- Email: [email protected]
Measurement of heart rate variations: influencing factors, normal values and diagnostic impact on diabetic autonomic neuropathy
Diabetes Research and Clinical Practice 29 (1995) 179-187
ELSEVIER
Measurement of heart rate variations: influencing factors, normal values and diagnostic impact on diabetic autonomic neuropathy H.-J. Braune*, U. Geisendiirfer Department
of Neurology, Philipps University Hospital, Rudolf-Bultmann-Stra$e
8, D-35033 Marburg, Germany
Received 16 November 1994;revision received 7 August 1995;accepted 15 August 1995
Abstract Different methods of heart rate variation (HRV) measurementare recommendedto assessdisturbances of cardiovascular reflexes in autonomic polyneuropathies. Influencing factors on the E/I-ratio, HRV variations during Valsalva’s maneuver, and change of position from lying down to standing up, as well as age, gender, time of day and satiety were investigated in healthy volunteers. Normal values were then calculated and a total of 100 diabetics with clinically manifest neuropathy of different stages (NO = 8, Nl = 15, N2a = 24, N2b = 28, N3 = 25), 26 of whom were insulindependent and 74 non-insulin-dependent, were examined. Apart from clinical examination, nerve conduction velocity (NCV) measurementof five nerves as well as amplitude measurementof evoked sensory and motor action potentials was performed. No significant influence of the time of day the test was performed could be found. Test reproducibility was good, day-to-day differences did not reach a significant level. This was also true for the test results before and after food intake. Gender-correlated test result differences could not be proven. Statistical analysis showed significant differences between the overall group of diabetics and controls. The difference was most accentuated in the E/I ratio. The further developed the clinical picture of neuropathy, the more pathological were the results obtained in different tests. The results suggestthat most changes leading to pathological values of NCV and HRV occur in a clinical stage in which no, or only very slight, clinical signs give evidence of diabetic neuropathy (NO-Nl). Therefore, theseexaminations should, especially, be performed in diabetics with no, or only slight, clinical signs of neuropathy in order to reveal those patients with neuropathic disturbances at an initial stage. Keywords:
Diabetic autonomic neuropathy; Heart rate variability; Bed-side test
1. Introduction Manifest autonomic neuropathy triples the lethality risk [ 1,2]. Since Rundles published the l Corresponding author, Tel.: +49 06421 286278; Fax.: +49 06421 288955.
first comprehensive report on autonomic disturbances in diabetics in 1945 [3] clinical symptoms have been frequently described, and the introduction of computer-aided analytical methods has promoted the investigation of these autonomic disturbances over the last decade. A variety of different testing procedures has been described to
0168-8227/95/SO9.500 1995 Elsevier Science Ireland Ltd. All rights re%rved SSDI 0168-8227(95)01133-X
180
H.-J. Braune. U. Geisendiirfer /Diabetes Research and Clinical Practice 29 (1995) 179-187
date, however, influencing factors, test optimization, test sensitivity and reproducibility have rarely been taken into consideration [4]. Most of the test procedures measurethe impairment of heart rate variability (HRV) in autonomic neuropathy. Different analytical strategies are implied: standard deviation (SD), difference between maximal and minimal heart rate in a given time period (min-max), ratio of minimal heart rate at expiration and maximal heart rate at inspiration (E/I-ratio), mean difference of consecutive RRintervals (MSD) and vector or spectral analysis (see overviews in Refs. [5,6]). Accentuation of HRV is provoked by different reflex tests, e.g., the Valsalva maneuver, position change from lying down to standing up [5] or the reverse [7], the hand grip test and the cough test [8]. Pharmacological testing with sympatho- or parasympathomimetic drugs and neuroendocrinological measurements give further information about the adaptability of the autonomic nervous system. The apparent clinical ‘patchiness’ may result from differences in organ susceptibility to ‘denervation’. More obviously, however, they may result from an uneven test sensitivity applied to different parts of the autonomic system [4]. Therefore, a single test has always been considered to be insufficient to either detect or deny the presenceof an autonomic involvement. We examined influencing factors on HRV variation such as age, gender, time of day and satiety, since little information is given about these in the literature [9, lo]. We then calculated normal values and examined a population of 100 patients with clinically manifest diabetic neuropathy. Apart from clinical examination, NCV measurement of live nerves as well as amplitude measurement of evoked sensory and motor action potentials was performed. 2. Patients and methods 2.1. Patients
The study was carried out in accordance with the Declaration of Helsinki as modified by the 35th World Medical Assembly, Venice, Italy, 1983. All persons gave their informed consent
prior to being included into the study. Measurement of HRV on deep breathing, after standing up and during Valsalva’s maneuver were performed on 100 consecutive outpatients with diabetes mellitus (WHO definition, Ref. [ 111)at the Dept. of Neurology of the University Hospital Marburg. The study was carried out over a period of 18 month (54 men, 46 women, age 18-82 years, mean 52.0 years). Patients with symptoms caused by other diseasesor casehistories other than diabetes, which could be the cause of abnormal test results for conventional neurophysiological testing such as, e.g., previous injury, lumbar discogenic disease, carpal-canal syndrome, inherited or alcoholic neuropathy, or other neurological diseases,as well as patients taking drugs which act on the cardiovascular systemwere excluded. A further exclusion criterion was the presence of acute focal neuropathy and acute painful neuropathic syndromes, which are probably caused by diabetes, but are rare and presumably of varied pathogenesis. Several studies show that these syndromes have a different clinical course compared to distally symmetrical and primarily sensory polyneuropathy [ 12,131.In accordancewith the clinical stage of neuropathy, which was independently assessed by two experienced neurologists, the patients could be divided into five subgroups, as recommended by Dyck et al. [13]: eight patients without any clinical signs of neuropathy (NO), 15 patients with only one clinical symptom of neuropathy (Nl), 24 patients with mild distally symmetrical sensory polyneuropathy (N2a), 28 patients with distally symmetrical sensorimotor polyneuropathy (N2b), and 25 patients with severe symmetrical sensorimotor polyneuropathy (N3). Autonomic symptoms included vesical incontinence (12 patients), constipation (nine patients), sexual dysfunction (six of 54 males) and orthostatic dysregulation (23 patients). Autonomic symptoms were regarded as mild (only one symptom) in 14 N2b-patients and as moderate (two symptoms) in 18 N3-patients. According to hospital tiles and information given by the patients, the duration of diabeteswas estimated: 20 patients with a duration of diabetes of less than 5 years, 31 patients with a duration of diabetes of more than 5 but less than 10 years, and 49 patients with a duration of more
H.-J. Braune, lJ. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
than 10 years, A total of 26 patients younger than 40 years suffered from IDDM, 74 from NIDDM. In 72 patients, metabolic control was good with normal fasting blood-sugar values and normal or only slightly HbAi, values elevated. It was moderate in 28 patients. Tests were also done on age- and sex-matched healthy volunteers of different ages (51 men, 49 women, age 18-83 years, mean 50.3 years) to define the normal range of values (~95th percentile). The influence of time of day of measuring was tested on six women (age 20-68 years) and four men (35-63 years). To examine the retestvariability, eight women (age 23-63 years) and two men (23 and 63 years old) underwent the HRV-testing procedure several times. The influence of satiety on results of HRV-testing was checked in seven women (24-72 years) and six men (27-69 years). 2.2. Methods The examinations were carried out on the basis of routine outpatient visits. On the day of the examination the patients had their usual breakfast and, if necessary, their morning insulin dose. Apart from clinical examination, all patients were submitted to conventional electroneurographical tests. Measurement of motor nerve conduction velocity (MNCV) was performed on nn. medianus, tibialis and peronaeus. Measurement of sensory nerve conduction velocity (SNCV) were done on nn. medianus and suralis at a standardized limb temperature of 35°C. Furthermore, we assessed axonal loss by measurement of amplitudes of evoked sensory and motor action potentials. The sums of the three amplitudes of motor nerves and the two sensory nerves were calculated for statistical evaluation and data reduction. Assessmentof HRV included measurement of the E/I-ratio after 5 min of deep rhythmical breathing. It was measured computer-aided by an nDX-test-unit (Fa Q-Med, 100 Metro Park South, Laurence Harbour, NJ 08878, USA). After an appropriate tryout time the persons tested were requested to breath regularly and deep with five exand inspirations per minute. Breathing was synchronized using an acoustic signal and the inter-
181
vention of surveying personnel if necessary. E/I-ratio is calculated as the mean of five ratios derived from the longest RR-interval during inspiration and shortest RR-interval during exspiration [5,6]. Furthermore, the mean circular resultant (MCR) represents the length of the main vector derived on a unit circle with the periodicity of one breathing cycle of ex- and inspiration. The heartbeats as events on time are given as dots on the unit circle. The main vector correlates with these dots [14]. The third parameter is the standard deviation (S.D.) of RR-intervals in the 5-min measuring time. To perform the Valsalva manouever the person tested is asked to exspirate forcefully against a resistance for 20 s [ 151.Three tests were carried out with a retest-interval of 2 min. The Valsalva-ratio is the ratio: longest RRinterval after exspiration/shortest RR-interval during the forced exspiration period. The mean value of the three tests performed was calculated. The 30/15-ratio is the ratio 30th/15th RR-interval after standing up from lying down position. The person was asked to relax in prone position for 5 min and then, unassisted, to stand up in a period of 5 s, and an ECG was registered for a further 60 s. Measurement of autonomic function was performed before noon in a quiet room at a temperature of 22°C and with dimmed lights after an ‘acclimatization period’ of 20 min. None of the subjectswas allowed to drink coffee or tea, smoke, or take medicaments in the hours before and during the visit. Considerable physical effort or hypo- or hyperglycaemic periods up to 3 days before the examination led to exclusion from the study [ 161.To assessthe retest variability, the testing procedure was performed on three consecutive days, between 4:00 and 5:00 p.m., on 10 healthy volunteers (eight women and two men). To examine circadian influences on the results of HRV-testing, the procedures were performed on 10 healthy volunteers on day one between 7:30 and 8:00 a.m. and on the following day between 4:00 and 5:00 p.m [9]. To test the influence of food intake, 13 healthy volunteers agreed to be tested between 7:30 and 8:OOa.m. after a period of more than 10 h without food, and again 20 min after consumption of
H.-J. Braune. U. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
182
a
b 1.8
140
17
.z y 1
16 1.5
3
.
14
.*
13
..
. .
11
. . .I . .
.
. .
.
.
:.
.
.,
:.
10
*
.
-* . .;--: . . . . . .. , _ *ii
30
4c
50
is I
---! . . . . . .
. 1’ -. ,,,,,,,,,
.
I,,,,
20
5 I
8. .
12
2 =
.
-.
60
.
70
80
.
L I
I.
30
20
.
.
40
50
60
70
80
d
C
.
2.5
160 140 E
80 70 60 50 40 30 20 10 0
.
120 100
'
80
is I
60 40 20 0
-.
. ..
.... * -1 .. . .: * . *:. *. . . 1 . . . 40 20 30
.
.o s 1 2 $ 2
.
.. * .... * .* a. -._ . *.-: y-..: 1 *. 50
63
70
8C
1.8 1.7 1.6 1.5 1.4 1.3 12 11 10
:
20
-.. *.
30
.
40
50
60
70
80
e
Fig. I. Results of heart rate variation (HRV) testing on normal control persons of different age (n = 100, for explanation of testing procedure see text): (a-c) 5 min deep rhythmical breathing: (a) E/I-ratio; (b) mean circular result (MCR); (c) standard deviation (SD.); (d) Valsalva maneuver; (e) position change from lying down to standing (30/I 5-ratio). The bold line represents the 95th percentile (95% of all normal values are greater). The dotted line represents the regression line, the influence of age is highly significant.
100 g water-free glucose (oral glucose tolerance test). The examinations of patients with diabetic neuropathy were carried out during a routine outpatient visit. On the day of the examination the patient had his usual breakfast and, if necessary, his morning insulin dose. Statistical exploration was performed with the parameter-free Mann-Whitney U-test, and graphi-
cal documentation was with Box-and-Whisker plots. The obtained values of controls were analysed to determine the 95th percentile. A minimal criterion for assumedabnormality in patients was a test result value > 95th percentile. Furthermore, Spearmanscorrelation coefficient and the partial correlation coefficient excluding the influence of age were calculated. Analysis of gender influence was calculated by Students r-test. Wilcoxon
H.-J.
Braune,
U. Geisendkfer
/ Diabetes
matched-pairs signed rank test was performed to assessinfluence of time of day and food intake. The level of significance of the day-to-day test result variations was defined by Friedmann’s twoway ANOVA.
Research
and
a
1.8
o .‘,
1.7 1.6
_t
1.5
>
1.4 I
3. Results
In accordancewith other studies, all parameters showed a decreasewith increasing age in the control group. Nevertheless,we defined single normal range for all agessince the control group was too small for defining age-dependent normal values, and since the resulting error was deemed to be negligible with respect to statistical analysis. Fig. la-e gives the results of HRV testing on controls, the influence of age is highly significant. Table 1 shows the normal ranges of the HRV tests, NCV measurements, and the amplitudes used in this study; the cut-off values were defined as 95thpercentiles. No significant influence of the time of day the test was performed could be found. Furthermore, Table 1 Synopsis of normal ranges established by examination of 100 normal subjects in our laboratory; test battery with electroneurography and various heart rate variation tests Test Nerve conduction velocity N. tibialis: mot. N. peronaeus: mot. N. suralis: sens. N. medianus: mot. N. medianus: sens. Sum of 3 evoked compound motor action potentials (nn. medianus, peronaeus, tibialis) Sum of 2 evoked compound sensory action potentials (nn. medianus, suralis) Heart rate variation E/I-Ratio S.D. MCR Valsalva-ratio 30/l 5-ratio
Normal range (>95th percentile) >40 >40 >40 >50 >50
2 =
183
Clinical Praclice 29 (1995) 179-187
0 0
1.3 1.2 1.1 1 .o
b 5 2 m 2 ’ 1 2
I
=
Diabetics;
Normals
Diabetics;
Normals
Diabetics;
Normals
2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1 .I
1.8
.-0 +
?
1.6
m/s m/s m/s n-h m/s
>18 mV
>lO pv
>I.06 >24 >I2 > 1.08
>I.04
Fig. 2. Group differences between diabetics and age- and sexmatchedcontrols with regard to heart rate variation (HRV): (a) 5 min of deep rhythmical breathing (E/I-ratio, and results for SD. and MCR are equivalent and therefore omitted here); (b) Valsalva-maneuver; (c) position change from lying down to standing (30115ratio). The bold line representsthe 95th percentile. Boxes, the range of values between the 25th and 75th percentiles; small T-lines (whiskers), the range between 10th and 90th percentiles. Open circles indicate extreme values. **Significant difference (P < 0.01); *slightly significant difference (P < 0.05); -, no significant difference (P > 0.05).
184
H.-J. Braune, IJ. Geisendiirfer / Diabetes Research and Clinical Practice 29 (I 995) I79- 187
the retest variability was good, day-to-day differencesdid not reach a significant level. This is also true for the testing before and after food intake. Gender-related test result differences could not be proven. Statistical analysis showed significant differencesbetween the group of diabetics and the control group. The difference was most accentuated in NCV, and clear in amplitude measurement of evoked sensoryand motor action potentials as well as in HRV tests (Fig. 2a-c). With respect to NCV of the sural nerve, eight of the Nl patients (n = 15), and with respect to the E/I-ratio (Fig. 3a) nine of the Nl patients, showed values below the normal range. Thus, NCV of the sural nerve and the E/I-ratio were the most sensitive parameters in detecting abnormalities in clinical stage Nl diabetics (only one clinical sign of neuropathy). In combination, the two methods showed pathological results in 72% of all diabetics (Nl-N3). For a quick survey, the combination of these two tests is advisable. Relying on this type of HRV test for examination of autonomic function alone, however, would have missed seven of the 59 subjects with abnormal test results in one or more of the other HRV testing procedures. In agreement with the recommendations of, e.g., Low [4], combination of these different HRV tests is thus advised. Even in summarizing and scoring the subtest results for sensorimotor neuropathy as well as for autonomic dysfunction, we found that a statistically significant influence of gender, age, metabolic control and duration or type of diabetes on the test results could not be found or was only present to a minor extent. This was rather surprising since age-relatedvariations and variations related to the duration of diabetes have been described in the literature [ 17,181.The failure to reproduce these already established data is probably due to the low number of patients tested in our study. An influence to some extent could be observed, however, data did not reach significant levels. Nevertheless, the clinical stage of diabetic neuropathy (NO-N3) exerted a significant influence on our study: the further developed the clinical picture of the disease, the more pathological were the different test results (Fig. 3a-c). It should
a t 0 1; 0
0
13 0
t-
b
LL t IT
o
C
Fig. 3. Influence of clinical stage of diabetic neuropathy on heart rate variation (HRV). (a) Five minutes of deep rhythmic breathing (E/I-ratio); (b) Valsalva maneuver; (c) position change from lying down to standing (30/15-ratio). Clinical staging as recommended by Dyck et al. (1992): NO, diabetics without any clinical sign of neuropathy (n = 8). Nl, diabetics with only one clinical sign of neuropathy (n = 15); N2a, diabetics with a mild distal symmetric sensory neuropathy (n = 24); N2b, diabetics with distal symmetric sensory motor neuropathy (n = 28); N3, diabetics with a severe symmetric sensory motor neuropathy (n = 25). Statistical analysis is performed with group-by-group comparison. The bold line represents the 95th percentile. Boxes, the range of values between the 25th and 75th percentiles; small T-lines (whiskers), the range between 10th and 90th percentiles; open circles, extreme values. **Significant difference (P < 0.01). *slight significant difference (P c 0.05); -, no significant difference (P > 0.05)
H.-J. Braune. I/. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
be stressedthat the greatest differences were found between diabetic neuropathy groups NO, Nl and N2a with no or only slight clinical signs. 4. Discussion
In recent years, several suggestionswith respect to a standardized test battery of autonomous nervous system function tests have been proposed, but an agreement has not yet been reached. The earliest and most common manifestations of autonomic neuropathy occur in the cardiovascular system in the form of orthostatic hypotension, resting tachycardia, painless myocardial ischemia and a ‘rigid pulse’. HRV is clearly involved in most forms of parasympathetic damage in autonomic neuropathy. A pathological result with the reflex test does not give any indication of the exact localization of the lesion: disturbances in the afferent, central, or efferent part of the reflex pathway, or even in the innervated organ itself, may be present [4]. Nevertheless, simple noninvasive tests permit accurate quantification and follow-up. Many reports on diabetic autonomic neuropathy measuredby simple HRV tests have been published [5,17,19-221 with a good correlation between the results [5] and a clear correlation to findings of somatic nerve function testing [23]. In accordance with the recommendations of the American Diabetes Association [24], three different testing procedures,namely deep rhythmic breathing, forced expiration against resistance, and position changefrom lying down to standing up, were combined in our study. The results of HRV testing at deep regular breathing were mathematically interpreted in three different ways (E/I-ratio, MCR, S.D.) to eliminate erroneous values due to variations of the intrinsic heart rate and extrasystoles WI. In order to obtain interpretable results, it is recommended to ascertain as many influencing factors as possible so that standardization of test conditions and procedures is feasible. The importance of the patient’s age is often stressedbut occasionally other factors are reported. The aim of this study was to investigate retest variability and influencing factors, such as age, gender, time of day the test was performed and satiety, on the results
185
of a common HRV test battery. Furthermore, the prevalence of pathological test results in a selected population with diabetes mellitus was examined. A strong correlation between age and test results existed in controls (Fig. la-e). Since the group was to enough to establish age-dependent normal ranges, these values were defined as 95th percentiles of the overall population. Further data sampling on control subjects will be helpful to make this testing battery even more reliable. Test reproducibility was good, day-to-day differences did not reach a significant level. In accordancewith other studies, no significant influence of the time of day the test was performed [26,27], the gender of tested persons [6,28,29], or the food intake before testing could be found [30-321. A statistically significant influence of gender, age, metabolic control, duration or type of diabetes on the test results could not be found, or was present only to a minor degree.This confirms already published findings [33,34], but is contrary to the results of Kuroda et al. [27] who found a relationship between age and autonomic neuropathy in diabetes mellitus. Failure to reproduce these already established data is probably due to the low number of patients tested in our study. An influence could be observed to some extent, however, data did not reach significant levels. After having established and validated this test battery, we are now in the process of setting up a more comprehensive study with recruitment of more patients. The most important parameter was the clinical stageof diabetic neuropathy (NO-N3): the further developed the clinical picture of the disease, the more pathological were the different test results. Of interest, however, is the fact that the greatest differences - regarding NCV and HFV - were found between diabetic neuropathy groups NO, Nl, and N2a (Fig. 3a-c). 5. Conclusion
The HRV test battery reviewed here is kind to patients, easy to carry out and is not time consuming. Influencing factors are controllable using a standardized test protocol. The results suggestthat
186
H.-J. Braune, U. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
most changes leading to pathological findings of NCV and HRV measurement occur in a clinical stage,in which no or only very slight clinical signs give evidence of diabetic neuropathy (NO-Nl, see also Refs. [35,36]). Apart from this, no really significant deterioration of NCV and HRV test results can be expected in patients with clinically manifest neuropathy (stagesN2a, N2b, N3). Since it has becomeclear that the 5-year lethality in patients with autonomic disturbances is three times as high as in patients without autonomic involvement, it is recommended that NCV and HRV examinations should be performed in diabetics with no or only slight clinical signs of neuropathy in order to reveal those patients with autonomic disturbances at an initial stage. References 111Ewing, D.J., Campbell, I.W. and Clarke, B.F. (1980)The
natural history of diabetics autonomic neuropathy. Quart. J. Med. 49, 95-99. 121O’Brian, LA., McFadden, J.P. and Corrall, R.J.M. (1991) The influence of autonomic neuropathy on mortality in insulin-dependent diabetes. Quart. J. Med. 79, 495-502. 131 Rundles, R.W. (1945) Diabetic neuropathy: general review with a report of 125 cases. Medicine 24, 1I I-160. 141 Low, P.A. (1993) Laboratory evaluation of autonomic failure. In: P.A. Low (Ed.), Clinical Autonomic Disorders - Evaluation and Management. Little, Brown and Company, Boston, Toronto, London, pp. 169-192. 151 Ewing, D.J., Borsey, D.Q., Bellavere, F. and Clarke, B.F. (1981) Cadiac autonomic neuropathy in diabetes: comparison of measures of RR-interval variation. Diabetologia 21, 18-24. WI Ziegler, D., Laux, G., Dannehl et al. (1992) Assessment of cardiavascular autonomic function: age-related normal ranges and reproducibility of spectral analysis, vector analysis, and standard tests of heart rate variation and blood pressureresponses.Diabetic Med. 9, 166-175. 171 Bellavere, F., Cardone, C., Ferri, M., Guarini, L., Piccoli, A. and Fedele, D. (1987)Standing to lying heart rate variation. A new simple test in the diagnosis of diabetic autonomic neuropathy. Diabetic Med. 4, 41-43. PI Cardone, C., Paiusco, P., Marchetti, G., Burelli, F., Fen&o, M. and Fedele, D. (1990) Cough test to assess cardiovascular autonomic reflexes in diabetes. Diabetes Care 13, 719-724. 191Malpas, KC. and Purdie, G.L. (1990) Circadian variation of heart rate variability. Cardiavasc. Res. 39, 210-213. (101Lawrence. G.P. (1992) . , Home. P.D. and Murrav. A..
Repeatability of measurementsand sourcesof variability in tests of cardiovascular autonomic function. Br. Heart 3. 68, 205-21 I. PII WHO Diabetes mellitus: Report of a study group, Technical report series,No. 727. World Health Organization, Geneva (1985). WI Behse,F., Buchthal, F. and Carlsen, F. (1977) Nerve biopsy and conduction studies in diabetic neuropathy. J. Neural. Neurosurg. Psychiatry 40, 1072-1082. 1131 Dyck, P.J., Kames, J.L., O’Brian, P.C., Litchy, W.J., Low, P.A. and Melton L.J. III (1992) The Rochester diabetic neuropathy study: reassessmentof tests and criteria for diagnosis and staged severity. Neurology 42, 1164-1170. 1141 Weinberg, CR. and Pfeifer, M.A. (1984) An improved method for measuring heart-rate variability: assessment of cardiac autonomic function. Biometrics 40, 855-861. 1151 Benarroch, E.E., Opfer-Gehrking, T.L. and Low, P.A., (1989)Analysis of Valsalva-maneuver in normal subjects. Ann. Neurol. 26, 186A. 1161Genovely, H. and Pfeifer, M.A. (1988) RR-variation: the autonomic test of choice in diabetes. Diab. Metab. Rev. 4, 255-271. 1171 Hilsted, J. and Low, P.A. (1993) Diabetic autonomic neuropathy - evaluation and management. In: Low, P.A. (Ed.), Clinical Autonomic Disorders. Little, Brown and Company, Boston, Toronto, London, pp. 423-443. The diabetes control and complication trial research group (1993)The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabets mellitus. New Engl. J. Med. 329, 977-986. f191Dyrberg, T., Benn, J., Christiansen, J.S., Hilsted, J. and Nerup, J. (1981) Prevalence of diabetic autonomic neuropathy measured by simple bedside tests. Diabetologia 20, 190-194. WI Clark, C.V. and Mapstone, R. (1986) Age-adjusted normal tolerance limits for cardiovascular autonomic function assessmentin the elderly. Age Ageing 15, 221-229. WI O’Brian, LA., O’Hare, J.P. and Corrall, R.J.M. (1986) Heart rate variability in healthy subjects: effect of age and derivation of normal ranges for tests of autonomic function. Br. Heart J. 55, 348-354. WI Wieling, W. and van Lieshout, J. (1990) The assessment of cardiovascular reflex activity: standardization is needed. Diabetologia 33, 182-183. 1231 Ewing, D.J., Burt, A.A., Williams, I.R., Campbell, I.W. and Clarke, B.F. (1976) Peripheral motor nerve function in diabetic autonomic neuropathy. J. Neurol. Neurosurg. Psychiat. 39, 453-460. 1241 American Diabetes Association (1992) Proceedings of a consensusdevelopment conferenceon standardized measures in diabetic neuropathy. Neurology 42, 1823-1839. 1251 Fleiss,J.L., Bigger, J.T. and Rolnitzzky, L.M. (1992)The correlation between heart rate period variability and mean period length. Satist. Med. 11, 125-129. WI Ewing, D.J., Borsey, D.Q., Travis, P., Bellavere, F.,
H.-J. Braune. U. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
Neilson, J.M.M. and Clarke, B.F. (1983) Abnormalities of ambulatory 24-h heart rate in diabetes mellitus. Diabetes 32, 101-105. [27] Kuroda, N., Taniguchi, H., Baba, S. and Yamamoto, M. (1990) Relationship between age and autonomic neuropathy in diabetes mellitus. Diabetes Res. Clin. Pratt. 9, 49-53. [28] Wieling, W., van Brederode, J.F.M., de Rijk, LG., Borst, C. and Dunning, A.J. (1982) Reflex control of heart rate in normal subjects in relation to age: a data base for cardiac vagal neuropathy. Diabetologia 22, 163-166. (291 Ewing, D.J., Neilson, J.M.M., Shapiro, C.M., Stewart, J.A. and Reid, W. (1991) Twenty four hour heart rate variability: effects of posture, sleep, and time of day in healthy controls and comparison with bedside tests of autonomic function in diabetic patients. Br. Heart. J. 65, 239-244. [30] Robertson, D., Wade, D. and Robertson, R.M. (1981) Postprandial alterations in cardiovascular hemodynamics in autonomic dysfunctional states. Am. J. Cardiol. 48, 1048-1052. [31] Ewing, D.J., Bellavere, F., Espi, F. et al. (1986) Correlation of cardiovascular and neuroendocrine tests of
187
autonomic function in diabetes. Metabolism 35, 349-353. [32] Micieli, G., Martignoni, E., Cavallini, A., Sandrini, G. and Nappi, G. (1987)Postprandial and orthostatic hypotension in Parkinson’s disease. Neurology 37, 386-393. 139 Bergstrom, B., Lilja, B., Osterlin, S. and Sundkvist, G. (1987) Autonomic neuropathy in type I diabetes: influence of duration and other diabetic complications. Acta Med. Stand. 222, 147-154. 1341 Veglio, M., Carpano-Maglioli, P., Tonda, L. et al. (1990) Autonomic neuropathy in non-insulin-dependent diabetic patients: correlation with age, sex, duration and metabolic control of diabetes. Diabet. Metabol. 16, 200-206. [35] Jermendy, G., Toth, L., Voros, P. et al. (1990) Impairment of caridorespiratory reflexes and its association with distal somatic neuropathy in diabetic patients free from clinical symptoms of autonomic neuropathy. Exp. Clin. Endocrinol. 96, 199-206. [36] Young, R.J., Zhou, Y.Q., Rodriguez, E., Prescott, R.J., Ewing, D.J. and Clarke, B.F. (1986) Variable relationship between peripheral somatic and autonomic neuropathy in patients with different syndromes of diabetic polyneuropathy. Diabetes 35, 192- 197.
ELSEVIER
Measurement of heart rate variations: influencing factors, normal values and diagnostic impact on diabetic autonomic neuropathy H.-J. Braune*, U. Geisendiirfer Department
of Neurology, Philipps University Hospital, Rudolf-Bultmann-Stra$e
8, D-35033 Marburg, Germany
Received 16 November 1994;revision received 7 August 1995;accepted 15 August 1995
Abstract Different methods of heart rate variation (HRV) measurementare recommendedto assessdisturbances of cardiovascular reflexes in autonomic polyneuropathies. Influencing factors on the E/I-ratio, HRV variations during Valsalva’s maneuver, and change of position from lying down to standing up, as well as age, gender, time of day and satiety were investigated in healthy volunteers. Normal values were then calculated and a total of 100 diabetics with clinically manifest neuropathy of different stages (NO = 8, Nl = 15, N2a = 24, N2b = 28, N3 = 25), 26 of whom were insulindependent and 74 non-insulin-dependent, were examined. Apart from clinical examination, nerve conduction velocity (NCV) measurementof five nerves as well as amplitude measurementof evoked sensory and motor action potentials was performed. No significant influence of the time of day the test was performed could be found. Test reproducibility was good, day-to-day differences did not reach a significant level. This was also true for the test results before and after food intake. Gender-correlated test result differences could not be proven. Statistical analysis showed significant differences between the overall group of diabetics and controls. The difference was most accentuated in the E/I ratio. The further developed the clinical picture of neuropathy, the more pathological were the results obtained in different tests. The results suggestthat most changes leading to pathological values of NCV and HRV occur in a clinical stage in which no, or only very slight, clinical signs give evidence of diabetic neuropathy (NO-Nl). Therefore, theseexaminations should, especially, be performed in diabetics with no, or only slight, clinical signs of neuropathy in order to reveal those patients with neuropathic disturbances at an initial stage. Keywords:
Diabetic autonomic neuropathy; Heart rate variability; Bed-side test
1. Introduction Manifest autonomic neuropathy triples the lethality risk [ 1,2]. Since Rundles published the l Corresponding author, Tel.: +49 06421 286278; Fax.: +49 06421 288955.
first comprehensive report on autonomic disturbances in diabetics in 1945 [3] clinical symptoms have been frequently described, and the introduction of computer-aided analytical methods has promoted the investigation of these autonomic disturbances over the last decade. A variety of different testing procedures has been described to
0168-8227/95/SO9.500 1995 Elsevier Science Ireland Ltd. All rights re%rved SSDI 0168-8227(95)01133-X
180
H.-J. Braune. U. Geisendiirfer /Diabetes Research and Clinical Practice 29 (1995) 179-187
date, however, influencing factors, test optimization, test sensitivity and reproducibility have rarely been taken into consideration [4]. Most of the test procedures measurethe impairment of heart rate variability (HRV) in autonomic neuropathy. Different analytical strategies are implied: standard deviation (SD), difference between maximal and minimal heart rate in a given time period (min-max), ratio of minimal heart rate at expiration and maximal heart rate at inspiration (E/I-ratio), mean difference of consecutive RRintervals (MSD) and vector or spectral analysis (see overviews in Refs. [5,6]). Accentuation of HRV is provoked by different reflex tests, e.g., the Valsalva maneuver, position change from lying down to standing up [5] or the reverse [7], the hand grip test and the cough test [8]. Pharmacological testing with sympatho- or parasympathomimetic drugs and neuroendocrinological measurements give further information about the adaptability of the autonomic nervous system. The apparent clinical ‘patchiness’ may result from differences in organ susceptibility to ‘denervation’. More obviously, however, they may result from an uneven test sensitivity applied to different parts of the autonomic system [4]. Therefore, a single test has always been considered to be insufficient to either detect or deny the presenceof an autonomic involvement. We examined influencing factors on HRV variation such as age, gender, time of day and satiety, since little information is given about these in the literature [9, lo]. We then calculated normal values and examined a population of 100 patients with clinically manifest diabetic neuropathy. Apart from clinical examination, NCV measurement of live nerves as well as amplitude measurement of evoked sensory and motor action potentials was performed. 2. Patients and methods 2.1. Patients
The study was carried out in accordance with the Declaration of Helsinki as modified by the 35th World Medical Assembly, Venice, Italy, 1983. All persons gave their informed consent
prior to being included into the study. Measurement of HRV on deep breathing, after standing up and during Valsalva’s maneuver were performed on 100 consecutive outpatients with diabetes mellitus (WHO definition, Ref. [ 111)at the Dept. of Neurology of the University Hospital Marburg. The study was carried out over a period of 18 month (54 men, 46 women, age 18-82 years, mean 52.0 years). Patients with symptoms caused by other diseasesor casehistories other than diabetes, which could be the cause of abnormal test results for conventional neurophysiological testing such as, e.g., previous injury, lumbar discogenic disease, carpal-canal syndrome, inherited or alcoholic neuropathy, or other neurological diseases,as well as patients taking drugs which act on the cardiovascular systemwere excluded. A further exclusion criterion was the presence of acute focal neuropathy and acute painful neuropathic syndromes, which are probably caused by diabetes, but are rare and presumably of varied pathogenesis. Several studies show that these syndromes have a different clinical course compared to distally symmetrical and primarily sensory polyneuropathy [ 12,131.In accordancewith the clinical stage of neuropathy, which was independently assessed by two experienced neurologists, the patients could be divided into five subgroups, as recommended by Dyck et al. [13]: eight patients without any clinical signs of neuropathy (NO), 15 patients with only one clinical symptom of neuropathy (Nl), 24 patients with mild distally symmetrical sensory polyneuropathy (N2a), 28 patients with distally symmetrical sensorimotor polyneuropathy (N2b), and 25 patients with severe symmetrical sensorimotor polyneuropathy (N3). Autonomic symptoms included vesical incontinence (12 patients), constipation (nine patients), sexual dysfunction (six of 54 males) and orthostatic dysregulation (23 patients). Autonomic symptoms were regarded as mild (only one symptom) in 14 N2b-patients and as moderate (two symptoms) in 18 N3-patients. According to hospital tiles and information given by the patients, the duration of diabeteswas estimated: 20 patients with a duration of diabetes of less than 5 years, 31 patients with a duration of diabetes of more than 5 but less than 10 years, and 49 patients with a duration of more
H.-J. Braune, lJ. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
than 10 years, A total of 26 patients younger than 40 years suffered from IDDM, 74 from NIDDM. In 72 patients, metabolic control was good with normal fasting blood-sugar values and normal or only slightly HbAi, values elevated. It was moderate in 28 patients. Tests were also done on age- and sex-matched healthy volunteers of different ages (51 men, 49 women, age 18-83 years, mean 50.3 years) to define the normal range of values (~95th percentile). The influence of time of day of measuring was tested on six women (age 20-68 years) and four men (35-63 years). To examine the retestvariability, eight women (age 23-63 years) and two men (23 and 63 years old) underwent the HRV-testing procedure several times. The influence of satiety on results of HRV-testing was checked in seven women (24-72 years) and six men (27-69 years). 2.2. Methods The examinations were carried out on the basis of routine outpatient visits. On the day of the examination the patients had their usual breakfast and, if necessary, their morning insulin dose. Apart from clinical examination, all patients were submitted to conventional electroneurographical tests. Measurement of motor nerve conduction velocity (MNCV) was performed on nn. medianus, tibialis and peronaeus. Measurement of sensory nerve conduction velocity (SNCV) were done on nn. medianus and suralis at a standardized limb temperature of 35°C. Furthermore, we assessed axonal loss by measurement of amplitudes of evoked sensory and motor action potentials. The sums of the three amplitudes of motor nerves and the two sensory nerves were calculated for statistical evaluation and data reduction. Assessmentof HRV included measurement of the E/I-ratio after 5 min of deep rhythmical breathing. It was measured computer-aided by an nDX-test-unit (Fa Q-Med, 100 Metro Park South, Laurence Harbour, NJ 08878, USA). After an appropriate tryout time the persons tested were requested to breath regularly and deep with five exand inspirations per minute. Breathing was synchronized using an acoustic signal and the inter-
181
vention of surveying personnel if necessary. E/I-ratio is calculated as the mean of five ratios derived from the longest RR-interval during inspiration and shortest RR-interval during exspiration [5,6]. Furthermore, the mean circular resultant (MCR) represents the length of the main vector derived on a unit circle with the periodicity of one breathing cycle of ex- and inspiration. The heartbeats as events on time are given as dots on the unit circle. The main vector correlates with these dots [14]. The third parameter is the standard deviation (S.D.) of RR-intervals in the 5-min measuring time. To perform the Valsalva manouever the person tested is asked to exspirate forcefully against a resistance for 20 s [ 151.Three tests were carried out with a retest-interval of 2 min. The Valsalva-ratio is the ratio: longest RRinterval after exspiration/shortest RR-interval during the forced exspiration period. The mean value of the three tests performed was calculated. The 30/15-ratio is the ratio 30th/15th RR-interval after standing up from lying down position. The person was asked to relax in prone position for 5 min and then, unassisted, to stand up in a period of 5 s, and an ECG was registered for a further 60 s. Measurement of autonomic function was performed before noon in a quiet room at a temperature of 22°C and with dimmed lights after an ‘acclimatization period’ of 20 min. None of the subjectswas allowed to drink coffee or tea, smoke, or take medicaments in the hours before and during the visit. Considerable physical effort or hypo- or hyperglycaemic periods up to 3 days before the examination led to exclusion from the study [ 161.To assessthe retest variability, the testing procedure was performed on three consecutive days, between 4:00 and 5:00 p.m., on 10 healthy volunteers (eight women and two men). To examine circadian influences on the results of HRV-testing, the procedures were performed on 10 healthy volunteers on day one between 7:30 and 8:00 a.m. and on the following day between 4:00 and 5:00 p.m [9]. To test the influence of food intake, 13 healthy volunteers agreed to be tested between 7:30 and 8:OOa.m. after a period of more than 10 h without food, and again 20 min after consumption of
H.-J. Braune. U. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
182
a
b 1.8
140
17
.z y 1
16 1.5
3
.
14
.*
13
..
. .
11
. . .I . .
.
. .
.
.
:.
.
.,
:.
10
*
.
-* . .;--: . . . . . .. , _ *ii
30
4c
50
is I
---! . . . . . .
. 1’ -. ,,,,,,,,,
.
I,,,,
20
5 I
8. .
12
2 =
.
-.
60
.
70
80
.
L I
I.
30
20
.
.
40
50
60
70
80
d
C
.
2.5
160 140 E
80 70 60 50 40 30 20 10 0
.
120 100
'
80
is I
60 40 20 0
-.
. ..
.... * -1 .. . .: * . *:. *. . . 1 . . . 40 20 30
.
.o s 1 2 $ 2
.
.. * .... * .* a. -._ . *.-: y-..: 1 *. 50
63
70
8C
1.8 1.7 1.6 1.5 1.4 1.3 12 11 10
:
20
-.. *.
30
.
40
50
60
70
80
e
Fig. I. Results of heart rate variation (HRV) testing on normal control persons of different age (n = 100, for explanation of testing procedure see text): (a-c) 5 min deep rhythmical breathing: (a) E/I-ratio; (b) mean circular result (MCR); (c) standard deviation (SD.); (d) Valsalva maneuver; (e) position change from lying down to standing (30/I 5-ratio). The bold line represents the 95th percentile (95% of all normal values are greater). The dotted line represents the regression line, the influence of age is highly significant.
100 g water-free glucose (oral glucose tolerance test). The examinations of patients with diabetic neuropathy were carried out during a routine outpatient visit. On the day of the examination the patient had his usual breakfast and, if necessary, his morning insulin dose. Statistical exploration was performed with the parameter-free Mann-Whitney U-test, and graphi-
cal documentation was with Box-and-Whisker plots. The obtained values of controls were analysed to determine the 95th percentile. A minimal criterion for assumedabnormality in patients was a test result value > 95th percentile. Furthermore, Spearmanscorrelation coefficient and the partial correlation coefficient excluding the influence of age were calculated. Analysis of gender influence was calculated by Students r-test. Wilcoxon
H.-J.
Braune,
U. Geisendkfer
/ Diabetes
matched-pairs signed rank test was performed to assessinfluence of time of day and food intake. The level of significance of the day-to-day test result variations was defined by Friedmann’s twoway ANOVA.
Research
and
a
1.8
o .‘,
1.7 1.6
_t
1.5
>
1.4 I
3. Results
In accordancewith other studies, all parameters showed a decreasewith increasing age in the control group. Nevertheless,we defined single normal range for all agessince the control group was too small for defining age-dependent normal values, and since the resulting error was deemed to be negligible with respect to statistical analysis. Fig. la-e gives the results of HRV testing on controls, the influence of age is highly significant. Table 1 shows the normal ranges of the HRV tests, NCV measurements, and the amplitudes used in this study; the cut-off values were defined as 95thpercentiles. No significant influence of the time of day the test was performed could be found. Furthermore, Table 1 Synopsis of normal ranges established by examination of 100 normal subjects in our laboratory; test battery with electroneurography and various heart rate variation tests Test Nerve conduction velocity N. tibialis: mot. N. peronaeus: mot. N. suralis: sens. N. medianus: mot. N. medianus: sens. Sum of 3 evoked compound motor action potentials (nn. medianus, peronaeus, tibialis) Sum of 2 evoked compound sensory action potentials (nn. medianus, suralis) Heart rate variation E/I-Ratio S.D. MCR Valsalva-ratio 30/l 5-ratio
Normal range (>95th percentile) >40 >40 >40 >50 >50
2 =
183
Clinical Praclice 29 (1995) 179-187
0 0
1.3 1.2 1.1 1 .o
b 5 2 m 2 ’ 1 2
I
=
Diabetics;
Normals
Diabetics;
Normals
Diabetics;
Normals
2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1 .I
1.8
.-0 +
?
1.6
m/s m/s m/s n-h m/s
>18 mV
>lO pv
>I.06 >24 >I2 > 1.08
>I.04
Fig. 2. Group differences between diabetics and age- and sexmatchedcontrols with regard to heart rate variation (HRV): (a) 5 min of deep rhythmical breathing (E/I-ratio, and results for SD. and MCR are equivalent and therefore omitted here); (b) Valsalva-maneuver; (c) position change from lying down to standing (30115ratio). The bold line representsthe 95th percentile. Boxes, the range of values between the 25th and 75th percentiles; small T-lines (whiskers), the range between 10th and 90th percentiles. Open circles indicate extreme values. **Significant difference (P < 0.01); *slightly significant difference (P < 0.05); -, no significant difference (P > 0.05).
184
H.-J. Braune, IJ. Geisendiirfer / Diabetes Research and Clinical Practice 29 (I 995) I79- 187
the retest variability was good, day-to-day differencesdid not reach a significant level. This is also true for the testing before and after food intake. Gender-related test result differences could not be proven. Statistical analysis showed significant differencesbetween the group of diabetics and the control group. The difference was most accentuated in NCV, and clear in amplitude measurement of evoked sensoryand motor action potentials as well as in HRV tests (Fig. 2a-c). With respect to NCV of the sural nerve, eight of the Nl patients (n = 15), and with respect to the E/I-ratio (Fig. 3a) nine of the Nl patients, showed values below the normal range. Thus, NCV of the sural nerve and the E/I-ratio were the most sensitive parameters in detecting abnormalities in clinical stage Nl diabetics (only one clinical sign of neuropathy). In combination, the two methods showed pathological results in 72% of all diabetics (Nl-N3). For a quick survey, the combination of these two tests is advisable. Relying on this type of HRV test for examination of autonomic function alone, however, would have missed seven of the 59 subjects with abnormal test results in one or more of the other HRV testing procedures. In agreement with the recommendations of, e.g., Low [4], combination of these different HRV tests is thus advised. Even in summarizing and scoring the subtest results for sensorimotor neuropathy as well as for autonomic dysfunction, we found that a statistically significant influence of gender, age, metabolic control and duration or type of diabetes on the test results could not be found or was only present to a minor extent. This was rather surprising since age-relatedvariations and variations related to the duration of diabetes have been described in the literature [ 17,181.The failure to reproduce these already established data is probably due to the low number of patients tested in our study. An influence to some extent could be observed, however, data did not reach significant levels. Nevertheless, the clinical stage of diabetic neuropathy (NO-N3) exerted a significant influence on our study: the further developed the clinical picture of the disease, the more pathological were the different test results (Fig. 3a-c). It should
a t 0 1; 0
0
13 0
t-
b
LL t IT
o
C
Fig. 3. Influence of clinical stage of diabetic neuropathy on heart rate variation (HRV). (a) Five minutes of deep rhythmic breathing (E/I-ratio); (b) Valsalva maneuver; (c) position change from lying down to standing (30/15-ratio). Clinical staging as recommended by Dyck et al. (1992): NO, diabetics without any clinical sign of neuropathy (n = 8). Nl, diabetics with only one clinical sign of neuropathy (n = 15); N2a, diabetics with a mild distal symmetric sensory neuropathy (n = 24); N2b, diabetics with distal symmetric sensory motor neuropathy (n = 28); N3, diabetics with a severe symmetric sensory motor neuropathy (n = 25). Statistical analysis is performed with group-by-group comparison. The bold line represents the 95th percentile. Boxes, the range of values between the 25th and 75th percentiles; small T-lines (whiskers), the range between 10th and 90th percentiles; open circles, extreme values. **Significant difference (P < 0.01). *slight significant difference (P c 0.05); -, no significant difference (P > 0.05)
H.-J. Braune. I/. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
be stressedthat the greatest differences were found between diabetic neuropathy groups NO, Nl and N2a with no or only slight clinical signs. 4. Discussion
In recent years, several suggestionswith respect to a standardized test battery of autonomous nervous system function tests have been proposed, but an agreement has not yet been reached. The earliest and most common manifestations of autonomic neuropathy occur in the cardiovascular system in the form of orthostatic hypotension, resting tachycardia, painless myocardial ischemia and a ‘rigid pulse’. HRV is clearly involved in most forms of parasympathetic damage in autonomic neuropathy. A pathological result with the reflex test does not give any indication of the exact localization of the lesion: disturbances in the afferent, central, or efferent part of the reflex pathway, or even in the innervated organ itself, may be present [4]. Nevertheless, simple noninvasive tests permit accurate quantification and follow-up. Many reports on diabetic autonomic neuropathy measuredby simple HRV tests have been published [5,17,19-221 with a good correlation between the results [5] and a clear correlation to findings of somatic nerve function testing [23]. In accordance with the recommendations of the American Diabetes Association [24], three different testing procedures,namely deep rhythmic breathing, forced expiration against resistance, and position changefrom lying down to standing up, were combined in our study. The results of HRV testing at deep regular breathing were mathematically interpreted in three different ways (E/I-ratio, MCR, S.D.) to eliminate erroneous values due to variations of the intrinsic heart rate and extrasystoles WI. In order to obtain interpretable results, it is recommended to ascertain as many influencing factors as possible so that standardization of test conditions and procedures is feasible. The importance of the patient’s age is often stressedbut occasionally other factors are reported. The aim of this study was to investigate retest variability and influencing factors, such as age, gender, time of day the test was performed and satiety, on the results
185
of a common HRV test battery. Furthermore, the prevalence of pathological test results in a selected population with diabetes mellitus was examined. A strong correlation between age and test results existed in controls (Fig. la-e). Since the group was to enough to establish age-dependent normal ranges, these values were defined as 95th percentiles of the overall population. Further data sampling on control subjects will be helpful to make this testing battery even more reliable. Test reproducibility was good, day-to-day differences did not reach a significant level. In accordancewith other studies, no significant influence of the time of day the test was performed [26,27], the gender of tested persons [6,28,29], or the food intake before testing could be found [30-321. A statistically significant influence of gender, age, metabolic control, duration or type of diabetes on the test results could not be found, or was present only to a minor degree.This confirms already published findings [33,34], but is contrary to the results of Kuroda et al. [27] who found a relationship between age and autonomic neuropathy in diabetes mellitus. Failure to reproduce these already established data is probably due to the low number of patients tested in our study. An influence could be observed to some extent, however, data did not reach significant levels. After having established and validated this test battery, we are now in the process of setting up a more comprehensive study with recruitment of more patients. The most important parameter was the clinical stageof diabetic neuropathy (NO-N3): the further developed the clinical picture of the disease, the more pathological were the different test results. Of interest, however, is the fact that the greatest differences - regarding NCV and HFV - were found between diabetic neuropathy groups NO, Nl, and N2a (Fig. 3a-c). 5. Conclusion
The HRV test battery reviewed here is kind to patients, easy to carry out and is not time consuming. Influencing factors are controllable using a standardized test protocol. The results suggestthat
186
H.-J. Braune, U. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
most changes leading to pathological findings of NCV and HRV measurement occur in a clinical stage,in which no or only very slight clinical signs give evidence of diabetic neuropathy (NO-Nl, see also Refs. [35,36]). Apart from this, no really significant deterioration of NCV and HRV test results can be expected in patients with clinically manifest neuropathy (stagesN2a, N2b, N3). Since it has becomeclear that the 5-year lethality in patients with autonomic disturbances is three times as high as in patients without autonomic involvement, it is recommended that NCV and HRV examinations should be performed in diabetics with no or only slight clinical signs of neuropathy in order to reveal those patients with autonomic disturbances at an initial stage. References 111Ewing, D.J., Campbell, I.W. and Clarke, B.F. (1980)The
natural history of diabetics autonomic neuropathy. Quart. J. Med. 49, 95-99. 121O’Brian, LA., McFadden, J.P. and Corrall, R.J.M. (1991) The influence of autonomic neuropathy on mortality in insulin-dependent diabetes. Quart. J. Med. 79, 495-502. 131 Rundles, R.W. (1945) Diabetic neuropathy: general review with a report of 125 cases. Medicine 24, 1I I-160. 141 Low, P.A. (1993) Laboratory evaluation of autonomic failure. In: P.A. Low (Ed.), Clinical Autonomic Disorders - Evaluation and Management. Little, Brown and Company, Boston, Toronto, London, pp. 169-192. 151 Ewing, D.J., Borsey, D.Q., Bellavere, F. and Clarke, B.F. (1981) Cadiac autonomic neuropathy in diabetes: comparison of measures of RR-interval variation. Diabetologia 21, 18-24. WI Ziegler, D., Laux, G., Dannehl et al. (1992) Assessment of cardiavascular autonomic function: age-related normal ranges and reproducibility of spectral analysis, vector analysis, and standard tests of heart rate variation and blood pressureresponses.Diabetic Med. 9, 166-175. 171 Bellavere, F., Cardone, C., Ferri, M., Guarini, L., Piccoli, A. and Fedele, D. (1987)Standing to lying heart rate variation. A new simple test in the diagnosis of diabetic autonomic neuropathy. Diabetic Med. 4, 41-43. PI Cardone, C., Paiusco, P., Marchetti, G., Burelli, F., Fen&o, M. and Fedele, D. (1990) Cough test to assess cardiovascular autonomic reflexes in diabetes. Diabetes Care 13, 719-724. 191Malpas, KC. and Purdie, G.L. (1990) Circadian variation of heart rate variability. Cardiavasc. Res. 39, 210-213. (101Lawrence. G.P. (1992) . , Home. P.D. and Murrav. A..
Repeatability of measurementsand sourcesof variability in tests of cardiovascular autonomic function. Br. Heart 3. 68, 205-21 I. PII WHO Diabetes mellitus: Report of a study group, Technical report series,No. 727. World Health Organization, Geneva (1985). WI Behse,F., Buchthal, F. and Carlsen, F. (1977) Nerve biopsy and conduction studies in diabetic neuropathy. J. Neural. Neurosurg. Psychiatry 40, 1072-1082. 1131 Dyck, P.J., Kames, J.L., O’Brian, P.C., Litchy, W.J., Low, P.A. and Melton L.J. III (1992) The Rochester diabetic neuropathy study: reassessmentof tests and criteria for diagnosis and staged severity. Neurology 42, 1164-1170. 1141 Weinberg, CR. and Pfeifer, M.A. (1984) An improved method for measuring heart-rate variability: assessment of cardiac autonomic function. Biometrics 40, 855-861. 1151 Benarroch, E.E., Opfer-Gehrking, T.L. and Low, P.A., (1989)Analysis of Valsalva-maneuver in normal subjects. Ann. Neurol. 26, 186A. 1161Genovely, H. and Pfeifer, M.A. (1988) RR-variation: the autonomic test of choice in diabetes. Diab. Metab. Rev. 4, 255-271. 1171 Hilsted, J. and Low, P.A. (1993) Diabetic autonomic neuropathy - evaluation and management. In: Low, P.A. (Ed.), Clinical Autonomic Disorders. Little, Brown and Company, Boston, Toronto, London, pp. 423-443. The diabetes control and complication trial research group (1993)The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabets mellitus. New Engl. J. Med. 329, 977-986. f191Dyrberg, T., Benn, J., Christiansen, J.S., Hilsted, J. and Nerup, J. (1981) Prevalence of diabetic autonomic neuropathy measured by simple bedside tests. Diabetologia 20, 190-194. WI Clark, C.V. and Mapstone, R. (1986) Age-adjusted normal tolerance limits for cardiovascular autonomic function assessmentin the elderly. Age Ageing 15, 221-229. WI O’Brian, LA., O’Hare, J.P. and Corrall, R.J.M. (1986) Heart rate variability in healthy subjects: effect of age and derivation of normal ranges for tests of autonomic function. Br. Heart J. 55, 348-354. WI Wieling, W. and van Lieshout, J. (1990) The assessment of cardiovascular reflex activity: standardization is needed. Diabetologia 33, 182-183. 1231 Ewing, D.J., Burt, A.A., Williams, I.R., Campbell, I.W. and Clarke, B.F. (1976) Peripheral motor nerve function in diabetic autonomic neuropathy. J. Neurol. Neurosurg. Psychiat. 39, 453-460. 1241 American Diabetes Association (1992) Proceedings of a consensusdevelopment conferenceon standardized measures in diabetic neuropathy. Neurology 42, 1823-1839. 1251 Fleiss,J.L., Bigger, J.T. and Rolnitzzky, L.M. (1992)The correlation between heart rate period variability and mean period length. Satist. Med. 11, 125-129. WI Ewing, D.J., Borsey, D.Q., Travis, P., Bellavere, F.,
H.-J. Braune. U. Geisendiirfer / Diabetes Research and Clinical Practice 29 (1995) 179-187
Neilson, J.M.M. and Clarke, B.F. (1983) Abnormalities of ambulatory 24-h heart rate in diabetes mellitus. Diabetes 32, 101-105. [27] Kuroda, N., Taniguchi, H., Baba, S. and Yamamoto, M. (1990) Relationship between age and autonomic neuropathy in diabetes mellitus. Diabetes Res. Clin. Pratt. 9, 49-53. [28] Wieling, W., van Brederode, J.F.M., de Rijk, LG., Borst, C. and Dunning, A.J. (1982) Reflex control of heart rate in normal subjects in relation to age: a data base for cardiac vagal neuropathy. Diabetologia 22, 163-166. (291 Ewing, D.J., Neilson, J.M.M., Shapiro, C.M., Stewart, J.A. and Reid, W. (1991) Twenty four hour heart rate variability: effects of posture, sleep, and time of day in healthy controls and comparison with bedside tests of autonomic function in diabetic patients. Br. Heart. J. 65, 239-244. [30] Robertson, D., Wade, D. and Robertson, R.M. (1981) Postprandial alterations in cardiovascular hemodynamics in autonomic dysfunctional states. Am. J. Cardiol. 48, 1048-1052. [31] Ewing, D.J., Bellavere, F., Espi, F. et al. (1986) Correlation of cardiovascular and neuroendocrine tests of
187
autonomic function in diabetes. Metabolism 35, 349-353. [32] Micieli, G., Martignoni, E., Cavallini, A., Sandrini, G. and Nappi, G. (1987)Postprandial and orthostatic hypotension in Parkinson’s disease. Neurology 37, 386-393. 139 Bergstrom, B., Lilja, B., Osterlin, S. and Sundkvist, G. (1987) Autonomic neuropathy in type I diabetes: influence of duration and other diabetic complications. Acta Med. Stand. 222, 147-154. 1341 Veglio, M., Carpano-Maglioli, P., Tonda, L. et al. (1990) Autonomic neuropathy in non-insulin-dependent diabetic patients: correlation with age, sex, duration and metabolic control of diabetes. Diabet. Metabol. 16, 200-206. [35] Jermendy, G., Toth, L., Voros, P. et al. (1990) Impairment of caridorespiratory reflexes and its association with distal somatic neuropathy in diabetic patients free from clinical symptoms of autonomic neuropathy. Exp. Clin. Endocrinol. 96, 199-206. [36] Young, R.J., Zhou, Y.Q., Rodriguez, E., Prescott, R.J., Ewing, D.J. and Clarke, B.F. (1986) Variable relationship between peripheral somatic and autonomic neuropathy in patients with different syndromes of diabetic polyneuropathy. Diabetes 35, 192- 197.