secretion stimulated by certain amino acids. However, the traditional formulas used to calculate the glycemic index and glycemic load, based exclusively on carbohydrate content, cannot predict these effects, often tending to overestimate the real glycemic impact of protein-rich meals (Meng et al., 2017) [22].

The same principle applies to dietary fiber and, to a variable extent, to fats. A higher presence of fiber slows down the digestion and absorption of glucose, reducing the speed at which it reaches the bloodstream. Although the extent of this effect depends on the type of fiber, its viscosity, the degree of refinement, and the food matrix, epidemiological evidence consistently demonstrates that a high fiber intake is associated with lower glycemic variability and a reduced risk of developing type 2 diabetes. Fats, on the other hand, seem to exert a more variable effect, likely influenced by the amount consumed and their qualitative composition, but they also contribute to modifying the glycemic response by slowing gastric emptying and carbohydrate digestion (Meng et al., 2017) [22] (Bakhshi et al., 2025) [21].

The importance of the glycemic load emerges with particular clarity when foods are observed in normal daily life. By analyzing over one million glycemic measurements obtained through continuous glucose monitoring in more than five hundred non-diabetic adults, it was demonstrated that meals characterized by a high glycemic load cause more intense and prolonged increases in blood sugar, especially after lunch and dinner. Furthermore, the glycemic response does not depend solely on the meal's content but is modulated by age, body mass index, sex, glycated hemoglobin, and circadian rhythm.

Aging, for example, prolongs the time blood sugar remains at high levels, while the physiological drop in insulin sensitivity in the evening hours amplifies the effect of carbohydrate-rich meals. These data demonstrate how the glycemic load maintains a significant predictive capacity even in real-life conditions, confirming that the metabolic response depends simultaneously on the characteristics of the food and those of the person consuming it (Calvo-Malvar et al., 2025) [23].

The distribution of carbohydrates throughout the day also plays an important role. A systematic review including eleven randomized clinical trials demonstrated that a breakfast characterized by a low glycemic index and low glycemic load significantly reduces postprandial blood sugar (blood sugar level after a meal) at 60, 90, and 120 minutes compared to high glycemic index breakfasts. The benefit is even more marked in subjects with type 2 diabetes, insulin resistance, hyperinsulinemia, or impaired glucose tolerance, in whom the slowing of glucose absorption limits the secretory load on pancreatic β-cells, improves insulin sensitivity, and reduces glycemic variability (Toh et al., 2020) [24].

Similar results emerge when the proportion of macronutrients in breakfast is modified. In subjects with impaired glucose regulation, a breakfast containing less than 45% of energy derived from carbohydrates produces glycemic responses practically identical to those observed in metabolically healthy subjects. As the percentage of carbohydrates increases, the glycemic peak, the duration of hyperglycemia, and glycemic variability progressively increase, while a higher presence of proteins, unsaturated fats, and fiber allows for a faster return to baseline values and reduces the metabolic load on β-cells (Kang et al., 2013) [25].

In type 2 diabetes, these effects appear even more evident. A randomized crossover study demonstrated that drastically limiting carbohydrates exclusively during breakfast reduces the main daily glycemic peak by 74%, significantly decreases overall exposure to hyperglycemia over 24 hours, and improves the MAGE parameter, without compromising the response to subsequent meals. The benefit therefore extends well beyond the single meal, confirming how controlling the quality and quantity of carbohydrates in the early hours of the day can favorably influence the entire daily glycemic profile (Chang et al., 2019) [26].

More recent research has also highlighted that the glycemic response to meals is not the same for everyone. Even when consuming identical foods, some people develop much higher glycemic excursions than others. Using continuous glucose monitoring, it was possible to identify individuals particularly sensitive to carbohydrates, defined as "high responders," in whom meals containing over 50% carbohydrates cause marked and prolonged glycemic increases.

Conversely, meals in which carbohydrates represent about 45% of total energy are not associated with either glycemic variability or average blood sugar during two weeks of monitoring. This suggests that the composition of the individual meal may be more important than the total amount of carbohydrates consumed throughout the day and reinforces the concept of personalized nutrition based on individual metabolic response (Song et al., 2023) [27].

This heterogeneity is further confirmed by analyses carried out in healthy subjects of different ages. Although normal aging does not significantly increase glycemic variability, a higher carbohydrate intake continues to correlate with wider glycemic oscillations in both the young and the elderly. Furthermore, the ratio between carbohydrates and fats significantly modifies the metabolic response, suggesting that the complexity of the meal influences blood sugar more than the simple carbohydrate content. Once again, it emerges that carbohydrate quality, macronutrient distribution, and individual characteristics must be considered simultaneously in nutritional planning (Köhlmoos and Dittmar, 2025) [28].

Finally, even in forms of diabetes characterized by severe insulin deficiency, the quality and especially the quantity of carbohydrates continue to represent one of the main determinants of glycemic stability. A low carb diet improves the coefficient of variation, reduces MAGE, increases the time spent in the optimal glycemic range, and decreases the need for insulin, with benefits particularly evident in patients with more severely compromised β-cell function (Song et al., 2026) [29]. These results confirm that intervening in the glycemic load of the diet does not just mean reducing blood sugar after a meal, but substantially modifying the entire dynamic of metabolic control. Understanding the role of the glycemic index and glycemic load therefore represents the necessary prerequisite for addressing the topic of carbohydrate restriction, which will be explored in the next paragraph dedicated to the effects of low-carbohydrate diets on glycemic variability.

The growing popularity of continuous glucose monitoring systems has allowed for precise observation of how different dietary strategies influence metabolic control throughout the day. While in the past the efficacy of a diet was evaluated almost exclusively through fasting glycemia or glycated hemoglobin, today it is also possible to analyze glycemic variability—that is, the set of fluctuations that blood glucose levels undergo over twenty-four hours. This shift in perspective has led to a profound re-evaluation of the role of low carb diets (LCDs), as their benefit does not seem to be limited to reducing average blood glucose, but primarily concerns the stability of the glycemic profile.

Reducing the main substrate responsible for the rise in blood glucose after meals means, in fact, attenuating postprandial excursions, decreasing the secretory load imposed on pancreatic β-cells, and reducing the amount of insulin required to maintain metabolic control, with effects that are favorably reflected in numerous indicators obtained through continuous glucose monitoring (Wang and Pei, 2025) [30].

These observations have been confirmed by a recent meta-analysis that brought together all studies in which the effects of low carb diets were evaluated using CGM. By analyzing a total of seven studies and over three hundred patients with type 2 diabetes, the authors observed a significant reduction in average twenty-four-hour blood glucose, with an estimated decrease of approximately 0.51 mmol/L, equivalent to a reduction in HbA1c of approximately 0.3 percentage points.

This improvement derives mainly from the reduction of the glycemic load of meals, lower chronic exposure to insulin, and the progressive improvement of insulin sensitivity, particularly in skeletal muscle. The meta-analysis also suggests that longer-term interventions could produce even more marked benefits, while emphasizing that the currently available evidence derives primarily from short-term studies and that the sustainability of the approach in the long term depends largely on the patient's adherence to the dietary program (Wang and Pei, 2025) [30].

The speed with which metabolism responds to the reduction of carbohydrates also emerges clearly from experimental studies. In a randomized crossover trial conducted on patients with type 2 diabetes treated with metformin, a diet containing approximately 30% of energy from carbohydrates produced, after just forty-eight hours, an improvement in virtually all indices of glycemic variability. In addition to the decrease in average blood glucose recorded over twenty-four hours, significant reductions were observed, indicating that blood glucose was not only lower but, above all, much more stable throughout the day. The particularly interesting aspect is that these benefits appeared well before any change in HbA1c, demonstrating how the stabilization of blood glucose represents one of the first metabolic effects of carbohydrate restriction (Thomsen et al., 2020) [31].

The reduction of glycemic oscillations also seems to be closely linked to an improvement in pancreatic endocrine function. In a subsequent study, carried out with an isocaloric diet (a diet in which the calories introduced correspond approximately to the body's energy requirements) aimed at maintaining body weight, carbohydrate restriction reduced postprandial glycemic excursion by approximately 60% and average blood glucose detected by CGM by 13%. At the same time, circulating proinsulin levels decreased, β-cell function indices improved, and the sensitivity of insulin secretion to glucose increased.

The authors also observed a slowing of gastric emptying, greater secretion of incretins, and an increase in satiety, suggesting that the effects of the diet do not depend exclusively on the reduction of glucose introduced with food, but also on profound modifications of digestive and hormonal physiology that contribute to making the glycemic response to meals more gradual (Skytte et al., 2021) [32].

An elegant crossover study demonstrated that even just limiting carbohydrates exclusively during breakfast favorably modified the entire daily glycemic profile. A breakfast containing less than 10% of calories in the form of carbohydrates reduced the main morning glycemic peak by 74%, significantly decreased the incremental area of blood glucose over twenty-four hours, and improved the MAGE parameter, without causing a worsening of glycemic responses after lunch or dinner. It is interesting to note that participants also reported less hunger in the following hours and a reduction in the desire for sweet foods, suggesting that a relatively simple and easily sustainable strategy can produce relevant metabolic benefits without requiring permanent carbohydrate restriction in all meals (Chang et al., 2019) [26].

The extent of the benefit seems, however, to depend on the degree of carbohydrate restriction. An observational study based on continuous glucose monitoring highlighted that a moderate reduction in carbohydrate intake improves body weight, triglycerides, HDL cholesterol, and reduces the time spent in hyperglycemia. However, more marked changes in glycemic variability, such as the improvement of Time in Range and the reduction of the High Blood Glucose Index, are more evident in subjects who follow dietary regimens characterized by greater carbohydrate restriction. The authors emphasize, however, that such strategies must maintain an adequate intake of fiber, vitamins, and minerals, avoiding that the reduction of carbohydrates compromises the overall nutritional quality of the diet (Osugi et al., 2023) [33].

The efficacy of low-carb diets appears even more evident when they are maintained for prolonged periods within structured programs. An important two-year prospective study demonstrated that a program including nutritional education, continuous remote monitoring, and nutrition oriented toward nutritional ketosis leads to persistent improvements in HbA1c, fasting glycemia, insulinemia, and the HOMA-IR index. At the same time, a consistent weight loss, a marked reduction in visceral fat, an improvement in blood pressure, hepatic steatosis, inflammatory markers, and the lipid profile are observed, accompanied by a significant reduction in pharmacological therapy and the achievement of diabetes remission in a significant proportion of patients. These results indicate that, when adherence is supported through adequate clinical support, carbohydrate restriction can produce lasting metabolic benefits that go well beyond simple blood glucose control (Athinarayanan et al., 2019) [34].

Adherence is, in fact, one of the decisive elements for the success of any nutritional intervention. In this context, continuous glucose monitoring assumes a role that goes beyond simple blood glucose measurement, transforming into a true educational tool. In a randomized pilot study, patients undergoing a low-calorie, low carb diet and an exercise program achieved better results when they could visualize the trend of their blood glucose in real time. Although the reduction in HbA1c and body weight was similar in the two groups, real-time monitoring led to a constant trend toward a greater reduction in MAGE, CONGA, coefficient of variation, and standard deviation of blood glucose, as well as a decrease in pharmacological requirements. Immediate feedback allows the patient to observe the effect of different foods and physical activity on their blood glucose, increasing awareness and facilitating the maintenance of lifestyle changes (Taylor et al., 2019) [35].

Further confirmation also comes from populations characterized by severe impairment of insulin secretion. In over three hundred and fifty patients with insulin-deficient diabetes, a low carb diet simultaneously improved the coefficient of variation, MAGE, standard deviation, Time in Range, and exposure to postprandial glycemic excursions, also reducing insulin requirements by approximately seven units per day, the number of daily injections, and the use of hypoglycemic drugs.

The greatest benefits were observed in subjects with the lowest C-peptide levels, suggesting that residual β-cell function may represent a useful biomarker for personalizing the intensity of carbohydrate restriction. Also interesting is the observation that higher concentrations of ketone bodies were associated with lower glycemic variability, opening new perspectives on the possible mechanisms through which carbohydrate restriction contributes to metabolic stabilization (Song et al., 2026) [29].

Although conducted in type 1 diabetes, a multicenter crossover study provides important pathophysiological confirmation of the relationship between glycemic load and glycemic stability. By reducing carbohydrate intake from 50% to 30% of total energy, researchers observed a significant decrease in average blood glucose, an increase of approximately sixty-eight minutes per day in Time in Range, and a reduction of approximately eighty-five minutes in time spent in hyperglycemia, without any increase in hypoglycemia, clinically relevant ketonemia (amount of ketone bodies present in the blood), or episodes of ketoacidosis (excessive accumulation of ketones that causes acidification of the blood). Although this is a different population from that with type 2 diabetes, these results confirm that a moderate and well-planned reduction of carbohydrates is able to improve glycemic control while maintaining a high safety profile (Sterner Isaksson et al., 2023) [36].

Overall, all this evidence converges toward the same concept: the main strength of low carb diets lies not only in the reduction of average blood glucose, but above all in the ability to make the glycemic profile more stable throughout the day. The decrease in postprandial peaks, the improvement of β-cell function, the reduction of insulin requirements, the increase in time spent in the optimal glycemic range, and the possibility of personalizing the intervention based on the metabolic characteristics of the individual make carbohydrate restriction one of the most promising nutritional tools for the control of glycemic variability. However, the same evidence emphasizes how the success of this approach depends not only on the extent of carbohydrate restriction, but also on the overall quality of the diet, long-term sustainability, and the ability to adapt the nutritional strategy to the clinical characteristics of each patient.

The onset of type 2 diabetes is commonly attributed to insulin resistance, which is the progressive reduction in the ability of muscle, liver, and adipose tissue to respond to the action of insulin. Although this alteration represents one of the main pathophysiological elements of the disease, it is now clear that, on its own, it is not sufficient to cause diabetes. Many individuals live for years with marked insulin resistance while maintaining perfectly normal glycemia. This is possible thanks to the extraordinary adaptive capacity of pancreatic β-cells, which increase insulin secretion in proportion to the body's growing demand.

As long as this compensatory response remains efficient, glucose homeostasis is preserved. Type 2 diabetes only appears when this mechanism begins to fail progressively, making the amount of insulin secreted insufficient relative to the degree of insulin resistance present. In other words, insulin resistance creates the need for a greater amount of insulin, but it is the loss of the compensatory capacity of β-cells that transforms this metabolic imbalance into a full-blown disease (Del Prato et al., 2002) [37].

To understand how this process develops, it is necessary to analyze the normal functioning of insulin secretion. β-cells do not release insulin in a continuous and uniform manner, but follow a highly organized pattern that allows the body to adapt rapidly to changes in glycemia. During fasting, they produce a relatively constant basal secretion, sufficient to limit hepatic glucose production, control lipolysis (the process of mobilizing and breaking down fats stored in the body), and keep glycemia stable between meals. After food intake, however, secretion takes on a characteristic biphasic pattern. In the first few minutes, the insulin already stored in secretory granules is rapidly released: this represents the so-called first phase. Subsequently, a second phase comes into play, which is slower and more prolonged, supported by the synthesis and release of new insulin. Although both are essential for glycemic control, the first phase is of particularly significant physiological importance, as it represents the initial signal that coordinates the metabolic response of the entire organism (Del Prato et al., 2002) [37]; (Del Prato, 2003) [38].

Contrary to what one might imagine, the main target of the first phase of insulin secretion is not skeletal muscle, but the liver. The sudden increase in insulin concentration in the portal vein determines a rapid suppression of glycogenolysis and gluconeogenesis, almost immediately interrupting endogenous glucose production. This mechanism prevents the glucose released by the liver from adding to that absorbed by the intestine after a meal, thus limiting the extent of the postprandial glycemic excursion. Subsequently, the second phase of insulin secretion promotes the entry of glucose into peripheral tissues and contributes to progressively returning glycemia to basal values. The entire metabolic response therefore depends on a perfect temporal synchronization between the two phases of insulin secretion, rather than on the absolute amount of insulin produced (Del Prato, 2003) [38].

This observation has profoundly changed the understanding of the natural history of type 2 diabetes. Experimental evidence shows that the first defect in β-cell function does not consist of a reduction in the total amount of insulin secreted, but rather in the loss of its timeliness. In the initial stages of the disease, the first phase of insulin secretion progressively attenuates until it disappears, while the second phase is initially maintained and often even enhanced as a compensatory attempt. Insulin continues to be produced, but it is secreted with a delay that profoundly alters metabolic control.

In the meantime, the liver continues to release glucose just as the glucose from food enters the bloodstream, leading to a marked increase in postprandial glycemia. Only later does the pancreas react by increasing insulin secretion, but this delayed response can no longer prevent the initial glycemic peak and favors the establishment of persistent compensatory hyperinsulinemia (Del Prato et al., 2002) [37]; (Del Prato, 2003) [38].

This mechanism represents the origin of a true vicious cycle. The repeated increase in postprandial glycemia exposes β-cells to chronic glucotoxicity, characterized by increased oxidative stress and alterations in mitochondrial function (the ability of mitochondria to produce energy).

In parallel, the excess of free fatty acids promotes lipotoxicity, with intracellular accumulation of toxic lipid metabolites, local inflammation, and progressive apoptosis of β-cells. In this way, the pancreas gradually loses the ability to compensate for insulin resistance, further accelerating metabolic deterioration and favoring the progression toward full-blown diabetes (Del Prato et al., 2002) [37].

To evaluate the real compensatory capacity of β-cells, the Disposition Index was developed, a parameter that integrates early insulin secretion with insulin sensitivity. This index is now one of the best tools for estimating β-cell function, as it takes into account the fact that the amount of insulin required varies according to the degree of insulin resistance.

In a study conducted on adolescents with overweight and obesity, the Disposition Index measured by hyperglycemic clamp (a test that evaluates the pancreas's ability to secrete insulin) decreased progressively, moving from high values in normoglycemic subjects to intermediate values in young people with impaired glucose tolerance, down to extremely low values in patients with type 2 diabetes. In addition, the Oral Disposition Index, calculated using the simple oral glucose tolerance test, showed a close correlation with the clamp, demonstrating that even an examination easily performed in clinical practice can identify the loss of β-cell compensatory capacity with good accuracy (Sjaarda et al., 2012) [39].

The importance of identifying this deterioration early emerges with particular clarity from a study conducted on over nine thousand individuals at high risk for diabetes. The authors observed that a surprisingly high proportion of subjects classified as normoglycemic according to traditional diagnostic criteria already had a one-hour glycemia during the oral glucose tolerance test of 155 mg/dL or higher. Despite still-normal values of fasting glycemia, two-hour glycemia, and glycated hemoglobin, these individuals showed a marked impairment of β-cell function.

The early insulinogenic index was reduced by approximately 45%, while the Oral Disposition Index decreased by over 50%, indicating a substantial loss of the pancreas's ability to respond rapidly to the increase in glycemia. Conversely, insulin sensitivity appeared only moderately reduced, suggesting that the main pathophysiological defect in the early stages of the disease is not yet represented by insulin resistance, but by the progressive deterioration of early insulin secretion. These results indicate that glycemia measured one hour after an oral glucose load represents a much more sensitive biomarker than traditional diagnostic criteria in identifying the first alterations of β-cell function (Nie et al., 2025) [40].

When the compensatory capacity of β-cells begins to decrease, the first parameter to change is not necessarily mean glycemia, but its stability throughout the day. This aspect was demonstrated by a study that evaluated subjects distributed along the entire metabolic continuum, from normal glucose tolerance to full-blown diabetes, using continuous glucose monitoring.

As β-cell deterioration increased, mean glycemia, standard deviation, and all related indices grew progressively. In parallel, the Oral Disposition Index decreased constantly, highlighting a close inverse relationship between the secretory capacity of the pancreas and glycemic stability. In other words, glycemic variability does not simply represent a consequence of hyperglycemia, but constitutes a direct manifestation of the progressive inability of β-cells to modulate insulin secretion rapidly in response to meals (Chen et al., 2013) [41].

This interpretation was reinforced by the study by Kohnert, which simultaneously analyzed postprandial β-cell function and glycemic variability indices obtained through continuous monitoring. The results demonstrated that the ability of β-cells to respond to a meal represents the main determinant of the MAGE parameter, explaining on its own almost half of the variability observed among patients. Conversely, factors commonly considered determinants, such as age, duration of diabetes, body mass index, glycated hemoglobin, insulin sensitivity, and even carbohydrate intake, showed a decidedly lower influence. The study also identified a critical threshold of β-cell function beyond which glycemic variability increases rapidly, suggesting that there is a pathophysiological turning point at which the pancreas definitively loses the ability to maintain stable glycemic control (Kohnert et al., 2009) [42].

Alterations in β-cell function can be observed long before the appearance of full-blown diabetes and even during adolescence. In young people with obesity, continuous glucose monitoring has shown that an increase in mean glycemia, standard deviation, and the coefficient of variation is associated with a higher intrahepatic fat content, a reduction in both first-phase and basal insulin secretion, a decrease in insulin sensitivity, and an increase in plasma lactate concentrations, an indicator of early alteration in energy metabolism.

Glycemic oscillations are also closely correlated with postprandial hyperglycemia, suggesting that the progressive loss of the early insulin response simultaneously favors the accumulation of liver fat, the worsening of insulin resistance, and the functional exhaustion of β-cells. These observations indicate that the pancreas, liver, and energy metabolism participate from the very early stages in a single systemic alteration that precedes the appearance of clinically manifest diabetes by many years (Barbieri et al., 2025) [43].

The same pathophysiological pattern also emerges in gestational diabetes. In these patients, the increase in the MAGE parameter is closely associated with the reduction of first-phase insulin secretion and only to a lesser extent with insulin resistance. This finding confirms that the loss of β-cell function represents the main determinant of glycemic instability even in a condition characterized by profound hormonal and metabolic changes of pregnancy, further reinforcing the central role of β-cells in maintaining glucose homeostasis (Su et al., 2013) [44].

Finally, definitive confirmation comes from the largest systematic review published to date on the subject, which analyzed seventy-one studies conducted on non-diabetic subjects undergoing continuous glucose monitoring. The authors observed that glycemic variability increases already in the very early stages of prediabetes, when mean glycemia and glycated hemoglobin may still be within normal limits. The most interesting aspect is that glycemic oscillations show a much stronger relationship with the deterioration of β-cell function than with obesity, body mass index, visceral adiposity, hepatic steatosis, insulin sensitivity, or other traditional cardiometabolic risk factors.

Furthermore, high glycemic variability is associated with greater coronary calcification, more rapid progression of atherosclerosis, and an increased risk of developing type 2 diabetes and cardiovascular events. This evidence suggests that the loss of β-cell function represents one of the earliest pathophysiological events in the entire natural history of diabetes and that continuous glucose monitoring can identify apparently healthy subjects who are already characterized by an increased cardiometabolic risk (Hjort et al., 2024) [45].

Overall, these studies change our understanding of type 2 diabetes. The disease does not begin when glycemia increases stably, but when the pancreas progressively loses the ability to respond rapidly to food stimuli. The disappearance of the first phase of insulin secretion represents the first measurable pathophysiological event; this is followed by the reduction of the compensatory capacity of β-cells, the increase in glycemic variability, and the establishment of glucotoxicity and lipotoxicity that further accelerate pancreatic deterioration.

To maintain relatively normal glycemia, β-cells initially try to compensate by increasing late insulin secretion, establishing a condition of compensatory hyperinsulinemia. This response, although initially effective, causes profound changes in insulin metabolism and represents the starting point of the subsequent pathophysiological step: the progressive reduction of insulin clearance, a topic that will be explored in the next paragraph.

In the previous paragraph, we saw how the progressive loss of β-cell function represents one of the first pathophysiological events in type 2 diabetes. For many years, it was believed that to compensate for this alteration and the growing insulin resistance, the pancreas was simply forced to produce ever-increasing amounts of insulin. Today, we know that this interpretation is only partially correct. The concentration of insulin in the blood, in fact, does not depend exclusively on how much is secreted, but also on how rapidly the body manages to eliminate it. This process is called insulin clearance and represents one of the main regulators of metabolic homeostasis. One can imagine insulin like water entering a tank: the water level certainly depends on the amount coming out of the tap, but also on the speed at which it is drained through the outlet. Similarly, high concentrations of insulin can result not only from increased pancreatic secretion but also from a reduced capacity of the body to remove the hormone from circulation (Koh et al., 2022) [46]; (Bergman et al., 2022) [47].

Under physiological conditions, the main organ responsible for eliminating insulin is the liver. After being secreted by pancreatic β-cells, insulin immediately reaches the portal circulation (the pathway that carries nutrient-rich blood to the liver), where approximately 50-80% is removed during the first pass through the liver. Only the remaining portion reaches the systemic circulation and is subsequently used primarily by skeletal muscle, the liver, and the kidneys. This means that the availability of insulin receptors simultaneously determines both the ability of tissues to respond to insulin and the speed at which it is cleared from the blood. Insulin sensitivity and clearance are therefore closely interdependent, representing two sides of the same biological process (Koh et al., 2022) [46]; (Bergman et al., 2022) [47].

For many decades, the reduction in insulin clearance was interpreted as a simple compensatory mechanism for insulin resistance. According to this view, when muscle, liver, and adipose tissue become less sensitive to the action of insulin, the body voluntarily reduces its elimination, maintaining higher plasma concentrations to try to preserve glycemic control. Although this explanation appears intuitive, numerous recent findings suggest a different interpretation. Several studies indicate that the reduction in clearance may represent a very early event, in some cases even preceding the development of insulin resistance. According to this new pathophysiological hypothesis, reduced hepatic elimination of insulin would lead to chronic exposure of peripheral tissues to high concentrations of the hormone. Over time, this persistent hyperinsulinemia would favor the down-regulation of insulin receptors, the progressive onset of insulin resistance, and, subsequently, the functional overload of pancreatic β-cells. In this perspective, reduced clearance would not represent the consequence of insulin resistance, but rather one of its possible initial determinants (Bergman et al., 2022) [47]; (Bergman et al., 2019) [48].

Supporting this hypothesis is the observation that insulin clearance shows enormous variability even among metabolically healthy individuals. Studies conducted using sophisticated clamp techniques (tests that allow for the precise study of the body's response to glucose and insulin) have shown that the portion of insulin eliminated by the liver can range from 20% to over 80%, even in the absence of any metabolic pathology. This characteristic is surprisingly stable in the individual and seems to be influenced by both genetic and environmental factors. Diet, in fact, is also capable of rapidly modifying the liver's ability to degrade insulin. These modifications increase the amount of insulin that reaches the systemic circulation without requiring a proportional increase in pancreatic secretion, thus contributing to the development of chronic hyperinsulinemia (Bergman et al., 2019) [48].

One of the most elegant demonstrations of the role of insulin clearance comes from a longitudinal study conducted in an animal model of insulin resistance induced by a moderately high-fat diet. After just one week of a high-fat diet, insulin sensitivity decreased by approximately 35%, while glycemia remained surprisingly normal. Initially, the pancreas responded by increasing first-phase insulin secretion; however, detailed analysis of the data showed that this adaptation explained only part of the observed hyperinsulinemia. Approximately half of the increase in peripheral insulin concentrations derived from a reduction in hepatic clearance, which decreased by about 50%. In other words, the liver allowed a much larger portion of insulin to pass into the systemic circulation, allowing for the maintenance of normoglycemia without immediately forcing β-cells into an excessive increase in hormone production. This refined compensation mechanism therefore allows the body to face the early stages of insulin resistance while temporarily preserving pancreatic function (Mittelman et al., 2000) [49].

This interpretation was further confirmed by a subsequent study by the same research group, which longitudinally followed the evolution of insulin resistance in dogs fed a moderately high-fat diet. After an initial phase characterized by an increase in insulin secretion, the β-cells progressively returned to secretory levels close to baseline values, while hyperinsulinemia persisted due to a marked reduction in hepatic insulin extraction during the first pass. The authors defined this phenomenon as a sort of "β-cell rest": by reducing hepatic clearance, the liver maintains high peripheral concentrations of insulin without forcing the pancreas to continuously increase secretion. This is likely a second line of physiological defense that, at least in the initial stages, contributes to preserving β-cell function. However, if this adaptation persists over time, chronic exposure of tissues to high concentrations of insulin can favor a further worsening of insulin resistance, establishing a vicious cycle destined to culminate in the development of type 2 diabetes (Kim et al., 2007) [50].

The pathophysiological importance of reduced insulin clearance also emerges with particular clarity in humans. A study conducted on children and adolescents with obesity showed that the presence of metabolic dysfunction-associated steatotic liver disease (MASLD) is accompanied by a marked reduction in the liver's ability to eliminate insulin even in the early stages of the disease. Compared to obese peers without liver involvement, subjects with MASLD showed significantly lower hepatic clearance, accompanied by increased insulin resistance and greater compensatory activity of pancreatic β-cells. In this phase, the pancreas still manages to maintain relatively effective glycemic control thanks to the combination of two adaptations: on one hand, it increases insulin secretion, and on the other, the liver eliminates a smaller portion of the hormone, allowing higher concentrations to remain in the bloodstream for longer. However, this compensatory strategy has limited effectiveness. In subjects who had already developed prediabetes, in fact, the reduction in clearance was even more marked, while the first signs of deterioration in β-cell function began to appear, indicating that the compensation cannot be maintained indefinitely (Jiang et al., 2024) [51].

The study also provided a convincing biological explanation for this phenomenon. In the murine model (a study conducted on mice) used by the authors, the accumulation of fat in the liver led to a reduction of approximately 50% in the expression of enzymes responsible for insulin degradation. Although this adaptation initially helps to limit the increase in glycemia, in the long term, it entails a high biological cost: chronic exposure to high concentrations of insulin stimulates hepatic lipogenesis, promotes fat accumulation in the liver, aggravates insulin resistance, and accelerates the progressive functional exhaustion of β-cells. Thus, a vicious cycle is established in which steatosis reduces insulin clearance, and reduced clearance, in turn, fuels the progression of steatosis and metabolic alterations (Jiang et al., 2024) [51].

Studies conducted in apparently healthy populations also confirm that insulin clearance represents an important determinant of plasma hormone concentrations. In an accurate physiological analysis performed on non-diabetic Black and White women, it was observed that women of African descent had significantly higher insulin concentrations after an oral glucose load, during a mixed meal, and even after intravenous glucose administration. Contrary to what was believed in the past, this difference was not explained by greater total insulin secretion nor by a different incretin response mediated by GLP-1 (a hormone produced by the gut after food intake). C-peptide concentrations, an index of pancreatic secretion, were in fact comparable between the groups, as were GLP-1 levels. Hyperinsulinemia instead derived mainly from the combination of reduced hepatic insulin clearance and a greater capacity of β-cells to rapidly release already available insulin granules. These results demonstrate that high insulin concentrations can develop even in the absence of an increase in pancreatic production, simply because the body eliminates the hormone more slowly (Chung et al., 2019) [52].

The dynamic nature of insulin clearance is further demonstrated by nutritional studies. In healthy, non-obese Japanese men, an isocaloric diet (a diet with calories adequate to the person's energy needs) low in carbohydrates and high in fat, followed for only three days, was able to significantly increase the metabolic clearance rate of insulin. This increase was accompanied by a reduction in plasma hormone concentrations, without evidence of a real worsening of insulin sensitivity after correcting the data for circulating insulin levels. The authors interpreted this phenomenon as a physiological adaptation to reduced glucose availability: when carbohydrate intake decreases, the body increases insulin elimination to favor the use of fatty acids as the main energy source. The study also highlighted notable individual variability in the response to the diet, suggesting that genetic and metabolic factors can profoundly influence the regulation of insulin clearance. These observations reinforce the concept that clearance does not represent a static process, but a dynamic parameter, modulated by the nutritional status and metabolic needs of the body (Suzuki et al., 2019) [53].

The most solid evidence on the role of reduced clearance in the pathogenesis of diabetes, however, comes from prospective studies. In a cohort of 570 non-diabetic Native American adults, followed for a median of nearly eight years, the reduction in the metabolic clearance rate of insulin proved to be an independent predictor of the development of type 2 diabetes. Subjects with a lower capacity to eliminate insulin showed greater adiposity, higher fasting insulin concentrations, a more intense compensatory insulin response, and reduced insulin sensitivity. However, even after statistically correcting the results for age, sex, adiposity, glycemia, insulin secretion, and insulin sensitivity, clearance maintained significant prognostic value: for every one-standard-deviation increase in insulin clearance, the risk of developing diabetes decreased by approximately 23%. This data suggests that the reduction in clearance does not simply represent a consequence of insulin resistance, but constitutes an autonomous pathophysiological factor capable of favoring the progression toward type 2 diabetes (Shah et al., 2021) [54].

Overall, these findings profoundly modify the traditional conception of hyperinsulinemia. For many years, it was believed that high insulin concentrations were exclusively the result of increased secretion by pancreatic β-cells. Today, it appears evident that the amount of insulin present in the circulation depends on the balance between two closely coordinated processes: pancreatic secretion and its elimination by the liver and other tissues. In the early stages of insulin resistance, the reduction in clearance likely represents a refined adaptation mechanism aimed at preserving glycemic control and reducing the secretory load imposed on β-cells. However, when this condition persists for years, chronic hyperinsulinemia fuels a vicious cycle characterized by progressive worsening of insulin resistance, increased hepatic steatosis, further reduction in clearance, and increasing metabolic overload of the pancreas. In this perspective, the reduction in insulin clearance constitutes not only a consequence of metabolic alterations but one of the main events linking β-cell dysfunction to the development of chronic hyperinsulinemia, a theme that will represent the focus of the next paragraph dedicated specifically to hyperinsulinemia (Fosam et al., 2020) [55].

In the previous paragraph, we saw how the reduction in insulin clearance is one of the main mechanisms through which the body modulates the amount of insulin present in circulation. This observation naturally leads to a fundamental question: why does hyperinsulinemia appear so early in type 2 diabetes? For many years, it was believed that high insulin concentrations were simply the consequence of insulin resistance. According to this model, when muscles, liver, and adipose tissue respond less effectively to the hormone, the pancreas progressively increases insulin secretion to keep glycemia stable. Although this compensatory mechanism truly exists, it is now evident that it represents only part of the story. Plasma insulin concentrations are, in fact, the result of a dynamic balance between two distinct but closely integrated processes: the amount of insulin secreted by pancreatic β-cells and the speed at which the liver and other tissues eliminate it from circulation. Hyperinsulinemia can therefore develop both because the pancreas produces more insulin and because the body degrades it more slowly. Understanding this distinction is fundamental, as it profoundly changes our interpretation of the pathophysiology of type 2 diabetes (Goodarzi et al., 2011) [56]; (Bergman et al., 2019) [48].

For a long time, fasting insulin has been used as a simple indicator of insulin resistance. However, an important analysis conducted on 470 Hispanic-American subjects demonstrated that this interpretation is reductive. Through the euglycemic-hyperinsulinemic clamp, the oral glucose tolerance test, and multivariate statistical models, the authors highlighted that the main determinant of basal insulinemia is not insulin sensitivity, but rather the metabolic clearance of insulin. The reduction in the speed at which the body eliminates the hormone showed a stronger association with insulinemia than insulin resistance itself, body mass index, waist circumference, and even fasting glycemia. This means that two individuals with identical insulin values can present completely different pathophysiological mechanisms: one may produce a lot of insulin, while the other may simply eliminate it more slowly. Consequently, interpreting hyperinsulinemia exclusively as an expression of pancreatic hypersecretion risks masking the determining role played by the liver in regulating insulin homeostasis (Goodarzi et al., 2011) [56].

Obesity also contributes to hyperinsulinemia through more complex mechanisms than previously thought. A study that simultaneously evaluated insulin secretion, β-cell function, insulin sensitivity, and insulin clearance in over one hundred subjects with different glycemic status demonstrated that the increase in adipose mass exerts a direct effect on pancreatic β-cells, increasing insulin secretion independently of the degree of insulin resistance. In parallel, in obese subjects, insulin clearance was reduced by approximately 20-30% compared to normal-weight subjects, further amplifying the plasma concentrations of the hormone. The combination of increased secretion and decreased elimination therefore allows for the maintenance of normoglycemia despite a moderate reduction in insulin sensitivity. However, this adaptation has a limited duration. When β-cell function begins to progressively deteriorate, insulin secretion can no longer compensate for the existing insulin resistance, and first impaired fasting glycemia and subsequently type 2 diabetes appear. The study therefore suggests that the hyperinsulinemia of obesity represents the combined product of pancreatic hypersecretion and reduced clearance, while the progression toward the disease is determined primarily by the progressive functional exhaustion of β-cells (Mittendorfer et al., 2022) [57].

Hyperinsulinemia can appear surprisingly early, even before the subject becomes obese. This possibility was demonstrated in a prospective study in which healthy young men underwent a modest, controlled weight gain, while still remaining within the normal range of body mass index. After about four months, the participants already showed a significant increase in basal insulin, C-peptide, HOMA-IR, and fasting glycemia. The most interesting aspect, however, concerned the mechanism responsible for this hyperinsulinemia. Through C-peptide modeling, the authors observed that insulin secretion increased only modestly, while insulin clearance decreased to a much greater extent. After a meal, during the OGTT, and even in the hyperglycemic clamp, insulin concentrations were significantly higher without any significant increase in pancreatic secretion. The increase in circulating insulin was therefore explained almost entirely by the reduced hepatic elimination of the hormone. In parallel, a significant reduction in insulin sensitivity already appeared, demonstrating that small increases in fat mass are enough to quickly trigger a vicious cycle characterized by reduced clearance, hyperinsulinemia, and insulin resistance (Erdmann et al., 2008) [58].

A further element of great interest concerns the role of hyperglycemia itself in fueling this process. In an experimental study conducted on metabolically healthy subjects, mild hyperglycemia maintained for only 72 hours, with average values approximately 50 mg/dL above normal, was sufficient to simultaneously cause a marked increase in circulating insulin, a significant reduction in β-cell function, a decrease in insulin sensitivity, and a consistent reduction in the metabolic clearance of insulin. Although insulin secretion initially appeared increased, the calculation of the disposition index showed that the β-cells were actually already losing efficiency. The greater share of hyperinsulinemia derived, in fact, from reduced insulin elimination rather than from true pancreatic hyperactivity. The authors therefore propose a pathophysiological model in which even modest hyperglycemia is able to quickly establish a vicious cycle: the increase in glucose reduces insulin clearance, hyperinsulinemia promotes insulin resistance, and the latter forces β-cells to further increase their workload, accelerating their functional deterioration (Merovci et al., 2021) [59].

The origin of hyperinsulinemia, however, is not identical in all individuals. Studies conducted in different populations have shown that the body can adopt different compensatory strategies to maintain glycemic homeostasis. Important research conducted on Caucasian, African-American, and Hispanic children highlighted that, in the face of a similar reduction in insulin sensitivity, African-American children developed significantly higher plasma insulin concentrations even without a corresponding increase in pancreatic secretion. The simultaneous analysis of insulin and C-peptide, in fact, demonstrated that the main mechanism responsible was a marked reduction in hepatic insulin extraction. In Hispanic children, by contrast, compensation occurred primarily through an increase in second-phase insulin secretion, while clearance remained relatively more preserved. This study represents one of the first experimental demonstrations that hyperinsulinemia does not constitute a uniform phenomenon, but can derive from different physiological adaptations in which the reduction of hepatic clearance plays a role as important as pancreatic secretion (Goran et al., 2002) [60].

These differences persist even in later ages and become even more evident during the progression toward diabetes. A longitudinal study conducted on normal-weight and obese adolescents demonstrated that the reduction in hepatic insulin clearance already appears in obese, insulin-resistant youth with normal glucose tolerance, therefore before the onset of hyperglycemia. Even more importantly, the liver's ability to eliminate insulin, measured at the beginning of the study, proved to be an independent predictor of the subsequent decline in β-cell function over the following two years. Subjects characterized by lower hepatic extraction, in fact, showed a significantly faster deterioration of pancreatic function, regardless of age, sex, obesity, and degree of insulin resistance. The liver therefore emerges not only as an organ responsible for insulin degradation but as a true regulator of the endocrine balance of the entire organism. When its ability to eliminate the hormone decreases, the pancreas and peripheral tissues are exposed for years to high concentrations of insulin, favoring the progressive functional exhaustion of β-cells (Galderisi et al., 2019) [61].

These observations have recently been confirmed in a larger population of young people with obesity distributed along the entire spectrum of glucose tolerance. In subjects with type 2 diabetes, the metabolic clearance of insulin was reduced by approximately 45% compared to normoglycemic peers, despite a marked reduction in first-phase insulin secretion and a collapse of the Disposition Index. In other words, fasting hyperinsulinemia persisted even in the presence of already severely compromised β-cell function. This apparent paradox was explained precisely by the reduced elimination of insulin, which allowed the hormone to remain in circulation longer. The study also highlighted that, in young African-Americans with dysglycemia, clearance was further reduced compared to white peers, associating with greater insulin resistance and a more intense residual secretory response. These results reinforce the concept that the reduction in clearance represents a fundamental compensatory mechanism during the progression of juvenile diabetes, but at the same time contributes to fueling the vicious cycle of chronic hyperinsulinemia (Cho et al., 2025) [62].

The importance of insulin clearance also emerges in the context of nutritional interventions. In obese adolescents subjected for twelve weeks to an isocaloric diet characterized by a low ratio of omega-6 to omega-3 polyunsaturated fatty acids, a significant increase in insulin clearance was observed both while fasting and after an oral glucose load. This improvement was accompanied by a marked reduction in hyperinsulinemia, even in the absence of changes in body weight, glycemia, and insulin secretion. The benefits were particularly evident in subjects who initially presented the most compromised metabolic profile, characterized by greater hyperinsulinemia, reduced insulin sensitivity, and lower clearance. The study therefore demonstrates that even the quality of dietary fats can profoundly influence insulin metabolism independently of weight loss, highlighting how hyperinsulinemia is not an inevitable phenomenon but a condition potentially reversible through appropriate nutritional interventions (Tricò et al., 2022) [63].

Taken together, these findings profoundly modify the traditional conception of hyperinsulinemia. The excess of circulating insulin does not simply represent a compensatory response to insulin resistance, but the result of the dynamic interaction between pancreatic secretion, hepatic clearance, and the sensitivity of peripheral tissues to the hormone. In the initial phases, this condition constitutes a physiological adaptation aimed at preserving normoglycemia; over time, however, chronic exposure to high concentrations of insulin further favors insulin resistance, stimulates hepatic lipogenesis, accelerates the deterioration of β-cells, and contributes to the progression toward type 2 diabetes. Hyperinsulinemia therefore assumes a dual role: initially, it represents a compensatory mechanism, but subsequently, it becomes an important pathogenetic factor itself. This new interpretation constitutes the pathophysiological prerequisite for understanding the subsequent step of the metabolic cascade, in which the chronic excess of insulin activates specific intracellular pathways, including that of mTOR, which is destined to play a central role in the worsening of insulin resistance and the progression of type 2 diabetes (Fosam et al., 2020) [55].

Over the last few decades, insulin resistance has been considered the primary pathophysiological event at the origin of type 2 diabetes. According to the traditional model, skeletal muscle, the liver, and adipose tissue progressively become less sensitive to the action of insulin, forcing the pancreas to increase the secretion of the hormone to maintain normal glycemia. When β-cells can no longer sustain this increased demand, prediabetes appears first, followed by overt diabetes. Although this model has been the dominant paradigm for many years, numerous scientific findings that have emerged in recent years suggest a more complex pathophysiological sequence. Indeed, increasing data indicate that chronic hyperinsulinemia, favored by increased pancreatic secretion and reduced hepatic insulin clearance, may precede the onset of insulin resistance and contribute directly to its development. In this perspective, insulin resistance would not represent the initial event of the disease, but rather a consequence of the prolonged exposure of tissues to high concentrations of insulin (Janssen, 2024) [64].

This interpretation finds solid foundations in the physiology of insulin action. When insulin binds to its receptor present on the cell membrane, it activates a complex cascade of intracellular signals that promotes the entry of glucose into cells, stimulates glycogen synthesis (the body's glucose reserve), promotes lipogenesis (the production and accumulation of fats by the body), and inhibits hepatic glucose production. Under physiological conditions, these signals are activated only for limited periods after meals, allowing receptors to quickly recover their sensitivity during fasting hours.

If, however, insulin remains chronically elevated for most of the day, tissues are subjected to continuous stimulation. As occurs with many other biological systems, the insulin receptor also tends to progressively reduce its response to protect the cell from excessive activation. This results in a decrease in insulin sensitivity that initially affects the liver, subsequently the skeletal muscle, and finally the adipose tissue. Insulin resistance therefore takes on the characteristics of a biological adaptation to a persistent hormonal stimulus rather than a primary defect that appeared spontaneously (Bkaily et al., 2025) [65].

One of the most important reevaluations of this concept was proposed as early as 2012 by researcher Barbara Corkey in the historic Banting Lecture (a prestigious scientific conference dedicated to diabetes research), in which she questioned the traditional paradigm according to which hyperinsulinemia is exclusively a compensatory response to insulin resistance. By analyzing numerous experimental and clinical studies, Corkey observed that in many populations, the increase in insulin concentrations preceded the alteration of insulin sensitivity by several years. According to this hypothesis, numerous environmental factors typical of the modern lifestyle, including caloric excess, highly processed foods, continuous stimulation of β-cells, alterations in intracellular metabolic signals, and reduced hepatic clearance, lead to persistent insulin hypersecretion.

Chronic exposure to high concentrations of the hormone subsequently induces a progressive desensitization of target tissues, similar to what is observed for other continuously stimulated endocrine systems. Insulin resistance would therefore represent a secondary adaptation aimed at limiting the excessive biological action of insulin rather than its initial causal event (Corkey, 2012) [66].

Clinical observations also support this new interpretation. In numerous apparently healthy subjects, hyperinsulinemia can be documented many years before the appearance of hyperglycemia and often precedes any alteration detectable by common tests for assessing insulin sensitivity. Longitudinal studies have shown that individuals with high fasting insulin concentrations have a significantly higher risk of subsequently developing insulin resistance, hepatic steatosis, and type 2 diabetes, regardless of initial glycemia.

This suggests that insulin is not just a biomarker of the disease, but a possible factor actively involved in its progression. This view also explains why interventions capable of early reduction of hyperinsulinemia, such as caloric restriction, improving the quality of diet, and reducing ectopic fat (accumulation of fat outside normal adipose tissue), can lead to a significant recovery of insulin sensitivity even before substantially modifying other metabolic parameters (Thomas et al., 2019) [67].

The central role of hyperinsulinemia becomes even more evident when analyzing the effects of chronic energy excess. A caloric intake consistently higher than requirements keeps β-cells in a state of continuous secretory stimulation and, in parallel, progressively reduces hepatic insulin clearance. The persistent increase in circulating concentrations of the hormone profoundly modifies lipid metabolism: it stimulates hepatic lipogenesis, inhibits lipolysis of adipose tissue (the process that mobilizes fat reserves), and promotes the accumulation of triglycerides in the liver, pancreas, and skeletal muscle. When the capacity of subcutaneous adipose tissue to store energy is exceeded, the lipid surplus is progressively deposited in metabolically active organs, leading to so-called ectopic fat.

Lipid infiltration of the liver further reduces hepatic insulin sensitivity and increases endogenous glucose production, while that of the pancreas progressively compromises β-cell function. A vicious cycle is thus established in which hyperinsulinemia, ectopic fat accumulation, and insulin resistance feed each other, accelerating the progression toward type 2 diabetes (Janssen, 2024) [64].

Further confirmation comes from studies that have analyzed the role of specific proteins involved in cellular homeostasis. Of particular interest is the BAG3 protein, which is involved in autophagy processes (the cell's internal cleaning and recycling system) and in maintaining the functionality of metabolically active tissues. In animal models fed a high-fat diet, reduced BAG3 expression is associated with a rapid worsening of insulin sensitivity, an increase in hepatic lipid accumulation, and greater activation of intracellular inflammatory pathways. Conversely, maintaining adequate levels of BAG3 preserves insulin receptor functionality, improves glucose metabolism, and limits the onset of insulin resistance, suggesting that the loss of normal cellular protection mechanisms is one of the factors that favor the progression of metabolic alterations induced by hyperinsulinemia (Damiani et al., 2023) [68].

The liver also occupies a central position in this network of events. In addition to being the main organ responsible for insulin clearance, it is the first district exposed to high concentrations of the hormone coming from the portal vein. Persistent hyperinsulinemia stimulates de novo lipogenesis, increases triglyceride synthesis, and promotes the progressive appearance of hepatic steatosis. The accumulation of lipids in hepatocytes (the main cells of the liver) compromises the ability of insulin to suppress hepatic glucose production, leading to progressive hepatic insulin resistance. In parallel, steatosis further reduces insulin clearance, allowing an even greater portion of the hormone to reach the systemic circulation. The liver therefore becomes one of the main amplifiers of the vicious cycle that links hyperinsulinemia and insulin resistance (Bergman et al., 2019) [48].

This pathophysiological model also helps explain one of the most surprising aspects that has emerged in recent years: the possibility of achieving remission of type 2 diabetes. If insulin resistance were the result of permanent cellular damage, short-term nutritional interventions should not produce significant improvements. On the contrary, numerous studies show that a drastic caloric restriction leads to a rapid reduction of liver fat within just a few days, followed by a progressive decrease in pancreatic fat and the recovery of β-cell function.

In subjects with diabetes diagnosed only a few years prior, the loss of approximately 15 kg of body weight can return glycemia to normal values without the need for pharmacological therapy, indicating that a significant part of insulin resistance is reversible when the metabolic stimuli that sustain it are removed. This observation further reinforces the hypothesis that insulin resistance represents a functional adaptation to chronic hyperinsulinemia rather than an irreversible process (Janssen, 2024) [64].

Further confirmation comes from studies on bariatric surgery. The metabolic improvements observed after gastric bypass or vertical gastrectomy (a procedure that reduces the size of the stomach) appear, in fact, long before substantial weight loss occurs. Already in the first days after surgery, a marked reduction in insulinemia, an improvement in hepatic insulin sensitivity, and a significant decrease in hepatic glucose production are observed. This rapid recovery suggests that the reduction of energy intake and the consequent decrease in hyperinsulinemia are the main drivers of the initial metabolic improvement, while subsequent weight loss helps to consolidate the effects over time. Once again, hyperinsulinemia appears as a central element in the pathophysiology of the disease, and its reduction constitutes one of the first steps toward the recovery of insulin sensitivity (Pories and Dohm, 2012) [69].

Taken together, these findings outline a pathophysiological model different from the traditional one. Insulin resistance no longer represents exclusively the starting point of type 2 diabetes, but one of the intermediate stages of a process that begins much earlier with chronic energy excess, reduced hepatic insulin clearance, and the appearance of persistent hyperinsulinemia. Prolonged exposure of tissues to high concentrations of insulin induces progressive receptor desensitization, promotes ectopic fat accumulation, alters liver function, and accelerates the deterioration of pancreatic β-cells. Insulin resistance thus becomes an integral part of a vicious cycle in which each alteration fuels the next, making the restoration of metabolic homeostasis progressively more difficult (Zhang et al., 2021) [70].

This new interpretation has important clinical implications. If hyperinsulinemia represents one of the initial events leading to the development of insulin resistance, therapeutic interventions should aim not only at reducing glycemia but also at normalizing circulating insulin concentrations through nutritional strategies, improving hepatic clearance, reducing ectopic fat, and recovering tissue sensitivity to the hormone. This perspective also allows us to understand the subsequent pathophysiological step of the metabolic cascade: chronic exposure to high concentrations of insulin, in fact, determines a persistent activation of the PI3K/AKT/mTOR signaling pathway, which represents one of the main molecular mechanisms through which hyperinsulinemia contributes to the progression of insulin resistance and type 2 diabetes. This will be the topic of the next paragraph dedicated to the relationship between hyperinsulinemia and mTOR (Janssen, 2021) [71].

In the previous paragraph, we saw how chronic hyperinsulinemia is not simply a compensatory response to insulin resistance, but can directly contribute to its onset through the progressive desensitization of target tissues. To understand how this process occurs, it is necessary to delve into one of the main molecular nodes of insulin signaling: the mTOR (mammalian Target of Rapamycin) complex. Under physiological conditions, mTOR activation is an indispensable part of the insulin response. After the hormone binds to its receptor, the PI3K/AKT cascade transmits the signal to mTOR, which coordinates protein synthesis, cell growth, energy metabolism, and nutrient utilization. This system is essential for allowing cells to adapt to the abundance of energy available after a meal. However, as with many biological mechanisms, what is beneficial when transient can become harmful if it remains constantly active. When hyperinsulinemia keeps the PI3K/AKT/mTOR pathway continuously stimulated, a regulatory circuit designed to limit the excesses of the insulin response progressively transforms into one of the main promoters of insulin resistance (Tremblay et al., 2007) [72].

One of the most surprising aspects emerging from experimental studies is the speed with which this system comes into play. In murine and human adipocytes, insulin activates mTORC1, the main effector of the S6K1 kinase, within a few minutes. This rapid activation demonstrates that mTOR does not only intervene in chronic metabolic changes, but participates from the very first stages of insulin signal transduction. At the same time, however, this same mTOR activation triggers a sophisticated negative feedback circuit intended to prevent excessive cell stimulation. Through the phosphorylation (the addition of a phosphate that modifies a molecule's activity) of specific residues of the IRS-1 protein, the main adapter of insulin signaling, mTOR progressively reduces the intensity of the signal directed toward AKT. Under physiological conditions, this represents a refined self-regulation system, analogous to a thermostat that automatically turns off the heating when the desired temperature is reached. However, when insulin remains high for many hours or years, this same mechanism continues to be stimulated without interruption, transforming from a physiological regulator into a promoter of insulin resistance (Tremblay et al., 2007) [72]; (Gual et al., 2003) [73].

The study by Gual and colleagues has clarified the molecular steps of this process with precision. The authors demonstrated that insulin induces the phosphorylation of different serine residues present on IRS-1, each characterized by a specific regulatory function. When exposure to insulin is prolonged, however, control progressively shifts to the PI3K/AKT/mTOR pathway. Rapamycin, a potent mTOR inhibitor, almost completely abolished IRS-1 phosphorylation in both adipocytes and in the muscle and adipose tissue of animals treated with insulin, demonstrating that mTOR represents the main regulator of this negative feedback. From a functional point of view, serine phosphorylation modifies the ability of IRS-1 to interact with the insulin receptor and PI3K, progressively attenuating the propagation of the signal toward AKT. In other words, the longer the mTOR pathway remains active, the greater the cell's difficulty in responding to subsequent insulin stimulations (Gual et al., 2003) [73].

The physiological importance of this circuit emerges even more clearly by observing the effects of pharmacological mTOR inhibition. In adipocytes treated with rapamycin, the ability of insulin to promote glucose uptake increased by approximately 45%, accompanied by a marked amplification of AKT activation. Surprisingly, this improvement did not depend on an increase in the translocation of GLUT4 or GLUT1 transporters to the cell membrane, nor on an increase in the activity of PI3K associated with IRS-1. Instead, the benefit derived from the preservation of IRS-1 functionality. These results demonstrate that mTOR acts as an authentic physiological "brake" on insulin signaling; however, when the brake remains constantly applied, as occurs in chronic hyperinsulinemia, the cell's ability to use glucose progressively decreases, favoring the development of insulin resistance (Tremblay et al., 2007) [72].

The consequences of chronic hyperactivation of the PI3K/AKT/mTOR pathway are not limited to the simple onset of insulin resistance, but also affect the cells' gene expression program. One of the most convincing demonstrations comes from an in-depth transcriptomic analysis of human skeletal muscle integrated with an experimental model of chronic hyperinsulinemia. The authors observed that as fasting insulin concentrations increase, the expression of the INSR gene, which encodes the insulin receptor, progressively decreases. Exposing myotubes (cellular structures that give rise to mature muscle fibers) to high insulin concentrations for just sixteen hours resulted in a marked reduction of INSR mRNA, a decrease in the total amount of the receptor, and a lower presence of it on the cell membrane. At the same time, AKT phosphorylation was attenuated during subsequent insulin stimulation, demonstrating that the cell had developed a true functional desensitization (Cen et al., 2022) [74].

Chronic activation of the PI3K/AKT/mTOR pathway produces even deeper consequences when it affects the liver. A recent study has shown that prolonged exposure of human hepatocytes to high insulin concentrations is sufficient to activate an authentic cellular senescence program. Through the use of specific pharmacological inhibitors, the authors demonstrated that this phenomenon depends directly on the activation of the insulin receptor → PI3K → AKT → mTOR cascade, since blocking any of these steps almost completely prevents the appearance of senescence markers.

Definitive confirmation comes from mice lacking the insulin receptor exclusively in the liver: despite presenting marked systemic hyperinsulinemia, their liver does not develop signs of senescence, demonstrating that it is not circulating insulin itself that causes the damage, but the persistent activation of its intracellular signaling pathway. These results suggest that chronic hyperinsulinemia may directly contribute to the progression of hepatic steatosis and steatohepatitis through the persistent activation of mTOR and cellular aging programs (Baboota et al., 2022) [75].

The consequences of chronic hyperactivation of insulin signaling also extend to the kidney, confirming how excess insulin can cause organ damage well before the onset of overt diabetes. In an elegant experimental model of the prediabetic phase, persistent hyperinsulinemia caused a progressive reduction of the insulin receptor in glomerular podocytes (kidney cells that help filter the blood), accompanied by a marked attenuation of the IRS1–PI3K–AKT pathway. The resulting loss of the inhibition exerted by AKT favored the hyperactivation of GSK3β, which phosphorylated IRS1, establishing a pathological feedback circuit that further amplified insulin desensitization. These observations demonstrate how chronic hyperinsulinemia, through the persistent activation of the insulin cascade and its feedback circuits, can initiate the onset of microvascular complications early, independently of glycemia (Chen et al., 2026) [76].

Taken together, these findings outline an extremely coherent pathophysiological framework. Chronic hyperinsulinemia keeps the PI3K/AKT/mTOR pathway constantly active, transforming a physiological self-regulation mechanism into a pathological circuit that progressively reduces insulin sensitivity, modifies the expression of genes involved in insulin signaling, favors oxidative stress, promotes cellular senescence, and contributes to the development of metabolic complications in the liver, muscle, and kidney. The central node of this process is represented by the negative feedback exerted by mTOR on IRS-1, a mechanism that initially protects the cell from excessive insulin stimulation but, when it remains active for years, becomes one of the main drivers of type 2 diabetes progression. This very molecular circuit will be the subject of the next paragraph, in which the relationship between mTOR, S6K1, and IRS-1—the true fulcrum of hyperinsulinemia-induced insulin resistance—will be explored in detail.

In the previous paragraph, we saw how chronic hyperinsulinemia keeps the PI3K/AKT/mTOR pathway persistently active, transforming a normal self-regulatory mechanism into one of the main drivers of insulin resistance. It remains to be understood what the molecular mechanism is through which this hyperactivation concretely compromises the cellular response to insulin. The evidence accumulated over the last twenty years converges in identifying a central protagonist of this process: S6 kinase 1 (S6K1), the main effector of the mTORC1 complex. Under physiological conditions, the activation of S6K1 represents an essential component of the anabolic response to insulin. After a meal, in fact, the increased availability of glucose, amino acids, and insulin progressively activates the IRS-1 → PI3K → AKT → mTORC1 cascade, culminating in the phosphorylation of S6K1. The latter stimulates protein synthesis, cell growth, and numerous metabolic processes essential for adaptation to energy abundance. To prevent this response from remaining excessively prolonged, the body has developed a sophisticated negative feedback circuit in which S6K1 itself progressively limits insulin signaling by acting on its main intracellular adapter, IRS-1. However, when the nutritional stimulus becomes chronic, this system loses its homeostatic function and turns into one of the main drivers of insulin resistance (Um et al., 2006) [77]; (Tzatsos and Kandror, 2006) [78].

The role of S6K1 as a sensor of nutritional abundance emerges clearly from the observation that this kinase does not respond exclusively to insulin. Amino acids, particularly leucine and other branched-chain amino acids, are also potent activators of the mTORC1 complex through a pathway independent of the classic PI3K/AKT cascade. In this way, the cell interprets high nutrient availability as a sufficient signal to promote growth, protein synthesis, and energy storage even when insulin stimulation is not the primary input. Under physiological conditions, this integration allows for the extremely efficient coordination of metabolism and energy availability. However, when glucose, amino acids, and insulin remain elevated for prolonged periods, all these signals converge on the persistent activation of mTORC1 and its effector kinase S6K1, causing the continuous activation of the negative feedback circuit toward IRS-1. This explains why chronic nutritional excess, regardless of its origin, is progressively translated into a reduction in insulin sensitivity (Um et al., 2006) [77]; (Tzatsos and Kandror, 2006) [78].

The main target of this circuit is the protein IRS-1 (Insulin Receptor Substrate-1), the molecular adapter that connects the insulin receptor to the activation of PI3K. When S6K1 remains constantly active, however, the situation changes radically. The kinase begins to phosphorylate IRS-1 on numerous serine residues, sites that profoundly modify the protein's conformation and reduce its ability to interact with the insulin receptor and PI3K. The final result is that the receptor continues to recognize and bind insulin, but the signal can no longer be effectively transmitted inside the cell (Harrington et al., 2004) [79]; (Tzatsos and Kandror, 2006) [78].

The functional consequences of this modification are profound. The reduction in PI3K recruitment leads to lower activation of AKT, the main regulator of the metabolic effects of insulin. Thus, glucose transport via GLUT4, glycogen synthesis, inhibition of hepatic gluconeogenesis, lipid metabolism, and numerous cell survival programs are compromised. It is important to emphasize that, initially, this represents a physiological mechanism intended to prevent excessive stimulation of the insulin pathway during periods of nutritional abundance. However, when mTORC1 and S6K1 remain constantly active, the phosphorylation of IRS-1 serines becomes persistent, transforming a temporary self-regulatory circuit into a stable condition of insulin resistance. Cells therefore continue to be exposed to high concentrations of insulin but progressively lose the ability to respond to the hormone, establishing the typical vicious cycle that characterizes type 2 diabetes (Khamzina et al., 2005) [80]; (Ueno et al., 2005) [83].

The functional inactivation of IRS-1, however, represents only the first stage of an even deeper process. Numerous studies have in fact shown that the persistent activation of the mTOR/S6K1 pathway is not limited to temporarily modifying the protein's activity, but promotes its progressive quantitative reduction. After the phosphorylation of inhibitory serines, IRS-1 is recognized by specific regulatory proteins, which promote its detachment from the insulin receptor, its transfer into the cytoplasm, and its subsequent degradation. In parallel, the chronic activation of S6K1 also reduces the transcription of the IRS-1 gene, leading to a progressive decrease in protein synthesis. Insulin resistance, therefore, does not derive exclusively from a temporary functional inactivation of the molecular adapter, but also from its progressive disappearance within the cell. Over time, the insulin receptor continues to be normally expressed and capable of binding insulin, but it has an ever-decreasing number of IRS molecules through which to transmit the signal toward PI3K and AKT. This dual mechanism, functional and quantitative, contributes to making insulin resistance progressively more stable and difficult to reverse (Tzatsos and Kandror, 2006) [78]; (Shah et al., 2004) [81]; (Harrington et al., 2004) [79].

A fundamental role in the control of this circuit is played by the TSC1/TSC2 complex, the main physiological regulator of mTORC1 activity. Under normal conditions, TSC1/TSC2 prevents excessive activation of mTOR, thus preserving correct insulin sensitivity. However, when this system is overwhelmed by chronic nutrient excess or loses its function, mTORC1 continues to stimulate S6K1 and the negative feedback toward IRS-1 becomes permanent. Elegant experimental models have shown that the simple loss of TSC1 or TSC2 is sufficient to drastically reduce the levels of IRS-1 and IRS-2, almost completely abolishing AKT activation despite the insulin receptor remaining perfectly functional. It is therefore evident that the defect does not lie in the binding between insulin and the receptor, but rather in the progressive disappearance of the adapter proteins essential for propagating the intracellular signal. The TSC1/TSC2–mTOR–S6K1–IRS axis therefore represents one of the main metabolic sensors through which the cell translates persistent energy abundance into a reduction in insulin sensitivity (Harrington et al., 2004) [79]; (Shah et al., 2004) [81].

The pathophysiological relevance of S6K1 has been further confirmed by studies conducted on animals genetically lacking this kinase. S6K1 knockout mice, despite producing lower amounts of insulin and presenting a reduced mass of pancreatic β-cells, maintain near-normal glycemia thanks to marked peripheral insulin sensitivity. Even more significant is the fact that these animals are largely protected from the development of age-induced or high-fat diet-induced obesity. The absence of S6K1 reduces the pathological phosphorylation of IRS-1, preserving PI3K recruitment, AKT activation, and GLUT4 translocation. In parallel, fatty acid oxidation and energy expenditure increase, limiting the accumulation of adipose tissue. Although a highly high-fat diet can still produce some metabolic alterations, the insulin response remains significantly superior compared to control animals. These results convincingly demonstrate that S6K1 is not simply a marker of insulin resistance, but one of the main causal mediators through which nutritional excess compromises insulin signaling (Um et al., 2004) [82].

The definitive confirmation of the existence of this pathological circuit derives from pharmacological inhibition studies. In different experimental models, treatment with rapamycin, a potent inhibitor of mTORC1, rapidly reduces S6K1 activity, decreases the inhibitory phosphorylation of IRS-1, restores PI3K and AKT activation, and significantly improves the insulin response in both the liver and skeletal muscle. Similarly, the activation of AMPK (the cell's energy sensor that helps maintain energy balance), obtained experimentally via metformin (a drug that helps reduce glycemia), antagonizes mTORC1 activity, reduces IRS-1 phosphorylation, and restores the efficiency of the PI3K/AKT cascade. Taken together, these observations demonstrate that the mTOR–S6K1–IRS-1 circuit does not represent an inevitable consequence of obesity or diabetes, but rather a dynamic and potentially reversible biological mechanism. Interrupting the persistent activation of mTOR early means preserving the integrity of IRS-1, maintaining efficient insulin signal transduction, and slowing the progression of insulin resistance (Khamzina et al., 2005) [80]; (Ueno et al., 2005) [83]; (Tzatsos and Kandror, 2006) [78].

Overall, the experimental evidence converges on a pathophysiological model that is now extremely solid. The chronic excess of glucose, amino acids, fatty acids, and insulin keeps the mTORC1/S6K1 pathway constantly active, transforming a physiological self-regulatory system into a self-sustaining pathological circuit. The resulting phosphorylation, inactivation, and progressive degradation of IRS-1 and IRS-2 interrupt communication between the insulin receptor and the PI3K/AKT cascade, progressively compromising all the main metabolic actions of insulin. In this way, hyperinsulinemia does not simply represent a compensatory response to insulin resistance, but contributes directly to amplifying it through a precise, reproducible molecular mechanism demonstrated in numerous experimental models. This circuit now constitutes one of the most solid biological paradigms for understanding the progression of type 2 diabetes and represents the ideal link to the next paragraph, dedicated to the mechanisms through which the loss of IRS-1/PI3K/AKT signaling translates into alterations in glucose transport and the metabolic function of peripheral tissues.

Insulin resistance is the result of a progressive imbalance between cellular signals that promote energy storage and those that, conversely, favor its consumption. While in the previous paragraphs we saw how chronic hyperactivation of the mTOR/S6K1 pathway represents one of the main mechanisms responsible for the loss of insulin sensitivity, it is equally important to understand the role of its main physiological antagonist: AMP-activated protein kinase (AMPK). This protein is universally considered the cell's energy sensor, as it continuously monitors the ratio between AMP and ATP (molecules involved in cellular energy management), adapting metabolism to the organism's actual energy availability. When energy is abundant, anabolic processes prevail, whereas when energy reserves decrease, AMPK is rapidly activated to restore metabolic balance. Its task is to limit all activities that require high ATP consumption, such as protein synthesis, lipogenesis, and cell proliferation, while favoring fatty acid oxidation, glucose uptake, mitochondrial biogenesis (the process by which the cell increases and renews its mitochondria), and energy production. In other words, while mTOR represents the main promoter of cell growth, AMPK constitutes the control system that prevents this growth from becoming excessive or metabolically unsustainable (Tao et al., 2010) [84]; (Saha et al., 2010) [85].

This functional opposition between AMPK and mTOR is fundamental to understanding the pathophysiology of insulin resistance. Under physiological conditions, the two pathways cooperate to adapt metabolism to the organism's needs: after a meal, mTOR activity prevails, stimulating the synthesis of glycogen, proteins, and lipids; during fasting, physical exercise, or any situation characterized by reduced energy availability, AMPK activity increases instead, temporarily suspending anabolic processes and directing the cell toward ATP production. This continuous alternation ensures remarkable metabolic flexibility. The problem arises when nutritional excess becomes chronic.

Persistent hyperinsulinemia, high glucose availability, and the accumulation of amino acids and saturated fatty acids keep the mTORC1 pathway constantly active, while AMPK activity progressively decreases. The loss of this important physiological brake allows mTOR and the S6K1 kinase to remain constitutively active, favoring the inhibitory phosphorylation of IRS-1 and the progressive attenuation of insulin signal transduction. Insulin resistance can therefore be interpreted not only as a condition of excessive mTOR activation but also as the direct consequence of the reduced capacity of AMPK to counterbalance such activation (Kwon and Querfurth, 2015) [86]; (Liu et al., 2012) [87].

From a molecular perspective, AMPK exerts this action through numerous intracellular targets. Among the most important are TSC2 and Raptor, two fundamental regulators of the mTORC1 complex. Their phosphorylation by AMPK leads to a marked reduction in mTOR activity. This step is of crucial importance because it interrupts the negative feedback loop that leads to the inhibitory phosphorylation of IRS-1, thus preserving the ability of the insulin receptor to correctly transmit the signal. In parallel, AMPK is also able to directly enhance the PI3K/AKT cascade, increasing AKT phosphorylation. The result is a dual effect: on one hand, one of the main obstacles to insulin signal transduction is eliminated, and on the other, the cells' response to insulin is simultaneously amplified, favoring the processes responsible for glucose uptake and the subsequent translocation of the GLUT4 transporter to the plasma membrane (Tao et al., 2010) [84].

The pathophysiological importance of these mechanisms emerges clearly from experimental studies that have analyzed the effects of nutritional excess on skeletal muscle. Prolonged exposure to high concentrations of glucose or leucine leads to a rapid reduction in AMPK activity, accompanied by a marked increase in mTOR phosphorylation. Simultaneously, a reduction in AKT phosphorylation in response to insulin and a progressive onset of insulin resistance are observed. Particularly interesting is the fact that nutrients belonging to different metabolic categories induce a substantially overlapping molecular response, suggesting that the cell interprets energy excess as a single metabolic signal regardless of its origin.

When AMPK is experimentally reactivated using AICAR (an experimental compound used to stimulate AMPK) or α-lipoic acid, all these alterations are largely prevented: mTOR activity decreases, AKT phosphorylation is restored, and insulin sensitivity returns to near-normal levels. These results demonstrate that the reduction in AMPK activity represents one of the initial events in the development of insulin resistance, rather than a simple consequence of already established metabolic damage (Saha et al., 2010) [85].

Further confirmation comes from studies that have experimentally modified amino acid availability. Leucine restriction, in fact, leads to a reduction of mTORC1 with a consequent increase in the AMPK pathway. This effect eliminates the negative feedback exerted on IRS-1, allowing for the recovery of PI3K/AKT pathway activity and significantly improving insulin sensitivity. Similarly, mouse models characterized by reduced activity of the AMPD1 enzyme, and therefore by greater AMPK activation, show a better response to insulin and a greater ability to adapt to nutritional excess, confirming that the AMPK–mTOR–S6K1 axis constitutes one of the main regulators of glucose and lipid homeostasis (Tandelilin et al., 2015) [88].

The biological relevance of AMPK is also confirmed by studies dedicated to physical activity, probably the most effective non-pharmacological intervention for improving insulin sensitivity. During exercise, ATP consumption by the muscle fiber increases and, consequently, the AMP/ATP ratio rises, leading to rapid AMPK activation. This response does not represent merely an adaptation aimed at meeting the increased energy demand, but also constitutes a powerful metabolic protection mechanism.

In experimental models of high-fat diet-induced obesity, training reactivates AMPK, reduces the activity of mTORC1 and the S6K1 kinase, increases fatty acid oxidation, and reduces lipid accumulation within the muscle. These changes favor the recovery of metabolic flexibility and significantly improve glucose tolerance and insulin sensitivity. It follows that a significant portion of the benefits of exercise does not depend solely on increased energy expenditure, but on the ability to restore the correct balance between AMPK and mTOR, directly counteracting the molecular mechanisms responsible for insulin resistance (Liu et al., 2012) [87].

The protective role of AMPK also emerges with particular evidence in models of lipotoxicity. Among saturated fatty acids, palmitate is considered one of the main mediators through which obesity promotes the onset of type 2 diabetes. Chronic exposure of muscle cells to palmitate leads to a marked reduction in AMPK phosphorylation, removing one of the most important control systems of the mTORC1 complex. Consequently, mTOR remains persistently active, increasing the inhibitory phosphorylation of IRS-1, and the response to insulin progressively decreases. In addition to compromising glucose metabolism, this condition favors cellular inflammation and the onset of metabolic alterations typical of lipotoxicity. Conversely, oleate, the main monounsaturated fatty acid in extra virgin olive oil, keeps AMPK active, thus preventing the hyperactivation of the mTORC1/S6K1 pathway. The protective effect persists even after the removal of oleate, suggesting that AMPK activation induces lasting metabolic adaptations; however, when AMPK is pharmacologically inhibited, this protection disappears completely, confirming the central role of this kinase (Kwon and Querfurth, 2015) [86].

The importance of AMPK is further confirmed by numerous natural compounds that, despite belonging to different chemical classes, all converge on the same molecular circuit. Among these, resveratrol is one of the most studied. In muscle cells exposed to chronic hyperinsulinemia, this polyphenol reactivates AMPK, reduces mTOR pathway activity, limits the inhibitory phosphorylation of IRS-1, and restores normal activation of the PI3K/AKT cascade. The recovery of insulin signaling translates into greater translocation of GLUT4 to the plasma membrane and an increase in glucose uptake by the muscle. These results demonstrate that resveratrol does not act simply by increasing glucose consumption, but directly corrects the molecular defect underlying hyperinsulinemia-induced insulin resistance (Vlavcheski et al., 2020) [89].

Similar results have also been obtained with other phytocompounds. Evodiamine, an alkaloid extracted from Evodia rutaecarpa, increases AMPK activity in adipose tissue, inhibits the mTOR/S6K pathway, reduces the inhibitory phosphorylation of IRS-1, and significantly improves glucose tolerance in animal models of obesity and diabetes (Wang et al., 2013) [90]. Similarly, HM-chromanone, a compound isolated from Portulaca oleracea, activates the AMPK pathway, reduces IRS-1 phosphorylation, and again favors the translocation of GLUT4 to the cell membrane, restoring normal glucose uptake by skeletal muscle (Park and Han, 2022) [91]. Although different in chemical structure and botanical origin, both molecules confirm that the recovery of AMPK activity constitutes one of the most effective mechanisms for counteracting the loss of insulin sensitivity.

Further confirmation comes from studies on anthocyanins (colored plant substances with antioxidant activity). Cyanidin-3-O-glucoside and its main metabolite, protocatechuic acid, activate AMPK, inhibit the mTOR/S6K pathway, increase the expression of GLUT1 and GLUT4 transporters, and improve glucose tolerance in experimental models of obesity. These effects are associated with a significant reduction in insulin resistance, demonstrating that even compounds naturally present in fruits and vegetables can favorably modulate the main circuits that regulate glucose metabolism (Talagavadi et al., 2016) [92].

However, the actions of AMPK are not limited to muscle and adipose tissue but also involve the liver, a central organ in the regulation of metabolic homeostasis. In experimental models of hepatic steatosis, eugenol, the main phenolic component of cloves, activates the AMPK pathway, reduces mTOR activity, and favors fatty acid oxidation. The final result is a marked reduction in triglyceride accumulation in hepatocytes, accompanied by an improvement in steatosis and a decrease in the main markers of liver fibrosis. These data highlight how AMPK activation produces metabolic benefits that extend well beyond glycemic control, contributing to the maintenance of the health of the entire organism (Jo et al., 2014) [93].

Taken together, all this evidence outlines an extremely coherent pathophysiological model. In healthy conditions, AMPK and mTOR operate as two complementary systems that allow the cell to continuously adapt its metabolism to energy availability. However, when nutritional excess, chronic hyperinsulinemia, and lipotoxicity progressively reduce AMPK activity, the main physiological brake exerted on the mTOR/S6K1 pathway is lost. This results in the inhibitory phosphorylation of IRS-1, reduced activation of the PI3K/AKT pathway, decreased GLUT4 translocation, and, finally, the development of insulin resistance. Conversely, apparently very different interventions, such as physical activity, the restriction of certain nutrients, pharmacological activation of AMPK, and numerous bioactive compounds of natural origin, all converge on the same molecular circuit: activation of AMPK, inhibition of the mTOR/S6K1 axis, recovery of IRS-1 function, reactivation of the PI3K/AKT cascade, and consequent improvement in insulin sensitivity.

This extraordinary convergence of experimental results identifies AMPK not only as a sensor of cellular energy status but as one of the most important regulators of metabolic homeostasis and one of the most promising therapeutic targets in the prevention and treatment of type 2 diabetes mellitus and other pathologies characterized by insulin resistance.

If insulin resistance represents the first pathological adaptation preceding type 2 diabetes, the progressive loss of pancreatic β-cells is the event that makes the disease clinically manifest and determines its inexorable progression. In the initial stages, the pancreas is still able to compensate for the reduced sensitivity of tissues to insulin by increasing the production of the hormone. This adaptation allows, at least temporarily, for the maintenance of blood glucose levels close to normal despite the increasing metabolic demand. However, this compensation comes at an extremely high biological cost. β-cells are forced to sustain continuous secretory hyperactivity for years, increasing insulin synthesis, energy consumption, and the workload of the endoplasmic reticulum. When this condition is prolonged over time, adaptive mechanisms begin to progressively exhaust themselves, and the compensatory capacity of the pancreas decreases until it can no longer meet the body's insulin demand (Butler et al., 2003) [94].

For many years, it was believed that this decline was mainly due to a simple loss of the secretory function of β-cells. However, evidence gathered over the last two decades has profoundly changed this interpretation. Today, it is clear that the pathophysiology of type 2 diabetes simultaneously involves two closely intertwined phenomena: on one hand, the progressive functional impairment of β-cells, and on the other, their gradual disappearance through programmed cell death processes. These two events do not represent independent stages of the disease, but rather different aspects of the same pathological process. Before undergoing apoptosis, β-cells go through a long phase during which they progressively lose the ability to synthesize, store, and secrete insulin in a physiological manner. It therefore becomes extremely difficult to distinguish with precision how much of the insulin deficit depends on functional reduction and how much on the anatomical loss of pancreatic cells, as both alterations develop in parallel and reinforce each other (Matveyenko and Butler, 2008) [96].

Anatomopathological observations conducted on the human pancreas clearly confirm this evolutionary model. In obese but metabolically healthy subjects, the pancreas reacts to insulin resistance by moderately increasing β-cell mass, in an attempt to increase insulin production and maintain glycemic homeostasis. This adaptation, however, is much less marked in humans compared to animal models and shows significant individual variability. Some subjects possess a greater capacity for β-cell mass expansion and manage to compensate for insulin resistance for many years, while others show limited pancreatic plasticity and develop glucose tolerance alterations early on. The failure of this adaptive response represents one of the main factors determining the transition from compensated insulin resistance to overt diabetes (Rahier et al., 2008) [95].

One of the most interesting aspects emerging from autopsy studies concerns the moment when the loss of β-cells actually begins. Contrary to what was believed in the past, the reduction in pancreatic mass does not appear suddenly after a diabetes diagnosis, but is already evident during prediabetes. Butler and colleagues demonstrated that in subjects with impaired fasting glycemia, β-cell mass is already reduced by about 40% compared to non-diabetic obese individuals, while in overt diabetes, the loss reaches values exceeding 60% in obese subjects and about 40% in normal-weight subjects. Even more important is the observation that neither cell proliferation nor the neogenesis of pancreatic islets is significantly compromised. The pancreas continues, in fact, to generate new β-cells even during the progression of the disease, but these are eliminated at a significantly higher rate through apoptosis. The progressive reduction of β-cell mass therefore derives primarily from an imbalance between production and cell loss, rather than from an inability to regenerate new endocrine cells (Butler et al., 2003) [94].

These conclusions are also confirmed by studies conducted on the European population. By analyzing over one hundred pancreases obtained from autopsies, Rahier and colleagues observed an average reduction in β-cell mass of about 39%, associated with a decrease in pancreatic insulin content of about 30%. However, one of the most significant results concerns the close relationship between disease duration and cell loss. In the first years after diagnosis, the reduction in β-cell mass appears relatively modest, while it increases progressively over time, suggesting that diabetes itself contributes to accelerating the death of pancreatic cells through phenomena of glucotoxicity, lipotoxicity, inflammation, and oxidative stress. This data profoundly changes the therapeutic perspective: the loss of β-cells does not represent only an inevitable consequence of the disease, but a dynamic process that can potentially be slowed down by intervening early on the mechanisms responsible for cell damage (Rahier et al., 2008) [95].

Subsequent studies have also confirmed that the simple quantitative reduction of β-cell mass is not enough, on its own, to explain the onset of diabetes. In fact, there are individuals who have undergone surgical resection of more than half of the pancreas who maintain normal blood glucose regulation for many years, demonstrating that the absolute number of β-cells does not represent the only determinant of insulin function. What really distinguishes the diabetic pancreas is the progressive deterioration of the so-called functional mass of β-cells, that is, the set consisting of both the number of cells and their secretory efficiency. In the initial stages of the disease, a functional impairment of insulin secretion therefore prevails, while the anatomical reduction of β-cell mass becomes progressively more important as chronic metabolic stress continues.

The transformation of physiological compensation into progressive pancreatic insufficiency depends mainly on the appearance of so-called glucolipotoxicity. Chronic exposure to high concentrations of glucose and free fatty acids forces β-cells into continuous biosynthetic activity, increasing the load on the endoplasmic reticulum, the production of reactive oxygen species, and the accumulation of misfolded proteins. In an initial phase, these mechanisms represent an adaptive response aimed at supporting the increased demand for insulin. However, when the metabolic stimulus becomes persistent, these adaptation systems progressively transform into powerful inducers of cell death, paving the way for the activation of the apoptotic pathways that characterize type 2 diabetes.

Continuing along this evolutionary line, it is evident that the death of β-cells is not the result of a single mechanism, but of the convergence of multiple forms of cellular stress that, over time, exceed the adaptive capacity of the pancreas. Among these, endoplasmic reticulum stress occupies a central position. β-cells are, in fact, highly specialized in the synthesis and correct folding of enormous quantities of insulin. At the same time, persistent metabolic overload causes significant mitochondrial dysfunction, reduces ATP production, and further amplifies cellular damage, creating a vicious cycle that accelerates the deterioration of β-cell function.

Associated with this picture is oxidative stress, now considered one of the main determinants of β-cell degeneration. Unlike many other tissues, pancreatic cells possess a relatively limited supply of antioxidants and are therefore particularly vulnerable to the accumulation of reactive oxygen species (ROS). Chronic hyperglycemia, lipotoxicity, and amyloid deposition progressively increase the production of free radicals, while simultaneously reducing the expression of important defense systems. The consequences are also evident at the histological level: in the pancreases of patients with type 2 diabetes, markers of DNA oxidative damage and lipid peroxidation (oxidative damage to fats present in cells) increase significantly. The close correlation observed between the increase in these markers and the reduction of β-cell mass suggests that oxidative damage does not simply represent a consequence of hyperglycemia, but one of the main drivers of the progressive loss of insulin-producing cells.

Taken together, anatomopathological studies now allow for the delineation of an extremely coherent pathophysiological model. In the early stages of insulin resistance, the pancreas increases insulin secretion and attempts, at least in part, to expand β-cell mass to maintain normoglycemia. With the continuation of chronic hyperinsulinemia, however, the continuous functional overload exposes these cells to glucotoxicity, lipotoxicity, oxidative stress, endoplasmic reticulum stress, amyloid deposition, and inflammation. These factors converge in the activation of apoptotic programs, leading to a progressive reduction of the functional mass of β-cells. The resulting decrease in insulin secretion promotes an increase in blood glucose, which in turn further amplifies all mechanisms of cellular damage, establishing a powerful self-feeding vicious cycle in which hyperglycemia and β-cell loss reinforce each other.

This vision also profoundly modifies the therapeutic approach to type 2 diabetes. For many years, the main goal was almost exclusively the reduction of blood glucose, considered the primary indicator of treatment efficacy. Current knowledge shows, however, that glycemic control represents only part of the therapeutic strategy. Preserving the vitality of β-cells, reducing their functional load, limiting oxidative and endoplasmic reticulum stress, countering glucolipotoxicity, and preventing the activation of apoptotic pathways are now equally important goals. Intervening early on these mechanisms could, in fact, slow down the loss of β-cell mass, prolong the residual secretory capacity of the pancreas, and concretely change the natural history of the disease, rather than limiting treatment to its metabolic consequences alone.

Type 2 diabetes mellitus therefore emerges as a dynamic pathology in which hyperinsulinemia represents the starting point, while the progressive exhaustion and death of β-cells constitute the decisive step toward the irreversible loss of glycemic control. Understanding this temporal sequence also means understanding that diabetes does not appear suddenly when blood glucose exceeds a certain diagnostic threshold, but begins many years earlier, during a long silent phase in which the pancreas fights daily to compensate for an ever-increasing insulin demand. It is precisely during this time window that the possibility of preventing or slowing the evolution of the disease lies. Protecting β-cells means preserving the main physiological system responsible for blood glucose control and intervening in the mechanism that, more than any other, determines the transition from metabolic adaptation to overt diabetes.