Cardiovascular autonomic neuropathy in insulin-dependent diabetes mellitus: prevalence and estimated risk of coronary heart disease in the general population


Ole May, Section of Cardiology, Department of Medicine, Herning Central Hospital, 7400 Herning, Denmark (fax: 99272053, e-mail:


Abstract. May O, Arildsen H, Damsgaard EM, Mickley H (Odense University Hospital, Odense, and Horsens Hospital, Horsens, Denmark). Cardiovascular autonomic neuropathy in insulin-dependent diabetes mellitus: prevalence and estimated risk of coronary heart disease in the general population. J Intern Med 2000; 248: 483–491.

Objectives. The aim of the study was to estimate the prevalence of cardiovascular autonomic neuropathy (CAN) in Type 1 diabetes mellitus in the general population and to assess the relationship between CAN and risk of future coronary heart disease (CHD).

Methods. The Type 1 diabetes mellitus population in the municipality of Horsens, Denmark, was delineated by the prescription method and a random sample of 120 diabetics aged 40–75 years was recruited. Type 1 diabetes mellitus was registered if fasting C-peptide was below 0.30 nmol L−1. The E/I ratio was calculated as the mean of the longest R–R interval in expiration divided by the shortest in inspiration during deep breathing at 6 breaths min−1 and taken to express the degree of CAN. A maximal symptom-limited exercise test was carried out and the VA Prognostic Score, indicating risk of cardiovascular death or non-fatal myocardial infarction, was computed. Additionally, the 10-year risk of CHD was calculated using the Framingham model.

Results. A total of 84 people responded, of whom 71 had Type 1 diabetes mellitus. The E/I ratio was measured in 69 people. The prevalence of CAN expressed as an E/I ratio below the normal 5th percentile was 38%. The E/I ratio was significantly reduced in old age, long duration of diabetes, female gender, high fasting blood glucose, triglyceride, systolic blood pressure and urinary albumin excretion. A high risk of future CHD calculated using the Framingham model was associated with a low E/I ratio (r = − 0.39, P = 0.001). Exercise capacity, rise in systolic blood pressure and heart rate were positively correlated with the E/I ratio. A high VA Prognostic Score was correlated with a low E/I ratio (r = − 0.58, P < 0.0005). The risks estimated by the two models were significantly correlated (r = 0.60, P < 0.0005).

Conclusion. The prevalence of CAN in the 40–75-year-old Type 1 diabetes mellitus population is estimated to be 38%. CAN is associated with exercise test parameters and a coronary risk factor profile indicating a high risk of future CHD.


In the last 20 years, cardiovascular autonomic neuropathy (CAN) has been known to imply a poor prognosis in diabetes [1]. The increased mortality in CAN has been attributed especially to sudden cardiac death [1, 2] and nephropathy [1–3]. The pathophysiological process behind the increased mortality in diabetic patients with CAN is, however, still largely unknown. Only a few studies addressing the prognostic effect of CAN in diabetes are available [1–5]. These studies have all been carried out in groups selected from diabetes clinics, and a possible confounding effect of other risk factors has not been fully addressed. The poor prognosis in diabetic patients with CAN could thus be caused by an association with other risk factors. Because coronary heart disease (CHD) is the major cause of death in diabetes [6], a possible association between CAN and predictors of CHD seems especially interesting.

The reported prevalence of CAN has ranged from 17 to 70% [7], the wide range illustrating large differences in the selection of study groups and the assessment of CAN. Only one population-based prevalence study exists. Neil examined 43 patients with insulin-dependent diabetes mellitus (Type 1 diabetes mellitus) and 202 patients with non-insulin-dependent diabetes mellitus (Type 2 diabetes mellitus) randomly selected in a geographically defined area in Oxford, and reported a 17% prevalence of CAN [8].

The aim of this study was to estimate the prevalence of CAN in the adult Type 1 diabetes mellitus population and to assess the relationship of CAN to risk of future coronary heart disease as expressed by exercise test parameters and conventional coronary risk factors.

Patients and methods


To be able to draw a random sample of the diabetic population in the municipality of Horsens, Denmark, the insulin-treated part of the diabetic population was delineated by the prescription method. From 1 November 1992 to 1 July 1993, 328 users of insulin were registered. The municipality had, at that point in time, 55 265 inhabitants, giving a prevalence of insulin-treated diabetes mellitus of 0.6%.

A randomly drawn age- and gender-stratified sample of 120 users of insulin in the age group of 40–75 years was asked to participate in the study. The diagnosis of diabetes mellitus according to the criteria of WHO [9] was verified in each participant. If fasting C-peptide was below 0.30 nmol L−1, persons were registered as having Type 1 diabetes mellitus; the remainder had insulin-treated Type 2 diabetes mellitus.


The autonomic nervous system function was assessed by the deep breathing test. This test ‘approaches the optimal test for the cardiovagal function’ according to a recent expert committee on non-invasive tests of the autonomic function [10]. The examination was carried out in the morning after a light meal. Each participant was instructed to breath deeply whilst sitting up [11], each inspiration and expiration lasting 5 s. During the procedure, a six-lead electrocardiogram (ECG) was obtained at 50 mm s−1, and the beginning of each inspiration and expiration marked on the ECG with a pencil. The ECG was blinded by a random number, before it was interpreted. In each respiratory cycle the shortest R–R interval during inspiration and the longest during expiration were selected using a pair of compasses, and the R–R intervals were measured by a ruler to the nearest 0.5 mm. The ECGs were examined by two different people, and if a reading differed by more than 0.5 mm, the ECG was measured once more to reach agreement. The ratio between the longest and the shortest R–R interval in each respiratory cycle, the E/I ratio, was calculated and averaged over three consecutive respiratory cycles. The variability of the R–R interval was expressed as the E/I ratio because this reading, as opposed to the difference between maximal and minimal heart rates, is independent of the resting heart rate in normals [12, 13]. The E/I ratio was defined as normal if the reading was higher than the age-adjusted 5th percentile in healthy non-diabetics. As previously reported, values of E/I ratio > 1 + exp(−1.12 − 0.0198 × age [years]) can be regarded as normal [13].

A 12-lead ECG taken at rest was blindly interpreted according to the Minnesota code [14] by two experienced technicians. Evidence of left ventricular hypertrophy was taken as suggested by the World Health Organization [14].

The participants were asked about myocardial infarction (MI), smoking habits, angina pectoris (Rose questionnaire [14]) and lower leg oedema or pulmonary congestion. From pre-coded forms, the questions were read aloud to each participant. A history of MI was only accepted for the analysis if verified by hospital records.

To avoid observer-induced bias, all blood pressure (BP) values were recorded with a Hawksley random zero manometer [15]. The BP was measured twice or more until the difference between two consecutive readings was ≤ 4 mmHg, the mean of the two was used in the calculations.

A fasting blood sample was taken at 08.00 h in the morning. Blood glucose was measured enzymatically (Cobas Mira, Roche), C-peptide was analysed radioimmunologically (antibody K6, Novo Nordisk), cholesterol was measured enzymatically (CHOD-PAP) and haemoglobin A1c was quantified photometrically (Waters HPLC). A morning clean-catch midstream urine specimen was tested for nitrite and leucocyte esterase activity with a dipstick (Boehringer–Mannheim, LN). If any of the two reagent squares responded, the urine was cultured and, in case of significant bacteriuria, antibiotics were given. As soon as the urine was dipstick-negative, three overnight urines were collected. The urine albumin excretion rate (UAE) was evaluated by turbidimetry (Cobas Mira).

The 10-year risk of future CHD was estimated by an algorithm developed in the Framingham Heart Study [16] based on gender, age, systolic BP, total/HDL cholesterol, left ventricular hypertrophy, diabetes and smoking status. The algorithm using systolic BP is recommended because systolic BP is more accurately determined, has a wider range and stronger predictive power compared with diastolic BP [16].

A maximal symptom-limited bicycle ergometer exercise test was accomplished. The workload was initially 25 watts (W), but increased by 25 W every 2 min. Prior to and each minute during the test, a standard 12-lead ECG was recorded, and the participants questioned about possible chest pain. If present, the pain was quantified according to the Borg scale [17]. Blood pressure was measured manually before, every 2 min during and 1, 2, 4 and 6 min after exercise. Horizontal or descending ST-segment depression of at least 0.1 mV measured 80 ms after the J-point in three consecutive cycles was considered a significant sign of ischaemia. If ST depression was present on the resting ECG, an additional 0.1 mV was required. The ECGs were blinded by a random number before interpretation.

From the results of the exercise test, the Veterans Affairs (VA) Prognostic Score was calculated [18]. This score was developed by Froelicher, who examined a consecutive series of routine clinical exercise tests in 3134 men. In the study by Froelicher, cardiovascular death or non-fatal MI was independently predicted by the exercise-induced change in systolic BP, ST depression, exercise capacity and the presence of congestive heart failure/use of digoxin. A score including these four parameters was constructed:

5 × CHF/digoxin [yes = 1]+ (ST depression [mm]) + systolic BP change score [0  > 40 mmHg; 1: 31 − 40 mmHg; 2: 21 − 30 mmHg; 3: 11 − 20 mmHg; 4: 0 − 10 mmHg; 5: < 0 mmHg]– METs

A low score (less than − 2) was consistent with an annual cardiovascular mortality below 2%.

Angina pectoris was present if the Rose questionnaire [14] was positive or chest pain was registered in the presence of ST depression during the exercise test.


The study complies with the Declaration of Helsinki II, and the study protocol was approved by the local ethics committee. Written informed consent was obtained from each participant after verbal and written information was given. The results of each individual examination were forwarded to the participant as well as the general practitioner, and in cases with abnormal results, the participant was offered a consultation at the hospital.


The associations between the E/I ratio and each coronary risk factor were expressed by Spearman's coefficient of correlation. Differences between two groups were tested using the Mann–Whitney test in cases of continuous variables and Fisher's exact test in cases of proportions. Divergence between more than two groups was examined by the Kruskal–Wallis test. Tests results with a two-sided P-value below 0.05 were taken as statistically significant.


Basic characteristics

From the randomly selected sample of 120 users of insulin, 84 diabetics (70%) responded (six dead, two dementia, six insulin for relatives and 22 refused to participate). No difference was found between responders and non-responders with regard to gender (P = 0.20), but the non-responders were on average older than the responders (P = 0.002). Thirteen diabetic patients had a C-peptide above 0.30 nmol L−1 and were excluded from further analyses. Thus, 71 subjects with Type 1 diabetes mellitus (29 F, 42 M) were included, with an age range from 40 to 74 years (Tables 1 and 2). Twenty-four patients were treated with some kind of cardiac medication (10 were on ACE inhibitors, one on beta-blockers, 13 on calcium blockers, 11 on diuretics and six on aspirin). Nobody used digitalis, nitrates or antiarrhythmics. The E/I ratio was obtained in 69 subjects (two were not sufficiently cooperative); the mean measured difference between the longest and the shortest R–R interval was 7.4 mm.

Table 1.  Correlation between E/I ratio and continuous risk factors
 Median (10th, 90th percentile)Spearman rP
  1. NS, non-significant, HbA1c, haemoglobin A1c; UAE, urinary albumin excretion rate; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; XT, exercise test; HR, heart rate.

E/I ratio1.17 (1.04, 1.51)
Age (years)51.5 (42, 69)−0.400.001
Length of diabetes (years)19.6 (4.6, 35)−0.250.042
Fasting blood glucose (mmol L−1)10.4 (4.7, 18)−0.300.012
HbA1c (%)7.9 (5.9, 9.1)−0.05NS
Body mass index (kg m−2)24.8 (21, 29)−0.05NS
Waist/hip ratio0.95 (0.85, 1.04)0.15NS
UAE (µg min−1)3.7 (2.5, 117)−0.270.024
Resting SBP (mmHg)143 (118, 179)−0.50< 0.0005
Resting DBP (mmHg)83 (69, 96)−0.23NS
Cholesterol (mmol L−1)5.8 (4.5, 7.7)−0.20NS
HDL cholesterol (mmol L−1)1.7 (1.1, 2.3)−0.22NS
LDL-cholesterol (mmol L−1)3.5 (2.4, 5.1)−0.08NS
Triglyceride (mmol L−1)1.15 (0.61, 1.8)−0.260.031
Exercise capacity (watt)121 (56, 197)0.62< 0.0005
Maximum SBP during XT (mmHg)208 (170, 246)0.270.026
Increase in SBP (mmHg)78 (44, 111)0.48< 0.0005
Maximum HR during XT (beats min−1)158 (122, 183)0.62< 0.0005
Increase in HR (beats min−1)73 (30, 99)0.60< 0.0005
Increase in HR × increase in SBP (mmHg min−1)5164 (1546, 10 466)0.63< 0.0005
Table 2.  Relationship between E/I ratio and categorical risk factors
Median E/I ratio
n with factor (%)Factor absentPFactor present
  1. NS, non-significant.

Male gender42 (61)1.100.0251.22
Currently smoking35 (51)1.13NS1.23
Angina pectoris5 (7)1.14NS1.21
Previous myocardial infarction4 (6)1.15NS1.25
Left ventricular hypertrophy5 (7)1.14NS1.27
Exercise-induced ST depression5 (7)1.15NS1.45

Prevalence of autonomic neuropathy

The measured value of the E/I ratio was below the age-adjusted 5th percentile obtained in healthy non-diabetics in 26 out of 69 patients. The prevalence of CAN in the Type 1 diabetes mellitus population was thus 38% (95% CI = 26–50%).

E/I ratio and coronary risk factors

The E/I ratio was reduced in old age, long duration of diabetes, high fasting blood glucose, UAE, systolic BP (Fig. 1), triglyceride and female gender (Tables 1 and 2). In normoalbuminuria (UAE < 20 µg min−1) the median E/I ratio was 1.23 (range 1.03–1.67), in microalbuminuria (20 µg min−1≤ UAE < 200 µg min−1) it was 1.11 (range 1.02–1.42), and in albuminuria (200 µg min−1≤ UAE) the median E/I ratio was 1.06 (range 1.03–1.12) (P = 0.01) (Fig. 1). No statistically significant relationship was detected between E/I ratio and smoking, total cholesterol, LDL cholesterol, HDL cholesterol, diastolic BP, body mass index, haemoglobin A1c, waist–hip ratio, angina pectoris, previous MI, ECG signs of left ventricular hypertrophy or exercised-induced ST depression (Tables 1 and 2).

Figure 1.

Deep breathing-induced E/I ratio against the 10-year risk of future coronary heart disease (Framingham model), risk score of future cardiovascular death or non-fatal myocardial infarction based on exercise test parameters (VA Prognostic Score), age, duration of diabetes, systolic blood pressure, fasting blood glucose, exercise capacity (linear regression line indicated) and urinary albumin excretion rate (UAE) (normoalbuminuria, UAE < 20 µg min−1; microalbuminuria, UAE = 20–200 µg min−1; overt albuminuria, UAE > 200 µg min−1) (median values indicated).

Based on gender, systolic BP, age, total/HDL cholesterol, ECG sign of left ventricular hypertrophy, diabetes and smoking status, the 10-year risk of CHD according to the Framingham Heart Study model was calculated (median = 12.1%, 10th, 90th percentile = 5.8, 29%). The magnitude of the E/I ratio was significantly reduced with increasing risk of CHD (r = − 0.39, P = 0.001) (Fig. 1). If individuals with angina pectoris, previous MI or a history of congestive heart failure (n = 13) were withdrawn from the analyses, a significant correlation between the calculated 10-year risk of CHD and the E/I ratio was still present (r = − 0.40, P = 0.003).

E/I ratio and exercise test

The exercise test was completed by all but one, in whom the left leg was amputated. The exercise capacity was significantly related to E/I ratio (r = 0.62, P < 0.0005) (Fig. 1). Patients with a low E/I ratio had a low increase as well as a low maximum heart rate and systolic BP during the test (P < 0.0005) (Table 1). No significant relationship was found between the E/I ratio and exercise-induced chest pain or ST depression (Table 2). A significant relationship between VA Prognostic Score and the E/I ratio was found (r = − 0.58, P < 0.0005) (Fig. 1). When females were withdrawn from the analysis, a significant inverse correlation between the score and E/I ratio in the males was still present (r = − 0.61, P < 0.0005, n = 41).

Comparison of risk estimated by the Framingham model and the VA prognostic score

The VA Prognostic Score was significantly correlated with the estimated 10-year risk of CHD according to the Framingham model (r = 0.60, P < 0.0005). The patients were further classified as either low- or high-risk according to the VA Prognostic Score (< − 2, ≥ − 2) as well as the Framingham model (above/below the median = 12.1%). Ten individuals had a high VA Prognostic Score, nine of whom were in the high half according to the Framingham model [9/34 (26%)], and one in the low half [1/34 (3%)] (P = 0.01).


We found that the prevalence of CAN was 38% (95% CI = 26–50) in a population-based group of Type 1 diabetes mellitus patients aged 40–75 years. Autonomic neuropathy in the Type 1 diabetes mellitus population has only been investigated in one survey before. Neil [8] examined 202 Type 2 and 43 Type 1 diabetes mellitus patients randomly drawn in a defined area of Oxford. The overall prevalence of CAN was 16.7%. In the Type 1 group, the prevalence was 14% (95% CI = 5–28) defined as an E/I ratio below the normal 2.5th percentile. If we had defined CAN as the lower 2.5th percentile in healthy persons, the prevalence proportion in our study would have been 31% (95% CI = 21–44), still significantly larger than the figure found in the smaller sample in Neil's study. This difference is not easily explained. An age-corrected normal limit was applied in both investigations, but this age correction may not have been perfect. The age difference between the groups included (mean age: Neil, 41 years; our study, 53 years) could therefore have played a role. In Neil's study, the duration of diabetes was not provided, but since the diabetics in our study were older, they probably also had a longer duration of diabetes. This may therefore be an explanatory factor behind the larger prevalence of CAN in the present study. The fact that the responders in our investigation were younger than the non-responders indicates that the non-respondent group had probably also been diabetic for a longer period of time, and thus could have a higher prevalence of CAN. This implies that the prevalence found in our study could be an underestimation.

The estimated risk of CHD based on the exercise test parameters was significantly increased in individuals with a reduced E/I ratio. A lower exercise capacity in diabetics with CAN is well known. In healthy persons a decrease in the parasympathetic activity takes place during the early stages of exercise, followed by an increase in sympathetic nervous system activity, and at maximal levels of exercise there is a release of adrenaline. This leads to increased cardiac output, BP and heart rate, thus increasing the supply of oxygen and nutrients needed by the muscles. In diabetics with CAN, a disturbed vagal/sympathetic balance induces a higher resting heart rate, the beta-receptors are downregulated due to a higher habitual noradrenaline level, and the capacity to increase noradrenaline during exercise is reduced [19]. Correspondingly, we found a smaller increase in, and a lower maximum level of, heart rate and BP, as well as their product in patients with autonomic neuropathy (Table 1). Thus, there is good biological evidence to support the notion that the reduced exercise capacity was caused directly by the presence of autonomic neuropathy. On the other hand, reduced exercise capacity is a well-known prognostic predictor in CHD [18]. Hence, the low exercise capacity in patients with CAN may well, in some of the patients, be due to coronary artery or small vessel disease possibly affected by a common causative mechanism behind both CAN and CHD. No significant correlation was found between ST depression and E/I ratio. This may, of course, be explained by the small number of patients who developed ST depression. But, additionally, ST depression is known to be a much weaker predictor of CHD than exercise capacity. In a review of papers addressing the prognostic value of exercise test parameters, only three out of nine reported a significant predictive value of ST depression [18]. The VA Prognostic Score was developed in male patients, and whether it works equally well in females may be disputable. The relationship between CAN and the score was, however, also present when the test was carried out in the male group only.

The risk estimation based on gender, age, systolic BP, total/HDL cholesterol, left ventricular hypertrophy, diabetes and smoking status using the Framingham algorithm was significantly correlated to the degree of CAN (P = 0.001). Because this model was developed mainly in non-diabetics, there may be some uncertainty involved when it is applied to purely Type 1 patients. Nevertheless, data from the diabetic part of the Framingham cohort showed that the relationship between risk factors and subsequent CHD was similar in diabetics and non-diabetics [20].

A significant association was present between the degree of CAN and UAE (P = 0.02), confirming the findings from previous studies in Type 1 diabetic patients selected from diabetes clinics [21–23]. The causal mechanisms behind this relationship or the pathogenesis of diabetic neuropathy per se are not well known. Long-standing hyperglycaemia is thought to be the primary factor causing neuropathy [24], but the pathophysiological process leading to neuropathy is not fully clarified. One theory claims that disturbances in the glucose metabolism in the presence of aldose reductase lead to nerve accumulation of sorbitol and nerve damage. However, the findings in human sural nerve biopsies have not lent much support to this hypothesis [25, 26], and clinical trials of aldose reductase inhibitors have been disappointing [27]. More promising seems the vascular hypothesis stating that the nerve damage is caused by hypoxia secondary to endoneurial microangiopathy. Endoneurial oxygen tension was found to be lower in patients with diabetic neuropathy compared with patients without [28], and endoneurial capillaries in patients with neuropathy are reduced in density and have profound basement membrane thickening [29]. If the vascular hypothesis is true, diabetic neuropathy is mainly a marker of microvessel disease, and the strong relation to UAE is well explained. The presence of increased UAE is an established indicator of future CHD in Type 1 diabetes mellitus [30] as well as in non-diabetics [31].

Also significantly increased with a low E/I ratio, but not included in any of the two models, were the duration of diabetes, fasting blood glucose and triglyceride. Each of these variables is a known predictor of CHD [32–34] and the inverse correlations between these variables and the E/I ratio thus give further support to the hypothesis that CAN is associated with an elevated risk of future CHD. Contrary to blood glucose, haemoglobin A1c was not related to E/I ratio. This is likely to be caused by the small number of participants.

In the DCCT investigation, intensive insulin treatment reduced the development of autonomic and peripheral neuropathy as well as established microvascular complications [35]. It is therefore important to identify diabetics with CAN, as they represent a high-risk group, and intensified insulin treatment may delay progression of autonomic neuropathy. Whether reducing autonomic neuropathy per se also improves the prognosis regarding CHD is an unsettled question.

In summary, this study showed that the prevalence of CAN in the 40–75-year-old Type 1 diabetes mellitus population was 38% (95% CI = 26–50%). We found that a reduced autonomic function was associated with a high risk of future CHD whether predicted from conventional coronary risk factors or from exercise test parameters. The prognostic value of a reduced autonomic function in unselected patients with Type 1 diabetes mellitus, when accounting for the confounding effect of coronary risk factors, remains to be seen in future follow-up studies.


The study was supported by The 1991 Pharmacy Foundation, The Danish Heart Foundation, The Danish Medical Association Research Fund, The Vejle County Medical Research Foundation and the Clinical Institute, Odense University, Denmark.

Received 6 July 2000; accepted 10 August 2000.