The double life of glucose metabolism: brain health, glycemic homeostasis, and your patients with type 2 diabetes

Abstract

The maintenance of cognitive function is essential for quality of life and health outcomes in later years. Cognitive impairment, however, remains an undervalued long-term complication of type 2 diabetes by patients and providers alike. The burden of sustained hyperglycemia includes not only cognitive deficits but also the onset and progression of dementia-related conditions, including Alzheimer’s disease (AD). Recent research has shown that the brain maintains an independent glucose “microsystem”—evolved to ensure the availability of fuel for brain neurons without interruption by transient hypoglycemia. When this milieu is perturbed, brain hyperglycemia, brain glucotoxicity, and brain insulin resistance can ensue and interfere with insulin signaling, a key pathway to cognitive function and neuronal integrity. This newly understood brain homeostatic system operates semi-autonomously from the systemic glucoregulatory apparatus. Large-scale clinical studies have shown that systemic dysglycemia is also strongly associated with poorer cognitive outcomes, which can be mitigated through appropriate clinical management of plasma glucose levels. Moreover, these studies demonstrated that glucose-lowering agents are not equally effective at preventing cognitive dysfunction. Glucagon-like peptide-1 (GLP-1) receptor analogs and sodium glucose cotransporter 2 inhibitors (SGLT2is) appear to afford the greatest protection; metformin and dipeptidyl peptidase 4 inhibitors (DPP-4is) also significantly improved cognitive outcomes. Sulfonylureas (SUs) and exogenous insulin, on the other hand, do not provide the same protection and may actually worsen cognitive outcomes. In the creation of a treatment plan, comorbid cognitive conditions should be considered. These efficacious treatments create a new gold standard of managing hyperglycemia—one which is consistent with the “complication-centric prescribing” mandates issued in type 2 diabetes treatment guidelines. The increasing longevity enjoyed by our populace places the onus on clinical care to play the “long game” in using targeted treatments for glucose control in patients with, or at risk for, cognitive decline to maintain cognitive wellness later in life. This article reviews critical emerging data for scientists and trialists and translates new enhancements in patient care for practitioners.

Background

The prevalence and prominence of type 2 diabetes and dementia-related disorders are increasing as our population enjoys greater longevity. This places a burden and a mandate for physicians to bring their practices in line with evolving gold standards for the management of these chronic diseases. In recent years, the ADA guidelines have recommended that practitioners prioritize comorbid conditions in the crafting of patient treatment plans (“complication-centric prescribing”) [1]. Although macrovascular and microvascular complications of diabetes are well attended, there is incomplete appreciation that cognitive dysfunction must be considered among other long-term complications, as type 2 diabetes is an independent high-risk factor for dementia, AD, and mild cognitive impairment [2, 3].

This review arose through a collaboration bridging the fields of diabetology (SS, MEH, MTHT) and dementia-related disorders (EB, DAB). This work aims to provide practitioners with cutting edge research about this important area. It also provides tools to effectively factor-in and manage cognitive wellness over the lifetime of the patient. It also informs the choice of therapy for patients with or at risk of comorbid cognitive impairment with type 2 diabetes.

Type 2 diabetes and cognitive impairment: prevalence and comorbidity

As of 2021, 38.4 million Americans, or 11.6% of the population, had diabetes [4]. Approximately half of individuals with type 2 diabetes in America are aged 65 years and older [4]. Approximately 20% of US adults who meet laboratory criteria for diabetes are unaware that they are diabetic; this number represents 3.4% of all US adults [5]. The frequency of type 2 diabetes among pediatric patients has increased significantly [3], highlighting that early onset subjects those individuals to decades or lifetime exposure to glucolipotoxicity.

The prevalence of AD and type 2 diabetes are increasing in tandem with longer lifespans and changing lifestyles. The US 2017 figures for AD, or mild cognitive impairment due to AD, reported that approximately 6 million people were impacted [6].

Insulin resistance and diabetes are factors strongly associated with cognitive impairment and AD [7]. Other risk factors include increasing age, smoking, high blood pressure, elevated cholesterol, obesity, depression, lack of physical exercise and social interaction, low education level, and the presence of the APOE4 gene variant [8–10]. In a recent analysis published in Nature, individuals who carry two APOE4 copies had a 60% chance of developing AD by age 85 [11], although the underlying abnormalities of AD began to appear as early as age 55, providing a window for early detection and treatment.

Dysglycemia, however, presents a risk factor independent of APO-E4 genetic influence. Sustained hyperglycemia exerts a direct and significant impact on cognitive function and integrity, a detriment that can be clinically mitigated through good glucose control. One study tracked 9666 participants over 20 years and reported strong associations between diabetes and slowed cognitive processing speed, among and parameters of cognitive function [12]. Other cardiometabolic risk factors, such as hypertension [13, 14] and nonalcoholic fatty liver disease (NAFLD), independently impact cognitive function [15]. In the Atherosclerosis Risk in Communities (ARIC) cohort, the risk of developing dementia was proportional to the severity of NAFLD, dementia was reciprocally a risk factor for NAFLD [15]. The improvement or resolution of NAFLD has also been linked to reducing the development and onset of type 2 diabetes [16].

Cognitive deficits can start as early as the fourth or fifth decade of life. Like the “silent” nature of hyperglycemia, which can exist for decades without awareness by individuals, individuals are often unaware of the onset of cognitive decline. Poor cognitive function also leads to suboptimal adherence by patients and contributes to poorer glycemic control and greater glycemic variability and incidence of hypoglycemic episodes. The bidirectional associations reported between hypoglycemia and dementia suggest an interplay between behavioral factors and underlying physiological parameters for these comorbid conditions [17, 18].

The composite of research in this area solidly evidences cognitive wellness to be among the physiological compromises and sequelae of diabetes and dysmetabolism. Early recognition and management are critical for improving short- and long-term cognitive outcomes for individuals with, or at risk of, comorbid cognitive decline and dementia-related disorders.

Type 3 diabetes?

The negative impact of sustained dysmetabolism on cognitive dysfunction includes memory deficits, decreased psychomotor speed, and reduced frontal lobe/executive function (e.g., planning, coordinating, sequencing, and monitoring of cognitive operations) [19–23]. Detriment begins to accrue in the prediabetic stage [24].

The surface of brain neurons is rich in insulin receptors. In addition to regulating metabolism and promoting the growth and development of nerve cells, insulin in the brain is important for learning and memory. A key link discovered between dysglycemia and cognitive function is the short-circuiting of neuronal insulin signaling [25–28]. In fact, one study using treatment with a thiazolidinedione to reverse insulin insensitivity was shown to improve memory in patients with type 2 diabetes [29].

The full burden of hyperglycemia on neuronal integrity extends beyond cognitive decline: dysglycemia is also a strong risk factor for the development of dementia-related disorders, including AD [24]. Dysglycemia contributes to the accumulation of amyloid beta-peptide, vascular complications, neuroinflammation, hyperglycemia in brain interstitial fluid, mitochondrial dysfunction, and cerebrovascular lesions [30] (Fig. 1). In a synergistic manner, these factors impair brain energy metabolism, a precedent of cognitive dysfunction [31, 32].

Fig. 1.

Risk factors associated with the development of brain insulin resistance and cognitive decline in type 2 diabetes

Impaired insulin signaling is associated with the development of extracellular senile plaques and the hyperphosphorylation of the tubulin-associated unit (TAU) protein and the characteristic neuronal tangles of AD—early indicators of onset of the disease (Fig. 1).

The relationship between dysglycemia and cognition is so prominent that a number of researchers have proposed reclassifying cognitive impairment coincident with type 2 diabetes as “type 3 diabetes” [33–35]. While the current nomenclature for type 2 diabetes and AD suffices well and avoids confusion, this new classification may gain momentum among the medical community.

Glucose’s double life: the brain’s divergent metabolic strategy

The brain is a gluttonous organ. While the brain comprises only 2% of the body’s mass, the exigencies of neuronal transmission and brain metabolism monopolize ~ 20% of our overall energy utilization [36]. Glucose is the principal fuel source for essential brain functions, including synaptic transmission, neuronal survival, and maintenance of the blood–brain barrier [37, 38].

Even short bouts of inadequate glucose or oxygen levels can upset brain bioenergetics, which interrupt neuronal and synaptic connections and can induce atrophy and cell death [28, 39]. The rapidity of neuronal damage following a stroke event exemplifies how quickly an energy crisis can occur in the brain. Notably, even moderate glucose excursions in the brain can cause “mini” insults to the cognitive apparatus that contribute to cognitive loss as well as to the development of dementiarelated diseases.

Insulin resistance in the brain contributes to neuronal pathologies

• Systemic hyperglycemia and systemic insulin resistance are well evidenced and, importantly, for translational medicine, represent modifiable risk factors for cognitive disorders.

• Neuronal function is subject to glycemic regulation by both the brain and the periphery.

• Unique to the brain, glycemic homeostasis is governed by local pressures and operates in an autonomous bioenergetic system. Dysmetabolism can lead to brain insulin resistance, hyperglycemia, and glucose hypometabolism.

• Brain insulin resistance appears to sit in the crosshairs of neurodegeneration. Brain insulin resistance is a possible chief insult that initiates and potentiates diabetes-related cognitive dysfunction and dementia-related disorders.

• One mechanism of redox dysregulation in the brain occurs through the waylaying of insulin signaling, that cells otherwise rely upon to fuel neuronal and cognitive processes.

Unique paradigms of glucose homeostasis and insulin action in the brain

The influx of glucose in the brain induces the transport of insulin across the blood–brain barrier. This is accomplished via insulin receptors on neurons and triggers a signaling cascade. While glucose homeostasis is maintained fairly tightly, cells in the periphery can withstand a certain amount of variability. Brain cells, in contrast, are much more susceptible to neuronal damage in the face of fleeting nutrient deprivation. To ensure a constant perfusion of glucose, the brain has a fail-safe mechanism; it maintains its own glycemic homeostatic “microsystem.” This localized milieu favors excess glucose (hyperglycemia) over the possible advent of momentary glucose depletion (short bouts of hypoglycemia) [40, 41].

If excess glucose is sustained in the brain, then brain hyperinsulinemia and brain insulin resistance can ensue. This produces a hypometabolic bioenergetic state in neurons (Fig. 2), which impairs neuronal function [42–44] and results in cognitive and learning deficits.

Fig. 2.

The sequelae of acute hyperglycemia under physiological conditions begin with acute elevations in glucose within the brain and lead to increased transport of insulin across the blood–brain barrier (BBB). This results in elevated levels of insulin available for brain cells. Insulin activates the intracellular signaling pathway that ultimately favors cognitive and learning functions. However, chronic hyperglycemia leads to persistent elevated levels of glucose and insulin in the brain. The resulting neuronal impairment manifests as cognitive and learning deficits that may culminate in dementia over time

Notably, the association of AD with brain insulin resistance is even greater than that of the most canonical hallmarks of AD, such as Ab oligomers and Tau proteins [45].

An important conceptual leap is that neurons and glial cells are subject to the same ravages of glucolipotoxicity as other cell types of the body. Glucolipotoxicity damages the proteins, lipids, and nucleic acids that otherwise support cell integrity, function, and viability [31, 37, 46–48]. As recently reviewed by Pignalosa and colleagues [46], glucolipotoxicity has been linked to outcomes including cerebral vasculature hemorrhage, damage to brain blood vessels, accumulation of amyloid beta peptide, and structural damage.

Sustained hyperglycemia also leads to overactivation of the Krebs cycle. Critically, this generates excessive reactive oxygen species [49]. Reactive oxygen species and oxidative stress are culprits of cell dysfunction and aging [31, 37, 38, 50]. Whether oxidative stress precedes brain insulin resistance, or vice versa, remains a chicken-and-egg conundrum. Brain insulin resistance and excess oxidative stress have been shown to proceed through a vicious cycle in which brain insulin resistance disrupts mitochondrial energy processes and increases oxidative stress. The production of reactive oxygen species perpetuates insulin resistance and so forth. The resulting neuronal impairment manifests as cognitive and learning deficits and can culminate in dementia [19, 25, 27, 38]. Intriguingly, preclinical studies have suggested that insulin sensitization (through exercise and treatment with the glucose-lowering agent, metformin) effectively counteracts the mitochondrial dysfunction caused by brain insulin resistance [51].

Like cognitive malfunction, the development of AD is strongly associated with brain insulin resistance, brain hyperglycemia, and brain glucose hypometabolism, each as an independent biomarker of AD [31, 37, 38, 44, 52]. The intervening mechanisms that have been shown experimentally include impairment of synaptic plasticity, cell signaling, microglial cell activation, Tau protein phosphorylation, cholinergic function, and cell autophagy, as examples [27, 37, 38, 53, 54].

Aberrant protein phosphorylation as an early event in the neuropathology of dementia

Cell signaling pathways are particularly vulnerable points through which oxidative radicals cause neuronal damage. The enzymes involved coincide, interestingly, with the energy-producing pathways of cells. The consequences of disrupted cell signaling include faltering synaptic transmission and diminished ATP synthesis [37, 38, 44, 45, 55]. This can even push neurons toward cell death. The kinases identified include insulin receptor substrate 1 (IRS1), enzymes within the PI3K/Akt axis, the energy-generating, glycogen synthase kinase-3β (GSK-3β) enzyme, and the mitochondrial-resident ATP synthase complex [28, 31, 37, 53, 56–58]. The extent of IRS1 malfunction was shown, in one study, to be proportional to the progression of AD pathology, as well as poorer performance of cognitive tasks [45]. The development of small oligomers of amyloid beta-peptide has similarly been linked to altered insulin receptor signaling. In animal models of AD, discoordination of IRS1/PI3K/Akt signaling sits upstream to Tau hyperphosphorylation in association with brain insulin resistance [59, 60]. Excessive phosphorylation of tau further exacerbates neurotoxicity, in a vicious cycle.

Barone and colleagues recently characterized two enzymes that appear to be key intermediaries between brain insulin resistance and resulting progression of AD [31]. Insulin-degrading enzymes (IDEs) are chief regulators of insulin levels by regulating the rate of degradation of these proteins. In addition to insulin, IDE targets glucagon, atrial natriuretic peptide, and beta-amyloid peptide, among other factors for degradation [31]. Oxidative or nitrosative modifications, critically, reduce the function of IDE. In the face of reduced IDE function, fewer Ab peptide oligomers are degraded, allowing these to accumulate.

The second enzyme recently discovered within the sequelae leading to AD is biliverdin reductase-A (BVR-A). Bilirubin is an endogenous free radical scavenger; BVR-A catalyzes the reduction of heme-derived biliverdin into bilirubin. BVR-A possesses both serine/threonine/tyrosine kinase and scaffold activities, through which it interacts with the insulin receptor kinase. By intersecting a complex array of signal transduction pathways, BVR is involved in the pathogenesis of neurodegenerative, metabolic, cardiovascular, and immune-inflammatory diseases as well as in cancer. The effect of BVR-A dysfunction on insulin receptor substrate 1 (IRS-1)-mediated regulation of insulin signaling amounts to the short circuiting of the ability of brain cells to fuel insulin-dependent energetic processes [26, 38].

This highlights an important theme: under conditions of nutrient excess, the brain’s adaptive mechanisms that ensure uninterrupted nutrient access for the brain becomes maladaptive rather than adaptive—with negative consequences to cognitive processes, learning, memory, and development of dementia (Fig. 2).

The omics of hyperglycemia and cognitive impairment

Several genes have strong links to AD. APOE4 is one of the most studied genetic links with AD [11]. A network clustering study identified three tightly connected gene clusters between cognition and diabetes. A total of 1381 genes among these three clusters were found to be common to both AD and type 2 diabetes; hundreds of genes mapped between type 2 diabetes with those involved in memory [61].

The genetic and epigenetic linkages between AD and type 2 diabetes were shown to involve brain insulin resistance; neuroinflammation; insults to compensatory mechanisms, and peripheral metabolic dysregulation. Intersections were found for these with insulin signaling, inflammation, and inflammasome activity; proteolysis, gluconeogenesis, and glycolysis; glycosylation; lipoprotein metabolism and oxidation; cell cycle regulation or survival; and the autophagic–lysosomal and energy pathways [62].

Using samples from the Framingham Heart Study cohort, Sarnowski et al. [63] conducted an epigenetic analysis leveraging both blood and brain omics in 3167 participants from this cohort. This analysis identified potentially distinct epigenetic regulatory mechanisms that underly insulin resistance and AD at several loci between the periphery and the dorsolateral prefrontal cortex. Systemic insulin resistance was associated with higher DNA methylation levels in blood in the periphery. In the brain, however, higher DNA methylation levels of these same markers were associated with increased AD risk. This new research begins to highlight the complex genetic mechanisms at work in these comorbidities.

Cognitive dysfunction: elevated to a complication of type 2 diabetes

Cognitive impairment joins the (growing) list of diabetes-related complications

Sustained hyperglycemia is deleterious to all cell types of the body; those organ systems most affected present clinically as diabetes-related long-term complications. Damage due to hyperglycemia begins to accrue early, that is, in the prediabetic state, which can be sustained over many years. Like the incremental, progressive damage caused by hyperglycemia to cells of the periphery, chronic hyperglycemia is detrimental to the integrity of the cognitive apparatus. Individuals with diabetes have been shown to be twice as likely to develop dementia as those without diabetes. Dysmetabolism of glucose and insulin in the brain energetic “microsystem” contributes to this as well as glucose dysmetabolism arising from the periphery.

The “classical” micro- and macrovascular complications (cardiovascular disease, peripheral artery disease, nephropathy, retinopathy, and neuropathy) have been considered dogma for decades but, in fact, includes a broader range of disorders. As reviewed by Schwartz et al. [64], these also include stroke, gastrointestinal problems, dental disease, immunocompromise, and even cancer. Cognitive function is another consequence. The remainder of this article focuses on practical therapies of choice for managing systemic hyperglycemia in patients with type 2 diabetes and comorbid cognitive impairment.

Management principles for addressing cognitive outcomes in patients with type 2 diabetes

• Cognitive dysfunction is now clearly recognized among the long-term outcomes of diabetes.

• ‘Complication-centric’ management of type 2 diabetes is recommended in the ADA guidelines and considers cognitive status and the risk of cognitive decline independent of patient age or frailty.

• ‘Complication-centric’ management should be incorporated early in the patient treatment plan.

• Screening for cognitive decline is readily available and easy to administer.

• Counseling includes diet, exercise, lifestyle choices, and supplements conducive to ‘cognitive wellness’.

• Tight glucose control is essential for cognitive wellness and encompasses attainment of A1c targets, as well as the control of glycemic variability and hypoglycemic excursions.

• Reaching A1c targets is increasingly achievable with current agents, especially through the avoidance of agents prone to hypoglycemia.

• Metformin, GLP-1 receptor analogs, DDP-4is, and SGLT2is are associated with a decreased risk of developing dementia. Exogenous insulin and SUs, on the other hand, do not show csimilar benefit and may actually increase the risk of dementia and should be avoided.

Cognition as a modifiable long-term complication of hyperglycemia: practical clinical management strategies

Screening for cognitive impairment

In the clinic, the use of a cognitive assessment tool is highly recommended by the ADA to diagnose and monitor neurodegenerative diseases. Several simple screening assessment tools are available, including the Mini-Cog [65] and the Montreal Cognitive Assessment [66]. The Saint Louis University Mental Status Examination (SLUMS) Cognitive Assessment [67] was developed to be possibly less influenced by socioeconomic factors.

A common challenge with brief assessments is a lack of sensitivity to detect mild cognitive deficits and changes in early dementia [68]. Comprehensive neuropsychological evaluations establish a cognitive baseline for individuals presenting with symptoms, facilitating future comparisons. These evaluations enhance diagnostic accuracy and identify key cognitive and behavioral symptoms for targeted nonpharmacologic interventions and caregiver education. The domains assessed by these tools include attention, episodic memory, language, visuospatial functions, and executive abilities. The impact of impairments in one domain on the performance of others can be captured and assessed [69].

Traditional treatments for dementia

Traditional treatments for dementia include cognitive training, vitamin D, and other supplements, and restrictions on alcohol consumption and smoking. Acetylcholinesterase inhibitors that reduce beta-amyloid deposits have been developed. Memantine is a widely used partial N-methyl-D-aspartate receptor antagonist[70]. Interestingly, vaccinations for influenza have been shown to strikingly reduce the odds ratio for developing dementia by as much as 50% [71], vaccination for varicella zoster has been shown to provide similar protection [72]. The mechanism(s) for this benefit are still to be resolved.

Although amyloid beta is a rational target for therapy, the clinical effectiveness of current anti-amyloid beta antibodies has been underwhelming. Lecanemab (Leqembi; Eisai/Biogen) was approved by the U.S. Food and Drug Administration for the treatment of early symptomatic AD in mid-2023; donanemab (Kisunla; Eli Lilly) was approved in June 2024. These agents follow the first-in-class anti-amyloid antibody aducanumab (Aduhelm; Biogen), which was discontinued in early 2024 because of meager sales. The modest effects of these agents in their clinical trials have left an open question for the community as to whether the benefits are clinically meaningful, especially given the adverse events sometimes associated with this class [73].

Individual antidiabetic agents and cognitive outcomes

Large diabetes registries (such as insurance claims databases) have allowed the records of hundreds of thousands of patients treated with glucose-lowering agents to be evaluated [74–76]. These studies complement smaller analyses [77] or prospective clinical trials [78]. Such studies have shown that not all diabetes drugs are equal in terms of their capacity to support cognitive wellness.

Wium-Andersen and coworkers evaluated hyperglycemia and dementia in approximately 170,000 patients with type 2 diabetes and reported that the use of metformin, DPP4is, GLP1 analogs, or SGLT2is was associated with lower odds of dementia after multiple adjustments [ORs of 0.94 (95% confidence interval (CI): 0.89–0.99), 0.80 (95% CI 0.74–0.88), 0.58 (95% CI: 0.50–0.67), and 0.58 (95% CI: 0.42–0.81), respectively], with a gradual decrease in the odds of dementia for each increase in daily defined dose [74]. The benefits accrued relatively quickly and were enduring. The REWIND Trial (NCT01394952) evaluated cardiovascular outcomes with treatment with the GLP-1 receptor agonist, dulaglutide. Cognitive outcomes were secondary endpoints in this large (n = 9901). The hazard of substantive cognitive impairment was reduced by 14% in those assigned dulaglutide as compared to placebo (HR 0·86, 95% CI 0·79–0·95; p = 0·0018) at year 5 [79].

A meta-analysis of twenty-seven studies reported that the risk reduction for incident dementia was evident at approximately 3 years of treatment with metformin and was maintained at 8 years of follow-up [76]. The reduced risks for dementia were most pronounced for SGLT-2is and GLP-1 receptor agonists, with dramatic odds ratios of 0.41 [95% CI 0.22–0.76] and 0.34 [95% CI 0.14–0.85], respectively. More modest benefits were reported for thiazolidinedione (OR 0.60 [95% CI 0.51–0.69]) and DPP-4 (OR 0.78 [95% CI 0.61–0.99]) [75].

Shin et al. evaluated 110,885 adults with type 2 diabetes aged 40–69 years in a Korean patient registry. These researchers reported a 35% lower risk of dementia associated with the use of SGLT-2is than with the use of DPP-4is at approximately 2 years. This trend persisted regardless of dementia type, age, sex, baseline cardiovascular risk, or concomitant use of metformin [80].

Exogenous insulin and SUs, on the other hand, do not appear to protect against dementia. SUs were shown to be deleterious rather than protective, with a striking odds ratio of 1.43 for developing dementia [95% CI 1.11–1.82] [75]. In one study, higher doses of exogenous insulin (> 144 IU/d) had a positive effect on cognitive function, especially in individuals with poorly controlled diabetes (A1c ≥ 9%). However, lower dosages showed no benefit [81].

There are concerns with the use of exogenous insulin, including a loss of sensitivity to insulin, a similar phenomenon to endogenous hyperinsulinemia. The development of insulin resistance with insulin treatment is well established in the periphery; it may also contribute to insulin resistance in the brain.

Hypoglycemia is an additional serious concern with the use of exogenous insulin [82–84]. Hypoglycemia, an independent risk factor for cognitive decline, causes direct and indirect damage to the nervous system and can lead to neuronal cell death [85]. Notably, insulin causes not only severe hypoglycemic episodes but also frequent, small, incremental “non-severe” hypoglycemic episodes [86]. In a prospective observational type 2 diabetes study across 17 UK centers, insulin therapy was associated with the highest incidence of hypoglycemia among antidiabetes treatments, with rates nearly five times greater than those associated with metformin. Treatment with sulphonylureas was also associated with increased hypoglycemia risk [87].

“Non-severe” hypoglycemic episodes have been shown to contribute to cognitive dysfunction; glycemic variability is an independent risk factor for cognitive decline [88, 89]. Hypoglycemic excursions and glycemic variability are difficult to avoid with some glucose-lowering regimens, particularly with exogenous insulin. These various downsides of insulin therapy cast a shadow on its risk/benefit equation. The present authors urge a reduced reliance on exogenous therapy in the management of type 2 diabetes. The latest data suggest that it is plausible that insulin therapy may contribute, rather than offset, cognitive dysfunction in patients with type 2 diabetes.

Proposed mechanisms of the cognitive-sparing effect of antidiabetic agents

By virtue of large patient outcome registries and clinical trials, we have learned that many of our antidiabetes medications currently in use possess important benefits beyond lowering glucose. The mechanisms studied in the preservation of cognitive function by antidiabetic agents include improvements in brain mitochondrial function, insulin resistance, inflammation, apoptosis, and neuronal plasticity, among others [90, 91]. At least some of these actions are accomplished pleiotropically. Some of the most notable effects of GLP-1 receptor agonists and SGLT2is are on the cardiorenal axis and are exerted through pathways that are independent of the glucose-lowering actions of these classes [92–94].

It is not surprising that GLP-1 receptor agonists have direct effects on brain neurons, given its biologic role to bridge communication between the gut and brain. It has been less anticipated that the therapeutic value of SGLT2is extend beyond the kidney. As recently reviewed by Packer [95], SGLT2is also has a direct influence on the brain, including in the re/programming of the mitochondrial response of neurons to nutrient deprivation, which is a complexly orchestrated evolutionary strategy that supports neuron survival under conditions of low nutrient availability [95]. SGLT2 alters a number of pathways, including inhibition of acetylcholinesterase, which is, coincidentally, a primary target of current AD therapies [96].

Preclinical research has been fruitful in characterizing these mechanisms [60, 97–99]. Rodent models of AD have demonstrated that liraglutide repairs disrupted insulin signaling pathways [60], perturbations that contribute to cognitive impairment and dementia-related processes. Other studies have shown that the phosphorylation of Tau and neurofilament proteins [26, 60] and mTOR cycle processes are modulated by antidiabetic agents [100].

The prescribing push-me, pull-me of tight glucose control

A quandary for prescribers has long been the attainment of A1c targets without introducing hypoglycemia. Nocturnal asymptomatic hypoglycemia has been shown to occur in approximately 25% of patients with diabetes treated with insulin therapy [101]. Elderly individuals are most likely to experience asymptomatic hypoglycemia [101].

“Hypoglycemia unawareness” is complicated by the standard age-related loss of the body’s natural counterregulatory response to hypoglycemia. In elderly individuals, there is an inherent aging deficit in the ability to stabilize plasma glucose levels, termed “hypoglycemia-associated autonomic failure.”

With today’s broader and “gentler” antidiabetic treatments, tight glucose control without the induction of hypoglycemia is realizable. Treatment adherence issues increase the risk of hypoglycemia, especially in patients with diabetes who are elderly, frail or have cognitive or mental issues [1]. These can be mitigated with use of patient education and diabetes counseling, the involvement of family and caregivers, and more frequent glucose monitoring, including continuous glucose monitoring. These approaches can also be used in patients at risk of cognitive issues.

Complication-centric prescribing in patients with diabetes and cognitive concerns

The knowledge that the choice of antidiabetic therapy can achieve more favorable cognitive outcomes is a key recent advance to care. Metformin, GLP-1 receptor analogs, or SGLT2is may be the most “brain-friendly” options, with pioglitazone or DDP-4 as other good options. Interestingly, large-scale randomized clinical trials on SGLT-2is and GLP-1 receptor agonists have shown that these agents also provide significant cardioprotective benefits [102]. SGLT2is reduced the risk of hospitalization due to heart failure by 30–37% [103]. The benefits of SGLT2is were observed within 1 month of treatment initiation and persisted with the continuous use of the treatment [103].

Despite their broad benefits to diabetes-related complications, the drug acquisition costs of GLP-1 receptor analogs and SGLT2is are prohibitive for some patients. Metformin is inexpensive, readily available, well tolerated, available in oral and extended-release formulations, and has been shown to help protect against cognitive decline.

We urge that SUs should be avoided, as these agents are inclined toward hypoglycemia. SUs are speculated to also cause undue wear and tear on pancreatic beta cells and might similarly do so on the cognitive apparatus. Despite low drug acquisition costs, the unfavorable risk/benefit profile and poorer short- and long-term outcomes of SUs makes these agents poorer choices among today’s antidiabetes armamentarium.

Exogenous insulin use may increase the risk of neuropathies, suggesting that insulin therapy is detrimental (or does not provide protection) as well as other agents for use in type 2 diabetes. Accordingly, we advocate for the reduced reliance on exogenous insulin in clinical practice. The use of high doses of insulin as short-term therapy can successfully remit uncontrolled hyperglycemia, but healthcare practitioners should not, however, neglect to subsequently reduce the insulin dose to the required minimum. Many patients can, in fact, achieve intensive glucose control with alternative antidiabetes agents to insulin therapy, allowing practitioners to omit or reduce reliance on exogenous insulin.

In patients for whom insulin resistance is believed to be a primary driver of hyperglycemia, pioglitazone or metformin may be the preferred treatment. In patients with weight issues, GLP-1 receptor agonists present an advantage, as well as for patients with comorbid cardiovascular disease. In patients with kidney disease, SGLT-2is may be the treatment of choice. These follow the current ADA recommendations for “complication-centric” management of hyperglycemia and allow room for cognition to be a target for therapy, among other comorbid conditions in the patient treatment plan.

Depending on the extent of comorbid cognitive issues or the onset of dementia or AD, referral to a specialist may be the best course of action for a given patient.

Ongoing research into antidiabetes agents for the management of cognitive outcomes

Studies are underway the evaluate the use of intranasal insulin preparations to improve memory. Earlier trials did not show adequate effectiveness, although the agents possess the mechanistic advantage of delivering exogenous insulin within the brain and bypassing the need to cross the BBB. Intranasal insulin preparations continue to be tested in healthy adults and AD patients and are hoped to resolve observed gender differences in efficacy in cognitive outcomes [104–108].

The various pleiotropic actions of glucose-lowering agents are an area of intensifying research. Cognitive outcomes are now gaining research momentum. Novo Nordisk has launched a clinical program, the ISAP Trial, assessing the effects of semaglutide on cognitive function and AD outcomes. The accumulation of cortical Tau protein and neuroinflammation will be evaluated in subjects with preclinical/prodromal AD [109].

Past findings have shown that intranasal insulin or dulaglutide has beneficial effects on cerebrovascular disease; these findings provide a rationale for assessing a regimen of semaglutide in combination with intranasal insulin. The 80-subject COMMETS study (NCT06072963) in patients with metabolic syndrome and mild cognitive impairment assesses global cognition, neurobiological marker, cerebral blood flow, cerebral glucose utilization, white matter hyperintensities, AD-related blood biomarkers, and the expression of insulin signaling proteins measured in brain-derived exosomes [110].

A second exciting area comes from the field of ophthalmology. Brain degeneration status is “reflected” in neurovascular changes in the eye. This allows a fundal exam to serve as a noninvasive diagnostic “portal” to detect cognitive disorders. Multispectral eye scans may be the future of detecting minute changes in eye neural structures accompanying the early development of brain neuropathies [111].

Conclusions

Serendipitous findings from long-term studies of individuals with type 2 diabetes fortuitously extended our roster of long-term complications to include cognitive dysfunction, cognitive decline, and dementia-related disorders. Cognitive outcomes need to become top priorities for practitioners to the same extent as classic microvascular and macrovascular complications. The preservation of cognitive function is essential for quality of life in later years. The extended longevity that our populace now enjoys means that practitioners need to play the “long game”—life management planning for “cognitive wellness” through good glucose control and other relevant strategies. These approaches are especially important, as brain cells naturally become less efficient at glucose utilization and homeostasis with advancing age [112]. Furthermore, these deficits are accelerated by uncontrolled hyperglycemia.

Current guidelines endorse the consideration of comorbid conditions in the creation and maintenance of patient treatment plans. “Complication-centric prescribing” represents a major advance in the routine care for patients with type 2 diabetes. Cognitive wellness is now a standard parameter to be considered in treatment plans.

To existing management principles for cognition wellness (i.e., enhanced lifestyle practices, screening and early detection, and neurological management approaches), glucose control needs to be added for individuals with type 2 diabetes (and perhaps for those with prediabetes). Among antidiabetic treatment options, metformin, SGLT-2is, GLP-1 receptor agonists, and DPP-4is have been shown to prevent and mitigate cognitive compromise. SUs and exogenous insulin, on the other hand, have not been consistently found to afford the same protection. Through frank or silent hypoglycemia and likely through a range of other mechanisms, SUs and insulin worsen cognitive outcomes. Therefore, the consideration and appropriate use of noninsulin antidiabetes therapy are essential for patients with type 2 diabetes and comorbid cognitive issues or those at risk of cognitive issues to minimize, rather than accelerate, cognitive decline and dementia-related disorders.

Abbreviations

AD

Alzheimer’s disease

ADA

American Diabetes Association

ARIC

Atherosclerosis Risk in Communities Study

BVR-A

Biliverdin reductase-A

BBB

Blood–brain barrier

DPP-4is

Dipeptidyl peptidase 4 inhibitors

GLP-1

Glucagon-like peptide-1

GSK-3β

Glycogen synthase kinase-3β

IDE

Insulin-degrading enzyme

IRS1

Insulin receptor substrate 1

NAFLD

Nonalcoholic fatty liver disease

SLUMS

Saint Louis University Mental Status Examination Cognitive Assessment

SGLT2is

Sodium glucose cotransporter 2 inhibitors

SUs

Sulfonylureas

TAU

Tubulin-associated unit

USPSTF

US Preventive Services Task Force

Authors’ contributions

S.S.S. and M.E.H. made substantial contributions to the conception and framework (design) for the article. M.T.H.T. lent a literature review and contributed to content development. S.S.S. and M.E.H. provided direction and citations related to type 2 diabetes; E.B. and D.A.B. made substantial contributions to the conception of portion pertaining to dementia-related disorders. D.A.B. substantively revised the manuscript. All authors approved the submitted version.

Funding

These authors state that they have received no direct or indirect funding in the creation of this manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.