Читать книгу Diabetic Neuropathy - Friedrich A. Gries - Страница 9

3 Epidemiology of Diabetic Neuropathy

J. E Shaw, P. Z. Zimmet, F. A. Gries, and D. Ziegler

Introduction

The epidemiology of a disease primarily describes the frequency with which it occurs, and determines the risk factors associated with it. The former informs the clinician about the likelihood that the patient in front of him or her has the condition and the public health authorities about the potential overall burden relating to the condition. The latter sheds light on etiological processes, although the associations described by epidemiological studies cannot alone be taken as proof of causality.

In order to understand the epidemiology of a disease properly, it is important to have well-validated diagnostic tests that can be used by a variety of investigators assessing different populations in a similar manner. Furthermore, it is important that the populations studied are representative of the total population being considered, and have not been subjected to significant selection biases. However, the study of the epidemiology of diabetic neuropathy has been beset by numerous problems relating both to diagnostic tests and to population selection.

Testing for Peripheral Neuropathy

Diabetic peripheral neuropathy is a complex disorder in which the disease process may affect different sets of nerve fibers to different degrees in different individuals. Thus, one individual may have an abnormality of large-fiber sensory function, which could be detected by measuring the vibration perception threshold (VPT), whilst another may have a predominantly small-fiber neuropathy that can only be detected by measuring the thermal perception threshold (TPT). This feature of neuropathy can cause problems in the selection of a single test with which to screen a population. The issue of measurement of neurological function is further complicated by the nature of the tests. Many are psychophysical tests, in which the subject is required to interpret the nature of an external stimulus (usually applied to the foot). This subjectivity can lead to relatively poor reproducibility of tests such as VPT, in which the subject has to differentiate light touch from vibratory sensation.

This lack of certainty over the value of individual tests for the assessment of diabetic neuropathy has led to recommendations that several different tests should be performed, and that diabetic neuropathy should only be diagnosed when more than one is abnormal [1]. Whilst this may make the diagnostic process more rigorous in any individual or in an individual study, it can make comparisons between studies more difficult. Such recommendations have been only patchily adopted, and where they have been put into practice, the selection of tests has not been uniform. Thus, we are faced with having to compare the prevalence when neuropathy is determined by a single test with the prevalence when neuropathy is diagnosed when any two out of perhaps three to five tests are abnormal. Increasing the number of tests that are performed will automatically increase the number of individuals in whom an abnormality is found, while requiring that more than one abnormality is present will then tend to decrease the prevalence. The overall effect is thus complex. The impact of varying the diagnostic testing procedure can be seen in the Diabetes Control and Complications Trial (DCCT) data (Table 3.1), where the prevalence of distal symmetrical polyneuropathy (DSP) at baseline in the conventional therapy cohort varied from 0.3% (abnormalities of reflexes, sensory examination, and neuropathic symptoms) to 21.8 % (abnormal nerve conduction in at least two nerves) [2]. Confirmed clinical neuropathy (abnormal history or examination, confirmed by abnormal nerve conduction or autonomic function) was the gold standard for the study, and was found in 2.1% of this cohort. This 73-fold difference in the prevalence of DSP, between the two extremes, but within a single population, highlights the difficulties of comparing studies with differing diagnostic criteria.

The influence of test selection on the understanding of etiological factors associated with neuropathy is likely to be considerably smaller than that on prevalence. As long as each diagnostic process is indeed measuring some aspect (or aspects) of neuropathy, it is likely that those who are rated as having neuropathy are genuinely more severely affected than those who are not so rated. Thus, associations with factors such as age or glycemic control may be easier to compare across different studies than are prevalences, which are highly dependent not only on the test (or tests) selected, but also on the diagnostic threshold that is used. For example, the groups selected either by absent ankle reflexes or by elevated VPT are both likely to be older and have poorer glycemic control than the groups who are normal on these respective tests. However, unless loss of ankle reflexes represents the same degree of neuropathy as does an abnormal VPT (according to the threshold selected - see below), the prevalences generated from the same population by these two tests could be quite different.

Table 3.1 Prevalence of polyneuropathy within the DCCT conventional therapy primary prevention cohort, according to diagnostic test

Diagnostic criteriaa	Neuropathy prevalence (%)
Sensory examination, reflexes, and symptoms	0.3
Symptoms and reflexes	0.5
Sensory examination and symptoms	1.1
Symptoms or sensory examination, confirmed by abnormal nerve conduction or autonomic functionb	2.1
Sensory examination and reflexes	2.1
Abnormal autonomic function	2.4
Symptoms alone	3.7
Any two out of symptoms, sensory examination, reflexes	4.5
Reflexes alone	5.8
Sensory examination alone	12.2
Nerve conduction (abnormal in at least 2 nerves)	21.8

Data adapted from [2]

^aAbnormalities required in the fields listed

^bCold standard measure of neuropathy for the study

Even when the same test has been used, the methods and the selection of thresholds have not been uniform. VPT is widely used, but some studies use agerelated thresholds (based on UK normal data) [3], while others have ignored age and used a single cutoff [4]. More recently, locally derived normal ranges have been applied, although in one study these were derived separately for different age groups [5], while another compared all diabetic subjects to a young nondiabetic population [6].

It is obvious from the discussion above that there is a long way to go in finding a definition of neuropathy that can be widely used, although it is clearly needed for accurate epidemiological study. This definition needs to be appropriate for the clinical forms that the disease may take (e.g., small fiber or large fiber involvement), related to hard endpoints, such as foot ulceration, agreed upon widely, and useful for a range of both clinical and research purposes. Multiple testing of different neurological functions, although recommended as the gold standard [1], can be difficult and expensive, especially in large field studies. Furthermore, recent data suggest complex patterns of sensation in DSP with decreased VPT (hypoesthesia) and heat stimulus-induced hyperesthesia (low thresholds), both being characteristic of mild DSP as they correlate with neuropathic symptoms and deficits, whereas panmodality hypoesthesia is typical of severe DSP [7]. The only studies that relate measures of neurological function (in a prospective manner) with subsequent foot ulceration and amputation have used simple quantitative sensory tests, such as VPT or monofilament sensitivity [8,9].

Population Selection

The second important obstacle in determining the prevalence of neuropathy is the selection of people who are going to be tested. Hospital clinic-based studies will usually be biased towards those with more serious problems, and are likely to overestimate the prevalence of neuropathy in the general diabetic population. Nevertheless, the value of hospital-based prevalence studies should not be underestimated. Whilst they do not directly address the issue of what the general risk of having neuropathy is in a person with diabetes, they precisely answer the question of how common neuropathy is amongst hospital clinic patients. This in itself may be a very important and valuable issue. The problem, however, is that one hospital-based population may not be the same as another - an issue that can be partially overcome with multicenter studies [10]. Another problem is that temporal trends in referral patterns from primary care can make today's hospital population quite different to tomorrow's.

Studies using registers of people with diabetes that are derived from primary care will overcome the hospital-based biases, and not surprisingly have reported lower prevalences [11]. However, this selection methodology remains dependent on the proportion of people with diabetes who are actually diagnosed. Since approximately 50% of all of those with diabetes are undiagnosed [12], this method also has flaws. Population-based epidemiological surveys, in which a high proportion of a representative sample of the general population is directly tested for diabetes (by blood glucose testing), and then those found to have diabetes (previously diagnosed and newly diagnosed) are screened for neuropathy, represent an alternative approach. This should be a more reliable and reproducible way of determining the prevalence of diabetic neuropathy, but only two such studies have been reported [5,6].

One drawback of this approach is that the group with “newly diagnosed” diabetes (i.e., diagnosed as part of the study by blood glucose testing) is usually identified by an abnormal glucose value measured on a single day. Since the diagnosis of diabetes requires verification of this on another day [13], it is likely that the group identified includes a number of individuals who would have a nondiabetic glucose value on repeat testing, and therefore do not actually have diabetes. Assuming that these people do not have neuropathy (since they do not have diabetes), their inclusion would dilute the sample and lead to an underestimate of the prevalence of neuropathy in the newly diagnosed group. Nevertheless, neuropathy prevalence estimates derived from population-based samples that include all of those people who have diabetes (both diagnosed and undiagnosed) should be the most accurate and reproducible approach.

How Common Is Diabetic Neuropathy?

Distal Symmetrical Polyneuropathy

Prevalence in Diabetes

Taking into account the issues discussed above, it is difficult to determine with any precision the prevalence of diabetic neuropathy. Several large studies have examined the prevalence in hospital-based populations. A number of these are in fairly close agreement, and reveal prevalences of DSP at approximately 30%, amongst both European and African populations [8,14-16]. However, other hospital-based studies have produced figures closer to 20% [3], and a prevalence of 50 % was reported from a US veterans population [17].The higher figure in this last study may relate to the age of the participants, and to the fact that they were almost all male.

Studies which have used primary care registers of people with diabetes, or which report the prevalence in people with previously diagnosed diabetes derived from population-based surveys, have mainly reported lower prevalences than those given in hospital-based studies. Some studies report prevalences of around 20% [6,11,18-20], although a figure as high as 54% in type 1 and 45% in type 2 diabetic patients was reported from a population based sample from Rochester [21], and 42% of a sample of 811 type 2 diabetic subjects drawn from 37 UK general practices were found to be neuropathic [22]. The Rochester study used electrophysiology as part of the neurological assessment. This is frequently abnormal in diabetes, even within a short time of diagnosis of type 1 diabetes, and has not been validated as a predictor of foot ulceration. The prevalence rates of DSP in type 1 and type 2 diabetic patients are summarized in Table 3.2.

When general populations are screened by glucose tolerance testing, and those identified as having diabetes are assessed for peripheral neuropathy, the prevalence has been relatively low (Table 3.3), In our own study from the Indian Ocean island of Mauritius (in which 70% of the population originates from India) the data showed the lowest prevalence of DSP yet reported [5]. The overall prevalence in the total diabetic population was 8.3%, i.e., 12.7% in subjects with known diabetes and just 3.6% among those with newly diagnosed diabetes. Consistent with the low neuropathy prevalence was a low prevalence of lower limb amputation - a finding that has been reported elsewhere for people who originate from the Indian subcontinent [36]. Neuropathy was based on VPT measurements, using locally derived age specific normal ranges. In the only similar study, from Egypt [6], 14% of the newly diagnosed diabetic population were found to have DSP. However, the reference range used for VPT in that study related to healthy adults under the age of 45, and therefore probably did not account for the normal age related rise in VPT. Thus, since the majority of those with diabetes were over the age of 45, the neuropathy prevalence in that study may be an overestimate. In two Native American populations, which are regularly screened for diabetes with glucose tolerance testing, the prevalence of DSP in the whole diabetic population was 19% [4,9].

The prevalence of diabetic neuropathy in general populations was evaluated in two door-to-door surveys that were restricted to a questionnaire which represents the most crude and simple screening instrument [37,38]. However, since no screening for diabetes had been performed, the exact prevalence of diabetes in these populations was unknown.

It can be seen from Tables 3.b and 3.3 that diagnostic methodology is far from uniform. Indeed, there are hardly any two studies that have used identical methods. Nevertheless, a pattern can be discerned in these studies. The prevalence in hospital-based populations is probably around 30%, falling to about 20% when people with established diabetes are selected from a community base. Amongst the complete diabetic population, including both diagnosed and undiagnosed diabetes, the prevalence of DSP is probably less than 20%.

Prevalence in Impaired Glucose Tolerance

Although the development of diabetic polyneuropathy is typically insidious over several years, it may occasionally be the presenting feature in type 2 diabetic patients [39]. It is not known whether this “early” nerve alteration has developed during a period of unrecognized diabetes or evolves gradually during a state of impaired glucose tolerance (IGT) prior to the transition to overt diabetes. In this context, it is unclear whether there is a glycemic threshold beyond which nerve dysfunction develops. If IGT constitutes such a threshold which needs to be passed, subjects with IGT should not have a degree of neuropathy higher than that of nondiabetic subjects [40] The results of several studies comparing the prevalence of nerve dysfunction in patientswith IGT and normal glucose tolerance (NGT) are summarized in Table 3.4. The hospital-based studies could not demonstrate a difference in peripheral nerve function and structure between patients with IGT and NGT, except for an abnormal expiration/inhalation ratio, a marker of cardiovascular autonomic neuropathy (CAN), which was found to be more frequent in IGT than in NGT [40]. In one population-based study ankle reflexes were detected less frequently in patients with IGT than in those with NGT [29]. Thus, there is preliminary evidence suggesting that peripheral neuropathy may be slightly more frequent in IGT than in NGT, but further studies are needed to confirm these findings.

Incidence and Natural History

Most of the data describing the frequency of diabetic neuropathy relate to prevalence. The long-term natural history and progression of diabetic neuropathy has been difficult to study owing to the following problems: (1) various nerve fiber populations may be affected at different rates, (2) expected changes may take place over several years. (3) minor changes may not be detected due to a low reproducibility of some methods, (4) measures of nerve function such as nerve conduction velocity (NCV) may deteriorate with age in nondiabetic subjects, and (5) glycemic control or risk factor profile may change and nerve dysfunction may be reversible over time. It has been emphasized that inception cohorts followed from the diagnosis onward are needed to determine the impact of diabetic neuropathy, because prevalence rates consider only the surviving cases and therefore might underestimate the true risk of acquiring the disease [43].

A long-term clinic-based study carried out in a large group of type 1 and type 2 diabetic patients, designed to investigate the relationship between the duration and degree of hyperglycemia and the prevalence of clinically overt neuropathy from the onset of diabetes, was performed by Pirart [44] over a period of 25 years. He defined neuropathy as a loss of Achilles and/or patellar reflexes combined with a reduced vibration sensation or presence of “a more dramatic polyneuropathy or monoor multi-neuropathy.” The prevalence of neuropathy increased from about 8% among the 1900 patients evaluated at the time of diagnosis to about 50% among 100 patients reassessed after a follow-up of 25 years. The incidence of neuropathy increased from three cases per 100 unaffected patients per year to about 19 cases per 100 patients after 25 years. However, Pirart's study had several drawbacks: it was not population-based, no standardized diagnostic measures were employed, the various forms of neuropathy were not differentiated, control subjects were not assessed, and the types of diabetes were not evaluated separately.

In a study focusing on 133 Finnish patients with newly diagnosed (following clinical presentation) type 2 diabetes, the10-year incidence of DSP among those who were free of neuropathy at baseline was 17-20% depending on whether electrophysiological or clinical parameters were used [33]. This gives an annual incidence of approximately 2%. However, the development of DSP and CAN was divergent in that approximately the same proportions of patients with and without peripheral neuropathy developed CAN, either with predominant parasympathetic or sympathetic involvement or both [45]. To explain this finding different underlying mechanisms involved in the evolution of peripheral neuropathy and CAN have been suggested: while hypoinsulinemia predicted the development of DSP, it was hyperinsulinemia that predicted the development of CAN [45].

Over the first 12 years of the UK Prospective Diabetes Study (UKPDS), which similarly studied people newly presenting with type 2 diabetes, neuropathy (as judged by a VPT >25 V) developed in approximately 20% of subjects irrespective of treatment arm, giving an annual incidence of just under 2% [46]. In the American San Luis Valley study [47], DSP was determined by a combination of symptoms and clinical signs. Over nearly five years, its average annual incidence was 6.1 % (6.1 per 100 person-years). A study of US veterans reported neuropathy (monofilament insensitivity) developing in 20% of subjects over approximately 2.5 years [17]. The higher incidence in the last two studies was probably due to the fact they only included people with established diabetes, who, consequent on their longer disease duration, would have had a higher baseline risk than the subjects with newly diagnosed diabetes who were included in the other two studies.

In the Rochester Diabetic Neuropathy Study (RDNS), after two years the Neuropathy Impairment Score (NIS) was unchanged in 81%, worse in 10%, and better in 9% of the patients and the Neuropathy Symptom Score (NSS) was unchanged in 92% and worse or better in 4% of the patients, respectively. The most prominent monotone (consistent) worsening over two years was found for a composite score combining neuropathic deficits and quantitative measures of DSP (NIS[lower limbs] + 7 tests), which worsened on average by 0.34 points per year in the entire cohort and by 0.85 points per year in patients with DSP [48].

The Epidemiology of Diabetes Complications Study (EDCS) reported on a six-year follow-up of 453 type 1 diabetic patients who were free of DSP at baseline. 15% of whom developed DSP in six years, giving an incidence of 2.8 per 100 person-years and a cumulative probability of 0.29 [49], DSP was defined as the presence of two or more of the following: symptoms, sensory and/or motor signs, and/or absent (or present only with reinforcement) tendon reflexes. In the DCCT, which included groups of highly selected individuals with type 1 diabetes, polyneuropathy (abnormal history or examination, confirmed by abnormal nerve conduction or autonomic function) developed in 9.6% of those who were conventionally treated, and in 2.8% of those in the intensively treated group, over five years [2].

Comparing to the data for type 2 diabetes, it appears that the incidence of neuropathy in type 1 diabetes (using the data from the conventionally treated group - approximately 2% per year) is similar to that in newly diagnosed type 2 diabetes, but less than that in previously diagnosed type 2 diabetes. However, age is an important factor in diabetic neuropathy (see below), and the considerably younger age of the subjects in the DCCT probably accounts for the differences seen. Indeed, when age has been corrected for, the prevalence of neuropathy does not differ between type 1 and type 2 diabetes [8,11].

In the EURODIAB Prospective Complications Study, which included 986 type 1 diabetic subjects without DSP at baseline, the cumulative prevalence after an average follow-up of 7.3 years was 25% [50].

Several clinic-based prospective studies examined the relationship between the natural history of abnormalities in nerve function tests and the degree of longterm glycemic control. Hillson et al [51], followed 71 patients who showed a slight deterioration in mean pedal VPT during the first five years after diagnosis of type 2 diabetes, which correlated significantly with increased mean fasting blood glucose. A more pronounced deterioration in sensory nerve function as assessed by thermal, vibration, and pressure perception thresholds has been observed over 1-3.5 years and two years in newly diagnosed and longer-term type 2 diabetic patients, respectively [35,52]. There was a correlation between the changes in the various sensory thresholds. [52], suggesting that small and large nerve fiber dysfunction may develop in parallel in type 2 diabetic patients.

Young et al. [53,54] studied 75 type 1 diabetic patients aged 16-19 years over an average of 2.4 years, 70 of whom were reassessed again after six years. During the first study period the deterioration in motor and sensory NCV was associated with poor glycemic control, but no patient had symptoms and only seven had minor signs, of neuropathy [53]. After six years 16 patients had Symptoms, 12 had “major signs,” 15 had “minor signs”, and 27 showed no symptoms of neuropathy. Baseline HbA₁ was significantly higher in patients who developed symptoms or major signs compared win hi those showing minor signs or no symptoms. In addition, peroneal motor NCV (MNCV) deteriorated if the symptomatic but not in the asymptomatic group [54]. On the other hand, it has been demonstrated that neuropathic pain may improve and resolve completely within three to four years of follow-up despite persistent poor glycemic control, but thermal perception thresholds continue to deteriorate [55]. Thus, small-fiber function tests do not seem to predict the evolution of painful symptoms.

Regarding the magnitude of changes in the quantitative measures of DSP, we have shown that over the first decade of type 1 diabetes the mean difference between poorly controlled and well-controlled patients in the annual rate of slowing of peroneal MNCV is 0.6 m/s per year and that of sural sensory NCV (SNCV) is 0.7 m/s per year. After 12 years from the diagnosis of type 1 diabetes, none of the well-controlled patients but 60% of those with long-term poor control developed polyneuropathy [56]. Macleod et al. [57] estimated the annual rate of change for the VPT on the great toe to be 0.4 V in healthy subjects and 2.5 V in those with diabetic neuropathy.

In a mixed cohort of type 1 and type 2 diabetic subjects followed over 9-16 years (mean 10 years) peroneal MNCV was: Progressively reduced with increasing duration of diabetes, but it stabilized once it reached a “plateau” between 36 and 38 m/s [58].

Focal and Multifocal Neuropathies

Prevalence

In the Rochester Diabetic Neuropathy Study the prevalence of neuropathies other than DSP was low: symptomatic carpal tunnel syndrome in 9% of type 1 and 4% of type2 patients, proximal asymmetric neuropathy in 1% each (of type 1 and type 2 patients, ulnar neuropathy in 2% each of type 1 and type 2 patients, peroneal neuropathy in one type 2 patient, lateral femoral cutaneous neuropathy of the thigh (meralgia paresthetica) in 1% each of type 1 and type 2 patients. Cranial neuropathy or truncal neuropathy was not present in any of the patients at the time of examination. Approximately 10% of diabetic patients had neurological deficits attributable to nondiabetic causes [21].

In a Japanese survey among 1961 diabetic patients, 19(1%) had cranial nerve palsies: 19 facial nerve, 6 oculomotor nerve, and 2 abducent nerve palsies. Among 3841 nondiabetic subjects only 5 (0.1%) showed cranial nerve palsies, all of which were facial palsies. Only 1 out of 9 patients with facial palsy compared to 7 out of 10 patients with ophthalmoplegia had diabetic complications, suggesting that ophthalmoplegia rather than facial palsy is related to diabetes [59].

Natural History

Proximal motor neuropathy (femoral neuropathy, diabetic amyotrophy) is characterized by pain with or without wasting in one thigh, or, more often, both thighs. It develops over several weeks, and recovery is the rule. The worst pain resolves within 6-12 months, while the remaining discomfort disappears within three years. Function recovers completely, although slight wasting may persist [39].

The onset of cranial nerve palsies is abrupt and oculomotor nerve involvement may be accompanied by supraorbital pain. Cranial nerve palsies resolve completely in three to six months, and relapses are infrequent [39].

Diabetic polyradiculopathy most commonly affects the trunk, where it causes unilateral or bilateral truncal pain and abdominal muscle bulging. It may be accompanied by profound weight loss. Spontaneous recovery within 4-30 months is the rule, but relapses may occur [60].

For the natural history of painful neuropathies see Chapter 5, page 212-213.

Cardiovascular Autonomic Neuropathy

Prevalence

Although the impact of autonomic neuropathy is increasingly being recognized, little information exists as to its incidence in representative diabetic populations. The difficulties encountered in epidemiological studies of diabetic neuropathy particularly affect the evaluation of autonomic dysfunction, which involves an even greater variety of nonspecific clinical manifestations that can be explored by numerous diagnostic approaches. Of the manifestations of diabetic autonomic neuropathy, cardiovascular autonomic neuropathy (CAN) has attracted most attention, because it is easily detected at subclinical stages by noninvasive cardiovascular autonomic function tests (AFTs) based on heart rate variability (HRV) and blood pressure responses, before the late sequelae such as resting tachy-cardia or orthostatic hypotension develop.

To date there have been two population-based studies that assessed the prevalence of autonomic dysfunction in diabetic patients. The Oxford Community Diabetes Study [61] included 29 873 adults aged at least 20 years. 412 of whom were known diabetic patients. Autonomic function was assessed in 43 type 1 and 202 type 2 diabetic patients using the standard deviation of HRV at rest, expiration/inhalation (E/l) ratio from a single deep breath, and Valsalva ratio. The overall prevalence of abnormal results in one or more of the three tests was 20.9% in type 1 patients and 15.8% in type 2 patients. Since this study did not include control subjects, the reported prevalence of autonomic dysfunction is difficult to interpret. The definition of one abnormality (<2.5th percentile of normal range) among three tests might yield one abnormal test result in up to 7.5% of a control population. Hence, the specificity of the test battery and the prevalence of CAN in that study were relatively low.

In the Pittsburgh Epidemiology of Diabetes Complications Study [62] autonomic neuropathy was evaluated in 168 type 1 diabetic subjects aged 25-34 years, representing 71% of those who were eligible. Cardiovascular autonomic function tests included the E/l ratio and mean circular resultant (MCR) to deep breathing. 30:15 ratio to standing, and Valsalva ratio. The results of the E/I ratio and MCR were significantly associated with the hypertension status, LDL and HDL cholesterol, and gender. The prevalence of autonomic symptoms, ranging from 0 to 8%, was relatively low except for hypoglycemia unawareness, which was present in 26% of the patients. However, this study neither reported the prevalence of abnormal test results nor did it take into account treatment with drugs that potentially influence autonomic function. Orthostatic hypotension, defined as a systolic blood pressure fall of at least 30 mmHg, was noted in 3.4% of the patients.

In a clinic-based multicenter study (DiaCAN) we have found prevalence rates of borderline CAN (2 out of 7 indices abnormal) or definite CAN (>3 out of 7 indices abnormal) of 8.5% and 16.8% respectively among 647 unselected type 1 diabetic patients, and 12.2% and 22.1% respectively in 524 type 2 diabetic patients [31]. The percentage of type 1 diabetic patients with definite CAN was identical with the 16.6% rate observed in 506 patients randomly selected from four hospital diabetic clinics in Bristol [63]. In that study CAN was diagnosed by HRV responses during rest and in response to a single deep breath. Valsalva maneuver, and standing, and was defined by abnormal test results in at least two of these four tests. In the EURODIABIDDM Complications Study [64], among the total of 3250 patients studied. 19.3% (range among centers: 7.8-51.8%) had abnormal HRV and 5.9% (range: 0-14.5%) had postural hypotension. In a Japanese clinic-based cohort of 886 type 2 diabetic patients the prevalence of orthostatic hypotension was 7% [65].

Since particularly in type 2 diabetes the actual onset of the disease may often be preceded by a long period of unrecognized metabolic abnormality, it appears likely that neural dysfunction can be detected at the time of diagnosis of diabetes. Two studies have reported the frequency of neuropathy in type 2 diabetic patients who were examined within four weeks following the diagnosis of the disease. Lehtinen et al. [66] evaluated 132 patients aged 45-64 years in the district of Kuopio University Central Hospital and 142 controls randomly selected from the population registers. The rates of abnormal E/l ratio were 6.3% and 1.4%, respectively. A similar study was performed by Ratzmann et al. [34] in 95 newly diagnosed type2 diabetic patients who were representative of the East Berlin community. The prevalence of abnormal maximum/minimum heart rate during deep breathing was 7.3%.

In a clinic-based study we determined the prevalence of CAN in 120 healthy subjects and 130 newly diagnosed type 1 diabetic patients within 3-49 days after the initiation of insulin treatment with stable mean blood glucose levels. The prevalence of CAN defined by the strict criterion of at least three abnormal tests out of six was 7.7% in the patients and 0 in the controls. Two abnormal tests out of six were found in an additional 9.2% of the patients but in only 1.7% of the controls [67].

Thus, CAN cannot be generally regarded as a late complication of diabetes, but it should be borne in mind that subclinical cardiovascular autonomic dysfunction may be detected even in newly diagnosed type 1 diabetic patients if sensitive indices of spectral analysis are applied in conjunction with tests based on standard analysis of HRV.

Incidence an d Natural History

Only a few studies have evaluated the natural history of CAN in inception cohorts. Töyry et al. [68] followed 133 newly diagnosed type 2 diabetic patients and 144 control subjects over 10 years in Kuopio, Finland. CAN was assessed by the E/l ratio as an index of parasympathetic activity and the systolic blood pressure drop on standing as a measure of sympathetic activity. The prevalence of “parasympathetic neuropathy” rose from 5% in diabetic patients vs 2% in controls at baseline to 65% vs 28% after 10 years, while the prevalence of “sympathetic neuropathy” increased from 7% vs 6% at five years to 24% vs 9% after 10 years. Poor glycemic control and high insulin levels were significant predictors of the development of CAN [68]. Moreover, the presence of CAN at five years was an independent risk factor for the development of stroke after 10 years [69]. Clinic-based studies in patients with variable duration of type 2 diabetes showed a gradual increase in the prevalence of AFT score abnormalities from 41% at baseline to 64% after four years [70], and worsening of the AFT score in 57% of the patients after five years of follow-up [71].

There are no population-based studies including inception cohorts at the time of diagnosis of type 1 diabetes. In a clinic-based study we have shown that the natural evolution of abnormal HRV during the first five years following diagnosis of type 1 diabetes is clearly related to the degree of glycemic control [72]. The rates of abnormalities in HRV at five years were 5% in well-controlled patients and 23% in those who were poorly controlled. In type 1 diabetic children the prevalence of abnormal HRV at rest increased from 27% at the time of diagnosis to 56% after 10 years, but abnormal HRV during deep breathing changed only from 12% to 14% [73].

In longer-term teenage type 1 diabetic patients Young et al. [53] found an increase in the rates of abnormal HRV from 19% to 28% within 20-35 months in poorly controlled type 1 diabetic teenagers. Sampson and colleagues [74] noted a relatively slow decline in HRV of about 1 beat/min per year, approximately three times faster than in healthy subjects [39], over a decade in patients with initially normal HRV. Postural hypotension was persistent, but did not deteriorate in the majority of patients. In contrast, a six-year study has reported the relatively rapid development of postural hypotension in a small group of long-term diabetic patients, which was preceded by a decrease in E/l ratio [75]. In two recent studies in type 1 diabetic patients, AFTs and autonomic symptoms and signs had not changed after 5 and 9 years, respectively [76,77]. In a prospective study over 10-11 years, CAN has been shown to predict future deterioration in glomerular filtration rate in type 1 diabetic patients. It has been hypothesized that nocturnal rise in intraglomerular pressure could be one possible link between CAN and the observed deterioration in glomerular filtration rate [78].

Gastrointestinal Neuropathy

There are no population-based studies reporting the prevalence or incidence of gastrointestinal symptoms or motor abnormalities in diabetic patients. A number of clinic-based studies reported relatively high rates of delayed gastric emptying in diabetic subjects, but most of these included small samples of highly selected populations subject to referral bias. Thus, the true prevalence of delayed gastric emptying is not known. Using the ¹³C-octanoic acid breath test, delayed gastric emptying has been shown in 16% of type 1 diabetic subjects without gastrointestinal symptoms, while the rate increased to 43% in those with these symptoms [79]. Several clinic-based studies reported the prevalence of gastrointestinal symptoms in diabetic and nondiabetic subjects [80–82] (Table 3.5). In type 1 diabetic patients only nausea, vomiting, early satiety, and fullness after meals occurred significantly more frequently than in the nondiabetic subjects in some studies. In type 2 diabetic patients nausea, early satiety, fullness after meals, diarrhea, and constipation were significantly more frequent than in the nondiabetic subjects. The prevalence of gastrointestinal symptoms such as dysphagia, heartburn, or bloating was comparable to that seen in nondiabetic subjects. However, there were considerable differences between the studies regarding the prevalence of the individual symptoms. For example, diarrhea was observed in 35% of the type 2 patients studied in Hong Kong vs 7% of those studied in Germany [80–82].

The gallbladder may become atonic in association with autonomic neuropathy, which increases the gallbladder volume by retention of bile. It is unknown whether these changes result in symptoms. It may be that biliary stasis explains the increased risk of gallstones shown by recent epidemiological data. In a population-based study from Tivoli. Rome, the prevalence of diabetes in subjects affected by gallstone disease was significantly higher than in controls matched for sex, age, and body mass index (12% vs 5%) [83].

“Gustatory sweating” refers to sweating over the face and scalp (and sometimes upper trunk), often accompanied by a flush, that follows the ingestion of food or drink. Diabetic gustatory sweating has been reported in a clinic-based study to occur in 69% of patients with nephropathy, 36% with peripheral neuropathy, and 4% of those without either complication [84]. Diabetic gustatory sweating is believed to be triggered by the taste buds as it is not evoked by chewing inert substances, thinking of food, smelling food or by placing food or alcohol in the stomach by a gastric tube. The tongue is the most sensitive area, and even cocainization fails to inhibit gustatory sweating completely. The etiology of diabetic gustatory sweating is uncertain. The hypothesis that it results from aberrant regrowth of parasympathetic fibers leading to a connection between salivation centers and facial sweat glands seems to have been disproved by a series of case reports of immediate cessation of gustatory sweating after renal transplantation [84].

Erectile Dysfunction

Prevalence

Male erectile dysfunction, defined as “the inability to achieve or maintain an erection sufficient for sexual intercourse” [85], is one of the most common sexual dysfunctions in men. Erectile dysfunction is more common with advancing age, and since the aged population is increasing, its prevalence will continue to rise [86]. Diabetes mellitus is the most frequent organic cause of erectile dysfunction. In the population-based Massachusetts Male Aging Study (MMAS), the age-adjusted prevalence of complete erectile dysfunction was 28% in men with treated diabetes. Minimal, moderate, and complete erectile dysfunction together had a prevalence of 64% in men with treated diabetes [87]. In a recent clinic-based survey in Italy involving 9868 men with diabetes, 45.5% of those aged over 59 years reported erectile dysfunction [88]. Table 3.6 shows that erectile dysfunction is encountered in 20-52% of type 1 and 36-54% of type 2 diabetic patients [61,87-93]. Once erectile dysfunction has developed, it is likely to persist in most patients. Risk factors and clinical correlates include duration of diabetes, glycemic control, each of the chronic diabetic complications, and smoking [88].

Incidence

The MMAS recently reported the incidence of erectile dysfunction after an average follow-up of 8.8 years in a population-based cohort of 847 men aged 40-69 years without erectile dysfunction at baseline. The crude incidence of erectile dysfunction was 26 cases per 1000 man-years (95% CI: 23-30) and increased with age, lower education, diabetes, heart disease, and hypertension. In diabetic patients the incidence of erectile dysfunction was twice as high, 51 cases per 1000 man-years. Population projections for men aged 40-69 years suggest that 617715 new cases of erectile dysfunction in the USA (white males only) are expected annually, a considerable proportion of whom will be diabetic subjects [94].

Etiological Factors Related to Diabetic Neuropathy

Whilst debate continues about the precise nature of the pathophysiological changes that eventually result in diabetic neuropathy, there is some agreement over the risk factors associated with its development. The risk factors and risk indicators for DSP and the relative degrees of their association with it are listed in Table 3.7.

Table 3.7 Risk factors and markers of diabetic polyneuropathy

	Type 1 diabetes	Type 2 diabetes
Age	+	+
Sex	-	-
Height	+	(+)
Weight	-	(+)
Hyperglycemia	++	++
Hypoinsulinemia	n.a.	+
Duration of diabetes	++	++
Smoking	+	(+)
Alcohol	(+)	(+)
Hyperlipidemia	(+)	(+)
Hypertension	++	(+)
Nephropathy	++	+
Retinopathy	++	+
CAN	++	++
Macroangiopathy	(+)	(+)

Association strong, ++; moderate, +; disputed, (+): not found. -; CAN, cardiovascular autonomic neuropathy: n.a., not applicable

Hyperglycemia

The central role of hyperglycemia has been demonstrated in a range of studies. Mean HbA_1c was approximately 1% higher in men with newly diagnosed type 2 diabetes who went on to develop DSP 10 years later, than in those who did not[33]. The risk of developing DSP (as measured by the odds ratio) has been calculated to rise by approximately 10-15% for every I mmol/l rise in fasting plasma glucose or every 1% rise in HbA_1c[6,17]. The importance of hyperglycemia has of course been confirmed in interventional studies. The DCCT demonstrated that intensive glycemic control led to a 64%reduction in the five-year risk of developing DSP [2] in patients with type 1 diabetes. In the similar but smaller Stockholm Diabetes Intervention Study, symptoms of DSP developed in only 14% of those who were intensively treated, compared to 32% in the conventional treatment arm [95]. The effect of glucose lowering in type 2 diabetes is less clear. In the large UKPDS study, strict glycemic control had no significant impact on the development of DSP over the first 12 years [46]. Among the relatively small number of subjects who were followed to 15 years, a significant risk reduction was apparent, but only in those in the main study. Amongst the overweight subjects, whose intensive therapy was primarily with metformin, there was no impact of intensive therapy at all on DSP [96].

In the Rochester Diabetic Neuropathy Study, mean HbA₁, severity of diabetic retinopathy, and a term calculated mean In (24-hour proteinuria multiplied by duration of diabetes) were the main covariates for severity of DSP at the 7-year follow-up [97].

Thus, the etiological link between hyperglycemia and DSP seems sound, but while it is clear that glucose lowering protects against the development of neuropathy in type 1 diabetes, the case is not yet proven in type 2 diabetes.

Diabetes Duration

DSP, like the other specific diabetic complications, is rare at the outset of type 1 diabetes. It is more common amongst people with established type 2 diabetes than among those with newly diagnosed diabetes [5,6]. This association between disease duration and the risk of diabetes is strong, has been confirmed in a variety of studies in both type 1 and type 2 diabetes, and remains after adjustment for age [5,14,17,49]. In a UK study, the prevalence of DSP rose from 21% in those with a diabetes duration of less than five years to 37% in people with a duration of over 10 years [8]. In a Spanish study, the prevalence rose from 14% at under five years' duration to 44% at a duration of more than 30 years [19]. Interestingly, in one study the association between DSP and duration of diabetes was not seen in those over the age of 54, but was strong in those aged under 54 [3]. This, however, has not been reported elsewhere.

Age

Advancing age has been widely reported to increase the risk of DSP. In a large sample (6487 subjects) of UK diabetic hospital outpatients, the prevalence of DSP rose from 5% in the 20-to 29-year age group to 44% in the 70- to 79-year age group [8]. Amongst a population-based sample from Egypt, DSP was detected in 8% of those aged 20-44, and in 23% of those aged45 years or over [6]. In this study, the risk of DSP independently attributed to age approximately doubled for every 10 year increment in age. A number of other studies have also documented age as an independent risk factor for DSP neuropathy [11,14,17,20]. However, the influence of age may not be straightforward. Neurological function, especially vibration perception threshold (VPT), deteriorates with advancing age even in the normal, nondiabetic population [98,99]. Whilst some studies have adjusted for the effects of age on VPT, this is not universal. Furthermore, diagnostic criteria for DSP often include other parameters, such as clinical signs, monofilament sensitivity and electrophysiology, which are not adjusted for age. Interestingly, in a study from France, where VPT was adjusted for age, height, and gender, there was no relationship between neuropathy and age [3]. Data from Mauritius confirmed the increase in VPT with age in both diabetic and nondiabetic subjects [5]. However, when neuropathy was defined as having a VPT above the normal range (mean plus 2 standard deviations in the nondiabetic population) defined separately in each of three age categories, there was no relationship between age and neuropathy in either cross-sectional or prospective analyses [5]. Thus, the influence of age may have been overestimated in some studies.

Hypertension

Hypertension is attractive as an etiological factor in DSP, as it could be viewed as lending weight to the vascular theory of the pathophysiology of DSP. Hypertension has been associated with DSP in several studies, most notably in the data from the Pittsburgh cohort of type 1 diabetes [49]. In this study, hypertension was the single strongest predictor of DSP, and was associated with an approximately four-fold risk of developing DSP over six years. In the EURODIAB Prospective Complications Study, systolic blood pressure was shown to be one of the predictors of the development of DSP after adjustment for age, duration of diabetes, and HbA_1c [50]. Reports in type 2 diabetes, however, have been conflicting [17,20,33,61,100], but mainly have not confirmed this association. In a small interventional study (including subjects with both type 1 and type 2 diabetes), treatment with the ACE inhibitor trandolapril reduced progression of electrophysiological parameters of DSP, and also lowered systolic blood pressure [101]. However, the UKPDS reported that intensive blood pressure lowering with a variety of agents had no effect in ameliorating the progression to DSP [102].

Hypoinsulinemia

Partanen et al. [33] have suggested a link between hypoinsulinemia and DSP, which they believed resulted from the possible beneficial effects of insulin and C-peptide on neuronal metabolism and function [103]. They found that baseline fasting and 2-hour insulin levels were lower in newly diagnosed Finnish male subjects with type 2 diabetes who developed DSP 10 years later than in those who remained free of DSP. In the San Luis Valley Study [47], these findings were also evident (using C-peptide), but this univariate association disappeared when diabetes duration was taken into account. Our own data from Mauritius [5] also suggested a similar association with hypoinsulinemia, but, perhaps because of the low prevalence of DSP in that study, the association was not apparent in all analyses.

Diabetes duration almost certainly has an important confounding effect on the assessment of this potential relationship, since it might be expected both to increase the prevalence of DSP and to reduce insulin levels. It is difficult to account for duration in type 2 diabetes, as the date of disease onset is not usually known. Only a study in a population in which the time of onset of type 2 diabetes can be accurately estimated will allow better insight into the role of hypoinsulinemia.

Hyperinsulinemia

Several population-based studies suggest that high insulin levels are associated with disturbed autonomic nervous system activity. High insulin levels were significant predictors of the development of CAN 10 years after the diagnosis of type2 diabetes, irrespective of obesity and glycemia [68]. In the Atherosclerosis Risk in Communities (ARIC) Study patients with the metabolic syndrome (hypertension, type 2 diabetes, and/or dyslipidemia) showed significantly lower HRV indices than did those without these disorders. An increase in fasting insulin of 1 SD was associated with 88% higher odds of having lower HRV [104]. In nondiabetic individuals an inverse association was observed between serum fasting insulin and parasympathetic activity [105]. In the Zutphen Elderly Study fasting C-pepride and glucose tolerance were associated with a QTc prolongation which itself is related to CAN [106].

Other Risk Factors

Cigarette smoking has been identified as an independent risk factor for DSP in two different studies of type 1 diabetes, in which it was associated with an approximate doubling of the risk of DSP [14,49]. Smoking was also found to carry an independent risk in the San Luis Valley study of type 2 diabetes [47], but was actually associated with a protective effect in US veterans [17], and had only a weak (and not independent) association in the study from Mauritius [5].

Alcohol consumption has been associated with DSP on a number of occasions [17], but it may be difficult, at least in epidemiological studies, to differentiate between diabetic neuropathy in which alcohol is a risk factor, and alcoholic neuropathy in a person with diabetes.

DSP, like other metabolic, nutritional, and toxic neuropathies, is a distal disease which is first manifest in the feet. It is therefore, clearly a length-dependent process, although the pathophysiology underpinning this phenomenon is not understood. It seems logical, therefore, that height, as a surrogate measure of the length of the longest nerves, might be associated with DSP. This hypothesis has been borne out in a number of different cross-sectional and prospective studies [5,14,17,49].

There are no consistent findings that would indicate a relationship between gender and the risk of DSP.

High total cholesterol [6] and elevated triglycerides [14] have been reported as independent risk factors for DSP (after adjustment for HbA_lc, age, and other potential confounders). Elevated LDL was also found to predict neuropathy in a study of type 1 diabetes, but the association was lost after adjustment for other risk factors [49]. In the EURODIAB Prospective Complications Study BMI, albuminuria, triglycerides, cholesterol, and systolic blood pressure were shown to be significant predictors of the development of DSP after adjustment for age, duration of diabetes, and HbA_lc[50]. In type 2 diabetic patients the link between DSP and dyslipidemia must, however, remain tentative at this stage, as several studies have failed to observe such a relationship [5,33].

Physiological factors for which an inverse relationship to HRV has been described include age, heart rate, female gender, body mass index, sitting or standing body position, blood pressure, and pregnancy. Pathophysiological correlates unfavorably affecting HRV include the duration of diabetes, any of the chronic diabetic complications, metabolic syndrome, peripheral vasomotor abnormalities, medial arterial calcification, peripheral edema, reduced muscle sympathetic nerve activity (MSNA), cardiovascular risk factors such as high LDL cholesterol and smoking, as well as various drugs [107]. The population-based Hoorn Study, including subjects with NGT, IGT, and type 2 diabetes, recently reported that the strongest determinants of cardiovascular autonomic function were age and use of antihypertensive drugs in subjects with NGT, insulin levels in those with type 2 diabetes [108], and albuminuria in subjects with IGT or type 2 diabetes [109].

Genetic Factors

Striking ethnic and racial differences in the prevalence of nephropathy and macroangiopathy have been reported, but no such effects have been observed in the population-based surveys of diabetic neuropathy [20,26]. Although hyperglycemia and duration of diabetes are generally accepted to represent major contributory factors to the prevalence of diabetic neuropathy, many diabetologists have come across patients who do not develop neuropathy despite long-term poor glycemic control. Whole-genome screening and candidate gene strategies can be applied to the genetics of type 1 diabetes complications [110].The most significant results were obtained regarding a role for polymorphisms of the renin-angiotensin system in diabetic nephropathy [110]. In diabetic polyneuropathy, reduced Na⁺/K⁺-ATPase activity and increased aldose reductase activity have been suggested to play an important pathogenetic role. Na⁺/K⁺-ATPase is encoded by various genes, of which the ATP1 Al gene is expressed predominantly in peripheral nerves and erythrocytes. A case-control study found that type 1 diabetic patients bearing a restriction fragment length polymorphism (RFLP) of the ATP1 Al gene carried a 6.5-fold (95% CI: 3.3-13) increased risk of peripheral neuropathy. Moreover, these patients showed reduced erythrocyte Na⁺/K⁺-ATPase activity [111]. Another case-control study showed increased susceptibility to peripheral neuropathy in type 1 diabetic patients who had a polymorphism at the 5 end of the aldose reductase (ALR2)gene [112]. In contrast, a study in Japanese type 2 diabetic patients found an association of the ALR2 gene polymorphism in the 5 region with retinopathy but not peripheral and autonomic neuropathy or nephropathy [113].

Thus, at present only a low level of evidence for a role of these candidate genes obtained from case-control studies is available. Only prospective controlled trials using strata selected along a candidate gene would clarify whether polymorphisms might have therapeutic relevance in future.

Prognosis

Distal Symmetrical Polyneuropathy

Available information on the mortality associated with DSP is sparse. In a cohort of 134 type 2 diabetic patients randomly selected from the register of the Helsinki Diabetes Association followed prospectively over 9 years, neuropathy as assessed by NCV was identified as an independent predictor of mortality after adjusting for macroangiopathy, albumin excretion rate, and HbA_1c [114]. In the Stockholm Diabetes Intervention Study 7.5-year mortality was predicted not only by albuminuria but also by sural sensory nerve action potential amplitude and sympathetic nerve function [115]. In a university hospital setting a 14-year observational study including 794 patients demonstrated that reduced VPT was the most significant predictor of mortality. Proteinuria and type 1 diabetes, but not HbA_1c and smoking, were other important risk factors [116]. In a clinic-based cohort of 583 type 1 diabetic patients who were followed over 1-11.5 years, both abnormal NCV and abnormal AFTs were associated with an increased risk of mortality. Mortality rates after 10 years in the groups with normal and highly abnormal NCV were 10% and 50% respectively, while the corresponding percentages were 10% and 43% respectively for the AFTs. Combining NCV and AFTs into a total neuropathy score resulted in an even stronger predictor of mortality than each of the two measures alone [117]. Thus, there is accumulating evidence to suggest that not only surrogate markers of microangiopathy, such as albuminuria, but also those used for polyneuropathy, such as NCV and VPT, may predict mortality in diabetic patients, but clearly further studies are needed to assess the prognostic role of polyneuropathy in diabetes. In diabetic patients with ultimate clinical endpoints of neuropathy such as foot ulcers the risk of death was increased to 12 per 100 person-years of follow-up, compared to 5 per 100 person-years in those without foot ulcers [118].

Cardiovascular Autonomic Neuropathy

A number of largely clinic-based prospective studies have demonstrated increased mortality among diabetic patients with symptomatic CAN or those with abnormal cardiovascular reflex tests. Their major findings are summarized in Table 3.8 [68,74,117,119 128]. The overall mortality rates over periods up to 10 years were about 26% in diabetic patients with CAN compared with 4% in those without evidence of CAN. However, it must be kept in mind that autonomic dysfunction may also be found in the absence of diabetes as a consequence of common cardiovascular diseases such as coronary artery disease, myocardial infarction, and heart failure. It has been shown that reduced HRV is an independent indicator of poor prognosis in these patients [129]. Since cardiovascular diseases represent the major cause of death in diabetic patients [130], the impact of diabetes and, for example, coronary sclerosis on the autonomic nervous system may overlap in a number of patients to such a degree that CAN may at least not be the only factor responsible for the increased mortality [127]. On the other hand, there is evidence that CAN contributes to the poor prognosis in diabetic patients as an independent factor [122].

The mechanisms by which CAN leads to the increased mortality remain a matter of debate, but two hypotheses have been suggested [123,131]. A number of studies have shown an association between CAN and QT interval prolongation, and a recent metaanalysis revealed a 2.3-fold increased risk of CAN in diabetic patients showing a prolonged QT interval [132], leading to the speculation that, in analogy to the QT prolongation encountered in the idiopathic long QT syndrome, which is characterized by recurrent episodes of syncope or cardiac arrest due to torsades de pointes [133], CAN might also predispose patients to malignant ventricular arrhythmias and sudden death. A recent five-year study from Italy showed a considerably increased risk of mortality in type 1 diabetic patients with QTc prolongation (odds ratio: 24.6 [95% CI: 6.5-92.9]) which requires further confirmation [128]. Although the studies listed in Table 3.8 were not sufficiently large to allow unequivocal conclusions regarding an increased incidence of sudden death in CAN, recent evidence indicates an important role of the autonomic nervous system in triggering sudden death in both nondiabetic and diabetic subjects with low HRV. A marked decrease in HRV in nondiabetic patients with coronary artery disease was present immediately before the onset of the ST shift precipitating ischemic sudden death, suggesting that sympathovagal imbalance may trigger fatal arrhythmias during acute myocardial ischemia [134]. HRV was also diminished in nondiabetic survivors of sudden death not associated with coronary artery disease, indicating that abnormal autonomic activity may trigger malignant arrhythmia independently of coronary artery disease [135]. In subjects with inducible ventricular tachycardia (VT), 42% of whom were diabetic, HRV was markedly lower than in those without evidence of VT [136]. In unselected patients a 2.6-fold higher relative risk of sudden death within two years (after adjustment for age, left ventricular dysfunction, and history of myocardial infarction) was observed among patients with reduced HRV than in those with normal HRV. Among the subjects who died suddenly. 9.3% were diabetic, whereas only 3.5% of the patients who survived had diabetes [137]. The Zutphen Study [138] recently showed that low HRV predicts mortality from all causes in the general population. The population-based Honolulu Heart Program demonstrated that orthostatic hypotension defined as a drop in systolic blood pressure of at least 20 mmHg or in diastolic blood pressure of at least 10 mmHg predicts four-year all-cause mortality in a cohort of 3522 elderly Japanese-American men aged 71-93. Four-year age-adjusted mortality rates were 57 and 39 per 1000 person-years in men with and without orthostatic hypotension, respectively [139].

The second hypothesis for the explanation of the increased mortality in CAN suggests that impaired central control of respiration rather than abnormal cardiovascular reflexes contributes to the poor prognosis [131]. This view is supported by studies reporting an increased prevalence of sleep apnea and nocturnal oxygen desaturation in diabetic patients with CAN [140,141]. Several studies have demonstrated impaired ventilatory responses to progressive hypercapnia or hypoxemia in CAN [142–144]. This impairment is not due to peripheral factors such as abnormal lung function or diaphragmatic muscle alterations, but to defective central control of respiration. This last suggestion is supported by the finding that naloxone, a specific opioid antagonist, produced no increase in CO₂ response in diabetic patients with CAN, in contrast to healthy subjects. However, a lack of effect of naloxone on CO₂ response was also observed in diabetic patients without CAN [143]. Hence, the question remains open as to whether the altered central control of respiration is specific to CAN or is rather a feature of diabetes itself. To complicate matters even more, an increased hypercapnic drive, indicating an exaggerated response of the central drive due to removal of the sympathetic inhibition, was recently found in patients with CAN and orthostatic hypotension [144].

There is little information available on the prognostic significance of the other manifestations of autonomic neuropathy. Diabetic gastroparesis diagnosed by scintigraphy was not associated with a poor prognosis in two studies performed over 3-5 and 9-14 years, respectively [145,146].

Conclusions

The study of the epidemiology of diabetic neuropathy remains clouded by lack of agreement over diagnostic criteria and variation in subject selection methods. It is essential that agreement is reached over diagnosis, although it is hard at the present time to see how this is going to come about. One issue that may be relevant in this context is the basis on which the diagnosis should be made. Is diabetic neuropathy a condition which predisposes to clinical endpoints such as foot ulceration and amputation, in which case quantitative sensory testing should suffice, or is it a condition in which neurological function differs from that in a healthy population, in which case diagnosis may require a more detailed assessment?

The available data indicate that DSP is present in approximately 30% of hospital clinic patients, 20% of patients in primary care, and 10% of the total diabetic population, including both diagnosed and undiagnosed diabetes. The major confirmed risk factors are poor glycemic control, diabetes duration and height, with possible roles for hypertension (probably only in type 1 diabetes), age, smoking, hypoinsulinemia, and dyslipidemia.

There are as yet relatively few epidemiological data on the various manifestations of autonomic neuropathy from representative cohorts of diabetic patients, except for erectile dysfunction. Estimates from the available studies suggest that CAN is encountered one of every 4-6 men and ED is observed in one of every 2-3 men. Symptomatic orthostatic hypotension is relatively uncommon. Gastrointestinal symptoms are common in both diabetic and nondiabetic subjects, suggesting that a considerable proportion of these symptoms in diabetic patients is due to causes other than autonomic neuropathy

Clinic-based data suggest that particularly CAN but possibly also DSP are associated with increased mortality in diabetic patients, but prospective population-based studies are required to confirm these findings.

References

[1] Statement C. Report and recommendations of the San Antonio Conference on diabetic neuropathy. Diabetes Care 1988; 37: 1000-4.

[2] The Diabetes Control and Complications Trial Research Group. The effect of intensive diabetes therapy on the development and progression of diabetic neuropathy. Ann Intern Med 1995; 122: 561-8.

[3] Delcourt C, Vauzelle-Kervroedan F, Cathelineau G, Papoz L. Low prevalence of long-term complications in non-insulin-dependent diabetes mellitus in France: a multicenter study. J Diabetes Complications 1998; 12: 88-95.

[4] Nelson RG, Gohdes DM, Everhart JE, Hartner JA, Zwemer FL, Pettitt DJ, Knowler WC. Lower-extremity amputations in NIDDM; 12-yr follow-up study in Pima Indians. Diabetes Care 1988; 11:8-16.

[5] Shaw JE, Hodge AM, Courten M de, Dowse GK, Gareeboo H, Tuomilehto J, Alberti KGMM, Zimmet PZ. Diabetic neuropathy in Mauritius: prevalence and risk factors. Diabetes Res Clin Pract 1998; 42: 131-9.

[6] Herman WH, Aubert RE, Engelgau MM, Thompson TJ, Ali MA, Sous ES, Hegazy M, Badran A, Kenny SJ, Gunter EW, Malarcher AM, Brechner RJ, Wetterhall SF, DeStefano F, Smith PJ, Habib M, abd el Shakour S, Ibrahim AS, el Behairy EM. Diabetes mellitus in Egypt: glycaemic control and microvascular and neuropathic complications. Diabet Med 1998; 15: 1045-51.

[7] Dyck PJ, Dyck PJB, Velosa JA, Larson TS, O'Bnen PC, the Nerve Growth Factor Study Group. Patterns of quantitative sensation testing of hypoesthesia and hyperalgesia are predictive of diabetic polyneuropathy. A study of three cohorts. Diabetes Care 2000; 23: 510-7.

[8] Young MJ, Breddy JL, Veves A, Boulton AJM. The prediction of diabetic neuropathic foot ulceration using vibration perception thresholds. A prospective study. Diabetes Care 1994; 17: 557-60.

[9] Rith-Najarian SJ, Stolusky T, Gohdes DM. Identifying diabetic patients at high risk for lower-extremity amputation in a primary health care setting. A prospective evaluation of simple screening criteria. Diabetes Care 1992; 15: 1386-9.

[10] Young MJ, Boulton AJ, MacLeod AF, Williams DR, Sonksen PH. A multicentre study of the prevalence of diabetic peripheral neuropathy in the United Kingdom hospital clinic population. Diabetologia 1993; 36: 150-4.

[11] Walters DP, Gatting W, Mullee MA, Hill RD. The prevalence of diabetic distal sensory neuropathy in an English community. Diabet Med 1992; 9: 349-53.

[12] Harris MI. Classification, diagnostic criteria, and screening for diabetes. In: National Diabetes Data Group, editors. Diabetes in America. 2nd ed. Bethesda Md: National Institutes of Health and Digestive and Kidney Diseases; 199NIH publ 1995; 95-1468: 15-36.

[13] American Diabetes Association. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 1997; 20: 1183-97.

[14] Tesfaye S, Stevens LK, Stephenson JM, Fuller JH, Plater M, lonescu-Tirgoviste C, Nuber A, Pozza G, Ward JD. Prevalence of diabetic peripheral neuropathy and its relation to glycaemic control and potential risk factors: the EURODIAB IDDM Complications Study. Diabetologia 1996; 39: 1377-84.

[15] Fedele D, Comi G, Coscelli C, Cucinotta D, Feldman EL, Ghirlanda G, Greene DA, Negrin P, Santeusanto F. A multicenter study on the prevalence of diabetic neuropathy in Italy. Italian Diabetic Neuropathy Committee. Diabetes Care 1997; 20: 836-43.

[16] Wikblad K, Smide B, Bergstrom A, Kessi J, Mugusi F. Outcome of clinical foot examination in relation to self-perceived health and glycaemic control in a group of urban Tanzanian diabetic patients. Diabetes Res Clin Pract 1997; 37: 185-92.

[17] Adler AI, Boyko EJ, Ahroni JH, Stensel V, Forsberg RC, Smith DG. Risk factors for diabetic peripheral sensory neuropathy. Results of the Seattle Prospective Diabetic Foot Study. Diabetes Care 1997; 20: 1162-7.

[18] Beghi E, Monticelli ML. Diabetic polyneuropathy in the elderly. Prevalence and risk factors in two geographic areas of Italy. Italian General Practitioner Study Group (IGPSG). Acta Neurol Scand 1997; 96: 223-8.

[19] Cabezas-Cerrato J. The prevalence of clinical diabetic polyneuropathy in Spain: a study in primary care and hospital clinic groups. Neuropathy Spanish Study Group of the Spanish Diabetes Society (SDS). Diabetologia 1998; 41: 1263-9.

[20] Franklin GM, Shetterly SM, Cohen JA, Baxter J, Hamman RF. Risk factors for distal symmetric neuropathy in NIDDM. San Luis Val Diabetes Study. Diabetes Care 1994; 17: 1172-7.

[21] Dyck PJ, Kratz KM, Karnes JL, Litchy WJ, Klein R, Pach JM, Wilson DM, O'Brien PC, Melton LJ 3d, Service FJ. The prevalence by staged severity of various types of diabetic neuropathy, retinopathy, and nephropathy in a population-based cohort: the Rochester Diabetic Neuropathy Study. Neurology 1993; 43: 817-24.

[22] Kumar S, Ashe HA, Parnell LN, Fernando DJ, Tsigos C, Young RJ, Ward JD, Boulton AJM. The prevalence of foot ulceration and its correlates in type 2 diabetic patients: a population-based study. Diabet Med 1994; 11: 480-4.

[23] Knuiman MW, Welborn TA, McCann VJ, Stanton KG, Constable IJ. Prevalence of diabetic complications in relation to risk factors. Diabetes 1986; 35: 1332-9.

[24] Fujimoto WY, Leonetti DL, Kinyoun JL, Shuman WP, Stolov WC, Wahl PW. Prevalence of complications among second-generation Japanese-American men with diabetes, impaired glucose tolerance, or normal glucose tolerance. Diabetes 1987; 36: 730-9.

[25] Maser RE. Steenkiste AR, Dorman JS, Kamp Nielsen V, Bass EB, Manjoo Q, Drash AL, Becker DJ, Kuller LH, Greene DA, Orchard TJ. Epidemiological correlates of diabetic neuropathy. Report from Pittsburgh Epidemiology of Diabetes Complications Study. Diabetes 1989; 38: 1456-61.

[26] Harris M, Eastman R, Cowie C. Symptoms of sensory neuropathy in adults with NIDDM in the US population. Diabetes Care 1993; 16: 1446-52.

[27] Nielsen JV. Peripheral neuropathy, hypertension, foot ulcers and amputations among Saudi Arabian patients with type 2 diabetes. Diabetes Res Clin Pract 1998; 41: 63-9.

[28] Verhoeven S, Ballegooie E van, Casparie AF. Impact of late complications in type 2 diabetes in a Dutch population. Diabet Med 1991; 8: 435-8.

[29] Neeling JND De, Beks PJ, Bertelsmann FW, Heine RJ, Bouter LM. Peripheral somatic nerve function in relation to glucose tolerance in an elderly Caucasian population: the Hoorn Study. Diabet Med 1996; 13: 960-6.

[30] Müller UA, Ross IS, Klinger H. Häufigkeit der Hypopallästhesie bei insulinspritzenden Diabetikern in einer unausgewählten Stadtpopulation, Diab Stoffw 1993; 2: 439-42.

[31] Ziegler D, Gries FA, Mühlen H, Rathmann W, Spüler M, Lessmann F, the DiaCAN Multicenter Study Group. Prevalence and clinical correlates of cardiovascular autonomic and peripheral diabetic neuropathy in patients attending diabetes centers. Diab Metab 1993; 19: 143-51.

[32] Veglio M, Sivieri R. Prevalence of neuropathy in IDDM patients in Piemonte, Italy. The Neuropathy Study Croup of the Italian Society for the Study of Diabetes. Piemonte Affiliate. Diabetes Care 1993; 16: 456-61.

[33] Partanen J, Niskanen L, Lehtinen J, Mervaala E, Siitonen O, Uusitupa M. Natural history of peripheral neuropathy in patients with non-insulin-dependent diabetes mellitus. N Engl J Med 1995; 333: 89-94.

[34] Ratzmann KP, Raschke M, Gander B, Schimke P. Prevalence of peripheral and autonomic neuropathy in newly diagnosed type II (noninsulin dependent) diabetes. J Diab Complications 1991; 5: 1-5.

[35] Wang WQ, Ip TP, Lam KS. Changing prevalence of retinopathy in newly diagnosed non-insulin dependent diabetes mellitus patients in Hong Kong. Diabetes Res Clin Pract 1998; 39: 185-91.

[36] Gujral JS, McNally PG, O'Malley BP, Burden AC. Ethnic differences in the incidence of lower extremity amputation secondary to diabetes mellitus. Diabet Med 1993; 10: 271-4.

[37] Bharucha NE, Bharucha AE, Bharucha EP. Prevalence of peripheral neuropathy in the Parsi community of Bombay. Neurology 1991; 41: 1315-7.

[38] Savettieri G, Rocca WA, Salemi G, Meneghini F, Grigoletto F, Morgante L, Reggio A, Costa V, Coraci MA, Di Perri R. Prevalence of diabetic neuropathy with somatic symptoms: a door-to-door survey in two Sicilian municipalities. Neurology 1993; 43: 1115-20.

[39] Watkins PJ. The natural history of the diabetic neuropathies. Q J Med 1990; 77: 1209-18.

[40] Eriksson KF, Nilsson H, Lindgärde F, Österlin S, Dahlin LB, Lilja B, Rosen I, Sundkvist G. Diabetes mellitus but not impaired glucose tolerance is associated with dysfunction of peripheral nerves. Diabet Med 1994; 11: 279-85.

[41] Sosenko JM, Kato M, Soto R, Goldberg RB. Sensory function at diagnosis and in early stages of NIDDM in patients detected through screening. Diabetes Care 1992; 15: 847-52.

[42] Sundkvist G, Dahlin L-B, Nilsson H, Eriksson K-F, Lindgärde F, Rosen I, Lattimer SA, Sima AAF, Sullivan K, Greene DA. Sorbitol and myo-inositol levels and morphology of sural nerve in relation to peripheral nerve function and clinical neuropathy in men with diabetic, impaired, and normal glucose tolerance. Diabet Med 2000; 17: 259-68.

[43] Melton III LJ, Dyck PJ. Epidemiology. In: Dyck PJ, rhomas PK, Asbury AK, Winegrad AI, Porte D Jr, editors. Diabetic neuropathy. Philadelphia, Pa: W. B. Saunders; 1987: 27-35.

[44] Pirart J. Diabetes mellitus and its degenerative complications: a prospective study of 4,400 patients observed between 1947 and 197DiabetesCare 1978; 252: 168-88.

[45] Töyry JP, Partanen JVS, Niskanen LK, Länsimies EA, Uusitupa MIJ. Divergent development of autonomic and peripheral somatic neuropathies in NIDDM. Diabetologia 1997; 40: 953-8.

[46] UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998; 352: 837-53.

[47] Sands ML, Shetterly SM, Franklin GM, Hamman RF. Incidence of distal symmetric (sensory) neuropathy in NIDDM. San Luis Val Diabetes Study. Diabetes Care 1997; 20: 322-9.

[48] Dyck PJ, Davies JL, Litchy WJ, O'Brien PC. Longitudinal assessment of diabetic polyneuropathy using a composite score in the Rochester Diabetic Neuropathy Study cohort. Neurology 1997; 49: 229-39.

[49] Forrest KY, Maser RE, Pambianco G, Becker DJ, Orchard TJ. Hypertension as a risk factor for diabetic neuropathy: a prospective study. Diabetes 1997; 46: 665-70.

[50] Tesfaye S, Chaturvedi N, Eaton SEM, Ward JD, Fuller JH. Cardiovascular risk factors predict the development of diabetic peripheral neuropathy. Diabetes 2000; 49 (Suppl 1): A34.

[51] Hillson RM, Hockaday TDR, Newton DJ. Hyperglycaemia is one correlate of deterioration in vibration sense during the 5 years after diagnosis of type 2 (non-insulin-dependent) diabetes. Diabetologia 1984; 26: 122-6.

[52] Sosenko JM, Kato M, Soto R, Bild DE. A prospective study of sensory function in patients with type 2 diabetes. Diabet Med 1993; 10: 110-4.

[53] Young RJ, Macintyre CCA, Martyn CN, Prescott RJ, Ewing DJ, Smith AF, Viberti G, Clarke BF. Progression of subclinical polyneuropathy in young patients with type 1 (insulin-dependent) diabetes: associations with glycaemic control and microangiopathy (microvascular complications). Diabetologia 1986; 29: 156-61.

[54] Young RJ, Macintyre CCA, Ewing DJ, Prescott RJ. Prediction of neuropathy over 5 years in young insulin-dependent diabetic patients. Diabet Med 1988; 5: 7.

[55] Benbow SJ, Chan AW, Bowsher D, MacFarlane IA, Williams G. A prospective study of painful symptoms, small-fibre function and peripheral vascular disease in chronic painful diabetic neuropathy. Diabet Med 1994; 11: 17-21.

[56] Ziegler D, Piolot R. Prevention of diabetic neuropathy by near-normoglycemia. A 12-year prospective study from the diagnosis of IDDM. Diabetes 1998; 47 (Suppl 1): A63.

[57] Macleod AF, Till S, Sönksen PHS. Discussion of the clinical trials of the aldose reductase inhibitor tolrestat. Int Proc J 1991; 4: 17-24.

[58] Negrin P, Zara G. Conduction studies as prognostic parameters in the natural history of diabetic neuropathy: a long-term follow-up of 114 patients. Electromyogr Clin Neurophysiol 1995; 35: 341-50.

[59] Watanabe K, Hagura R, Akanuma Y, Takasu T, Kajinuma H, Kuzuya N, Irie M. Characteristics of cranial nerve palsies in diabetic patients. Diabetes Res Clin Pract 1990; 10: 19-27.

[60] Longstreth GF. Diabetic thoracic polyradiculopathy: ten patients with abdominal pain. Am J Gastroenterol 1997; 92: 502-5.

[61] Neil HA, Thompson AV, John S, McCarthy ST, Mann JI. Diabetic autonomic neuropathy: the prevalence of impaired heart rate variability in a geographically defined population. Diabet Med 1989; 6: 20-4.

[62] Maser RE, Pfeifer MA, Dorman JS, Kuller LH, Becker DJ, Orchard TJ. Diabetic autonomic neuropathy and cardiovascular risk. Pittsburgh Epidemiology of Diabetes Complications Study III. Arch Intern Med 1990; 150: 1218-22.

[63] O'Brien IAD, O'Hare JP, Lewin IG, Corrall RJM. The prevalence of autonomic neuropathy in insulin-dependent diabetes mellitus: a controlled study based on heart rate variability. Q J Med 1986; 61: 957-67.

[64] Stephenson J, Fuller JH, and EURODIAB IDDM Complications Study Group. Microvascular and acute complications in IDDM patients: the EURODIAB IDDM Complications Study, Diabetologia 1994; 37: 278-85.

[65] Tsutsu N, Nunoi K, Yokomizo Y, Kikuchi M, Fujishima M. Relationship between glycemic control and orthostatic hypotension in type 2 diabetes mellitus–a survey by the Fukuoka Diabetes Clinic Group. Diabetes Res Clin Pract 1990; 8: 115-23.

[66] Ziegler D, Dannehl K, Spüler M, Mühlen H, Gries FA. Prevalence of cardiovascular autonomic nerve dysfunction assessed by spectral analysis and standard tests of heart rate variation in newly diagnosed IDDM patients. Diabetes Care 1992; 15: 908-11.

[67] Lehtinen JM, Uusitupa M, Siitonen O, Pyörälä K. Prevalence of neuropathy in newly diagnosed NIDDM and nondiabetic control subjects. Diabetes 1989; 38: 1307-13.

[68] Töyry JP, Niskanen LK, Mäntysaari MJ, Länsimies EA, Uusitupa MIJ. Occurrence, predictors, and clinical significance of autonomic neuropathy in NIDDM, Ten-year follow up from the diagnosis. Diabetes 1996; 45: 308-15.

[69] Töyry JP, Niskanen LK, Lansimies EA, Partanen KPL, Uusitupa MIJ. Autonomic neuropathy predicts the development of stroke in patients with non-insulin-dependent diabetes mellitus. Stroke 1996; 27: 1316-8.

[70] Mustonen J, Uusitupa M, Mäntysaari M, Länsimies E, Pyörälä K, Laakso M. Changes in autonomic nervous function during the 4-year follow-up in middle-aged diabetic and nondiabetic subjects initially free of coronary heart disease. J Intern Med 1997; 241: 227-35.

[71] Quadri R, Ponzani P, Zanone M, Maule S, Grotta A La, Papotti G, Valentini M, Matteoda C, Chiandussi L, Fonzo D. Changes in autonomic nervous function over a 5-year period in non-insulin-dependent diabetic patients. Diabet Med 1993; 10: 916-9.

[72] Ziegler D, Mayer P, Mühlen H, Gries FA. The natural history of somatosensory and autonomic nerve dysfunction in relation to glycaemic control during the first 5 years after diagnosis of type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1991; 34: 822-9.

[73] Solders G, Thalme B, Aguirre-Aquino M, Brandt L, Berg U, Persson A. Nerve conduction and autonomic nerve function in diabetic children. Acta Paediatr 1997; 86: 361-6.

[74] Sampson MJ, Wilson S, Karagiannis P, Edmonds M, Watkins PJ. Progression of diabetic autonomic neuropathy over a decade in insulin-dependent diabetics. QJ Med 1990; 75: 635-46.

[75] Sundkvist G, Lilja B. Autonomic neuropathy in diabetes mellitus: a follow-up study. Diabetes Care 1985; 8: 129-33.

[76] Donaghue KC, Fung ATW, Fairchild JM, Howard NJ, Silink M. Prospective assessment of autonomic and peripheral nerve function in adolescents with diabetes. Diabet Med 1996; 13: 65-71.

[77] Levitt NS, Stansberry KB, Wynchank S, Vinik AI. The natural progression of autonomic neuropathy aund autonomic function tests in a cohort of people with IDDM. Diabetes Care 1996; 19: 751-4.

[78] Sundkvist G, Lilja B. Autonomic neuropathy predicts deterioration in glomerular filtration rate in patients with IDDM. Diabetes Care 1993; 16: 773-9.

[79] Ziegler D, Piolot R, Pour Mirza A, Karallus M, Feng B, Schommartz B, Schadewaldt P. ¹³C-octanoic acid breath test to assess diabetic gastroparesis: patterns of gastric emptying in newly diagnosed and long-term type 1 and type 2 diabetic patients. Clin Auton Res 1997; 7: 246.

[80] Enck P, Rathmann W, Spiekermann M, Czerner D, Tschöpe D, Ziegler D, Strohmeyer G, Cries FA. Prevalence of gastrointestinal symptoms in diabetic patients and non-diabetic subjects. Z Gastroenterol 1994; 32: 637-41.

[81] Schvarcz E, Palmer M, Ingberg CM, Aman J, Berne C. Increased prevalence of upper gastrointestinal symptoms in long-term type 1 diabetes mellitus. Diabet Med 1996; 13: 478-81.

[82] Ko GTC, Chan W-B, Chan JCN, Tsang LWW, Cockram CS. Gastrointestinal symptoms in Chinese patients with type 2 diabetes mellitus. Diabet Med 1999; 16: 670-4.

[83] Santis A De, Attili AF, Corradini SG, Scafato E, Cantagalli A, Luca C De, Pinto G, List D, Capocaccia L. Gallstones and diabetes: a case-control study in a free-living population sample. Hepatology 1997; 25: 787-90.

[84] Shaw JE, Parker R, Hollis S, Gokal R, Boulton AJM. Gustatory sweating in diabetes mellitus. Diabet Med 1996; 13: 1033-7.

[85] NIH consensus development panel on impotence. Impotence. JAMA 1993; 270: 83-90.

[86] Wagner G, Saenz de Tejada I. Update on male erectile dysfunction. BMJ 1998; 316: 678-82.

[87] Feldman HA, Goldstein I, Hatzichristou DG, Krane RJ, McKinlay JB. Impotence and its medical and psychosocial correlates: results of the Massachusetts male ageing study. J Urol 1994; 151: 54-61.

[88] Fedele D, Coscelli C, Santeusanio F, Bortolotti A, Chatenoud L, Colli E, Landoni M, Parazzini F. Erectile dysfunction in diabetic subjects in Italy. Diabetes Care 1998; 21: 1973-7.

[89] Kolodny RC, Kahn CB, Goldstein HH, Barnett DM. Sexual dysfunction in diabetic men. Diabetes 1974; 23: 306-9.

[90] Klein R, Klein PEK. Lee KE, Moss SE, Cruickshanks KJ. Prevalence of self-reported erectile dysfunction in people with long-term IDDM. Diabetes Care 1996; 19: 135-41.

[91] McCulloch DK, Campbell IW, Wu FC, Prescott RJ, Clarke BF. The prevalence of diabetic impotence. Diabetologia 1980; 18: 279-83.

[92] Cavan DA, Barnett AH, Leatherdale BA. Diabetic impotence: risk factors in a clinic population. Diabetes Res 1987; 5: 145-8.

[93] Hackett GI. Impotence–the most neglected complication of diabetes. Diabetes Care 1995; 28: 75-83.

[94] Johannes CB, Araujo AB, Feldman HA, Derby CA, Kleinman KP. McKinlay JB. Incidence of erectile dysfunction in men 40 to 69 years old: longitudinal results from the Massachusetts male aging study. J Urol 2000; 163: 460-3.

[95] Reichard P, Pihl M, Rosenqvist U, Sule J. Complications in IDDM are caused by elevated blood glucose level: the Stockholm Diabetes Intervention Study (SDIS) at 10-year follow up. Diabetologia 1996; 39: 1483-8.

[96] UK Prospective Diabetes Study (UKPDS) Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet 1998; 352: 854-65.

[97] Dyck PJ, Davies JL, Wilson DM, Service JF, Melton III. LJ, O'Brien PC. Risk factors for severity of diabetic polyneuropathy. Diabetes Care 1999; 22: 1479-86.

[98] Wiles PC, Pearce SM, Rice PJ, Mitchell JM. Vibration perception threshold: influence of age, height, sex, and smoking, and calculation of accurate centile values. Diabet Med 1991; 8: 157-61.

[99] Bloom S, Till S, Sonksen P, Smith S. Use of a biothesiometer to measure individual vibration thresholds and their variation in 519 non-diabetic subjects. Br Med J (Clin Res Ed) 1984; 288: 1793-5.

[100] Savage S, Estacio RO, Jeffers B, Schrier RW. Urinary albumin excretion as a predictor of diabetic retinopathy, neuropathy, and cardiovascular disease in NIDDM. Diabetes Care 1996; 19: 1243-8.

[101] Malik RA, Williamson S, Abbott C, Carrington AL, Iqbal J, Schady W, Boulton AJ. Effect of angiotensin-converting-enzyme (ACE) inhibitor trandolapril on human diabetic neuropathy: randomised double-blind controlled trial. Lancet 1998; 352: 1978-81.

[102] UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes. UKPDS 3BMJ 1998; 317: 703-13.

[103] Ido Y, Vindigni A, Chang K, Stramm L, Chance R, Heath WF, DiMarchi RD, Di Cera E, Williamson JR. Prevention of vascular and neural dysfunction in diabetic rats by C-peptide. Science 1997; 277: 563-6.

[104] Liao D, Sloan RP, Cascio WE, Folsom AR, Liese AD, Evans GW, Cai J, Sharrett AR. Multiple metabolic syndrome is associated with lower heart rate variability. The Atherosclerosis Risk in Communities Study. Diabetes Care 1998; 21: 2116-22.

[105] Liao D, Cai J, Brancati FL, Folsom A, Barnes RW, Tyroler HA, Heiss G. Association of vagal tone with serum insulin, glucose, and diabetes mellitus–the ARIC Study. Diabetes Res Clin Pract 1995; 30: 211-21.

[106] Dekker JM, Feskens EJM, Schouten EG, Klootwijk P, Pool J, Kromhout D. QTc duration is associated with levels of insulin and glucose tolerance. Zutphen Elderly Study. Diabetes 1996; 45: 376-80.

[107] 10Ziegler D. Diabetic cardiovascular autonomic neuropathy: prognosis, diagnosis, and treatment. Diabetes Metab Rev 1994; 10: 339-83.

[108] Gerritsen J, Dekker JM, TenVoorde BJ, Bertelsmann FW, Kostense PJ, Stehouwer CDA, Heine RJ, Nijpels G, Heethaar RM, Bouter LM. Glucose tolerance and other determinants of cardiovascular autonomic function: the Hoorn Study. Diabetologia 2000; 43: 561-70.

[109] Smulders YM, Jager A, Gerritsen J, Dekker JM, Nijpels G, Heine RJ, Bouter LM, Stehouwer CDA. Cardiovascular autonomic function is associated with (micro-)alburninuria in elderly Caucasian subjects with impaired glucose tolerance or type 2 diabetes. Hoorn Study. Diabetes Care 2000; 23: 1369-74.

[110] Marre M. Genetics and the prediction of complications in type 1 diabetes. Diabetes Care 1999; 22 (Suppl 2): B53-58.

[111] Vague P, Dufayet D, Coste T, Moriscot C, Jannot MF, Raccah D. Association of diabetic neuropathy with NA/K-ATPase gene polymorphism. Diabetologia 1997; 40: 506-11.

[112] Heesom AE, Millward A, Demaine AG. Susceptibility to diabetic neuropathy in patients with insulin dependent diabetes mellitus is associated with a polymorphism at the 5′ end of the aldose reductase gene. J Neurol Neurosurg Psychiatry 1998; 64: 213-6.

[113] Ichikawa F, Yamada K, Ishiyama-Shigemoto S, Yuan X, Nonaka K. Association of an (A-C)n dinucleotide repeat polymorphic marker at the 5′-region of the aldose reductase gene with retinopathy but not with nephropathy or neuropathy in Japanese patients with type 2 diabetes mellitus, Diabet Med 1999; 16: 744-8.

[114] Forsblom CM, Sane T, Groop P-H, Tötterman KJ, Kallio M, Saloranta C, Laasonen L, Summanen P, Lepäntalo M, Laatikainen L, Matikainen E, Teppo A-M, Koskimies S, Groop L. Risk factors for mortality in type II (non-insulin-dependent) diabetes: evidence of a role for neuropathy and a protective effect of HLA-DRDiabetologia 1998; 41: 1253-62.

[115] Reichard P, Pihl M. Mortality and treatment side-effects during long-term intensified conventional insulin treatment in the Stockholm Diabetes Intervention Study. Diabetes 1994; 43: 313-7.

[116] Coppini DV, Bowtell PA, Weng C, Young PJ, Sönksen PH. Showing neuropathy is related to increased mortality in diabetic patients - a survival analysis using an accelerated failure time model. J Clin Epidemiol 2000; 53: 519-23.

[117] Navarro X, Kennedy WR, Aeppli D, Sutherland DER. Neuropathy and mortality in diabetes: influence of pancreas transplantation. Muscle Nerve 1996; 19: 1009-16.

[118] Boyko EJ, Ahroni JH, Smith DG, Davignon D. Increased mortality associated with diabetic foot ulcer. Diabet Med 1996; 13: 967-72.

[119] Ewing DJ, Campbell IW, Clarke BF. The natural history of diabetic autonomic neuropathy. Q J Med 1980; 49: 95-108.

[120] Hasslacher C, Bässler G. Prognose der kardialen autonomen Neuropathie bei Diabetikern. Münch Med Wochenschr 1983; 125: 375-7.

[121] Navarro X, Kennedy WR, Loewenson RB, Sutherland DER. Influence of pancreas transplantation on cardiorespiratory reflexes, nerve conduction, and mortality in diabetes. Diabetes 1990; 39: 802-6.

[122] O'Brien IA, McFadden JP, Corral RJM. The influence of autonomic neuropathy on mortality in insulin-dependent diabetes. Q J Med 1991; 79: 495-502.

[123] Ewing DJ, Boland O, Neilson JMM, Cho CG, Clarke BF. Autonomic neuropathy. QT interval lengthening, and unexpected deaths in male diabetic patients. Diabetologia 1991; 34: 182-5.

[124] Jermendy G, Toth L, Vörös P, Koltai MZ, Pogatsa G. Cardiac autonomic neuropathy and QT interval length. A follow-up study in diabetic patients. Acta Cardiol 1991; 46: 189-200.

[125] Rathmann W, Ziegler D, Jahnke M, Haastert B, Cries FA. Mortality in diabetic patients with cardiovascular autonomic neuropathy. Diabet Med 1993; 10: 820-4.

[126] Luft D, Rak R, Renn W, Konz K, Eggstein M. Diabetische autonome Neuropathie: Verlauf und prognostische Bedeutung kardiovaskulärer Reflexteste. Diab Stoffw 1993; 2: 239-44.

[127] Orchard TJ, Lloyd CE, Maser RE, Kuller LH. Why does diabetic autonomic neuropathy predict IDDM mortality? An analysis from the Pittsburgh Epidemiology of Diabetes Complications Study. Diabetes Res Clin Pract 1996; 34 (Suppl): S165-171.

[128] Veglio M, Sivieri R, Chinaglia A, Scaglione L, Cavallo-Perin P. QT interval prolongation and mortality in type 1 diabetic patients: a 5-year cohort prospective study. Neuropathy Study Group of the Italian Society of the Study of Diabetes, Piemonte Affiliate. Diabetes Care 2000; 23: 1381-3.

[129] Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability, standards of measurement, physiological interpretation, and clinical use. Circulation 1996; 93: 1043-65.

[130] Jacobi RM, Nesto RW. Acute myocardial infarction in the diabetic patient: pathophysiology, clinical course and prognosis. J Am Coll Cardiol 1992; 20: 736-44.

[131] Page MMcB, Watkins PJ. Cardiorespiratory arrest and diabetic autonomic neuropathy. Lancet 1978; i: 14-6.

[132] Whitsel EA, Boyko EJ, Siscovick DS. Reassessing the role or QTc in the diagnosis of autonomic failure among patients with diabetes. A meta-analysis. Diabetes Care 2000; 23: 241-7.

[133] Anonymous. Neural mechanisms in sudden cardiac death: insights from long QT syndrome. Lancet 1991; 338: 1181-2.

[134] Pozzati A, Pancaldi LG, Di Pasquale G, Pinelli G, Bugiardini R. Transient sympathovagal imbalance triggers “ischemic” sudden death in patients undergoing electrocardiographic Holter monitoring. J Am Coll Cardiol 1996; 27: 847-52.

[135] Fei L, Anderson MH, Katritsis D, Sneddon J, Statters DJ, Malik M, Camm AJ. Decreased heart rate variability in survivors of sudden cardiac death not associated with coronary artery disease. Br Heart J 1994; 71: 16-21.

[136] Bikkina M, Alpert MA, Mukerji R, Mulekar M, Cheng B-Y, Mukerji V. Diminished short-term heart rate variability predicts inducible ventricular tachycardia. Chest 1998; 113: 312-6.

[137] Algra A, Tijssen JGP, Roelandt JRTC, Pool J, Lubsen J. Heart rate variability from 24-hour electrocardiography and the 2-year risk for sudden death. Circulation 1993; 88: 180-5.

[138] Dekker JM, Schouten EG, Klootwijk P, Pool J, Swenne CA, Kromhout D. Heart rate variability from short electrocardiographic recordings predicts mortality from all causes in middle-aged and elderly men. Zutphen Study. Am J Epidemiol 1997; 145: 899-908.

[139] Masaki KH, Schatz IJ. Burchfiel CM, Sharp DS, Chiu D, Foley D, Curb JD. Orthostatic hypotension predicts mortality in elderly men: the Honolulu Heart Program. Circulation 1998; 98: 2290-5.

[140] Rees PJ, Cochrane GM, Prior JG, Clark TJH. Sleep apnea in diabetic patients with autonomic neuropathy. J R Soc Med 1981; 74: 192-5.

[141] Neumann C, Martinez D, Schmid H. Nocturnal oxygen desaturation in diabetic patients with severe autonomic neuropathy. Diabetes Res Clin Pract 1995; 28: 97-102.

[142] Sobotka PA, Liss HP, Vinik AI. Impaired hypoxic ventilatory drive in diabetic patients with autonomic neuropathy. J Clin Endocrinol Metab 1986; 62: 658-63.

[143] Wanke T, Abrahamian H, Lahrmann H, Formanek D, Merkle M, Auinger M, Zwick H, Irsigler K. No effect of naloxone on ventilatory response to progressive hypercapnia in IDDM patients. Diabetes 1993; 42: 282-7.

[144] Tantucci C, Scionti L, Bottini P, Dottorini ML, Puxeddu E. Casucci G, Sorbini CA. Influence of autonomic neuropathy of different severities on the hypercapnic drive to breathing in diabetic patients. Chest 1997; 112: 145-53.

[145] Chaudhuri TK, Fink S. Prognostic implications of gastroparesis in patients with diabetes mellitus. Clin Auton Res 1992; 2: 221-4.

[146] Kong M-F, Horowitz M, Jones KL, Wishart JM, Harding PE. Natural history of diabetic gastroparesis. Diabetes Care 1999; 22: 503-7.