Journal of the American College of Cardiology
Cardiac sympathetic denervationmodulates the sympathoexcitatoryresponse to acute myocardial ischemia
Author + information
- Received January 30, 2001
- Revision received October 10, 2001
- Accepted October 31, 2001
- Published online February 6, 2002.
Author Information
- Shuji Joho, MD∗,
- Hidetsugu Asanoi, MD∗,* (hidetugu{at}ms.toyama-mpu.ac.jp),
- Junya Takagawa, MD∗,
- Tomoki Kameyama, MD∗,
- Tadakazu Hirai, MD∗,
- Takashi Nozawa, MD∗,
- Katsumi Umeno, BS†,
- Masashi Shimizu, MD‡,
- Hikaru Seto, MD‡ and
- Hiroshi Inoue, MD, FACC∗
- ↵*Reprint requests and correspondence:
Dr. Hidetsugu Asanoi, Second Department of Internal Medicine, Toyama Medical and Pharmaceutical University, 2630 Sugitani, Toyama 930-0194, Japan.
Abstract
Objectives This study was designed to elucidate the influence of cardiac sympathetic denervation on the sympathoexcitatory response to acute myocardial ischemia during balloon coronary occlusion (BCO) in humans.
Background Alterations of cardiac sympathetic nerve function could modulate sympathetic reflexes originating from the ischemic area.
Methods In 23 patients with angina pectoris, we quantified the baseline cardiac sympathetic denervation of the ischemia-related area by iodine-123 metaiodobenzylguanidine (123I-MIBG), and transient changes in sympathetic activity during BCO by wavelet analysis of RR interval variability.
Results Balloon coronary occlusion resulted in a transient augmentation of low-frequency (LF: 0.04 to 0.14 Hz) spectral components of RR interval variability in 4 of 12 patients with cardiac denervation and in 8 of 11 patients without denervation (p < 0.01 by the chi-square test). Consequently, the increase in LF components was significantly less during BCO in patients with cardiac denervation (34%) than in those without denervation (273%) (interaction: p < 0.05). In seven patients with severe ischemia provoked by a fall of ≥10% in the left ventricular ejection fraction, LF components increased by 506% during BCO, regardless of the condition of cardiac denervation. In contrast, in patients with mild ischemia provoked by a fall of <10% in the ejection fraction, changes of LF components during BCO were significantly less in patients with denervation than in those without denervation (84 vs. 344%, p < 0.05).
Conclusions These findings suggest that if the provoked ischemia is not severe, cardiac sympathetic denervation could prevent ischemia-induced sympathoexcitation.
Cardiac sympathetic afferent fibers, which travel parallel to the coronary artery on the cardiac surface, are activated by experimental myocardial ischemia, resulting in sympathoexcitatory responses (1–3). These reflexes have been shown to elucidate ventricular tachyarrhythmia during myocardial ischemia in experimental settings (4). This ischemia-induced sympathoexcitation was also documented in humans (5,6). Using wavelet analysis, we previously examined dynamic alterations in hemodynamic variability during short periods of balloon coronary angioplasty and found that the transient augmentation of low-frequency (LF) components of RR interval variability was not uniform among the patients studied (7). The mechanisms for this heterogeneity could involve not only the severity of myocardial ischemia but also the denervation of sympathetic nerve by myocardial infarction (MI) or repetitive ischemic insults (8,9). Experimentally, phenol application on the myocardium or brief myocardial ischemia resulted in cardiac sympathetic denervation and attenuated the sympathoexcitatory response to myocardial ischemia (10,11). However, the relationship between cardiac sympathetic denervation and the sympathoexcitatory response to myocardial ischemia has not been examined in clinical settings.
Therefore, the purpose of the present study was to elucidate the role of cardiac sympathetic denervation in the sympathoexcitatory response to acute myocardial ischemia during balloon coronary occlusion (BCO). In the present human study, we used iodine-123 metaiodobenzylguanidine (123I-MIBG) to detect regional sympathetic denervation of the heart and wavelet analysis of RR interval variability to assess dynamic autonomic responses to transient myocardial ischemia.
Methods
Patient group
The study group consisted of 23 patients (17 men and 6 women; mean age 63 ± 9 years). All patients had effort angina and an electrocardiographically positive, symptomatic exercise stress test. Eleven patients showed electrocardiographic (ECG) evidence of a previous MI but had relatively preserved cardiac function (left ventricular ejection fraction [LVEF] 56 ± 13%). Elective coronary angioplasty was performed at 23 target lesions (11 in the left anterior descending coronary artery [LAD], 5 in the left circumflex coronary artery [LCx] and 7 in the right coronary artery [RCA]). All patients provided written, informed consent, and the protocol was approved by our institute’s Committee on Clinical Investigation. All procedures were successfully performed, and no complications were encountered.
123i-MIBG imaging
To quantify cardiac sympathetic denervation, 123I-MIBG scintigraphy (111 to 148 MBq) was performed in all patients (12,13)within one week before percutaneous coronary angioplasty. Images were obtained 4 h after administration of 123I-MIBG. Computed tomographic polar maps were divided into 24 segments and normalized by the segment with the highest isotope counts. Cardiac sympathetic denervation was determined to be present when the 123I-MIBG minimal counts in the area of the target coronary artery were <52.2%, which is a mean value −2 SD determined in 14 normal subjects (8 men and 6 women; mean age 52 ± 17 years). Twelve patients demonstrated regional cardiac denervation on the 123I-MIBG scan, where the minimal counts ranged from 50% to 19% (35 ± 11%). In the remaining 11 patients, the minimal counts ranged from 77% to 53% (64 ± 8%). Because detection of cardiac sympathetic denervation crucially depends on the maximal MIBG counts of the reference region, we excluded patients with three-vessel disease, diabetes or other neurologic diseases that influence the sympathetic innervation of the intact myocardium.
Data acquisition during coronary angioplasty
A 6F conductance catheter (Sentron, Roden, The Netherlands), together with a 2F micromanometer-tipped catheter (Millar Instruments, Houston, Texas), was inserted from the left femoral artery and placed along the long axis of the left ventricle (LV) to the apex (7,14). The volume catheter was connected to a Sigma-5/DF signal conditioner processor (Leycom Cardiodynamics Inc., Zoetermeer, The Netherlands), which used a dual excitation algorithm. Segmental volumes were summed to yield total conductance, which was calibrated by the actual volume derived from the left ventriculogram. Coronary angioplasty was performed through the right femoral sheath, whereas the ECG and LV pressure and volume were serially recorded on digital tape (RD-130TE, TEAC Corp., Tokyo, Japan) for later analyses. Patients were advised to breathe regularly and refrain from talking during BCO.
Data analysis
We excluded the ECG and LV pressure–volume data obtained during the first balloon inflation from the present analysis, because myocardial ischemia had already been provoked by the first positioning of the balloon catheter through the target lesion (15). The ECG, micromanometer pressure and conductance catheter volume were played back from the digital tape and digitized at 1,000 Hz per channel by an analog-to-digital converter (DT31-EZ, Data Translation Inc., Marlboro, Massachusetts) and stored on a hard disk memory system using a computer (Optiplex-GXMT, Dell Computer Corp., Round Rock, Texas). Hemodynamic variables derived from digital pressure and volume data were analyzed with software developed in our laboratory, as reported previously (7,14,16). The beat-to-beat RR interval was interpolated at 2 Hz to ensure equidistant sampling in each time series (17). The very-low-frequency nonstationarity (<0.01 Hz) and baseline trends were removed using a moving average function, which did not affect the higher frequencies. The beat-to-beat spectral estimation analysis was based on the discrete wavelet transform using Gabor basis wavelet (18)and 256-event data. The frequency contents were then classified as high-frequency (HF: 0.15 to 0.5 Hz) and LF (0.04 to 0.14 Hz) components. As in the previous study, augmentation of power spectra was defined to be present if the peak power of fluctuation exceeded a mean value 2 SD of the baseline power (7).
Statistical analysis
Data are expressed as the mean value ± SD. Comparisons between the two groups were performed by using the unpaired ttest or chi-square test. Two-way repeated measures analysis of variance was used to compare the responses to BCO between the two groups. Multiple comparisons were performed by using the Student-Newman-Keuls method. Analyses were performed using statistical software (SigmaStat, version 2.03, SPSS Inc., Chicago, Illinois). The level of statistical significance was set at p < 0.05.
Results
The patients with cardiac denervation (mean age 62 ± 10 years) had target lesions that were in the LAD in 8 patients, LCx in 1 and RCA in 3. The patients without cardiac denervation (mean age 63 ± 7 years) had target lesions in the LAD in 3 patients, LCx in 4 and RCA in 4. No significant differences in age or the distribution of coronary target lesions were found between the two groups. Previous MI was documented in 1 of 11 patients without denervation and in 10 of 12 patients with denervation (p < 0.01 by the chi-square test). The average number of denervated segments was 6.2 ± 3.6 (26% of the LV) in patients with denervation. There were no differences in the baseline hemodynamic variables between the 12 patients with denervation and the 11 patients without denervation (Table 1). The LVEF tended to be higher in patients without denervation than in those with denervation, but the difference did not reach statistical significance. In patients without denervation, coronary occlusion increased LV end-diastolic and end-systolic volumes by 9% (p < 0.05) and 35% (p < 0.05), respectively, resulting in a decline in LVEF by 9% (p < 0.05). The peak rate of rise in LV pressure (+dP/dt) and peak rate of fall in LV pressure (−dP/dt) fell by 10% (p < 0.05) and 12% (p < 0.05), respectively, with an increase in LV end-diastolic pressure (51%, p < 0.05). In patients with cardiac denervation, however, these detrimental effects of coronary occlusion were less than those in patients without denervation. Although LV end-systolic volume and LV minimal pressure were significantly increased during coronary occlusion, peak +dP/dt, peak −dP/dt and LV end-diastolic volume did not appreciably change in patients with cardiac denervation. Furthermore, the increase in LV end-diastolic pressure was significantly less in patients with denervation than in those without denervation (interaction: p < 0.05).
Effects of BCO on Left Ventricular Function
Figure 1illustrates the polar map of 123I-MIBG and the three-dimensional diagram of wavelet analysis in two representative patients with severe ischemia provoked by a fall of ≥10% in LVEF. The LF components began to rise at ∼60 s and reached a peak at ∼100 s after balloon inflation. Augmentation of the LF components was found in both patients with cardiac denervation and those without cardiac denervation. Figure 2shows the polar map of 123I-MIBG and the three-dimensional diagram of wavelet analysis in two representative patients with mild ischemia provoked by a fall of <10% in LVEF. Augmentation of the LF components was observed only in the patient without cardiac denervation, but not in the patient with denervation.
Polar map of iodine-123 metaiodobenzylguanidine (123I-MIBG) images and three-dimensional diagram of wavelet analysis of RR interval variability in two representative patients with severe myocardial ischemia provoked during balloon coronary occlusion (BCO). These patients had a substantial fall in left ventricular ejection fraction (≥10%) by BCO. In the patient without cardiac denervation (left), the low frequency (LF) components began to rise at ∼60 s and reached a peak at ∼100 s after balloon inflation. Similar augmentation of the LF components was found during BCO in the patient with cardiac denervation (right). Inf = balloon inflation; Def = balloon deflation.
Polar map of 123I-MIBG images and three-dimensional diagram of wavelet analysis of RR interval variability in two representative patients with mild myocardial ischemia provoked during BCO. These patients had a fall in LVEF (<10%) by BCO. The LF components were augmented in the patient without cardiac denervation (left), whereas no change in the spectral components was found in the patient with cardiac denervation (right). Abbreviations as in Figure 1.
Baseline LF and HF components of RR interval variability were comparable between the two groups (Table 2). During coronary occlusion, the LF components were augmented in 8 (73%) of 11 patients without denervation and in only 4 (33%) of 12 patients with denervation. The prevalence of augmentation of LF components was significantly different between the two groups (p < 0.05), and the average increase in LF components was significantly greater in patients without denervation than in patients with denervation (interaction: p < 0.05). The HF components did not change appreciably in either group during BCO. Consequently, the ratio of LF to HF components increased significantly in patients without denervation (p < 0.05), although it remained unchanged in patients with denervation. However, these changes did not reach statistical significance between the two groups.
Effects of BCO on Power Spectral Components of RR Intervals Determined With Wavelet Analysis
To examine the influence of the severity of ischemia and cardiac denervation on the autonomic nervous responses to myocardial ischemia, patients were classified into two groups. In seven patients, severe ischemia was provoked during BCO, as evidenced by a fall of ≥10% in LVEF. In the remaining 16 patients, the provoked ischemia was considered to be mild because LVEF fell by <10%. The seven patients who had severe ischemia during BCO showed an increase of 506 ± 366% (p < 0.05) in the LF components, regardless of the condition of cardiac sympathetic innervation (Fig. 3). In contrast, the patients with mild ischemia during BCO had variable responses, depending on the condition of sympathetic innervation. That is, the increase in LF components was smaller in patients with cardiac denervation than in those without denervation (84 ± 72% vs. 344 ± 317%, p < 0.05), and cardiac function was similarly depressed.
Changes in LVEF by BCO (ordinate) were plotted against the minimal 123I-MIBG counts of the artery of interest in all lesions (abscissa). Six of seven patients with a substantial fall in LVEF (≥10%) had augmentation of the LF components, regardless of the conditions of cardiac sympathetic innervation. In contrast, in 16 patients with mild ischemia provoked by a fall in LVEF of <10%, the prevalence of augmentation of LF components was significantly less in lesions with cardiac denervation (minimal count of MIBG <52.2%) than in those without cardiac denervation. Solid squares= patients with augmentation of LF components; open circles= patients without augmentation of LF components. Abbreviations as in Figure 1.
Discussion
In the present study, we quantified the baseline sympathetic innervation by 123I-MIBG scintigraphy, and the transient, autonomic changes during balloon coronary angioplasty by time-varying spectral analysis of RR interval variability. The major findings are as follows: 1) patients with cardiac denervation showed less increase in LF components during BCO, compared with those without denervation; 2) in patients with severe myocardial ischemia that was provoked, the LF components of RR interval variability were increased, independent of cardiac sympathetic denervation; and 3) in patients with mild ischemia that was provoked, the changes in LF components depended on sympathetic innervation, however. The increase in LF components was significantly less in patients with cardiac denervation than in those without denervation. These findings suggest that cardiac sympathetic denervation could attenuate the sympathoexcitatory response to myocardial ischemia, if the myocardial ischemia provoked is not too severe. These findings may have implications for the life-threatening arrhythmias that often accompany acute myocardial ischemia.
Effects of cardiac denervation on sympathoexcitation
The extent of cardiac innervation is assumed to play a crucial role in the sympathetic responses to mechanical and chemical stimuli produced during myocardial ischemia. However, no clinical studies have examined the relationship between cardiac neural innervation and the autonomic response to transient myocardial ischemia. This is partly related to a lack of an appropriate method to detect transient and dynamic changes in autonomic function during myocardial ischemia in humans. In this regard, the wavelet analysis employed in the present study allows a valid and serial evaluation of transient changes in spectral components. We have previously reported the higher sensitivity of wavelet analysis in detecting spectral changes during BCO, as compared with the traditional spectral analysis (7). In the present study, wavelet analysis clearly demonstrated a predominant influence of myocardial ischemia on augmentation of LF components in patients with severe ischemia provoked, as well as a significant role of cardiac sympathetic innervation in the genesis of augmentation of LF components in those with mild ischemia provoked. Acute myocardial ischemia activates cardiac sympathetic afferent fibers and could elicit a sympatho-sympathetic reflex, with resultant augmentation of myocardial oxygen consumption (19), leading to a vicious cycle of exaggerating myocardial ischemia and a propensity toward the onset of ventricular arrhythmias (4). Under these conditions, myocardial denervation preconditioned already by previous MI or repetitive myocardial ischemia would attenuate the reflex sympathoexcitation (8,11)and help to protect the myocardium from further augmentation of ischemia. There is already clinical evidence for the concept that cardiac sympathetic denervation can reduce the incidence of life-threatening arrhythmias (20). In a randomized clinical trial in high-risk patients who had an MI, left cardiac sympathetic denervation reduced the two-year incidence of sudden death from 21% to 3.5%. When myocardial ischemia is severe, baroreceptor reflexes could become dominant, resulting in sympathetic activation, independent of the degree of sympathetic denervation. In addition, the increase in LV filling pressure caused by severe myocardial ischemia could elicit a sympatho-sympathetic reflex through activation of sympathetic efferent fibers mechanically (21).
Study limitations
The heterogeneity of the patients could affect the results of the present study. Most of the patients with cardiac sympathetic denervation had a previous MI, which might provoke a less severe myocardial ischemia during BCO. To exclude the different effect of the grade of myocardial ischemia, we compared the LF components among patients with similar ischemic insults, as reflected by a deterioration of LVEF. If the ischemia provoked was not severe, cardiac denervation attenuated the sympathoexcitatory response to myocardial ischemia. Although statistical difference was not found in the distribution of the affected vessels between the two groups, a higher proportion of target lesions in the LAD in patients with cardiac denervation might influence the result. Generally, reflex sympathoexcitation is stronger with anterior wall myocardial ischemia (22), as shown by augmentation of LF components observed in all patients without denervation. Nevertheless, this reflex was clearly attenuated when the anterior wall had already been denervated in patients with mild myocardial ischemia. Therefore, the heterogeneity of target lesions probably had little effect on the present conclusions.
The criteria of cardiac sympathetic denervation that we used were based on the variation of 123I-MIBG counts measured in normal subjects. However, it is unknown whether the cut-off level of 52% could faithfully reflect the actual denervated condition of the heart. Nonetheless, the cardiac denervation assessed by these criteria correlated well with the defect of the sympathoexcitatory response to acute myocardial ischemia. This finding could support the feasibility of the denervation criteria employed in the present study.
Conclusions
The wavelet analysis of RR interval variability during brief periods of BCO clearly showed a dynamic profile of the sympathoexcitatory response in humans. Not only the severity of myocardial ischemia, but also the cardiac sympathetic denervation modified the sympathoexcitatory response to myocardial ischemia. Further studies are warranted to examine whether cardiac denervation could prevent the ischemia-related serious arrhythmias in clinical settings, through attenuation of the sympatho-sympathetic cardiac reflex.
Footnotes
☆ This study was supported by a Grant-in Aid for General Scientific Research (no. 09670706) from the Ministry of Education, Science and Culture of Japan.
- Abbreviations
- +dP/dt
- peak rate of rise in left ventricular pressure
- −dP/dt
- peak rate of fall in left ventricular pressure
- BCO
- balloon coronary occlusion
- ECG
- electrocardiogram or electrocardiographic
- HF
- high frequency
- LAD
- left anterior descending coronary artery
- LCx
- left circumflex coronary artery
- LF
- low frequency
- LV
- left ventricle or ventricular
- LVEF
- left ventricular ejection fraction
- MI
- myocardial infarction
- 123I-MIBG
- iodine-123 metaiodobenzylguanidine
- RCA
- right coronary artery
- Received January 30, 2001.
- Revision received October 10, 2001.
- Accepted October 31, 2001.
- American College of Cardiology
References
- ↵
- Zipes D.P
- Thames M.D,
- Kinugawa T,
- Dibner-Dunlap M.E
- ↵
- Schwartz P.J,
- Foreman R.D,
- Stone H.L,
- Brown A.M
- ↵
- ↵
- Joho S,
- Asanoi H,
- Remah H.A,
- et al.
- ↵
- Barber M.J,
- Mueller T.M,
- Davies B.G,
- et al.
- Inoue H,
- Skale B.T,
- Zipes D.P
- ↵
- Barber M.J,
- Mueller T.M,
- Davies B.G,
- et al.
- ↵
- Dae M.W,
- O’Connell J.W,
- Botvinick E.H,
- et al.
- Mitrani R.D,
- Klein L.S,
- Miles W.M,
- et al.
- Remah H.A,
- Asanoi H,
- Joho S,
- et al.
- ↵
- Perry R.A,
- Seth A,
- Hunt A,
- et al.
- Asanoi H,
- Ishizaka S,
- Kameyama T,
- et al.
- ↵
- Ishise H,
- Asanoi H,
- Ishizaka S,
- et al.
- ↵
- ↵
- Malliani A,
- Schwartz P.J,
- Zanchetti A
- ↵
- ↵
- Kurz M.A,
- Wead W.B,
- Roberts A.M,
- et al.
- ↵
- Webb S.W,
- Adgey A.A.J,
- Pantridge J.F