ORIGINAL ARTICLE 
Annals of Nuclear Medicine Vol.l3, No.2, 95-100,1999 
Diabetic cardiac autonomic dysfunction: parasympathetic versus sympathetic 
Akihiko UEHARA, Chinori KURATA, Toshihiko SUGI, Tadashi MIKAMI and Sakae SHOUDA 
Department of Medicine III, Hamamatsu University School of Medicine 
Background: Diabetic cardiac autonomic dysfunction often causes lethal arrhythmia and sudden cardiac death. 123I-Metaiodobenzylguanidine (MIBG) can evaluate cardiac sympathetic dysfunction, and analysis of heart rate variability (HRV) can reflect cardiac parasympathetic activity. We examined whether cardiac parasympathetic dysfunction assessed by HRV may correlate with sympathetic dysfunction assessed by MIBG in diabetic patients. 
Methods and Results: In 24-hour electrocardiography, we analyzed 4 HRV parameters: high-frequency power (HF), HF in the early moming (EMHF), rMSSD and pNN50. MIBG planar images and SPECT were obtained 15 minutes (early) and 150 minutes (late) after injection and the heart washout rate was calculated. The defect score in 9 left ventricular regions was scored on a 4 point scale (0 = normal - 3 = severe defect). In 20 selected diabetic patients without congestive heart failure, coronary artery disease and renal failure, parasympathetic HRV parameters had a negative correlation with the sum of defect scores (DS) in the late images (R = -0.47 - -0.59, p < 0.05) and some parameters had a negative correlation with the washout rate (R = -0.50 - -0.55, p < 0.05). In a total of 64 diabetic patients also, these parameters had a negative correlation with late DS (R = -0.28 - -0.35, p < 0.05) and early DS (R = -0.27 - -0.32, p < 0.05).
Conclusions: The progress of diabetic cardiac parasympathetic dysfunction may parallel the sympathetic one. 
Key words: diabetic mellitus, autonomic nervous system, metaiodobenzylguanidine, heart rate variability
INTRODUCTION 
PATIENTS with diabetes mellitus frequently have autonomic neuropathy, which often limits their lives by causing arrhythmia and sudden cardiac death.1,2 The progression of diabetic cardiac autonomic dysfunction may be intervened by using intensive insulin therapy and some medication.3 It is therefore important to detect cardiac autonomic nervous system dysfunction in diabetic patients as early as possible. Many investigators have assessed diabetic autonomic neuropathy by means of 
Received Nobember 26, 1998, revision accepted February 8, 1999. 
For reprint contact: Akihiko Uehara. M.D., Department of Medicine III, Hamamatsu University School of Medicine, 3600 Handa-cho. Hamamatsu 431-3192, JAPAN. 
E-mail: hiro26@hama-med.ac.jp 
heart rate variability (HRV)4 and myocardial scintigraphy with 123I-metaiodobenzylguanidine (MIBG).5 
Some investigators6,7 have reported that myocardial MIBG uptake predicts autonomic dysfunction in diabetic patients and is related to sympathetic neural modification of HRV such as low-frequency power (LF). In addition, one6 of them reported that MIBG scintigraphy is the most accurate quantitative measurement of diabetic autonomic dysfunction and that subjects with clinical autonomic dysfunction had the same values in HRV parasympathetic parameters as those without autonomic dysfunction. Because of this fact, they concluded that the parasympathetic component of HRV is less affected by diabetes than the sympathetic component, but LF reflects both sympathetic and parasympathetic influences, indicating that a decrease in LF does not necessarily imply a reduction in sympathetic activity.8 Earlier studies9,10 also demonstrated that defects in parasympathetic innervation were more frequent and occurred relatively earlier in diabetes mellitus and that defects in sympathetic innervation were less frequent and occurred later. If diabetic autonomic para-sympathetic dysfunction precedes or parallels a sympathetic one, it would be wise to examine HRV at first to evaluate diabetic autonomic dysfunction because HRV has some advantages over MIBG in costs and irradiation. 
The purpose of our study was to examine how the parasympathetic component of HRV may relate to myocardial MIBG washout and distribution and whether cardiac parasympathetic dysfunction assessed by HRV may follow a sympathetic one assessed by MIBG in diabetic patients. 
MATERIALS AND METHODS 
Study population. Of 598 consecutive patients who underwent myocardial scintigraphy with MIBG as a routine clinical examination in our laboratory from January, 1993 to July, 1997, 65 had diabetes mellitus and underwent a 24-hour ambulatory electrocardiographic recording. Since one of the 65 patients was excluded because of frequent extrasystoles (> 5% of all beats), the study population comprised the remaining 64 patients (age 60 +- 13 years, 28 females and 36 males). The median between HRV and MIBG studies was 18 days. None of the patients had tricyclic-antidepressants. Other characteristics including medications are summarized in Table 1. We further studied a subgroup of 20 patients (age 53 +- 12 years, 8 females and 12 males) in whom complications such as coronary artery disease, heart failure and renal failure on dialysis were excluded. Coronary artery disease was defined by history, stress perfusion scintigraphy and/ or coronary angiography. Heart failure and renal failure on dialysis were defined by history, medical records and laboratory examinations. The subgroup characteristics are also summarized in Table 1. 
MIBG. Myocardial scintigraphy with MIBG was performed by a previously reported method.11,12 Anterior images of 180 seconds acquisition were obtained over the whole thorax 15 and 150 min after 111 MBq MIBG injection. We measured mean counts per pixel in the heart (H) and upper mediastinum (M) in the early and late images. The washout rate (%) from the heart was calculated with the following formula:[(early H -early M)-(late H - late M)] x 100/(early H - early M). 
Immediately after the early and late planar imaging, single-photon emission computed tomography (SPECT) was acquired by using a three-headed SPECT system (PRISM3000. Picker International, Inc., Ohio) equipped with low-energy, general purpose, parallel-hole collimators and interfaced with a computer (Odyssey. Picker International, Inc., Ohio). Twenty projections over 120deg per head were acquired in a 64 x 64 matrix for 30 sec per projection. No attenuation or scatter correction was used. Transverse slices were reconstructed with a filtered backprojection algorithm after pre-processing of the projection images with a low pass filter. Vertical long-axis, short-axis and horizontal long-axis tomograms were reconstructed from the transverse slices. The MIBG defect score of was visually determined for each of a total of 9 myocardial segments (basal anterior, basal septal, basal inferior, basal lateral, apical anterior, apical septal, apical inferior, apical lateral walls and apex) according to a 4-point scoring system: 0 = normal, 1 = mildly reduced, 2 = moderately reduced and 3 = greatly reduced. Early and late defect scores (DS) were defined as the sum of the defect scores for the 9 segments in the early and late images, respectively. 
HRV. We performed 24-hour ambulatory electrocardiographic recording with Marquette 2-channel recorders (Marquette Electronics, Milwaukee, WI).11 With a fast Fourier transform algorithm, we analyzed the time series of normal RR intervals for the entire 24 hours. High-frequency power (HF, ms2) was computed for the 0. 15-0.40 Hz band with a Marquette HRV program (version 002A). We calculated HF in the early morning (EMHF, ms2), that is, from 2 a.m. to 5 a.m. to assess a parasympathetic component in the relatively stationary state. In addition, we calculated the square root of the mean of the squared differences between adjacent normal RR intervals (rMSSD, ms), and the proportion of adjacent normal RRS more than 50 ms different (pNN50,%). We considered HF, EMHF, rMSSD and pNN50 to be parasympathetic parameters. 
Data analysis. All measures were expressed as the mean +- SD. Comparisons of continuous data in the two series were made by unpaired Student's t-test. Correlation between two variables was examined by linear regression analysis. Logarithm transformation was used before statistical analysis of some variables to make them normally distributed. Statistical significance was defined as p < 0.05. 
RESULTS 
The late defect score could not be evaluated because of greatly reduced myocardial MIBG uptake in 2 of 20 selected patients as well as in 8 of the total of 64 patients. In the total population, HF, EMHF, rMSSD and pNN50 had no significant correlation with the MIBG washout rate, but had a significant negative correlation with both the early and late DS (Table 2). Representative figures are shown in Figure 1 . In the 20 patients selected, there was a significant negative correlation of the washout rate with rMSSD and pNN50, of late DS with HF, EMHF, rMSSD and pNN50, and of early DS with pNN50 (Table 2 and Figure 2). As a result, HF and EMHF similarly correlated with MIBG parameters in both the selected population and the total population in this study. 
A representative case with a discrepancy between MIBG and HRV is shown in Figure 3. The early and late MIBG images showed only a minimal defect in the apex and the washout rate was not increased (3.9%), although HRV was greatly decreased; for example, In(HF) and rMSSD were 3.13 In(ms2) (normal range 5.26 +- 0.70) and 15 ms (normal range 34 +- 12), respectively. 
DISCUSSION 
This study demonstrated that HF, EMHF,rMSSD and pNN50 negatively correlated with MIBG late DS in diabetic patients and that.rMSSD and pNN50 negatively correlated with the MIBG washout rate in diabetic patients without coronary artery disease, heart failure or renal failure on dialysis. In all the diabetic patients, HF, EMHF, rMSSD and pNN50 negatively correlated with MIBG early and late DS. 
Both rMSSD and pNN50 have been accepted as para-sympathetic measures, but 24-hour HF is inappropriate for the evaluation of cardiac parasympathetic activity since the physiological interpretation of the spectral components calculated over 24 hours is difficult.8 Therefore EMHF (HF from 2 a,m. to 5 a.m.) is more suitable for assessing parasympathetic modification than 24-hour HF because of its stationariness.8 In this study, however, the same tendency was obtained with HF and EMHF. On the other hand, MIBG scintigraphy may reflect cardiac sympathetic innervation.13 Some investigators have reported that diabetic patients with autonomic dysfunction generaliy show low late myocardial uptake of MIBG with rapid washout, which is prominent in the inferior wall in the early stage and progresses to the adjacent areas of the ventricle 14-16 An increased washout rate was assumed to be related to increased noradrenaline activity in the sympathetic nerves of diabetic patients with autonomic neuropaty.15,17,18 In our study, the increased washout rate of MIBG correlated closely the impairment of parasympathetic nerves. Some investigators14,15,19 showed that the ratio of heart to mediastinum (H/M) correlated with sympathetic impairment in diabetic patients. Although it is not shown in this paper, however H/M did not correlate with parasympathetic parameters of HRV in our study. The insignificant correlation may be due to the fact that sympathetic impairment begins with a local area which is indicated by a defect score and progresses to the whole heart, as indicated by H/M. Hattori et al.15 reported that H/M showed less sensitivity than the indices of regional uptake in evaluating diabetic autonomic neuropathy by MIBG. Our observations therefore suggest that cardiac parasympathetic neuropathy assessed by HRV may parallel the sympathetic one assessed by MIBG in diabetes mellitus. 
Some studies6,7 reported that LF had a negative correlation with the late MIBG defect score. One6 of them also reported that 24-hour HF, rMSSD and pNN50 were similar but MIBG defect size and LF in diabetic patients with and without clinically evident autonomic dysfunction were significantly different, and it suggested that parasympathetic nerves are less affected by diabetes than sympathetic ones. There are, however, some limitations to their suggestion. First, LF is a parameter that includes both sympathetic and parasympathetic influences, indicating that a decrease in LF does not necessarily imply a reduction in sympathetic activity.8 Second, diabetic patients "without autonomic dysfunction" in the study6 were likely to already have parasympathetic neuropathy, since their HF, rMSSD and pNN50 decreased compared to normal subjects (Table 3).20-22
The plots in Figure 2A, B,C, and D are fitted linearly, suggesting that sympathetic neuropathy may deteriorate at the same rate as parasympathetic neuropathy. Earlier studies9,10 demonstrated that defects in parasympathetic innervation were more frequent and occurred relatively earlier in diabetes mellitus than defects in sympathetic innervation. In fact, the plots in Figure 2A, B and C may be fitted to a concave curve, and it is therefore possible that sympathetic neuropathy develops slowly in the early stage and accelerate in the late stage, whereas parasympathetic neuropathy may develop early and deteriorate gradually. Nevertheless, the number of plots in these figures is not enough to support this hypothesis because one or two data points would easily change an apparent concave curve into a linear relationship. Diabetic cardiac parasympathetic neuropathy may therefore not follow a sympathetic one, but may parallel or precede it. In other words, HRV analysis can detect cardiac autonomic neuropathy at least as early as MIBG scintigraphy, although these methods evaluate different aspects of autonomic neuropathy. 
Comparing HRV and MIBG, HRV has some advantages over MIBG. First, the cost of HRV is cheaper than that of MIBG. In Japan the cost of ambulatory electrocardiographic recording is $115 (\15,000), but that of MIBG scintigraphy is $507 (\65,922). Second, HRV can be undergone without irradiation, but MIBG cannot. On the other hand, HRV has a disadvantage: it cannot be evaluated in the patients with frequent extra beats and atrial fibrillation. 
Although we demonstrated a significant correlation between MIBG and HRV parameters, the correlation is not very close. We should bear in mind that HRV cannot replace MIBG and that these methods evaluate different aspects of autonomic neuropathy. 
Study limitations 
First, our study population was relatively small but a consistent tendency was obtained. Second, medications capable of interfering with the assessment of autonomic function were administered to some patients in this study, but all medications were withheld on the day of the examination. Third, the prognostic studies are strictly necessary to elucidate whether cardiac parasympathetic dysfunction assessed by HRV may follow cardiac sympathetic dysfunction assessed by MIBG in diabetic patients; for that reason further studies are needed in the future. 
Clinical implication 
Our study demonstrated that HRV analysis can detect cardiac autonomic neuropathy at least, as early as MIBG scintigraphy, and HRV analysis also has some advantages over MIBG, such as costs and irradiation. In patients without frequent arrhythmia, therefore, the examination of HRV parasympathetic parameters can be the gatekeeper of an MIBG study. 
Conclusion 
In summary, a reduction in HRV may parallel or precede abnormalities in myocardial MIBG uptake in patients with diabetic mellitus. 
REFERENCES 
1. Kahn JK, Sisson JC, Vink AI. QT interval prolongation and sudden death in diabetic autonomic neuropathy. J Clin Endocrinol Metab 64: 751-754, 1987. 
2. Ewing DJ, Campbell IW, Clarke BF. Mortality in diabetic autonomic neuropathy. Lancet 1 : 601-603, 1976. 
3. Kurata C, Okayama K. Wakabayashi Y, Shouda S, Mikami T, Tawarahara K, et al. Cardiac sympathetic neuropathy and effects of aldose reductase inhibitor in streptozotocin-induced diabetic rats. J Nucl Med 38: 1677-1680, 1997. 
4. Pagani M. Malfatto G, Pierini S, Casati R, Masu AM, Poli M, et al. Spectral analysis of heart rate variability in the assessment of autonomic diabetic neuropathy. J Auton Nerv Syst 23: 143-153, 1988. 
5. Mantysaari M. Kuikka J, Mustonen J, Tahvanainen K, Vanninen E, Lansimies E, et al. Noninvasive detection of cardiac sympathetic nervous dysfunction in diabetic patients using [123I]metaiodobenzylguanidine. Diabetes 41 : 1069-1075, 1992. 
6. Freeman MR, Newman D, Dorian P. Barr A, Langer A. Relation of direct assessment of cardiac autonomic function with metaiodobenzylguanidine imaging to heart rate variability in diabetic mellitus. Am J Cardiol 80: 247-250, 1997. 
7. Murata K. Sumida Y, Murashima S, Matsumura K. Takeda H. Nakagawa T, et al. A novel method for the assessment of autonomic neuropathy in type 2 diabetic patients: a comparative evaluation of 123I-MIBG myocardial scintigraphy and power spectral analysis of heart rate variability. Diabet Med 13: 266-272, 1996. 
8. Task Force of European Society of Cardiology and North American Society of Pacing and Electrophysiology. Heart rate variability: standards and measurement, physiological interpretation, and clinical use. Circulation 93: 1043-1065, 1996. 
9. Ewing DJ, Campbell IW, Clarke BF. Assessment of cardiovascular effects in diabetic autonomic neuropathy and prognostic implications. Ann Intern Med 92 (Pt 2): 308-311, 1980. 
10. Oikawa N, Umetsu M, Toyota T, Goto Y, Quantitative evaluation of diabetic autonomic neuropathy by using heart rate variations: relationship between cardiac parasympathetic or sympathetic damage and clinical conditions. Tohoku J exp Med 148: 125-133, 1986. 
11. Kurata C, Wakabayashi Y, Shouda S. Okayama K, Yamamoto T, Ishikawa A, et al. Enhanced clearance of 123I-metaiodobenzylguanidine from the heart in patients with chronic renal failure on dialysis. J Nucl Med 36: 2037-2043,1995. 
12. Kurata C, Shouda S, Mikami T, Wakabayashi Y, Nakano T, Sugiyama T, et al. Comparison of 123I-metaiodobenzylguanidine kinetics with heart rate variability and plasma norepinephrine. J Nucl Cardiol 4: 515-523, 1997. 
13. Kreiner G, Wolzt M, Fasching P, Leita T, Edlmayer A, Korn A, et al. Myocardial m-[123I]iodobenzylguanidine scintigraphy for the assessment of adrenergic cardiac innervation in patients with IDDM. Comparison with cardiovascular reflex test and relationship to left ventricular function. Diabetes 44: 543-549, 1995. 
14. Nagamachi S, Jinnouchi S, Nakahara H, Flores LG 2nd, Ohnishi T, Hoshi H, et al. 123I-MIBG myocardial scintigraphy in diabetic patients: relationship to autonomic neuropathy. Nucl Med Commun 17: 621-632, 1996. 
15. Hattori N, Tamaki N, Hayashi T, Masuda I, Kudoh T, Tateno M, et al. Regional abnormality of iodine-123-MIBG in diabetic hearts. J Nucl Med 37: 1985-1990, 1996. 
16. Schnell O. Kirsch CM, Stemplinger J, Haslbeck M, Standl E. Scintigraphic evidence for cardiac sympathetic dysinnervation in long-term IDDM patients with and without ECG-based autonomic neuropathy. Diabetologia 38: 1345-1352, 1995. 
17. Dubois EA, Kam KL, Somsen GA, Boer GJ, de Bruin K, Batink HD, et al. Cardiac iodine-123 metaiodobenzylguanidine uptake in animals with diabetes mellitus and/or hypertension. Eur J Nucl Med 23: 901-908,1996. 
18. Ganguly PK, Dhalla KS, Innes IR, Beamish RE, Dhalla NS. Altered norepinephrine turnover and metabolism in diabetic cardiomyopathy. Circ Res 59: 684-693, 1986. 
19. Miyanaga H, Yoneyama S. Kamitani T, Kawasaki S, Takahashi T, Kunishige H. Clinical usefulness of 123I-metaiodobenzylguanidine myocardial scintigraphy in diabetic patients with cardiac sympathetic nerve dysfunction. Jpn Circ J 59: 599-607, 1995. 
20. Chakko S, Mulingtapang RF. Huikuri HV, Kessler KM, Materson BJ, Myerburg RJ. Alterations in heart rate variability and its circadian rhythm in hypertensive patients with left ventricular hypertrophy free of coronary artery disease. Am Heart J 126: 1364-1372, 1993. 
21. Counihan PJ, Fei L, Bashir Y, Farrell TG, Haywood GA, McKenna WJ. Assessment of heart rate variability in hypertrophic cardiomyopathy. Association with clinical and prognostic features. Circulation 88: 1682-1690, 1993. 
22. Bigger JT Jr, Fleiss JL, Steinman RC. Rolnitzky LM. Schneider WJ. Stein PK. RR variability in healthy, middleaged persons compared with patients with chronic coronary heart disease or recent acute myocardial infarction. Circulation 91 : 1936-1943, 1995.