Neuropathy in prediabetes

Diabetic neuropathies are the most frequent complications of diabetes, contributing to high morbidity, excess mortality, reduced quality of life, and increased healthcare costs. Prediabetes is characterised by glucose levels within an intermediate range above normoglycaemia yet below the diagnostic threshold for diabetes. In 2021, 10.6% and 6.2% of adults worldwide were estimated to have impaired glucose tolerance and impaired fasting glucose, respectively, the majority of whom are unaware of having prediabetes. Evidence has accumulated suggesting that prediabetes is a predictor of cardiovascular disease (CVD) and increased mortality. The past 2 decades have witnessed an extensive debate, particularly among diabetologists and neurologists, as to whether prediabetes is associated with peripheral neuropathy. In this review, we elaborate on the current evidence, particularly from population‐based studies supporting an increased risk of distal sensorimotor polyneuropathy (DSPN) and cardiovascular autonomic neuropathy (CAN) in people with prediabetes. Moreover, we discuss whether lifestyle interventions showing efficacy in preventing or delaying the transition from prediabetes to diabetes in persons with prediabetes may also exert favourable effects on the development and progression of DSPN and CAN. This review should help in raising the awareness of and translating the current knowledge on neuropathies in people with prediabetes into clinical practice and public health. The current recommendation that adults who are overweight or obese should be screened for prediabetes and referred to or offered preventive interventions should ultimately culminate in preventing not only CVD but also prediabetic neuropathy.


| PREDIABETES: CLASSIFICATION, EPIDEMIOLOGY, IMPACT AND SCREENING 1.| Classification
The term "prediabetes" refers to individuals whose glucose levels lie within an intermediate range above normoglycaemia below the diagnostic threshold for diabetes but feature impaired glucose metabolism.Laboratory parameters to identify prediabetes comprise fasting blood glucose (FBG), 2-h post-load blood glucose, or HbA1c. 1 The concept of prediabetes dates back to 1979, when the National Diabetes Data Group defined impaired glucose tolerance (IGT) using 2-h post-glucose load values in a 75-g oral glucose tolerance test (OGTT) of 140-199 mg/dL or 7.8-11.1 mmol/L.This definition was adopted by the American Diabetes Association (ADA) and the World Health Organization (WHO).In 1997 the ADA and in 1998 the WHO implemented impaired fasting glucose (IFG) as a further category based on FBG levels of 110-125 mg/dL or 6.1-6.9 mmol/L (IFG WHO ).
In 2003, the ADA reduced the lower fasting plasma glucose (FPG) threshold, widening the range from 110 to 125 mg/dL to 100-125 mg/dL or 5.6-6.9mmol/L (IFG ADA ).Subsequently, in 2009, an additional prediabetes definition based on HbA1c levels from 6.0% to 6.4% or 42-46 mmol/mol was proposed by the International Expert Committee (IEC) (HbA1c IEC ), and in 2010 the ADA widened the range to 5.7%-6.4% or 39-46 mmol/mol (HbA1c ADA ).However, both the lower diagnostic thresholds for FPG and HbA1c were not endorsed by the WHO and IEC. 2,3Thus, five definitions currently exist to diagnose prediabetes (Table 1).Apart from these categorical obstacles, the various glycaemic measures do not identify the same individuals at risk.A recent report from the U.S. National Health and Nutrition Examination Survey 2005-2016 estimated that mismatches of clinically significant magnitude could lead to HbA1c-related misdiagnosis in up to 30 million people in the USA with possible adverse consequences from inappropriate management. 4e pathophysiology underlying IFG and IGT shows several differences.While IFG is primarily characterised by reduced hepatic insulin sensitivity, stationary defective beta-cell function and/or chronic low beta cell mass, IGT shows reduced peripheral insulin sensitivity, near-normal hepatic insulin sensitivity, and progressive loss of beta-cell function. 5Individuals with combined IFG + IGT exhibit severe defects in both peripheral and hepatic insulin sensitivity as well as a progressive loss of beta cell function.Furthermore, IFG is predominantly related to genetic factors, smoking and male sex, while IGT is predominantly related to physical inactivity, unhealthy diet, and short stature. 5More recently, evidence has emerged indicating that the current classification of diabetes may not sufficiently account for its pathophysiological heterogeneity.In this context, the recent years have witnessed efforts to introduce novel diabetes subgroup categories to refine the classification of diabetes beyond the current definition of diabetes types aimed at optimising the risk stratification and treatment in terms of precision medicine to ultimately improve patient outcomes. 6,7For example, individuals with recent-onset type 2 diabetes (diabetes duration ≤1 year) had an increased risk of distal sensorimotor polyneuropathy (DSPN) if they were severely insulin deficient, 8 while those who were either insulin resistant or insulin deficient had an increased risk of erectile dysfunction (ED), a frequent manifestation of autonomic neuropathy in diabetes. 9Of note, a recent meta-analysis reported that ED is associated with prediabetes independent of the men's age. 10 Further evaluation of the present approaches and novel integrated methods to redefine diabetes subtypes will be required to determine when and how best to introduce these approaches into clinical practice. 11More recently, a first such effort has been undertaken in persons carrying an increased risk of type 2 diabetes (history of prediabetes [overall 41.2%: IFG 18.7%, IGT 10.9%, IFG + IGT 11.6], family history of diabetes or gestational diabetes, or body mass index (BMI) >27 kg/m2 ) from the TUEF/TULIP study.Six distinct subphenotypes (clusters) were identified using deep phenotyping, including body fat distribution, liver fat content and genetic risk: (1.) low risk, (2.) very low risk, (3.) ß-cell failure, (4.) low-risk obese, (5.) high-risk insulin-resistant fatty liver, and (6.) high risk visceral fat nephropathy.While clusters 3, 5 and 6 were associated with higher blood glucose at baseline than the other clusters, only clusters 3 and 5 were prospectively associated with an increased risk of type 2 diabetes.Moreover, clusters 3, 5 and 6 were associated with an increased risk of nephropathy, while clusters 5 and 6 were linked to increased all-cause mortality.Cluster 5 was identified as having the highest risk of type 2 diabetes, renal and vascular disease, and all-cause mortality.The results were replicated using simpler markers of similar anthropometric and glycaemic constructs in the Whitehall II cohort.While such a subphenotyping approach is currently not aimed to classify patients in clinical practice, it is helpful to decipher the metabolic heterogeneity prior to the development of type 2 diabetes, 12 and its further validation and fine tuning could represent one step forward towards precision medicine.

| Epidemiology
According to the International Diabetes Federation Diabetes Atlas, 13 in 2021 an estimated 537 million people had diabetes, and this number is projected to reach 783 million by 2045.However, 44.7% of adults living with diabetes (20-79 years old) were found to be highest in low-income countries (12.7%), followed by high-and middle-income countries (10.4% and 10.0%).In contrast, the ageadjusted prevalence estimates of IFG in 2021 were similar across high (5.7%),middle (5.7%), and low-income (5.8%) countries. 13cording to the Centers for Disease Control and Prevention 2020 National Diabetes Statistics Report, an estimated 34.5% of all US adults (18 years or older) met the criteria for prediabetes with older people being affected more frequently. 14The degree of unawareness is very high as only 15% of people with prediabetes reported being told by a healthcare professional to have that condition (statistics report).A large cohort study including 77,107 individuals with prediabetes reported that the risk of developing diabetes increased with increasing HbA1c levels and BMI. 15 Other risk factors include older age, family history, history of gestational diabetes, history of polycystic ovarian syndrome, and dietary and lifestyle factors. 16Estimates of the risk of the transition from prediabetes to diabetes vary considerably due to differences in the definition or the heterogeneity of prediabetes. 17The risk of progression from prediabetes to type 2 diabetes has been assessed by a Cochrane review including 103 prospective cohort studies.Overall, the cumulative progression to type 2 diabetes increased with the length of follow-up.
The cumulative incidence rates and hazard ratios of type 2 diabetes are shown in Supplemental Table 2.These data suggest that the cumulative progression to type 2 diabetes is relatively high during the first 5 years, while after 10 years there is a further slight increase only for IFG and HbA1c ADA , whereas the incidence decreases for the remaining categories.Subgroup analyses did not reveal any clear evidence for differences between geographic regions. 2 Thus, the risk of developing type 2 diabetes according to the prediabetes definition is generally higher when using the more stringent definitions.On the other hand, regression from prediabetes to normoglycaemia within one to five years ranged from 33% to 59% and from 17% to 42% for 6-11 years of follow-up in the 47 studies reporting the corresponding data, suggesting that prediabetes is also reversible in a considerable proportion of individuals. 2recent meta-analysis including 16 studies reported that when using the WHO criteria, the incidence of type  18 Thus, the combined IGT + IFG category was associated with the highest incidence of type 2 diabetes when compared to the single categories and higher for the WHO definition.

| Impact
There is accumulating evidence suggesting that prediabetes is an independent predictor of both cardiovascular disease (CVD) and mortality.A recent umbrella review including 95 meta-analyses of prospective studies reported that in the general population, prediabetes was associated with a 6%-10% increased risk of all-cause mortality and the incidence of cardiovascular outcomes, coronary heart disease, stroke, heart failure, atrial fibrillation, chronic kidney disease as well as total cancer, liver cancer, breast cancer and allcause dementia with moderate certainty of evidence.The risk of all-cause mortality was generally higher for prediabetes defined by IGT than for prediabetes defined by IFG and HbA1c. 19Another metaanalysis reported that in the general population, prediabetes was associated with an increased risk of all-cause mortality (RR 1.

| DIABETIC POLYNEUROPATHY
Diabetic neuropathy affecting the somatic and/or autonomic components of the peripheral nervous system develops as a consequence of diabetes and cannot be attributed to other causes of peripheral neuropathy.9][30] Chronic peripheral neuropathic pain is defined as persistent or recurrent pain lasting ≥3 months caused by a lesion or disease of the peripheral somatosensory nervous system. 31Neuropathic pain due to diabetes is defined as pain arising as a direct consequence of abnormalities in the somatosensory system in people with diabetes after exclusion of other causes. 28,32Chronic painful DSPN is observed in up to one fourth of people with diabetes, while up to 50% of those with DSPN may be asymptomatic.
Sensory symptoms and deficits due to DSPN are associated with reduced quality of life and increased healthcare costs and predict all-cause mortality and future neuropathic foot ulcerations as well as cardiovascular morbidity and mortality in diabetes patients. 30,33dside tests to screen for large nerve fibre dysfunction due to DSPN include vibration and pressure sensation and ankle reflexes, while small nerve fibre dysfunction is detected by examining pain and temperature perception.Distal sensorimotor polyneuropathy diagnosis can be confirmed by nerve conduction studies (NCS) as an objective measure of large fibre dysfunction, while small nerve fibre pathology is confirmed by measuring intra-epidermal nerve fibre density (IENFD) obtained from skin biopsy.Small and large fibre damage most frequently coexist in patients with DSPN. 30Timely testing of both aspects is equally important since parallel damage to small and large fibres with some progression yet potential regression occurs early in the course of diabetes. 34In the German Diabetes Study (GDS), we demonstrated that recent-onset type 2 diabetes (diabetes duration up to 1 year) is characterised by an array of abnormalities in small and large fibre function and structure.Thus, DSPN can no longer be viewed as a "late" but an "early" diabetic complication. 35

| DIABETIC CARDIOVASCULAR AUTONOMIC NEUROPATHY
Diabetic autonomic neuropathy may affect any organ innervated by the autonomic nervous system (ANS).The heart receives input from both the sympathetic and parasympathetic systems regulating HR, rhythm, and contractility.The cardiac parasympathetic activity is mediated through the vagus nerve, which originates in the medulla oblongata, while sympathetic innervation to the heart originates mainly from the right and left stellate ganglia.The ANS regulates both the cardiovascular system and energy balance and is disturbed in diabetes and obesity.The sympathetic nervous system is a key player linking the development and progression of CVD with obesity. 35rdiovascular autonomic neuropathy (CAN), defined as the impairment of cardiovascular autonomic control in the setting of diabetes after exclusion of other causes, affects approximately 20% of people with diabetes.It is a serious complication that may result in sinus tachycardia, severe postural hypotension, exercise intolerance, enhanced intra-operative instability as well as increased incidence of silent myocardial infarction or ischaemia and may even predict the development of stroke. 36According to a meta-analysis, the risk of mortality is two-fold increased in diabetes patients with CAN compared with those without CAN. 37Risk factors of CAN include higher age, obesity, hypertension, diabetes duration, poor glycaemic control, and smoking. 38In addition, recent evidence from the GDS points to a role of insulin resistance, hepatic steatosis, abnormal lipid metabolism, and blunted cardiorespiratory fitness (CRF) in the early development of diminished cardiovagal tone in recent-onset type 2 diabetes. 35Subclinical inflammation also plays a role. 39Interleukin 18, total and high-molecular-weight adiponectin, soluble intercellular adhesion molecule-1, soluble E-selectin, serum C-reactive protein and interleukin-6 were found to be associated with parameters of CAN. 39e hallmark and earliest indicator of CAN in diabetes is

| PREDIABETES IN NEUROPATHIES
4][45][46][47][48][49][50][51][52] The majority of these studies claimed that such an association could be explained by the high rates of prediabetes among the groups with these idiopathic neuropathies, but several authors cast doubt on this notion.It has been criticised that there is persisting uncertainty about this association, given its retrospective and uncontrolled nature. 45,50Caution has also been urged as to the use of historical reference groups, which may not be suitable for comparison 43,50 given the variable prevalence rates of prediabetes in the general population, 47 concluding that the type, severity, frequency, and timing of neuropathy associated with prediabetes remain essentially unknown. 46For these reasons, studies reporting the prevalence of prediabetes in individuals with idiopathic neuropathies were not considered in this review because of the different topic and multiple potential sources of bias.However, given the increased prevalence of DSPN in the general population with prediabetes (see next section), routine screening for IFG and IGT in patients with "idiopathic" neuropathy is recommended in clinical practice. 53,54

| POLYNEUROPATHY IN PREDIABETES
A recent systematic review including 29 studies (9351 participants) aimed to determine the prevalence of DSPN in adults with prediabetes and to evaluate how prevalence estimates are influenced by the method of neuropathy assessment.The DSPN prevalence estimates for the population with prediabetes in each study are shown in Figure 1. 55Prevalence estimates varied widely from 2% (95% CI: 0%-4%) in a study conducted in the USA in women 56 to 77% (95% CI: 54%-100%) in a study conducted in Brazil. 571][62][63] The overall conclusion was that most studies reported a higher prevalence of peripheral neuropathy in prediabetes, primarily of a small nerve fibre origin, than would be expected in the background population. 55In contrast to the present review, Kirthi et al. 55 differentiated neither between hospital-based and population-based studies nor between the different categories or definitions of prediabetes.However, such a differentiation is crucial due to selection bias inherent in hospital-based studies and the possible impact of prediabetes categories on DSPN prevalence.

| Population-based studies
The details of 16 studies in populations with prediabetes, diabetes, and normal glucose tolerance (NGT) are shown in chronological order in Table 2. 51,56,60,[62][63][64][65][66][67][68][69][70][71][72][73][74] Since there are no uniform diagnostic criteria for both DSPN and painful DSPN, the definitions and methods used for this purpose varied considerably across studies as did the criteria for the diagnosis of prediabetes.For example, if DSPN was defined as "possible" by the Toronto criteria, 28 in the study by Kurisu et al.
(2019) the prevalence reached 17.4% in the NGT population, but if it was defined by more stringent criteria as "probable" or "confirmed" the prevalence dropped to 4.2% for these definitions combined. 69other example is the study by de Neeling et al. (1996) in which bilateral absence of ankle reflexes was used as a single criterion to diagnose DSPN resulting in a prevalence as high as 16% in the NGT population, 65 but it is known that the ankle reflex may be absent in a considerable proportion of the elderly healthy population. 75Another potential source of bias may be the variable definitions of prediabetes using different cut-off values (e.g.American Diabetes Association vs. World Health Organization) without a clear distinction between IFG and IGT.This is particularly problematic, since the two conditions show differences in their underlying pathophysiology. 5spite these hurdles and discrepancies and the fact that the prevalence rates ranged from 1% to 16% in the NGT population ( The only study to report separately on i-IGT and IFG + IGT was the KORA F4 study 66 in which the prevalence of DSPN was 14.8% in the i-IGT and 23.9% in the IFG + IGT population, respectively.The KORA surveys 66,70,72 demonstrated markedly increased prevalence of painful DSPN in IGT � IFG, which was only slightly increased in i-IFG as compared to NGT, again suggesting that it is post-prandial hyperglycaemia and IGT rather than IFG that is associated with an increased risk.However, it must be kept in mind that some IGT cases also had IFG.Overall, there is limited evidence only from the KORA F4 study suggesting that DSPN prevalence was particularly high in combined IFG + IGT similar to the k-DM population.The prevalence of DSPN was approximately 3-fold higher in the n-DM group than in the NGT group.Notably, the prevalence of painful DSPN was markedly higher in IGT � IGT, whereas it was only slightly higher in i-IFG compared with the NGT population. A few studies found no evidence for an association of prediabetes with DSPN.In the Rotterdam Study 67 , type 2 diabetes, metabolic syndrome, waist circumference, and elevated triglycerides but not IFG were associated with DSPN.However, OGTT was not performed.In the OC IG survey, prediabetes was not associated with increased prevalence of typical or atypical DSPN or atypical neuropathy.Impaired fasting glucose and IGT were not considered separately, and the same applies to the two DSPN types. 64Likewise, based on positive sensory and pain symptoms and QST, this study found no increase in small fibre sensory DSPN in prediabetes. 51Thus, future studies will require a clear distinction between the different prediabetes categories.In the future, the OC IG study, 46 the 7-year KORA F4-FF4 follow-up, 76  nerve damage. 77Finally, in the Maastricht study, 78 hyperglycaemia (even at levels lower than diabetes mellitus (DM)) was linked with early nerve fibre dysfunction.Waist circumference, inflammation, hypertension and smoking appeared to contribute to this dysfunction.

| Hospital-based studies
A number of hospital-based studies point to an increased prevalence of DSPN in prediabetes (Figure 1 82 Another small study using diffusion tensor imaging of the sciatic nerve showed proximal nerve damage in individuals with diabetes and prediabetes. 83These preliminary findings need validation in larger scale studies.

| CARDIOVASCULAR AUTONOMIC NEUROPATHY IN PREDIABETES
The question as to whether prediabetes is associated with CAN has received less attention than DSPN in this context, 84 85 Overall, this systematic review concluded that there is a higher than expected prevalence of CAN in prediabetes.Unlike Eleftheriadou et al., 85 here we focus on the prevalence estimates of both CAN and OH reported in three population-based studies.
Moreover, we considered seven population-based cross-sectional studies reporting mean values of both HRV and BRS indices in people with prediabetes.Hospital-based studies reporting the prevalence of CAN in diabetes are not taken into account due to selection bias.

| Population-based studies
Table 3 shows the details of the three population-based crosssectional studies reporting the prevalence of CAN and OH.The KORA S4 study 86  factors associated with reduced HRV, including HR, BMI, hypertension, smoking, serum creatinine, and use of drugs suppressing HRV, were integrated in a simple score to screen for CAN in clinical practice. 86In contrast, the prevalence of OH was not increased in people with prediabetes. 87,88More large-scale studies are needed to confirm the KORA S4 data.A prospective 14-year KORA S4-FF4 follow-up will determine the long-term natural history of CAN in prediabetes.
8][89][90][91][92][93] All studies used time domain HRV indices, while 3 studies each used additional frequency domain parameters [89][90][91] and spontaneous BRS. 89,93,94All studies reported significantly lower mean values at least for some of the time and frequency domain HRV indices in IFG or IGT compared to NGT, while this was the case for BRS in one study only. 94One study reported only a significant trend across three groups. 89However, the majority of these studies did not correct for multiple testing.Altogether, these data support the notion that prediabetes is associated with blunted cardiovagal activity rather than baroreflex dysfunction and also supports the aforementioned findings of the KORA S4 study. 86However, prospective follow-up of these cohorts is not available.
Given that early diabetic CAN is associated with CRF, 95 it is important to consider CRF in the context of prediabetes.Examining participants who had CRF measured from young adulthood to middle age, the CARDIA study 96 demonstrated in young adults that CRF was associated with a lower risk of developing prediabetes or diabetes after 20 years, even when adjusting for BMI, emphasising the importance of continued lifelong fitness in reducing the (pre-)diabetes risk.

| Hospital-based studies
Various hospital-based studies have reported increased prevalence rates of CAN in persons with prediabetes, but as with DSPN, the interpretation of these studies is hampered due to selection bias, small sample sizes, variable definitions of prediabetes and CAN, and varying methods to define and diagnose CAN. 85

| PATHOPHYSIOLOGY OF NEUROPATHY IN PREDIABETES
There is accumulating evidence to suggest that the putative pathophysiologic mechanisms relevant to neuropathy in prediabetes are likely similar to those postulated for early DSPN in type 2 diabetes. 97In brief, hyperglycaemia and dyslipidaemia result in a substrate excess in mitochondria leading to mitochondrial dysfunction and overproduction of reactive oxygen species (ROS) and reactive carbonyls.Reactive oxygen species and carbonyl stress-mediated nuclear DNA damage activate key pathways implicated in the development of microvascular damage in diabetes and presumably also in prediabetes (polyol pathway, hexosamine pathway, protein kinase C activity, advanced glycation end products [AGEs]).In addition, overproduction of ROS and reactive carbonyls causes endoplasmic reticulum stress, while hyperinsulinaemia and inflammation cause impaired insulin signalling.These events converge in the activation of stress and inflammatory pathways, leading to widespread changes in gene expression and upregulation of cytokines/chemokines, apoptotic signalling, and soluble adhesion molecules. 42,76Dyslipidaemia is characterised by an abnormal plasma lipid profile with Abbreviations: BRS, baroreflex sensitivity; DM, diabetes mellitus; GM, geometric mean; HF, high frequency power in the R-R interval spectrum between 0.12 � 0.40 Hz; HRV, heart rate variability; i-IFG isolated impaired fasting glucose; IFG-1, fasting glucose 5.6-6.0 mmol/L; IFG-2, fasting glucose 6.1-6.9 mmol/L; IFG + IGT combined impaired fasting glucose and impaired glucose tolerance; i-IGT isolated impaired glucose tolerance; k-DM, known diabetes mellitus; LF, low frequency power in the R-R interval spectrum between 0.04 � 0.12 Hz; n-DM, newly detected diabetes mellitus; NGT, normal glucose tolerance; RMSSD, square root of the mean of the squares of successive differences in normal-to-normal intervals; SDANN, standard deviation of 5-min averaged normal-to-normal intervals; SDNN, standard deviation of normal-to-normal intervals; 30:15 R-R, ratio between 30th and the 15th R-R interval after standing up from supine position.The PLIS study has recently demonstrated that lifestyle intervention for 1 year reduces post-challenge glucose, cardiovascular risk, and liver fat in individuals with prediabetes showing a high-risk phenotype based on insulin secretion, insulin sensitivity, and liver fat content, suggesting that individualised risk phenotype-based approach could be beneficial for diabetes prevention of diabetes. 100e Diabetes Prevention Programme (DPP) reported a greater reduction in diabetes incidence over about 3 years with a lifestyle programme than with metformin, as compared with placebo (58 vs. 31%).Translated into numbers needed to treat, about 7 persons would need to be treated with the lifestyle intervention or about 14 with metformin to prevent 1 case of diabetes over about 3 years. 101e 15-year follow-up of DPPOS found reduction in diabetes incidence for persons in the lifestyle programme than for those taking metformin, although it found a decline in between-group difference (27% vs. 18%).In the lifestyle arm of the DPP, persons with prediabetes who did not develop diabetes had a lower prevalence of DSPN than those who developed diabetes. 102In contrast, in the Da Qing study, lifestyle intervention for 6 years in participants with IGT at baseline did not result in a reduction in DSPN incidence after 20 years. 103However, in both studies DSPN assessment was crude using the 10-g Semmes-Weinstein monofilament as a single measure.

| Lifestyle and pharmacological interventions
Several small uncontrolled studies focussing on the effects of lifestyle intervention on small nerve fibre pathology showed an increase in IENFD in heterogeneous groups with prediabetes or metabolic syndrome with and without type 2 diabetes. 104It has been suggested that exercise may improve cutaneous nerve regenerative capacity, neuropathic pain, and task-specific functional performance measures of gait, but trials are limited in scale. 104tervention trials studying the effect of exercise on cardiovascular autonomic function to prevent or delay the development of CAN are also scarce.In the lifestyle arm of the DPP after 3.2 years, HR and QT interval indexes decreased, and HRV increased.The magnitude of the decline in HR and QT duration was substantially smaller in the other arms, whereas HRV did not increase.Baseline HR was the only index associated with incident diabetes.Reductions in HR and QT indices and increases in HRV over time were associated with a lower risk of developing diabetes. 105Numerous trials in obese individuals with and without type 2 diabetes have shown favourable effects of various modes of exercise training on cardiovagal function, 40,106 suggesting by analogy that exercise could also be beneficial specifically in persons with prediabetes.
On the basis of the aforementioned evidence, the USPSTF recommends behavioural weight loss interventions to prevent obesityrelated morbidity and mortality in adults with a BMI ≥30 kg/m 2 . 16e ADA recommends that adults with overweight/obesity at high risk of type 2 diabetes should be referred to an intensive lifestyle behaviour change programme to achieve and maintain 7% loss of initial body weight, and increase moderate-intensity physical activity (such as brisk walking) to at least 150 min/week.Moreover, a variety of eating patterns can be considered to prevent diabetes in persons with prediabetes. 107

| Bariatric surgery
Bariatric surgery leads to pronounced sustained weight loss and is associated with other important health outcomes, among which the effect on remission of type 2 diabetes is substantial. 35A recent metaanalysis comprising 32,756 obese participants with type 2 diabetes reported a reduced incidence of DSPN (odds ratio 0.27; 95% CI, 0.22-0.34)compared with nonsurgical treatments. 108Two recent small uncontrolled studies showed improvements particularly in small fibre pathology assessed by CCM after 12 months in obese individuals with or without diabetes who underwent bariatric surgery, 109,110 but larger-scale controlled studies specifically in persons with prediabetes are needed to confirm these preliminary findings.
The potential benefit of bariatric surgery should be weighed against the risk of subacute axonal neuropathy caused by micronutrient deficiencies, particularly low thiamine levels and autonomic dysfunction including OH. 35

| Pharmacotherapy of painful neuropathy
A pilot study showed pain relief by 36% after 1 month of singleblind treatment with pregabalin in individuals with prediabetes and painful DSPN.After a further 1-month double-blind treatment, additional reduction of pain was observed for pregabalin but not placebo. 111In the absence of randomised clinical trials (RCTs), specifically in people with prediabetes and neuropathic pain, pharmacological treatment remains off-label in clinical practice using drugs approved for painful diabetic neuropathy, preferentially antidepressants, anticonvulsants, and capsaicin 8% patch, while pathogenetic treatment with a-lipoic acid and benfotiamine may also be considered. 30EGLER ET AL.
-13 of 18 Both DSPN and CAN in diabetes and prediabetes remain underdiagnosed and undertreated.The available evidence points to an increased risk of both DSPN and CAN in people with prediabetes at the general population level, particularly in the group with combined IFG and IGT, which also carries the highest risk of progression to diabetes.However, the true magnitude of the impact of prediabetes on DSPN and CAN is difficult to estimate due to major hurdles primarily related to small sample sizes, variable definitions of prediabetes and neuropathy, methods of prediabetes and neuropathy assessment, and selection bias in the hospital-based studies.A major unmet need is the lack of large-scale prospective studies.
The rising burden of prediabetes constitutes important public health challenges at both regional and global levels.For example, European countries have made variable progress towards investing in and implementing of comprehensive strategies for the prevention and treatment of diabetes.3][114] To reduce the burden resulting from prediabetic DSPN, adequate consideration and implementation of strategies aimed at early detection and prevention of the condition in national diabetes plans is imperative.
The USPSTF recently identified numerous research needs and gaps in the prevention of diabetes and stated that more research is needed on the effects of screening on health outcomes that enrol populations reflective of the prevalence of diabetes in the US, particularly racial and ethnic groups that have a higher prevalence of diabetes than persons of European descent.More data are also needed on the effects of lifestyle interventions and medical treatments for screen-detected prediabetes and diabetes on health outcomes over a longer follow-up period, particularly in populations reflective of the prevalence of diabetes and on how best to increase the uptake of lifestyle interventions, especially among populations at highest risk of progression to diabetes and adverse health outcomes.
Novel studies suggest that novel risk factors such as air and noise pollution may not only increase the incidence of type 2 diabetes but also of DSPN, [115][116][117] so that ubiquitous environmental determinants also merit further investigation in this context.Clinical trials and additional modelling studies are needed to better decipher the optimal frequency of screening and the age at which to start and stop screening.Finally, more research is needed on the natural history of prediabetes, including the identification of factors associated with the risk of progression to diabetes or reversion to normoglycaemia. 16 remains to be established whether novel subphenotyping approaches are helpful to elucidate the metabolic heterogeneity prior reduced heart rate variability (HRV), that is, the magnitude of HR fluctuations around the mean HR.Following the introduction of cardiovascular reflex tests based on changes in HRV and blood pressure regulation into clinical practice, it became evident that CAN may be frequently detected at early stages in asymptomatic patients with diabetes.According to the Toronto Consensus, 38 diagnosis of CAN should be based upon the use of HR response to deep breathing, standing, and Valsalva manoeuvre, and blood pressure response to standing.The following criteria for the diagnosis and staging of CAN were proposed: (1) the presence of one abnormal cardiovagal test result identifies the condition of possible or early CAN to be confirmed over time; (2) ≥2 abnormal cardiovagal test results are required for a definite or confirmed diagnosis of CAN; and (3) the presence of orthostatic hypotension (OH) in addition to abnormal HRV identifies severe or advanced CAN.Resting short-term or 24h circadian HRV can be assessed using statistical indices of R-R interval changes (time-domain) or using spectral analysis (frequency-domain) of the electrocardiogramm recording.Different indices correspond either to both or predominantly to one of the sympathetic or parasympathetic branches of the ANS.The spectral power consists of three major components: (1) The very-low-frequency band (VLF), related to fluctuations mediated primarily by sympathetic tone; (2) The low-frequency band, associated with the baroreceptor reflex and both parasympathetic and sympathetic activity; (3) The high-frequency band, influenced by respiratory activity and parasympathetic modulation. 40More recently, methods derived from nonlinear dynamics have provided new insights into HRV changes in various diseases.Examples for measures derived from nonlinear dynamics include (a) the Poincaré plot, a quantitative visual technique, whereby the shape of the plot is categorised into functional classes; (b) entropy measures which assess the regularity/irregularity or randomness of heartbeat fluctuations; or (c) symbolic dynamics which allows a simple description of the coarse-grained dynamics of HR fluctuations by a limited amount of symbols. 41The baroreflex system plays an important role in regulating short-term fluctuations of arterial blood pressure detected by arterial baroreceptors, which modulate the efferent autonomic neural control, resulting in the changes in the HR, cardiac contractility, and peripheral vascular resistance.Non-invasive measurement of spontaneous baroreflex sensitivity (BRS) is used to quantify baroreflex regulation of the cardiovascular system and represents an important prognostic index in cardiovascular diseases.Indeed, blunted BRS is predictive of an increased cardiovascular risk in patients with heart failure and those who underwent myocardial infarction.Evidence has emerged suggesting early BRS dysfunction in the course of type 2diabetes.40,42 is the only one to report prevalence rates of CAN separately for the various categories of prediabetes (5 categories).A total of 120 HRV variables (time domain, 15 indices; frequency domain, 15 indices; non-linear dynamics, 90 indices using eight different methods) were determined.To obtain the prevalence of abnormalities for a combination of HRV indices, four single indices from four different HRV classes showing the best discrim-inatory value between the NGT and the other groups were selected.As compared to the NGT group, the rates of CAN (≥2 of 4 HRV indices abnormal) were significantly higher in the i-IFG, IFG + IGT, n-DM, and k-DM groups.These data suggest that in the general population aged 55-74 years, compared to NGT, the prevalence of CAN is 1.8-fold and 2.5-fold increased in persons with i-IFG and IFG + IGT, the latter group approaching the prevalence found in n-DM.86Moreover, modifiable cardiovascular risk ZIEGLER ET AL.
to the development of type 2 diabetes.Their further validation and fine tuning could represent one step forward towards precision medicine in clinical practice. 11More research is needed to define the exact prevalence of DSPN and CAN in the various categories of prediabetes and to decipher the natural history of these conditions in clearly defined metabolic subpopulations and the corresponding determinants of their development and progression.Moreover, RCTs are needed to define the role of agents used for pathogenetically oriented and symptomatic pharmacotherapy and nonpharmacological treatment options such as lifestyle interventions for prediabetic DSPN using a state-of-the-art diagnostic armamentarium and standardised surrogate and clinical endpoints.

Table 1 .
13iteria for the definition of prediabetes. of having the condition.In 2021, 541 million adults, or 10.6% of adults worldwide, were estimated to have IGT.By 2045, this figure is projected to increase to 730 million adults or 11.4% of all adults.In addition, an estimated 319 million adults or 6.2% of the global adult population had IFG.An estimated 441 million adults or 6.9% of the global adult population are projected to have IFG in 2045.13Theage-adjusted prevalence estimates of IFG and IGT in the individual geographical regions in 2021 are shown in Supplemental The age-adjusted prevalence estimates of IGT in 2021 were WHO 110-125 6.1-6.9 140-199 7.8-11.1 n.a.IEC n.a.n.a.6.0-6.4 42-46 Abbreviations: ADA, American Diabetes Association; FPG, fasting plasma glucose; IEC, International Expert Committee; n.a., not applicable; 2-h PG, plasma glucose after 2 h in 75-g oral glucose tolerance test (OGTT); WHO, World Health Organization.unaware NFG ADA ), and 6.10 [4.72-7.88]forIGT + IFG ADA .
1 Summary of prevalence estimates of peripheral neuropathy in prediabetes.Prevalence estimates are reported with 95% CIs and primary method used to assess peripheral neuropathy.ATR, Achilles tendon reflex; CCM, corneal confocal microscopy; CPT, current perception threshold; DNES, Diabetic Neuropathy Examination Score; ESC, electrochemical skin conductance; MNSI, Michigan Neuropathy Screening Instrument; NCS, nerve conduction studies; NCT, nerve conduction test; NDS, Neuropathy Disability Score; NIS, Neuropathy Impairment Score; NSS, Neuropathy Symptom Score; NTSS, Neuropathy Total Symptom Score; PPS, Pressure Perception Score; PS, pressure sensation; PTR, patellar tendon reflex; QST, quantitative sensory testing; QTS, quantitative tactile stimulation; TCSS, Toronto Clinical neuropathy Scoring System; TDT, thermal discrimination threshold; TRI, Thermal Recovery Index; VP, vibration perception; VPT, vibration perception threshold; VS, vibration sensation.Reproduced with permission according to CC BY from. 54T A B L E 2 and other prospective studies should offer new insights into the natural history of DSPN in prediabetes.
111arkably, the Chinese PRECISE study, investigating 254 persons with combined IFG + IGT, 951 with i-IFG or i-IGT, 504 with diabetes, and 509 persons with NGT, recently reported cerebral white matter microstructural abnormalities using diffusion tensor imaging only in those with IFG + IGT and diabetes, confirming the notion that the combined abnormality carries the highest risk ofT A B L E 2 (Continued) Note: Table modified from.111"a"indicates the significance level (p) for the difference in the DSPN prevalence rates between the groups, e.g. last line: 41.9 is significantly different from 17.4, 19.2, and 7.7, etc.Abbreviations: DM, diabetes mellitus; DSPN, distal sensorimotor polyneuropathy; i-IFG, isolated impaired fasting glucose; IFG + IGT, combined impaired fasting glucose and impaired glucose tolerance; i-IGT, isolated impaired glucose tolerance; k-DM, known diabetes mellitus; MNSI, Michigan Neuropathy Screening Instrument; n.a., not available; NC, nerve conduction; n-DM, newly detected diabetes mellitus; NDS, Neuropathy Disability Score; NGT, normal glucose tolerance; NSS, Neuropathy Symptom Score; OR, odds ratio; PBP, postural blood pressure; PNSS, positive neuropathic sensory symptoms; PS, pressure sensation; VPT, vibration perception threshold.
111valence of cardiovascular autonomic neuropathy (CAN) and orthostatic hypotension (OH) in the general population with prediabetes compared to those with normal glucose tolerance (NGT) and those with diabetes.HRV, heart rate variability; i-IFG, isolated impaired fasting glucose; i-IGT, isolated impaired glucose tolerance; IFG+IGT, combined impaired fasting glucose and impaired glucose tolerance; k-DM, known diabetes; n-DM, mellitus; newly, detected diabetes mellitus; NGT, normal glucose tolerance; SBP, diastolic blood pressure; ∆SBP/DBP, drop in systolic or diastolic blood pressure after standing up.Population-based studies evaluating the mean differences in indices of heart rate variability (HRV) in individuals with prediabetes compared to those with normal glucose tolerance (NGT) and diabetes.Table modified from.111 T A B L E 3