Pathobiology of Prediabetes: Understanding and Interrupting Progressive Dysglycemia and Associated Complications

Abstract

Diabetes currently affects ∼37 million adults in the U.S. and 537 million people worldwide, with type 2 diabetes (T2D) accounting for 90%–95% of the diabetes burden. The transition from normal glucose regulation (NGR) to T2D is via an intermediate stage of prediabetes, characterized by impaired fasting glucose (IFG) and impaired glucose tolerance (IGT). Prediabetes affects ∼98 million adults in the U.S.; worldwide, more than 541 million adults have IGT and 319 million adults have IFG. Prediabetes is associated with increased risks of developing vascular and neuropathic complications, besides the risk of progression to T2D. Discussed herein are the demographic, anthropometric, biobehavioral, biochemical, and molecular factors associated with the transition from NGR to prediabetes. The natural history of prediabetes predicts time-dependent progression to T2D, as sustained recovery from prediabetes is uncommon without intervention. Lifestyle modification and certain medications interrupt the progression to T2D and may restore NGR. The landmark intervention trials are discussed, with an interpretive focus on their limitations and the need for novel approaches for durable reversal of prediabetes.

Article Highlights

• Prediabetes, an intermediate stage in the pathogenesis of type 2 diabetes (T2D), affects ∼98 million U.S. adults and >800 million people worldwide.

• Prediabetes shares common pathophysiological mechanisms with T2D and progresses to T2D at variable rates based on risk factor burden.

• Less well studied is the initial transition from normoglycemia to prediabetes. Discussed herein are behavioral, clinical, biochemical, and molecular factors associated with development of prediabetes.

• Current approaches to prediabetes management are discussed, with an interpretive focus on their limitations and the need for novel interventions for durable reversal of prediabetes and avoidance of related complications.

Introduction

The term “prediabetes” entered the medical literature in 1918, when Auer and Kleiner suggested that partially pancreatectomized dogs “may legitimately be considered as being in a prediabetic state” (1). Currently, prediabetes is diagnosed using fasting plasma glucose (FPG) levels between 100 and 125 mg/dL (impaired fasting glucose [IFG]), 2-h postload plasma glucose (2hrPG) levels between 140 and 199 mg/dL (impaired glucose tolerance [IGT]), or hemoglobin A_1c (HbA_1c) levels between 5.7% (39 mmol/mol) and 6.4% (46 mmol/mol) (2). Individuals with normal glucose regulation (NGR) are those with FPG levels <100 mg/dL and 2hrPG levels <140 mg/dL. The Centers for Disease Control and Prevention estimated that 97.6 million U.S. adults aged ≥18 years (34.5% of the adult U.S. population) had prediabetes in 2022 (3). Worldwide, more than 541 million adults have IGT and 319 million adults have IFG (4).

The transition from NGR to type 2 diabetes (T2D) is punctuated by a variable interlude of impaired glucose regulation (diagnosed as prediabetes, IFG, or IGT). Once T2D develops, it can be controlled but not cured. However, the development of T2D can be prevented or delayed by intervention at the stage of prediabetes. With the rare exception of monogenic diabetes, which can present at birth, most individuals who develop T2D were born with NGR and later progressed to prediabetes and thence to T2D. The present article focuses on the initial escape from NGR and transition to impaired glucose regulation and prediabetes. Data from cross-sectional and prospective studies are synthesized to distill a comprehensive accounting of the risk factors and mechanisms that trigger the initial escape from normoglycemia toward dysglycemia. Emerging knowledge of microvascular and macrovascular complications in people with prediabetes is summarized. Finally, strategies for prevention of T2D and reversal of prediabetes are discussed and placed in the context of risks, benefits, and sustainability.

The Intertwined Pathophysiology of T2D and Prediabetes

IGT and IFG

Prediabetes and T2D represent different stages in the continuum of dysglycemia and share a common pathophysiology. Components of DeFronzo’s “ominous octet” of pathophysiological defects underlying T2D manifest in people with prediabetes (5), the latter including insulin resistance and β-cell dysfunction (6–8), increased lipolysis (9), impaired incretin response (10), impaired glucagon suppression (10), decreased hepatic glucose uptake (11), and impaired postprandial suppression of hepatic glucose production (11). Also documented in the setting of insulin resistance and prediabetes are increased renal sodium reabsorption (12) and altered dopaminergic signaling in brain and adipose tissue (13–15). The dopamine agonist bromocriptine, approved for treatment of T2D, improves glycemic control presumably via augmentation of hypothalamic dopaminergic tone (16). Figure 1 shows comparative values for insulin sensitivity and secretion across glycemic strata defined by FPG and 2hrPG levels.

Figure 1.

Insulin sensitivity and β-cell function (disposition index) in people with low-normal FPG (Low-NFG), high-normal FPG (High-NFG), IFG, combined IFG and IGT (IFG+IGT), and newly diagnosed T2D (Newly dx T2D). Disposition index was calculated as the product of insulin sensitivity and acute insulin response to glucose (reprinted from Dagogo-Jack [32]; original data from Dagogo-Jack et al. [84]). *P = 0.04; **P = 0.02; ***P < 0.0001.

Transition From Normal to Impaired Glucose Regulation and Prediabetes

In randomized controlled trials (RCTs) with enrollment of populations with prediabetes (predominantly IGT), annual diabetes incidence rates of 7%−18% were reported in the placebo arm (17–21). Less well-known is the initial rate of transition from NGR to prediabetes. Reported annual rates of progression from NGR to prediabetes in prospective studies include 9.5% among Pima Indians (22), 6.2% among the predominantly (96%) White Baltimore Longitudinal Study of Aging (BLSA) cohort (23), and 11.2% in the Pathobiology of Prediabetes in a Biracial Cohort (POP-ABC), with enrollment of African American and European American adults with parental T2D (24). A study in Japanese adults reported an annualized incidence rate of 9% for progression to prediabetes by at least one criterion during a 5-year follow-up period (25). Together, findings of these studies, conducted in different populations at different times across a broad age range (18–96 years), are of a surprisingly narrow range of 6%–11% as the annual rate of progression from normoglycemia to prediabetes (22–25).

In the BLSA 2.5% of participants enrolled with NGR progressed to diabetes compared with 62% who progressed to prediabetes, over 10 years (23). In the POP-ABC study, 2.7% of participants enrolled with NGR progressed to diabetes and 29.3% progressed to prediabetes during a mean follow-up period of 2.62 years (24). These data support current understanding of prediabetes as a canonical intermediate stage in the pathogenesis of T2D (22–25). Better understanding is needed of the factors associated with the initial progression from NGR to prediabetes. As the only prospective study with enrollment of a diverse cohort, prespecification of incident prediabetes as the primary outcome, and exploration of pathophysiological mechanisms, the POP-ABC study has been a source of pertinent information (22–25). One limitation of the POP-ABC study, however, is that all participants had parental history of T2D, so the findings may not be generalizable (24).

Progression From NGR to Prediabetes

Demographic Factors

Older age, male sex, and maternal history of diabetes are some demographic factors associated with increased risk of incident prediabetes (24,26). Race/ethnicity was not a predictor of incident prediabetes or glycemic progression among African American and European American offspring of parents with T2D (24,27) (Fig. 2), consistent with national U.S. data showing ethnic and racial differences in the prevalence of prediabetes (2,3,28). Similarly, ethnic and racial differences in the rate of progression from prediabetes to diabetes were nonexistent or modest in the initial and 22-year reports from the Diabetes Prevention Program (DPP) (19,29). Thus, race and ethnicity may not be major determinants of glycemic progression among offspring of parents with T2D or high-risk individuals who already have developed prediabetes.

Figure 2.

Kaplan-Meier plot of prediabetes survival probability (A), proportions with 10-mg (B) and 20-mg (C) increases in FPG, and percentile distribution of changes in FPG (D) and 2hrPG (E) during 5.5-year follow-up of initially normoglycemic African American (red) and European American (blue) adults with parental history of T2D (log-rank P = 0.7855) (data from Razavi et al. [27]). Yr, year.

Behavioral, Anthropometric, and Inflammatory Factors

Weight gain, even modest amounts, increases risk of prediabetes (6,7,24). Among Pima Indians (6) and POP-ABC participants (30), weight gain of ∼1.5 kg/year increased risk of prediabetes, whereas a change of <1.0 kg/year was protective for incident prediabetes. The effects of body weight for incident prediabetes are mediated in part by adipocytokines: proinflammatory cytokines increase risk of prediabetes, whereas the anti-inflammatory cytokine adiponectin decreases risk of prediabetes (24,31,32). Each 1-SD (∼5 μg/mL) higher baseline plasma adiponectin level predicted a 52% decrease in incident prediabetes (31). Other inflammatory markers associated with increased prediabetes risk include hepatic steatosis (33) and albuminuria (34).

Physical inactivity and unhealthy dietary patterns are associated with increased risk of prediabetes (32,34). Consumption of fruits, vegetables, dairy products, and coffee (3–4 cups/day), and decreased intake of fats and sugar-sweetened beverages, may be associated with decreased risk of prediabetes (32). The decreased risk of prediabetes associated with coffee consumption has been attributed to improved insulin sensitivity and beneficial effects of chlorogenic acids in coffee (32).

Hemodynamic Factors

Systolic and diastolic blood pressures correlate positively with FPG and 2hrPG levels, and prehypertension and hypertension predict increased prediabetes risk compared with risk associated with normal blood pressure status (35,36). Additionally, arterial stiffness and wide pulse pressure are associated with increased risks of prediabetes and diabetes (37,38). Environmental factors associated with dysregulation of both blood pressure and blood glucose include overweight/obesity, unhealthy diet, physical inactivity, and psychosocial stress. Genetic mechanisms for increased dual risks for T2D and hypertension have been proposed (39). The effects of arterial stiffness and pulse pressure on diabetes risk were mediated substantially (55%) by FPG, suggesting increased autonomic tone as a plausible mechanism linking hemodynamics to dysglycemia (37,40–42). Figure 3 shows the relationships between blood pressure and blood glucose and the association of blood pressure strata and insulin sensitivity/adiposity phenotypes with prediabetes risk.

Figure 3.

A and B: Relationships between baseline blood pressure and blood glucose values among normoglycemic African American (●) and European American (○) adults with parental history of T2D. C and D: Kaplan-Meier plots of prediabetes survival probability by baseline blood pressure categories (C) and baseline insulin sensitivity/adiposity status (D) among initially normoglycemic African American and European American adults with parental history of T2D during 5.5 years of follow-up. The cumulative incidence of prediabetes was 26.3% in the normal blood pressure (BP) group, 35.7% in the prehypertension group, and 42.3% in participants with baseline hypertension (log-rank P = 0.0015) (data from ref. 37). The cumulative incidence of prediabetes was 26.0% in insulin-sensitive nonobese (ISN), 30.9% in insulin-sensitive obese (ISO), 47.1% in insulin-resistant obese (IR0), and 48.7% in insulin-resistant nonobese (IRN) participants (log-rank P = 0.0001) (data from Edeoga et al. [37] and Owei et al. [43]).

Insulin Resistance and β-Cell Dysfunction

Insulin sensitivity and insulin secretion predicted progression to prediabetes in longitudinal studies (6,22,24). With stratification of normoglycemic adults by insulin sensitivity and obesity status, insulin-sensitive individuals had lower risk of incident prediabetes than insulin-resistant individuals, regardless of obesity status (43) (Fig. 3D). Pancreatic β-cell dysfunction, impaired disposition index, decreased β-cell glucose sensitivity, and lower insulin clearance all are associated with increased risk of prediabetes (6–9,22,44). In the POP-ABC study, African American participants had lower insulin sensitivity but markedly higher insulin secretion in comparison with their European American counterparts (44) (Fig. 4). The robust insulin secretion likely compensated for the lower insulin sensitivity among African American participants and contributed to the lack of ethnicity disparities in glycemic progression (24,32,44).

Figure 4.

A–D: Plasma glucose and insulin levels during hyperinsulinemic-euglycemic clamp (A–C) and insulin sensitivity values (D) in normoglycemic African American (red) and European American (blue) adults with parental history of T2D. E–H: Plasma glucose and insulin levels during intravenous glucose tolerance test following a 20-g dextrose bolus (E–G) and disposition index (H) in the same study population. The disposition index was calculated as the product of insulin sensitivity and acute insulin response (data from Edeoga et al. [44]). *P = 0.02, **P = 0.002, ***P = 0.0013.

Metabolites and Progression to Prediabetes

Amino Acids

Owei et al. (45) reported that each 1-SD higher baseline fasting plasma level of aspartic acid/asparagine was associated with a 2.7-fold higher risk in incident prediabetes during a 5.5-year follow-up. Additionally, each 1-SD increase in baseline histidine level predicted 10% lower risk of incident prediabetes (45). The association of histidine with lower prediabetes risk may be related to its reported anti-inflammatory, antioxidant, and satiety effects (32). The bidirectionality in the association of amino acids with dysglycemia probably is mediated by differential associations with insulin sensitivity and secretion (45) (Fig. 5).

Figure 5.

A–F: Association of selected plasma amino acid levels with insulin sensitivity (A, C, and E) and insulin secretion (B, D, and F). G: “Vicious” cycle between circulating branched chain amino acids (BCAA) and other amino acids (AAs) and development of insulin resistance. Plasma amino acid levels reflect the balance between dietary delivery/synthesis and clearance/catabolism. Insulin promotes protein synthesis and inhibits proteolysis, thus lowering circulating amino acid levels. States of impaired insulin secretion or action are permissive of hyperaminoacidemia, which can further worsen insulin resistance via activation of the molecular target of rapamycin (mTOR), AMP kinase (AMPK), and other signaling pathways. Insulin resistance, in turn, impairs amino acid catabolism and augments hyperaminoacidemia, thus triggering a vicious cycle that increases risk of dysglycemia. Gastric bypass (GBP) surgery decreases plasma branched chain amino acid levels and improves glucose tolerance independently of weight loss (data from Dagogo-Jack [32], Owei et al. [45], Wang et al. [46], Lu et al. [47], and Laferrère et al. [48]).

The specific aromatic and branched chain amino acids associated with diabetes risk in the Framingham cohort (46) were not associated with prediabetes in the more diverse POP-ABC cohort (45). Thus, the amino acid signatures could differ, based on the outcome of interest (diabetes or prediabetes) and the population studied. Insulin, a potent anabolic hormone, promotes protein synthesis and inhibits proteolysis; thus, states of impaired insulin action or secretion favor hyperaminoacidemia. Hyperaminoacidemia can aggravate insulin resistance via activation of the molecular target of rapamycin, adenosine monophosphate kinase, and other signaling pathways (32). Insulin resistance, in turn, impairs amino acid catabolism, further increasing circulating levels in a vicious cycle that leads to dysglycemia (47) (Fig. 5). Gastric bypass surgery decreases plasma branched chain amino acid levels and improves glucose tolerance independently of weight loss (48).

Lipids, Fatty Acid Metabolites, and Glucose Dysregulation

Plasma lipid profiles are associated with incident prediabetes risk: the relative risk (per 1 SD of baseline level) was 1.97 (95% 1.07–3.65) for LDL cholesterol, 1.63 (95% 1.03–2.57) for triglycerides, and 0. 46 (95% 0.23–0.91) for HDL cholesterol (49). HDL cholesterol decreases prediabetes risk by improving insulin sensitivity and secretion (49). Fatty acid derivatives (ceramides and other sphingolipids) increase risk of T2D by inducing insulin resistance and β-cell dysfunction. In the POP-ABC study, the saturated–to–monounsaturated ceramides C18:0/C18:1 and sphingomyelins C26:0/C26:1 ratio in baseline fasting plasma significantly predicted incident prediabetes, with odds ratio (per 0.1 unit of baseline ratio) of 1.236 (95% CI 1.042–1.466) and 2.273 (95% CI 1.172–4.408), respectively (50). These potential sphingolipid biomarkers for prediabetes showed significant associations with adiposity, insulin sensitivity, and insulin secretion (50).

Long-chain fatty acyl-CoA molecules derived from dietary fat are converted to acylcarnitines and transported into the mitochondrial matrix for oxidation. Plasma acylcarnitines levels reflect efflux from mitochondria during states of mitochondrial overload or incomplete fatty acid oxidation. Distinct patterns of circulating acylcarnitines in people with T2D and prediabetes reflect dysregulation of fatty acid oxidation in the setting of impaired insulin action or secretion (32). Lower baseline plasma levels of C4-OH (β-hydroxy butyryl) carnitine and higher levels of C8:1 (octenoyl) carnitine predicted progression from NGR to prediabetes during 5.5 years of follow-up (51). Plasma C8:1 (octenoyl carnitine) levels correlated inversely with insulin sensitivity and positively with insulin secretion (51).

Insights From “omics” Studies

Genome-wide association studies have identified >500 T2D-associated variants, including the rs7903146 risk allele in TCF7L2 (32,52). Homozygosity for the TCF7L2 rs7903146 risk allele (TT) confers approximately twofold increased odds of T2D versus homozygosity for the wild type (CC) (32,52). The rs7903146 risk variant conferred a risk for progression from NGR to prediabetes (odds ratio 2.648 [95% CI 1.326–5.291]) similar to the risk it conferred for progression from prediabetes to T2D (odds ratio 2.221 [95% CI 1.046–4.718]) in a prospective multiethnic population (52). A polygenic risk score built from T2D variants also was able to identify individuals at high risk for prediabetes (53).

For participants in the DPP placebo group with the TCF7L2 rs7903146 TT genotype there was 81% higher diabetes hazard versus for participants with the CC genotype (hazard ratio 1.81 [95% CI 1.21–2.70], P = 0.004) (54). Remarkably, lifestyle intervention markedly attenuated the diabetes hazard from the TT genotype (hazard ratio 1.15 [95% CI 0.68–1.94], P = 0.60) (54). The TT genotype is associated with impaired insulin secretion; thus, lifestyle-mediated improvement in insulin sensitivity reduced insulin demand in carriers of that genotype (54). Transcriptomic analyses indicate a direct correlation of global miRNA quantity with plasma glucose and HbA_1c levels, and specific miRNA expression profiles have been associated with increased risks for diabetes and prediabetes, via mechanisms that involve pancreatic β-cell survival (32).

The Gut Microbiome in Prediabetes

Differences in gut flora have also been reported among individuals with different glycemic states, with a lower abundance of Verrucomicrobia and a higher abundance of Anaerostipes being associated with prevalent prediabetes in comparison with NGR (32,55). Gut microbiota interact with host metabolic pathways in ways that could alter susceptibility to obesity, prediabetes, and diabetes. Gut bacteria play active roles in energy metabolism, generate branched chain amino acids and short-chain acylcarnitines, and modulate bile acid abundance and signaling via fibroblast growth factor-19 receptors. Through these activities, gut flora influence the pathogenesis of obesity, dysglycemia, and cardiometabolic disorders (32,55).

Progression From Prediabetes to T2D

Investigators for clinical trials with enrollment of high-risk populations with prediabetes (predominantly IGT) reported annual diabetes incidence rates of 7%−18% in the placebo arm during (17–21). Those high rates probably reflect the multiple risk factor burden of participants in the diabetes prevention trials (17–21). In the BLSA, with enrollment of participants with lower risk factor burden, the annualized rate of T2D among participants with IFG-IGT at baseline was 3.93% (23). Without intervention, progression to T2D is the likely outcome for most people with prediabetes, as long-term, sustained remission is uncommon. (17–21). After 30 years of follow-up of the participants enrolled with IGT in the Da Qing study, the cumulative incidence of T2D was 95.9% in the control group and 88.7% in the diet and exercise intervention groups (56). After 15 years of follow-up of DPP participants enrolled with prediabetes, the cumulative incidence of T2D was 62%, 55%, and 56% in the placebo, lifestyle intervention, and metformin treatment groups, respectively (57). The risk factors for progression from prediabetes to T2D include higher baseline FPG and 2hrPG, overweight and obesity, family history of diabetes, insulin resistance, and impaired insulin secretion (2,6–8,17,18). In a longitudinal study of Pima Indians, progression from IGT to T2D was associated with weight gain of 13 kg (vs. 6 kg in nonprogressors), ∼30% decline in insulin sensitivity, and >50% decline in insulin secretion during a 5-year follow-up (6). Table 1 summarizes some factors associated with transition from NGR to prediabetes.

Table 1.

Predictors of incident prediabetes among initially normoglycemic adults

Demographic/anthropometric/behavioral
Older age
Male sex
Overeating
Physical inactivity
Higher BMI/body fat
Higher waist/abdominal fat
Metabolomic factors
LDLc, HDLc, triglycerides
Amino acids (aspartic acid, glutamic acid, histidine)
Acylcarnitines (C8:1 and C4-OH carnitines)
Sphingomyelins C26:0-to-C26:1 ratio
Ceramides C18:0-to-C18:1 ratio
Insulin sensitivity and secretion
Upper-normal FPG and 2hrPG
Lower insulin sensitivity
Impaired insulin secretion
Lower disposition index
β-Cell glucose insensitivity
Decreased insulin clearance
Hemodynamic/inflammatory factors
Higher blood pressure
Higher pulse pressure
Higher C-reactive protein
Lower adiponectin
Albuminuria
Hepatic steatosis, ALT, AST
“omics”
GWAS diabetes-associated variants
Transcriptomics (miRNAs)
Gut microbiome

Data are from references 32–41,43–45,49–55. GWAS, genome-wide association studies; HDLc, HDL cholesterol; LDLc, LDL cholesterol. 2hrPG from oral glucose tolerance test.

Microvascular and Macrovascular Complications in People With Prediabetes

The microvascular complications of diabetes (retinopathy, neuropathy, and nephropathy) can present in people with prediabetes. Approximately 7%–15% of people with prediabetes have evidence of diabetes retinopathy, 10% have chronic kidney disease, and 8%–16% have peripheral polyneuropathy (2,32,58). Although the estimates for microvascular complications in people with prediabetes emanated from studies that lacked normoglycemic control groups, systematic reviews/meta-analyses provide supportive data (59,60). In a meta-analysis of nine cross-sectional, population-based studies (N = 14,751 adults, of whom 3,847 [26.1%] had prediabetes), the odds ratio for retinopathy was 1.55 (95% CI 1.10–2.20) for the prediabetes groups versus normoglycemic control (60). Increases in the risks of macrovascular complications (coronary artery disease, myocardial infarction, congestive heart failure, stroke, peripheral vascular diseases, and cardiovascular death) have been reported in the prediabetes state in comparison with normoglycemia (2,32,58).

Role of Hyperglycemia

The mechanisms linking hyperglycemia to the microvascular complications of diabetes include alterations in the polyol, hexosamine, and protein kinase C (PKC) pathways, advanced glycosylation end products, glomerular hyperfiltration, and inflammatory and oxidative stress, among others (61). Activation of PKC by hyperglycemia induces downstream toxic pathways that result in endothelial dysfunction, increased permeability, extracellular matrix deposition, prothrombotic state, inflammation, and generation of reactive oxygen species (62). Theoretically, activation of these toxic processes at subdiabetes glucose levels might explain the occurrence of “diabetes complications” in susceptible people with prediabetes (58–61). However, the susceptibility factors remain to be elucidated. People with prediabetes often have comorbidities, such as overweight/obesity, hypertension, dyslipidemia, and proinflammatory state, that increase the risk for macrovascular complications independently of glycemia (58).

Interventions for Prediabetes

Prevention or Delay of T2D

The efficacy of lifestyle modification in preventing progression from prediabetes to T2D has been demonstrated in RCTs (17–21). In these RCTs individuals with prediabetes were enrolled and primary results were reported after an active intervention period ranging from approximately 3 to 6 years. The lifestyle intervention was focused on dietary modification, increased physical activity (∼150 min/week), and weight loss in participants with overweight/obesity and resulted in 30%–58% relative reduction in T2D risk (17–21). Weight loss was not uniformly reported across the prevention trials. Mean BMI of the participants in the Asian studies was ∼26 kg/m², and no weight loss was reported after lifestyle intervention (17,20,21). Despite the lack of weight loss, diabetes incidence was reduced by ∼30% in the lifestyle intervention group versus control. Elements of lifestyle intervention, including dietary modification, improved fitness, and body fat redistribution, probably contributed to the favorable results. After the active intervention phase, long-term follow-up studies documented sustained benefits of prior lifestyle intervention (29,56,57). A meta-analysis of 44 RCTs of lifestyle intervention in 14,742 participants with prediabetes reported a dose-response relationship between weight loss and decreased risk of progression to T2D (63). The participants with prediabetes in the 44 studies were selected with use of different criteria (IFG or HbA_1c in 2 studies, IFG in 5, IGT in 19, and IFG or IGT in 18), but there were no significant differences in outcomes based on the definition of prediabetes (63). In another meta-analysis, of 11 RCTs (N = 5,224 adults with prediabetes), investigators found that lifestyle intervention decreased risk of T2D by 36% in comparison with usual care (63).

Regarding medications, the efficacy of acarbose, metformin, and orlistat in preventing progression to diabetes was weaker than that of lifestyle intervention, whereas the efficacy of thiazolidinediones and glucagon-like peptide 1 receptor agonists (GLP-1RA) matched or exceeded that of lifestyle intervention (2,32,64–68). Unlike the association of weight gain with thiazolidinediones, GLP-1RA induce weight loss. In exploratory analyses of data from obesity intervention studies with enrollment of individuals without diabetes, treatment with GLP-1RA (vs. placebo) along with lifestyle modification was associated with substantial reductions in risks of incident prediabetes and T2D (64–68). Those glycemic benefits were sustained during chronic GLP-1RA treatment, and the medications were well tolerated, the most frequent adverse events being gastrointestinal symptoms (64–68). Treatment with tirzepatide (a dual agonist of GLP-1 and glucose-dependent insulinotropic peptide [GIP] receptors) decreased risk of T2D in individuals with both obesity and prediabetes (69). Study participants (n = 1,032 adults) were assessed during 176 weeks of active treatment. Tirzepatide induced dose-dependent weight loss ranging from 12.3% (5-mg dose) to 19.7% (15-mg dose); T2D incidence was 1.3% in the tirzepatide groups vs. 13.3% in the placebo group (69). Gastrointestinal symptoms were the most frequent adverse events (69). The risk reduction for prediabetes and T2D associated with GLP-1 and dual GLP-1/GIP receptor agonists was almost certainly mediated by weight loss, which ranged from 8% to 20% (64–69). The sodium–glucose cotransporter 2 (SGLT2) inhibitors showed a potential for diabetes prevention (∼21% risk reduction) in subgroups analysis of RCTs with enrollment of individuals with prediabetes and heart failure or chronic kidney disease (70). Apart from modest weight loss, the mechanism(s) for the apparent diabetes prevention effect are unclear, as HbA_1c levels were unchanged during the trials (70).

In studies that included a medication “washout” phase, glycemic rebound and decay in efficacy for diabetes prevention were noted (2,32,68,71–73). In the DPP, among participants in the metformin treatment arm who had not developed diabetes at the end of the study, glycemic rebound occurred, and new cases of diabetes occurred equally in the metformin and placebo groups within 1–2 weeks of stopping metformin (71). A similar loss of efficacy has been reported after stopping rosiglitazone and pioglitazone treatment in studies where those drugs had shown potent reductions in the risk of progression from prediabetes (72,73). Cessation of treatment with GLP-1 and dual GLP-1/GIP receptor agonists was associated with weight regain and glycemic rebound (64–69). In one report, for nearly half of the initial responders there was a rebound in weight and relapse of prediabetes within 26 weeks of stopping the GLP-1RA semaglutide (68). Further, T2D incidence increased slightly, from 1.3% (on therapy) to 2.4% (off therapy), in the tirzepatide group during a 17-week washout study, while remaining stable in the placebo group (from 13.3% to 13.7%) during the same period (69). The “washout” studies indicate that continuous administration of medications might be needed to maintain prevention, or delay, of diabetes. The cumulative costs and adverse effects (including weight gain, fluid retention, gastrointestinal upset, pancreatitis, heart failure, skeletal fracture risk) of medications make them unappealing as the initial or sole strategy for prevention of T2D.

Impact on Microvascular and Macrovascular Complications

After 15–21 years of follow-up of participants in the Diabetes Prevention Program Outcomes Study (DPPOS), no significant differences were observed in the incidence of aggregate microvascular or major cardiovascular events in the lifestyle intervention or metformin-treated groups versus placebo (57,74). However, after 30 years of follow-up of participants in the Da Qing study, significant reductions in microvascular and cardiovascular complications were observed in the lifestyle modification groups versus the usual care group (56). Thus, prolonged follow-up may be required for the full benefits of lifestyle intervention to evolve. Investigators in a secondary prevention trial targeting insulin resistance with pioglitazone in individuals with prediabetes and prior history of ischemic stroke reported significant decreases in the risks of recurrent stroke, myocardial infarction, and progression to T2D in the drug-treated group versus placebo control (75).

Reversal of Prediabetes and Restoration of NGR

Based on data from RCTs, most people with prediabetes are likely to develop T2D over the long term (56,57). During shorter-term follow-up, however, many will have persistent prediabetes and some may even revert to NGR (76). People with persistent prediabetes face significant risks of developing vascular and neuropathic complications (2,32,58). Among participants who received lifestyle intervention for incident prediabetes in the Pathobiology and Reversibility of Prediabetes in A Biracial Cohort (PROP-ABC) study, 42.8% reverted to NGR, 50% had persistent prediabetes, and 7.2% progressed to T2D during 5 years of follow-up (77). Diabetes prevention interventions have generally been more successful at delaying progression to T2D than inducing regression to NGR (76) (Table 2). Data from the DPPOS showed that participants who attained NGR, albeit transiently, experienced a 56% reduction in 6-year incidence of diabetes (78). There was a dose-response effect: reaching NGR once, twice, or three times during follow-up was associated with 47%, 61%, or 67% reduction, respectively, in the rate of incident diabetes (78).

Table 2.

Prevention of T2D and reversal of prediabetes in RCTs using lifestyle and/or medication interventions

Study	Duration (years)	Interventions	T2D RRR	% with prediabetes reversal
Malmo	10	Lifestyle vs. placebo	63%	52.2%
Da Qing	6	Lifestyle vs. placebo	42%	ND
FDPS	3	Lifestyle vs. placebo	58%	ND
DPP	2.8	Lifestyle vs. metformin vs. placebo	Lifestyle 58%, metformin 31%	Lifestyle 40%, metformin 20%
STOP-NIDDM	3.3	Acarbose vs. placebo	25%	35%
DREAM	3	Rosiglitazone vs. placebo	60%	50.5%
ACT NOW	2.4	Pioglitazone vs. placebo	72%	48%
IDPP-1	3	Lifestyle ± metformin vs. placebo	Lifestyle 28.5%, metformin 26.4%	ND
CANOE	3.9	Rosiglitazone + metformin vs. placebo	65.5%	79.6%
SCALE	3	Lifestyle + liraglutide vs. lifestyle + placebo	80%	66%
STEP 10	1	Lifestyle + semaglutide 2.4 mg s.c. vs. lifestyle + placebo	66.7%	81%

Data from references 2,17–21,32,65, and 67. ACT NOW, Actos Now for the prevention of diabetes; CANOE, CAnadian Normoglycemia Outcomes Evaluation;  DREAM, Diabetes REduction Assessment with ramipril and rosiglitazone Medication; FDPS, Finnish Diabetes Prevention Study; IDDP-1, Indian Diabetes Prevention Programme-1; Malmo, Malmö feasibility study; ND, no data; RRR, relative risk reduction vs. placebo; SCALE, SCALE Obesity and Prediabetes Trial; STEP 1, Semaglutide Treatment Effect in People with obesity 1; STOP-NIDDM, Study to Prevent Non-Insulin-Dependent Diabetes Mellitus.

The predictors of reversal to NGR included weight loss, lower baseline FPG and 2hrPG, younger age, and insulin secretion and sensitivity (78,79). Regression from prediabetes to NGR decreased the risks of microvascular and macrovascular complications and mortality (79–81). In exploratory analyses, weight loss of ∼15% following treatment with semaglutide was associated with reversion to normoglycemia in 80%–90% of participants with prediabetes (67,68). However, ∼50% of responders experienced a rebound in weight and relapse of prediabetes within 26 weeks of stopping semaglutide (68). Despite the impressive results of lifestyle and pharmacological intervention in diabetes prevention trials, the prediabetes state persisted in ∼50% of participants (76,79). Because persistent prediabetes increases risk of vascular complications, the goal of intervention ought to shift from diabetes prevention to prediabetes reversal.

Limitations of Current Approaches to Diabetes Prevention

Limitations of Lifestyle Intervention

The landmark randomized controlled studies of lifestyle intervention for diabetes prevention were designed as efficacy trials that involved frequent in-person visits by participants and a sizeable multidisciplinary team of interventionists and other research staff (17–21). Although substantial resources were consumed by the efficacy trials, the findings have been successfully translated to community settings with fewer resources (32,76,82,83). However, maintaining the 5%–7% weight loss required for diabetes prevention long-term is challenging. Weight regain due to physiological adaptations to the initial weight loss and nonadherence to behavioral interventions could negate the diabetes prevention benefits of lifestyle modification.

Limitations of Medications

Cost considerations, decay in efficacy following cessation of medications, and drug-related adverse events constitute the main argument against recommending medications as the primary or sole intervention for diabetes prevention. The rebounds in weight and glycemia observed after withdrawing agents from the different medication classes indicate a lack of fundamental impact on the underlying pathophysiology of prediabetes (5–15). The cumulative costs of uninterrupted administration of medications could be prohibitive, particularly for low- and middle-income countries.

Although current realities do not support the use of drugs as a first-line approach to diabetes prevention, there is a need for safe and effective medications for diabetes prevention, given the limitations of lifestyle intervention. The ideal medication for diabetes prevention should be efficacious, well tolerated, safe, and impactful on the key pathophysiological defects in prediabetes (i.e., insulin resistance and β-cell dysfunction). Importantly, the effects of such a drug should endure after drug therapy is discontinued, indicating a fundamental impact on pathophysiology. As no drug currently meets these ideal properties, there is an opportunity for novel drug discovery, possibly through the deployment of artificial intelligence.

Proposal for Lifestyle Intervention Plus Medication Strategy

Previous trials combining lifestyle intervention with metformin (250 mg twice daily) or pioglitazone (30 mg once daily) in Indians with prediabetes failed to show additive benefits (20,21). However, a strategy of cyclical use of the more potent GLP-1 and GLP-1/GIP receptor agonists in combination with lifestyle intervention is conceivable. Under such a therapeutic strategy, combined lifestyle intervention and GLP-1RA or a dual GLP-1/GIP receptor agonist is used to induce weight loss and regression from prediabetes to NGR. Thereafter, the drug is withdrawn, and the foundational lifestyle intervention is continued. Individuals are then observed for evidence of continued remission or relapse of prediabetes during short- to medium-term (3–6 months) follow-up. Individuals with suboptimal response receive a second cycle of medication superimposed on continued lifestyle intervention. Two to three such cycles could be tried before abandoning the strategy. A study for evaluating the efficacy of such a combined regimen comprising foundational lifestyle modification plus cyclical use of medication would be a valuable contribution to the field of diabetes prevention.

Discussion

Prediabetes is associated with the risks of progression to T2D and the development of vascular and neuropathic complications. Factors associated with development of prediabetes include dietary and physical activity habits, adiposity, insulin resistance, β-cell dysfunction, inflammation, hemodynamics, circulating metabolites, gut microbiome, and genomic and transcriptomic profiles (Fig. 6). These and other emerging risk factors could be useful for generating risk engines that predict glycemic trajectories among individuals with NGR. The increased nosological insight could advance personalized medicine by facilitating the design of targeted interventions for diabetes prevention, reversal of prediabetes to NGR, and avoidance of dysglycemia complications.

Figure 6.

In genetically susceptible individuals, the development of prediabetes is associated with adiposity, dietary and physical activity habits, impairments in insulin sensitivity and pancreatic islet β-cell dysfunction, and several interrelated factors, including hemodynamics, circulating amino acids, lipids, fatty acid metabolites, and gut microbiome, among others (32). BP, blood pressure; HDLc, HDL cholesterol; LDLc, LDL cholesterol.

Funding Statement

The author is supported, in part, by research grants from the National Institute of Diabetes Digestive and Kidney Diseases (grant R01 DK128129). The POP-ABC study was supported by a grant from the National Institute of Diabetes Digestive and Kidney Diseases (R01 DK067269).