- Email: [email protected]
Sustained augmentation of parasympathetic tone with angiotensin-converting enzyme inhibition in patients with congestive heart failure
..Sustained Augmenta Converting Enzyme Heart Failure PHILIP
F
BINKLEY,
ENRICO
NUNZIATA.
ROBERT
i. CODY.
Cohnbus.
Ohio
MD,
FACC.
MSBME, MD,
CARRIE
PATRLCLA
MD.
RANDALL
CNMT,
URL
C. STARLIhG. V. LEIER.
MD, FAX, MD.
FACC.
FACC
Objectives. The objeclive of this inve@aiion the changes in parasym:~thetic
J. HAAS, A. HAnON.
lone associated
was to evaluate with long-term
aogioteasin-converting euzyme inhibitor therapy iu palients with CWgcStiW heart failure. Backgwmd, Angioieasin-coevsrting enzyme inhibitors pnr vide hemudynamic aad symptomatic benefit and are associated
with impmved survive! in patients with conprtive heart failure. Angiotensin II, whusepmduction is ultimatety inhibited by these agents, exerts significant regulatory iaflueace on a variety of target organs induding the central and peripheral nervous sy” tems. Accwdiagty, it wautd he anticipated that angiotearinmnvertiu# enzyme iuhibitcux would siguiticentty alter the autoo&e lmhalan~c characteristic of patients with congestive heart failure and that Ihb lniiuence over neural mechanirms al cardtovascutar cuntrtd may s@ificantly coatribute to the hetttudynatnic heacflt and improved survival assuciated with angioteusinconverting enzyme inhibitor therapy. Methods. In the current bwstipation, changes in autrmomtc Ione &ted with luug-term dministmtion of an augiotensinconvertiog etwyme inhibitor were measured using spectral auatysis of heart me variability ia 13 palients with eeugestive heart faiture who WM enmlled ia a double-blind randomized placebw~ cwtrotled trial of the angiotensin-convertiagenzyme inhibitor
Aneiotensin-convertine enzvme inhibition has heen shown to be an effective strategy fdr the symptomatic management and improved survival of patients with congestive heart failure (l-8). Agents that inhibit angiotensin-converting enzyme have been shown to possess a diverse army of prop-
21. 1992, a&ted Avgvrt 24. 1992. m: Philip F. Binkky. MD. Ohio Srale Umversky Hospital, Rwm 610 Means Hall, 1654 Upham Drive. Columbus. Ohm 43210.
zorenoptil. Both placebo and treatment gsoslg were batsned ilt baseline study in terms uf fuactional rlass. ventrictiar performance and rutooomic toa. Resub. After II weeks al therapy with placebo, there was MI change in total heart rate variability, pua~mrpnthet~lty gowerncd h&b frequevxy bean rate variability 01 sympatbeticatty influenced Ion. frequency IwW rate variability. ln contrast, therapy with zofeoupril was ass&ted with a 50% itxrease in total heart rate vatiabillty (p = 0.09) and a rignificaat &I = 0.03) twofold increase in high frequency heart mte vnrLbility, tndicsting a signitirant augmentation of parasym+hetic tone. Conchsiom.
ment
These results dtmuuslrate
ihntl@@+‘m treat-
of patients having congestive heart Failure with aa aagiotensiusonvertiag eazyate inhibitor is a.wcSdcd with u IS toratton
of sutoncmtc
sustained agmentatiin
hatance,
shtch
derives
of purasympnlhctii
in part from P
tone. Such nugmen-
tation of vagal tone is kuowt to be proteclive Pgaimt mnlign;gnant ventricular nrrhy~hmtpr in patter& with &&em& heart d&w and therefore may have similar hen& in the setting of wWi&u failure. thus contrihuling to the improved survival a.wcilted with aogiutenshwuuverting enzyme inhibitor therapy in @ien& with
congestive heart Mute. (.I Am Cull Cwdioli993;21.~55%51)
erties that may contribute to varying degrees to the improved hcmodynamic profile, functional status and survival associated with these compounds (9-15). Given evidence derived from animal models that demonstrates that angiotensin II produces a central suppression of vagal outflow as well as an increase in sympathetic drive (16-24). it would bc anticipated that angiotensin-convening enzyme inhibitors would significantly alter the autonomic imbalance that is characteristic of patients with congestive heart failure (2533) and that may be an imporlanl determinant of disease progression and mortality in this patient group (34-W Previous studies (13,15.37) with angiotensin-converting cnzymc inhibitors have suggested that they may significantly reduce Ihe incnased sympathetic drive of patients with congestive heart failure as evidenced by a reduction of plasma norepmephtine and at least partial restoration of
JACCVol. 21, No. 3 March I. 1993:655-61
abnormal reflex responses in these patients. However, the potentially important role of increased parasympathetic tone resulting from angiotensin-converting enzyme inhibition could not previously be evaluated. Spectral analysis of heart rate variability permits a noninvasive in viva measure of parasympathetic tone that may be applied in a serial fashion to define the manner in which angiotensin-converting enzyme inhibition modulates the w gal component of iie autonomic nervous system in the setting of venlricuku falfure (33,38-47). The current inves. t&ion used a randomized double-blind placebo-controlled design to test the hypothesis that long-term therapy for congestive heart failure with an angiotensin-converting enzyme inhibitor will augment parasymparhetic tone, which has previously been shown to be significantly attenuated in patients with ventricular failure (32.33).
Methods Study group. Thirteen patients with congestive heart failure who were referred to the Ohio State University Heart Failure Clinic and enrolled in a randomized double-blind plti:ebo.controlled trial of the angiotensin-converting enzyme inhibitor zofenopril constituted the study group. All patients had symptomatic congestive heart failure due to left ventricular sgstolic dysfunction, which was confirmed by echocardiographically derived percent fmctional shortening of the minor axis of the left ventricle of ~25%. Ventricular systolic dysfunction was due to idiopathic dilated cardiomyopathy in 10 patients and to significant epicardial coronary artery stenosis in 3. None of these patients had uncontrolled angina or had experienced a myocardial infarction in the preceding 6 months. Eight patients (four men and four women) aged 39 to 65 years (mean L SD 46.8 2 8.3) were randomized to treatment with zofenopril and five patients (two men and three women) aged 47 to 67 years (mean 56 i: 8.5) were randomized IO placebo. The mean age of the two groups did nor significantly differ. The percent fractional shortening of the left ventricle for all I3 patients was 14 i 4% and did not differ between the Iwo groups (I6 + 4% for the placebo group and I3 t 4% for the treatment group). The two groups were balanced in terms of symptomatic status (with mean New York Heart Association functional class 2.4 + 0.4 for the placebo group and 2.4 c 0.4 for the treatment group). All patients provided informed consent to the investigation in accordance with the Human Subjects Review Committee of the Ohio State University, which reviewed and approved this protocol. Study design. All patients were enrolled in a 2-week placebo run-in period during which time any vasodilator medications that may have been prescribed wcrc wilhdrawn and doses of diurelic drugs and digoxin were adjusted to maintain optimal control of congestive heart failure symptoms with use of these agents alone. All patients received a single daily placebo capsule during this run-in period. At the end of this period, patients were treated with digitalis in
II
the form of digoxin. with doses ranging from 0.125 to 0.25 mglday, and 2 were treated with digitoxin, 0.1 rnd day. Eight patients were treated with furoseimide, 40 to 160 mgiday. One patient was treated with both fnrosemide and a hydmchlorothiazide-triamterene combination. One oatient was treated with a hvdmchlorothiazide-triamterene combination alone, two pat& were treated with bumetsnide, 2 to 3 md&y, and one patient did not require diuretic dtugs. At the conclusion of this 2-week period, patients underwent recording of ‘he electrocardiogram (ECG) for assessment of heart rate variability. After this baseline measurement was obtained, patients randomized tn rhe control group received daily placebo in addition to stable doses of a digitalis preparation and diuretic drugs as established in the run-in period. Patients randomized to angiotensinconvening enzyme inhibitor therapy received a total daily dose of I5 mg of nofenopril in addition to the digitalis preparation and diuretic drug dose established in the run-in phase. Doses of digitalis preparations and diuretic drugs were maintained constant throughout the entire K&week treatment phase in both groups. At the conclusion of this 12.week period, the ECG was again recorded for analysis of heart rate variability. Measurement of heart rate vtiabiliIy. Recordings of the ECG at baseline study and after the 12-week treatment period wore obtained in all patients in the supine posture after a 20&n period of equilibration. All examinations were performed between 9 and II AM and all medications were withheld on that day until the recordings were obtained. Electrocardiographic electrodes for recording standard limb leads were attached and a lead demonstrating predominant positive or negative QkS de&lions (facilitating QRS recognition by the signal-processing algorithm) was selected for recording and analysis. The selected ECG signal was recorded for a minimum of 4 min with use of a RACAL V STORE FM tape recorder at a speed of 3.7%in. (9.53 cm%. If somatic artifact or frequent ectopic beats were noted at any time, the recording was halted and reinitiated after such interference had resolved. The ECG signal was subsequently analyzed for generation of the power density spectrum of heart rate variabiliiy. Spectral analysis of benri rate variability. Spectral analysis of heart rate variability was performed as previously reported by our laboratory (33) by an observer who had no knowledge of the assignment of the subject lo placebo or zofenopril treatment. The tape-recorded ECG signal was preprocessed with use of an antialiasing filter and digitized by means of a Itbit analog/digital converter board (Metrobyte Co.) installed in au IBM/AT computer at a sampling rate of 512 Hz. Once digitized. the ECG signal was passed through a digital band-pass filter having a central frequency of 85 Hz. A dynamic user-interactive threshold technique was applied to the filtered signal to detect the R waves and compute the RR interval sequence. Subsequently, the RR interval sequences were passed through a statistical filter to
eliminate rapid transitions due 10 signal detection faults or ectopic beats. Data points ourside the 95% confidence inferval
of the previous
IO points
were
etiminatrd
and
a point
by linear interpolation of the preceding and following points was substituted 133). A plc! of the instantaneous heart rate versus rime (heart rate variability signal) war generated from the RR interval ;squence (48). The heart rate versus time series was then passed through a Farzen window and the power spectrum density of heart rate variability WBF generated with use of the modified pcriodogran method of derived
Welch
(49).
This
method
is based
on the multiple
14
2.4
3”
35
3-381 ACE Inhibition
computa-
tion and average of the fast Fourier tmnsform of overlappLri data segments (49). With this method. the variance of rhe estimated power spectral density is reduced by 1 fxtor proportional to the number of data segments used. The power spectral density was then normalized so that the total power was equal to the signal mean square. A plot of the values afthe power spectrum density agamst frequency WB~ then generated. An area under the curve method KBS used to quantify the power within specified frequencies. Specifically, the total area, the area under the low frequency (0.02 IO 0.1 Hz) region of the curve. which is influenced bs sympathetic nerve activity, and the high frequency PO.1 Hz) region, which reflects parasympathetic tone. were calculated (33,39.40.42,46,47). Thus, the magnil~+ of parasympathetic tone before and after the treatment period in the two groups could be quantihed by this system of analysis. Stalistical analysis. All data are presented as mean value A SD. A Student I tesf for unpaired data was used to compare baseline variables of ventricular function, functional class. age and autonomic tone in the treatment and placebo groups. Analysis of variance for repeated measures was used to evaluate intergroup and intrdgroup changes in variables of autonomic tone. functional class. heart rate. respiratory rate and blood pressure over the treatment period. Statistical significance was defined at the p < O.OS
I
level.
Results Spectral analysis of heart rate variability at baseline study. The power density spectrum of heart rate variability for both the placebo group and the group treated with the angiotensin-convening enzyme inhibitor zofenopril (Fig. and 2) was typical of that previously reported (33) in congestive heart failure with little variability noted in the parasympathetically mediated high frequency region (20.1 Hz) and the preponderance of variability noted in the sympathetically mediated low frequency region (~0. Hz). Furthermore, rhe baseline characteristics of the autonomic profile of the placebo and treatment groups as reflected by the power density spectra did not differ. Specifically. no sit#icast diBrences between groups were noted in the low (placebo 1.4 x lo-’ + 1.0 x IO-’ Ibeats/minl’; zofenopril 1.0 X lO.3 2 1.0 x 10m3 [beatslmin]‘) or high (placebo 0.5 X IO-’ 2 0.J x lo-’ Lbeatsirmn]‘; zoreenoprtl U.6 x IV” t
I
i
20
Frequency (Hz) Figure I. Power spectral delmity of heart mre variability recorded in !J pz!ien! 1~1th covzesrive heart failure at baseline study (top) and aller I? weeks of therapy with the angiotensin-convening enzyme (ACE) inhibitor zofenopril (bottom). Both we platted on the s8me scale. The baseline power density spectrum is typical of that previously reported for patients with congestive heart ftilure with little heart rdte variability @ower) in the parasympathetically mediated frequency band 20.1 Hz consistentwith il marked attenuation of parasympathetic tone. After therapy with the angiotensinconverting enzyme inhibitor. there is a marked augmentation or heart rate vanabilily in the high frequency region. indicating an augmentation of parasympadwic tone. In this example patient. the %ympathc!ically influenced low frequency band is also reduced in megnitude. Fnggestinga decrease in sympalktic drive (see Discussion). BPM’ = Ibeatslmir#.
0.6 x IO-’ [beatsimin]‘) frequency bands of heart rate variability. Accordingly, total heart rate variability as quantified by the total arev under the power density spectrum of heart rate variability did not differ between groups (placebo 2.4 x IO-’ 2 I.6 x IO-’ [beatslminj’; zofenopril 2.2 x IO“ 2 1.4 x lo-’ [beatslmin]?. Speml
analysk or heal l-ate vatibiiity aRETtre%tment.
Total heart rate variability did not change in the placebo control group. with the total area bounded hy the power density $prLtrurn measuring 2.7 x IO-’ f I.3 x IO-’ (beats/min? ?! !hc end of the 3-month period compared with 2.4 X IO-’ + 1.6 K IO-’ lbeatslmin)” measured at baseline (Fig. ?, In contrast, a significant t=nd (p = 0.09) toward an increase in total area was noted in the angiotensinconverting enzyme inhibitor treatment group, with a 50% increase in area from the baseline value of 2.2 x low9 k 1.4 x IO-’ Enmination
totat 0.03) high (Fig.
to 3.4 x IO-’
r
1.4 X 10m3 (beatslmin)‘(Fig.
3).
of the specific frequency bands comprising heart rate varlahilitv demonstrated a sienifcant (u = withm-group incrke in parasympatheti~ally mediated frequency heart rate variability in the treatment group 4) as shown by a twofold increase in the high frequency
2.33
Placebo Easellne
1.75
0.00 1. 02 2.33
08
.I4
.2o
.24
.3o
.35 ACE INHIBlTtON
Placebo - 12 Weeks
1.75
PLACEBO
Figure 4. Mean and SD of parasympathetic&
1.17 0.58 o.oaI4-.o2 .DB .14
.2o
.24
.3o
.35
Frequency (Hz) FigureZ. Power spectral density ofheart rate variability recorded in a patient with congestive heart failure at baseline (lop) and after 12 weeks of therapy with placebo (baltom) plotted on the same SC& The spectra have virWally identical config~n&ns. There is no change in the magnitude of variability in the parasympathetically mediated region >O.l Hz between the baseline and treatment periods. BPM* = (beatsiminl’.
area from 0.6 x IO-’ t 0.6 x IO-’ IO 1.3 x IO-’ ? 0.9 x IV3 (beats/mitt:‘. However, no change in high frequency area was noted i” the placebo control group. with an area of 0.4 X 10e3 -C 0.3 x IO~‘(beats/min)* notedafter 12 weeks as compared with the baseline value of 0.5 x IO-’ + 0.5 x 10m3 (beatsimin)*. No change in the low frequency area was noted in either group. A numeric but nonsignificant increase in low
tone
frequency area was observed in the placebo group, with a value of 1.9 x IO-’ + 1.2 x Km3 (beatslmin)* recorded at the end of 12 weeks. The low frequency area in the zofenopril treatmentgroup did not change from the baseline value, with an area of I.0 x IO-’ i. 1.0 x IO-’ (beats/min)* measured at the end of the treatment period. Accordingly, the proportion of high frequency heart rate variability increased in the zofemxxil treatment arouD as reflected by the ratio of hiab freque& to total heart ‘rate variability; which ittcreasad from 22 r 15% to 34 + 16% (p = 0.1) but did not change in the placebo group at baseline vs. 15 + at the end of the treatment period). These changes in the variables of hean rare variability are exemplified in the representative power density spectra demonstrates findings displayed in Figures and 2. Figure in a patient in the zofenopril treatment group. The baseline spectrum (Fig. I, top) is characteristic of oatients with
(19t 9% 1
Figure 3. Mean and SD of the tolal area under the power spectral density of heart rate variability in patients receiving an angiotensinconverting enzyme inhibitor (ACE INHIBITION) or placebo. A
50% increase (p = 0.09) in total heart rate variability (power) is noted after 12 weeks of treatment (hatched bars) with the angioten
sin-convortiq enzyme inhibitor. In contrast, there is no change in
total heart rale variability in the placebo contml group. bpm2 = (beatsimin)‘.
ACE tNHlBlTtON
PLACEBO
mtdiatcd high
frequency hean rare variability in patients refeiviagan angiotensinor placebo. A converting enzyme inhibitor [ACE INHIBtTION) significant (p = 0.03) twofold increase in high frequency area is noted after I2 weeks of treatment (hatched bars) with the angio’ensin-convertina enzyme inhibitor. indicatina an increase in oarasvmpathetic tone. No-change in high frequ&y area is notid in-the p!Zebo group, indicating conrinucd atlenuation of parasympathelic typical of palients with congestive heart failure. bpm2 = WatslminY.
II%
I
witi an r&kHz) heart rate variability consistent tion of parasympathetic tone (33,461. After the treatment period, there are striking differences in the power density spectrum, with a marked increase in variability in the band >O.i Hz. In this particular example, there is also a reduction in the sympathetically intIuenced frequency band in the region ~0. I Hz. In contrast, the power density spectrum for the patient in the placebo group (Fig. 2) shows a consistent configuraticn, with no change between baseline and the I’Lwerk treatment period. Clhtbl response. None of the patients in either the placebo group or the group treated with the angiotensinconvening enzyme inhibhor zofenopril required in-hospital treatment of decompensated consstive heart failure. As outlined in the study protocol, all patients were successfully managed with stable doses of diuretic drugs and a digitalis preparation as established in the run-in period. Mean sys-
temic blood pressure did not change from the baseline value of 87 t 9 mm Hg in the placebo group and 88 + 10 mm Hg in the zofenopril treatment group. Mean heart rake did cot change in the placebo group (81 t tcatsimin al baseline
intravenous infusion of angiotensin II. Considering the elevation of angictensin 11 in patiems with congestive heart failure. the characteristic blunting of ;he arterial baroreflex rcsponsc sod the attenuation of parasympathetic tone ob=
and 81 c 6 beatsimin at the end of the placebo period). A
perved in thete patients may derive in part from angiotennin
numeric decrease in mean heart rate thal did nor attain
ing to 85 L IO beatslmin. Respiratory rate measi;red at rest immediately before the ECG recording petiod did not change h iithergroup, measuring20 2 2 breathslmin at baseline and
It-mediated central suppression of vagal tone. Accordingly, a reduction of angiotensin II concentration. as occurs with angiorensin-converting enzyme inhibition, would reverse central suppression of vagal outflow, resulting in an augmentation of parasympathetic tone. The enhanced parasympathetic tone with angioteedn.convcrting enzyme jnhibition
21 i 2 breath&in
observed in the torrent
II
statistical significance was noted in the zofenoprii treatment group with the baseline value of 89 T IO beats/mm decreas-
after 12 weeks in the zofenopril rreatmenr
group. In the placebo group, respiratory breathslmin
rate ~3s ?I * 4
at baseline and 19 + 2 brearhslmin
weeks (bath p = NS).
New York
afler
I2
Heart Association func-
tional class significantly (p = 0.007) improved in the zofeno-
pril treatment group, decreasing from 2.4 + C.4 to I .8 ? 0.8. No change in functional class was noted in the placebo @‘oUp, with a mean vatue of 2.3 z 0.4 measored at the end of the study as compared with 2.4 * 0.4 recorded at baseline.
investigation
woilld therefore
he
predicted on the basis of these fundamental mech:aisns. The current investigation cannot verify that the observed augmentation
of parasympathetic
tone does indeed derive
from the angiotensin II-related mechanisms discussed. It is conceivable that some propeny of the anpiolensinconverting cnryme inhibitor unrelated to aogiotensin IImedialcd mechanisms may have produced the notedchanges in parasympathetic tone. Accordingly. it is not possible from the current dats to verify whether this vagotonic influence is
Discussion The current investigation demonstrate, that the autonomic imbalance in patients with congestive heart failure can be significantly altered by long-term rhenpy with an angiotensin-converting enzyme inhibitor. Autonomic tone was assessed with the technique of spectral analysis ofheart rate variability,
which has been shown to provide a nonin-
vasive in viva assessment of autonomic activity given that
I Hz)
variations in heart rate are exclusive& under the control of the parasympathetic nervous system and that low frequency heart rate variability (co.1 is influenced to a significant extent by sympathetic drive (39,40,47). Speciiically, with use of a randomized double-blind placebo-controlled stttdy design, it was shown high frequency 00.
Hz)
that 12 weeks of therapy with the aogiotensin-converting enzyme inhibitor zofeoopril produced an increase in total
(Fig. 3). This increase was due to a significant twofold increase in high frequency heart rate variability (Fig. 4), which is specifically a direct manifestation of a significant augmentation of parasympathetic tone. These findings contrast markedly with those in the placebo control group, which showed no change in variables of heart rate variability
autonomic tone after the la-week
treatment period (Fig. 3
a class effect of all angiotensin-converting enzyme inhibitors or is specific to zofenopril. However. if interruption of angiorensin [I-mediated suppression of vagal tone is indeed the operarive mechanism for the increased parasympathetic tone noted in lhis investigation. it would h predicted that any angiotensin-convening enzyme inhibitor tha1 reduces central nervous system concentntions of angiotenain II would produce a similar augmentation
of parasympathetic
heart failure. Clinical implications. Heart rate variability has been shown to be perhaps the most powerful independent predictorof malignant ventricular arrhythmias and sudden death in patients with ischemic heart disease (50-56). These clinical observations confam basic investigations in animal models thal have shown that the absence ofvagal tone as manifested by diminished high frequency heart rate variability is associated with a reduction of ventricular fibrillation threshold (53) and that restoration of vagai tone reduces this propensity to ventricular arrhythmias (57). if these principles pertain to patients with dilated cardiomyopathy, then the attenuaibn of parasympathetic tone thal is evident by the signiticant decrease in high frequency heart rate variability tone when administered
to patients with congestive
may contribute to the pmarrhythmic
environment that pre-
disposes this patient group to sudden cardiac death. There-
and 4). Proposed mechanisms for vagolonic effectsof angidensinconverting enzyme inhibition in coageslive heart f&w. The mechanism for the observed effects of angiotensinconverting enzyme inhibition are speculative at this time but are suggested by prior observations regarding the influence of angiotensin 11 on autonomic tone (16-24). Angiotensin
fore. it may be speculated that the augmentation of parasympathetic tone reported in this investigation may contribute to the improved survival associated with angioteasinconverting enzyme inhibitors in the setting of congestive
mediates a central suppression ofvagal o&low
strates that the technique of spectral analysis of heart rate variability may he successfully used for the prospective
when admin-
istered in a variety of animal models (16.23). Lumbers et al. (16) demonstrated a significant reduction of parasympathetic
activity recorded directly from vagal e&rent
fibers during
heart faiture (6,8). Methadolagic nmsiderations.
This investigation
demon-
serial assessment of changes in autonomic tone associated with long-term treatment of congestive heart Failure. The
660
RINKLEY ET AL. AuGMENT*TwN OF FARASYMPATHETIC
JACC Vol. 21. No. 3 hhrcb 1. lYw655-61
TONE
consistency of the measures obtained in the placebo contml group at baseline study and after a 12-week period in terms of total heart rate variability and its low and high frequency components demonstrates the reproducibility of this technique even over an extended period of time (Fig. 2 to 4). A consistent change in low frequency heart rate variability was not observed in the angiotensin-converting enzyme inhibitor trcatmeni group and, as a consequence, the mean value for this frequency band did not change after the treatment period. Because angiotensin II has been shown in some models to stimulate central sympathetic outflow (17), it may be anticipated that angiotensin-converting enzyme inhibition would reduce sympathetic tone and thus the low frequency band of hearl rate vaizbility. However. low frequency heart rate variability is not devoid of some degree of influence by the parasympathetic nervous system (40,47). Therefore, an increase in parasympathetic tone may result in some increase in low ficqtiency heart rate variability that would offset decreases in this frequency band caused by reduced sympathetic drive. The net result would be no change in the low frequency band as observed in the current investigation. Respiratory rate can influence the magnitude and frequency distribution of heart rate variability and thus produce changes in the frequency bands of interest independent of any primary change in autonomic tone ($59). However, in the current investigation, there was no change in respiratory rate in either group after the treatment period, a finding that further substantiates the conclusion that the observed changes in high frequency heart rate variability were indeed due to an augmentation of parasympathetic tone in the zofenopril treatment group. Conclusions. In this randomized double-blind placebocontrolled trial, a significant augmentation of parasympathetic tone was associated with long-term administration of the angiotensin-converting enzyme inhibitor wfenapril in patients with congestive heart failure. Whether in patients with ventricular failure such augmentation of vagal tone confers a protective effect against malignant ventricular arrhythmias and sudden death, as is seen in patients with ischemia, requires further study. Likewise, whether the changes in autonomic tone observed in this report are a nonspecific response to vasodilation or are specific to angiotensin-converting enzyme inhibition provides the foundation for future investigations. We ‘bank Trichia Zavilla for expen nssislance in Ihe preparation of the manl!script. .-
References I. Capbwil Multicenter Research Group. A placebcwntmlled lrial of caploprjl irarefractory chmnic congestive heart failure. I Am Colt Cardiat 1983;2:755-63. 2. Crcager MA, Massie BM. Faxon DP. et al. Acue and long-term effeleclsnf enaiapril un the cardmvascular response lo exercise and exercise loler-
3. Levine TB. Olivsri Ml. Garberg V. Sbarkey SW. Cohn JN. Hemodynamic and clinical response Lo enalapril. P longs&g convening-enzyme InhIbitor. in patients with congeslive hcarl fadurc. Circulalion 1984~69. 548-53. 4. Packer ht, Medina N. Yushak M. Comparative hrmodynamz and Clinical effects or long-term tratment with prazosin and captopnl for severe chronic congestive heart failure secondary to coronary a&y disease or idiopathic ddafed cardiomyopathy. Am 5 Cardiol 1986:5?: t32&7. 5. Cody RI. Cowl AB. Schacr CL. La@ JH. Evaluatmn of a long-acting converlinp enzyme inhibitor lenalapril) for Ihe freatmenl of chronic congeslive heart failure. I Am Co8 Cardiol 1983:1:1154-9. 6. The CONSENSUSTrial Study Oruup. EDctsofenalaprit on monalhy in severe congcslive heari f&x: rcrukr of Ihe Cooperative N*dh Scandinavian Enalapril Survival Sludy ICONSENSUSL N Engl J Mcd 1987: 316:1429-x 7. TheSOLVD lnvesli@ors. EIlecrof enalapril on survival in patiems whh reduced left ventricular eJection fractions and conge~Iwe hrart fadurc. N En@ J Mcd 1991:325:193-302. 8. fohn IN, Johnson G. tieschc S. et al. A compansan of enalapnl with hydratazine-irorbide dmrlrale in the rramentofchmn~c congerrive hean failure. N Engl J bled 1991:325:30~10. 9. tiollr J. Busse R, Sommer 0,Basrenge E. Dilatation ofepicardial arteries in consciaus dags induced by angiotenrin-convening enzyme Inhibition with enalaarilat. J Cardiovasc F’harmacal 1987:9:348-25. IO. Lindpaintner K. Ganten D. The cardiac reniwangiotensin ayeem: w rwraisal of Present extwimenlal and clinical evidence. Circ Res 1991: 68:%5-21. It. Pfeffer JM. PIa& MA. Mirsky I. Braunwald E. Regression of ten venrricular hypcrtrophy andprevention of kfr venlricular dy&r&n by caploprit in the sponlaneourly hypatensive rat. Proc Nad Acad ki USA 1982;?9:3310-4. 12. Tarazi RC. Fouad Phi. Ceimo JK. Brava EL. Renin. Pdorterane and cardiac decomperwtion: studies with an oral convertingenryme inhibitor in heart failure. Am J Cardtol 1979:44X113-8. 13. Cody RI, Franklin KW. Kluger 1. Laragb JH. Sympathetic rerponra and plasma norepinepbrine during therapy of chronic congestive heti failure with captopril. Am J Med 1%2;72:791-7. 14. Cody RJ, Bravo EL, Fouad FM. Tarui RC. Cardiovascular r&exer durine chronic convenina enzvme inhibition and sodium de&&m: the rest&e lo lilt in hyprt&& palientr. Am I Med 1981;71:~22-6. IS. Cody RI. Franklin KW. Klugerl. La& JH. Mechanismsgoverning the partural response and barnreceptor abnormalities in chroeic ccmgsstivr heanfaiture: eects of acule and tong-term convtrting~nzyme inhibibon. Clnulation 1982.66: 1X-42. 16. Lumbers ER. McClorkcy DI. Potter EK. tnlmbition by angiolenir, It OF baroreceptor-evoked aclivity in cardiac vagrd &snl nerves in the dog. I Phvriol 1979:194:#-80. I?. B&non RK. Buckley JP. Evidence for a centnt mechanism in an& leusin induced hypertension. Proc SOF Exp Biol Med 1%1.106:834-86. 18. Ismay MJA. Lumbers ER. Stevens AD. The action ofangiotrosin II on the baroreilex response of the conscious we and the wuscious foetus. J Physiol 1979:288:467-81. t9. Scrap GC. I owe RD. Efkrent pathways of the cardiovarculaF respome 10 vertebral infusions of smgiclcnsin in the dog. Clin Sci l%v37: M1!-19. 20. Sweel CS. GXIY MJ. Central inhibition of reRex vascdilaetion by angiolcnsin atid reduced renal treasure. Am J Phvriol l970:219:1751-8. 21. R&r GP. Raw JH. Aogisienrin II potential& the vasownstrictive e&t of norrpinephrine in normotemive rxrd hypertenrive man. J Cbn Hyperrcns t987:3:610-6. 22. Guo 08. Abboud FM. Angiotensin II ~ltenulltes barar&x conlrol of Am J Phvsiol 1984:24&H&9.
2‘tery
37. Horn EM. Cowm SJ. Sler,be~?SF. et al Reducedlymphoryle ~I~muktory gupniw nucleotide reg.ula~oryprotein and bera-adrenelgic receplarr in congestive heart failure and revenal wlh ;mgiolensin converllng enzyme inhibitor therapy. Circuletiun 1988.78:1371-Y. 38. Appel ML. Belger RD. Saul I?. Smith JM. Cohen RJ. Brat to bea variability in cardiovascularvariables: noise or music? f Am Coil Csrdlol 191L9.1d.11w9-lR 39. Pomemnz 8. Macauley RJB. Caudill MA. et al Asaessmenlafsutwomlc ;;;p; in humansby hean nle spectral a.dIws. Am J Physiol IYSN:
._“_ ,....._.
40. Aksebod S. Go&o D. Shannon DC. Bargcr AC. Cohci, RJ Puwcr sger~~m analysis of hean rate Ructualian: a quantitative probe of bat-to-ba cardwascubr control. Science 1981:?13:230-? 41. Kamna PG. Jh F. Respiratoryvnuranhylhmw: noninvaiivc mrawre of pwarym~helic cardiac contml. J Appl Phyrml 197S:;Y1801-S. 42 Abrrkod S, Gordon D. MadwedJR. Sn,dman NC, ShannonDC. Cohen RJ. Hemcdynamic rwlalion. inverdgarion by w:ml analyrir Am J ^.
.^^IX.. __-r
mgordrdlal mfuctmn m consciousdogs at hgh and I& nrk at sudden dcalh J ,AmCell Cadiol 1wO:16.97S-S~. ?5 Kleipr RE. Mdlcr JP. B~pgerJT. Moss.AJ, for the Muhicenter PostInhrwon Group. Cwrearcd hcxl rate vtiabildy and ds awxiallon with increased monalily afler acute myoadial infarclian. Am J Cardid ,987.59:2.i6-62. 56. B~wcrJT. Klcipr RE. FleibsIL, Rolnhrky LM. SleinmanRC, MillerJP. far the hluft~cenler Ponl.lniarclion Gwup. Components 01 hean raIe vorwbiliry mcrrureddurinp healtngofacute m)_aardml mfarclion. Am J Cardiol IRR;61:2MI-IS. 57. Billman GE. ERect of carbachol and cyclic GMP on sussepl!bilily to vrntncul.r Cbnllation. FASEB J 1990:1:1668-73. 58. Hlmh JA. Bishop B. Respiratory sinus anhythmia in humans: how hrcalhiq palrun modulatesheart rale. Am J Physic41981:?JI:H62U-9. 19. Pwni M. Lvmbardi F. CuEelli S. er sl. Powerrp~rnl analysis.ofhrar!
F
BINKLEY,
ENRICO
NUNZIATA.
ROBERT
i. CODY.
Cohnbus.
Ohio
MD,
FACC.
MSBME, MD,
CARRIE
PATRLCLA
MD.
RANDALL
CNMT,
URL
C. STARLIhG. V. LEIER.
MD, FAX, MD.
FACC.
FACC
Objectives. The objeclive of this inve@aiion the changes in parasym:~thetic
J. HAAS, A. HAnON.
lone associated
was to evaluate with long-term
aogioteasin-converting euzyme inhibitor therapy iu palients with CWgcStiW heart failure. Backgwmd, Angioieasin-coevsrting enzyme inhibitors pnr vide hemudynamic aad symptomatic benefit and are associated
with impmved survive! in patients with conprtive heart failure. Angiotensin II, whusepmduction is ultimatety inhibited by these agents, exerts significant regulatory iaflueace on a variety of target organs induding the central and peripheral nervous sy” tems. Accwdiagty, it wautd he anticipated that angiotearinmnvertiu# enzyme iuhibitcux would siguiticentty alter the autoo&e lmhalan~c characteristic of patients with congestive heart failure and that Ihb lniiuence over neural mechanirms al cardtovascutar cuntrtd may s@ificantly coatribute to the hetttudynatnic heacflt and improved survival assuciated with angioteusinconverting enzyme inhibitor therapy. Methods. In the current bwstipation, changes in autrmomtc Ione &ted with luug-term dministmtion of an augiotensinconvertiog etwyme inhibitor were measured using spectral auatysis of heart me variability ia 13 palients with eeugestive heart faiture who WM enmlled ia a double-blind randomized placebw~ cwtrotled trial of the angiotensin-convertiagenzyme inhibitor
Aneiotensin-convertine enzvme inhibition has heen shown to be an effective strategy fdr the symptomatic management and improved survival of patients with congestive heart failure (l-8). Agents that inhibit angiotensin-converting enzyme have been shown to possess a diverse army of prop-
21. 1992, a&ted Avgvrt 24. 1992. m: Philip F. Binkky. MD. Ohio Srale Umversky Hospital, Rwm 610 Means Hall, 1654 Upham Drive. Columbus. Ohm 43210.
zorenoptil. Both placebo and treatment gsoslg were batsned ilt baseline study in terms uf fuactional rlass. ventrictiar performance and rutooomic toa. Resub. After II weeks al therapy with placebo, there was MI change in total heart rate variability, pua~mrpnthet~lty gowerncd h&b frequevxy bean rate variability 01 sympatbeticatty influenced Ion. frequency IwW rate variability. ln contrast, therapy with zofeoupril was ass&ted with a 50% itxrease in total heart rate vatiabillty (p = 0.09) and a rignificaat &I = 0.03) twofold increase in high frequency heart mte vnrLbility, tndicsting a signitirant augmentation of parasym+hetic tone. Conchsiom.
ment
These results dtmuuslrate
ihntl@@+‘m treat-
of patients having congestive heart Failure with aa aagiotensiusonvertiag eazyate inhibitor is a.wcSdcd with u IS toratton
of sutoncmtc
sustained agmentatiin
hatance,
shtch
derives
of purasympnlhctii
in part from P
tone. Such nugmen-
tation of vagal tone is kuowt to be proteclive Pgaimt mnlign;gnant ventricular nrrhy~hmtpr in patter& with &&em& heart d&w and therefore may have similar hen& in the setting of wWi&u failure. thus contrihuling to the improved survival a.wcilted with aogiutenshwuuverting enzyme inhibitor therapy in @ien& with
congestive heart Mute. (.I Am Cull Cwdioli993;21.~55%51)
erties that may contribute to varying degrees to the improved hcmodynamic profile, functional status and survival associated with these compounds (9-15). Given evidence derived from animal models that demonstrates that angiotensin II produces a central suppression of vagal outflow as well as an increase in sympathetic drive (16-24). it would bc anticipated that angiotensin-convening enzyme inhibitors would significantly alter the autonomic imbalance that is characteristic of patients with congestive heart failure (2533) and that may be an imporlanl determinant of disease progression and mortality in this patient group (34-W Previous studies (13,15.37) with angiotensin-converting cnzymc inhibitors have suggested that they may significantly reduce Ihe incnased sympathetic drive of patients with congestive heart failure as evidenced by a reduction of plasma norepmephtine and at least partial restoration of
JACCVol. 21, No. 3 March I. 1993:655-61
abnormal reflex responses in these patients. However, the potentially important role of increased parasympathetic tone resulting from angiotensin-converting enzyme inhibition could not previously be evaluated. Spectral analysis of heart rate variability permits a noninvasive in viva measure of parasympathetic tone that may be applied in a serial fashion to define the manner in which angiotensin-converting enzyme inhibition modulates the w gal component of iie autonomic nervous system in the setting of venlricuku falfure (33,38-47). The current inves. t&ion used a randomized double-blind placebo-controlled design to test the hypothesis that long-term therapy for congestive heart failure with an angiotensin-converting enzyme inhibitor will augment parasymparhetic tone, which has previously been shown to be significantly attenuated in patients with ventricular failure (32.33).
Methods Study group. Thirteen patients with congestive heart failure who were referred to the Ohio State University Heart Failure Clinic and enrolled in a randomized double-blind plti:ebo.controlled trial of the angiotensin-converting enzyme inhibitor zofenopril constituted the study group. All patients had symptomatic congestive heart failure due to left ventricular sgstolic dysfunction, which was confirmed by echocardiographically derived percent fmctional shortening of the minor axis of the left ventricle of ~25%. Ventricular systolic dysfunction was due to idiopathic dilated cardiomyopathy in 10 patients and to significant epicardial coronary artery stenosis in 3. None of these patients had uncontrolled angina or had experienced a myocardial infarction in the preceding 6 months. Eight patients (four men and four women) aged 39 to 65 years (mean L SD 46.8 2 8.3) were randomized to treatment with zofenopril and five patients (two men and three women) aged 47 to 67 years (mean 56 i: 8.5) were randomized IO placebo. The mean age of the two groups did nor significantly differ. The percent fractional shortening of the left ventricle for all I3 patients was 14 i 4% and did not differ between the Iwo groups (I6 + 4% for the placebo group and I3 t 4% for the treatment group). The two groups were balanced in terms of symptomatic status (with mean New York Heart Association functional class 2.4 + 0.4 for the placebo group and 2.4 c 0.4 for the treatment group). All patients provided informed consent to the investigation in accordance with the Human Subjects Review Committee of the Ohio State University, which reviewed and approved this protocol. Study design. All patients were enrolled in a 2-week placebo run-in period during which time any vasodilator medications that may have been prescribed wcrc wilhdrawn and doses of diurelic drugs and digoxin were adjusted to maintain optimal control of congestive heart failure symptoms with use of these agents alone. All patients received a single daily placebo capsule during this run-in period. At the end of this period, patients were treated with digitalis in
II
the form of digoxin. with doses ranging from 0.125 to 0.25 mglday, and 2 were treated with digitoxin, 0.1 rnd day. Eight patients were treated with furoseimide, 40 to 160 mgiday. One patient was treated with both fnrosemide and a hydmchlorothiazide-triamterene combination. One oatient was treated with a hvdmchlorothiazide-triamterene combination alone, two pat& were treated with bumetsnide, 2 to 3 md&y, and one patient did not require diuretic dtugs. At the conclusion of this 2-week period, patients underwent recording of ‘he electrocardiogram (ECG) for assessment of heart rate variability. After this baseline measurement was obtained, patients randomized tn rhe control group received daily placebo in addition to stable doses of a digitalis preparation and diuretic drugs as established in the run-in period. Patients randomized to angiotensinconvening enzyme inhibitor therapy received a total daily dose of I5 mg of nofenopril in addition to the digitalis preparation and diuretic drug dose established in the run-in phase. Doses of digitalis preparations and diuretic drugs were maintained constant throughout the entire K&week treatment phase in both groups. At the conclusion of this 12.week period, the ECG was again recorded for analysis of heart rate variability. Measurement of heart rate vtiabiliIy. Recordings of the ECG at baseline study and after the 12-week treatment period wore obtained in all patients in the supine posture after a 20&n period of equilibration. All examinations were performed between 9 and II AM and all medications were withheld on that day until the recordings were obtained. Electrocardiographic electrodes for recording standard limb leads were attached and a lead demonstrating predominant positive or negative QkS de&lions (facilitating QRS recognition by the signal-processing algorithm) was selected for recording and analysis. The selected ECG signal was recorded for a minimum of 4 min with use of a RACAL V STORE FM tape recorder at a speed of 3.7%in. (9.53 cm%. If somatic artifact or frequent ectopic beats were noted at any time, the recording was halted and reinitiated after such interference had resolved. The ECG signal was subsequently analyzed for generation of the power density spectrum of heart rate variabiliiy. Spectral analysis of benri rate variability. Spectral analysis of heart rate variability was performed as previously reported by our laboratory (33) by an observer who had no knowledge of the assignment of the subject lo placebo or zofenopril treatment. The tape-recorded ECG signal was preprocessed with use of an antialiasing filter and digitized by means of a Itbit analog/digital converter board (Metrobyte Co.) installed in au IBM/AT computer at a sampling rate of 512 Hz. Once digitized. the ECG signal was passed through a digital band-pass filter having a central frequency of 85 Hz. A dynamic user-interactive threshold technique was applied to the filtered signal to detect the R waves and compute the RR interval sequence. Subsequently, the RR interval sequences were passed through a statistical filter to
eliminate rapid transitions due 10 signal detection faults or ectopic beats. Data points ourside the 95% confidence inferval
of the previous
IO points
were
etiminatrd
and
a point
by linear interpolation of the preceding and following points was substituted 133). A plc! of the instantaneous heart rate versus rime (heart rate variability signal) war generated from the RR interval ;squence (48). The heart rate versus time series was then passed through a Farzen window and the power spectrum density of heart rate variability WBF generated with use of the modified pcriodogran method of derived
Welch
(49).
This
method
is based
on the multiple
14
2.4
3”
35
3-381 ACE Inhibition
computa-
tion and average of the fast Fourier tmnsform of overlappLri data segments (49). With this method. the variance of rhe estimated power spectral density is reduced by 1 fxtor proportional to the number of data segments used. The power spectral density was then normalized so that the total power was equal to the signal mean square. A plot of the values afthe power spectrum density agamst frequency WB~ then generated. An area under the curve method KBS used to quantify the power within specified frequencies. Specifically, the total area, the area under the low frequency (0.02 IO 0.1 Hz) region of the curve. which is influenced bs sympathetic nerve activity, and the high frequency PO.1 Hz) region, which reflects parasympathetic tone. were calculated (33,39.40.42,46,47). Thus, the magnil~+ of parasympathetic tone before and after the treatment period in the two groups could be quantihed by this system of analysis. Stalistical analysis. All data are presented as mean value A SD. A Student I tesf for unpaired data was used to compare baseline variables of ventricular function, functional class. age and autonomic tone in the treatment and placebo groups. Analysis of variance for repeated measures was used to evaluate intergroup and intrdgroup changes in variables of autonomic tone. functional class. heart rate. respiratory rate and blood pressure over the treatment period. Statistical significance was defined at the p < O.OS
I
level.
Results Spectral analysis of heart rate variability at baseline study. The power density spectrum of heart rate variability for both the placebo group and the group treated with the angiotensin-convening enzyme inhibitor zofenopril (Fig. and 2) was typical of that previously reported (33) in congestive heart failure with little variability noted in the parasympathetically mediated high frequency region (20.1 Hz) and the preponderance of variability noted in the sympathetically mediated low frequency region (~0. Hz). Furthermore, rhe baseline characteristics of the autonomic profile of the placebo and treatment groups as reflected by the power density spectra did not differ. Specifically. no sit#icast diBrences between groups were noted in the low (placebo 1.4 x lo-’ + 1.0 x IO-’ Ibeats/minl’; zofenopril 1.0 X lO.3 2 1.0 x 10m3 [beatslmin]‘) or high (placebo 0.5 X IO-’ 2 0.J x lo-’ Lbeatsirmn]‘; zoreenoprtl U.6 x IV” t
I
i
20
Frequency (Hz) Figure I. Power spectral delmity of heart mre variability recorded in !J pz!ien! 1~1th covzesrive heart failure at baseline study (top) and aller I? weeks of therapy with the angiotensin-convening enzyme (ACE) inhibitor zofenopril (bottom). Both we platted on the s8me scale. The baseline power density spectrum is typical of that previously reported for patients with congestive heart ftilure with little heart rdte variability @ower) in the parasympathetically mediated frequency band 20.1 Hz consistentwith il marked attenuation of parasympathetic tone. After therapy with the angiotensinconverting enzyme inhibitor. there is a marked augmentation or heart rate vanabilily in the high frequency region. indicating an augmentation of parasympadwic tone. In this example patient. the %ympathc!ically influenced low frequency band is also reduced in megnitude. Fnggestinga decrease in sympalktic drive (see Discussion). BPM’ = Ibeatslmir#.
0.6 x IO-’ [beatsimin]‘) frequency bands of heart rate variability. Accordingly, total heart rate variability as quantified by the total arev under the power density spectrum of heart rate variability did not differ between groups (placebo 2.4 x IO-’ 2 I.6 x IO-’ [beatslminj’; zofenopril 2.2 x IO“ 2 1.4 x lo-’ [beatslmin]?. Speml
analysk or heal l-ate vatibiiity aRETtre%tment.
Total heart rate variability did not change in the placebo control group. with the total area bounded hy the power density $prLtrurn measuring 2.7 x IO-’ f I.3 x IO-’ (beats/min? ?! !hc end of the 3-month period compared with 2.4 X IO-’ + 1.6 K IO-’ lbeatslmin)” measured at baseline (Fig. ?, In contrast, a significant t=nd (p = 0.09) toward an increase in total area was noted in the angiotensinconverting enzyme inhibitor treatment group, with a 50% increase in area from the baseline value of 2.2 x low9 k 1.4 x IO-’ Enmination
totat 0.03) high (Fig.
to 3.4 x IO-’
r
1.4 X 10m3 (beatslmin)‘(Fig.
3).
of the specific frequency bands comprising heart rate varlahilitv demonstrated a sienifcant (u = withm-group incrke in parasympatheti~ally mediated frequency heart rate variability in the treatment group 4) as shown by a twofold increase in the high frequency
2.33
Placebo Easellne
1.75
0.00 1. 02 2.33
08
.I4
.2o
.24
.3o
.35 ACE INHIBlTtON
Placebo - 12 Weeks
1.75
PLACEBO
Figure 4. Mean and SD of parasympathetic&
1.17 0.58 o.oaI4-.o2 .DB .14
.2o
.24
.3o
.35
Frequency (Hz) FigureZ. Power spectral density ofheart rate variability recorded in a patient with congestive heart failure at baseline (lop) and after 12 weeks of therapy with placebo (baltom) plotted on the same SC& The spectra have virWally identical config~n&ns. There is no change in the magnitude of variability in the parasympathetically mediated region >O.l Hz between the baseline and treatment periods. BPM* = (beatsiminl’.
area from 0.6 x IO-’ t 0.6 x IO-’ IO 1.3 x IO-’ ? 0.9 x IV3 (beats/mitt:‘. However, no change in high frequency area was noted i” the placebo control group. with an area of 0.4 X 10e3 -C 0.3 x IO~‘(beats/min)* notedafter 12 weeks as compared with the baseline value of 0.5 x IO-’ + 0.5 x 10m3 (beatsimin)*. No change in the low frequency area was noted in either group. A numeric but nonsignificant increase in low
tone
frequency area was observed in the placebo group, with a value of 1.9 x IO-’ + 1.2 x Km3 (beatslmin)* recorded at the end of 12 weeks. The low frequency area in the zofenopril treatmentgroup did not change from the baseline value, with an area of I.0 x IO-’ i. 1.0 x IO-’ (beats/min)* measured at the end of the treatment period. Accordingly, the proportion of high frequency heart rate variability increased in the zofemxxil treatment arouD as reflected by the ratio of hiab freque& to total heart ‘rate variability; which ittcreasad from 22 r 15% to 34 + 16% (p = 0.1) but did not change in the placebo group at baseline vs. 15 + at the end of the treatment period). These changes in the variables of hean rare variability are exemplified in the representative power density spectra demonstrates findings displayed in Figures and 2. Figure in a patient in the zofenopril treatment group. The baseline spectrum (Fig. I, top) is characteristic of oatients with
(19t 9% 1
Figure 3. Mean and SD of the tolal area under the power spectral density of heart rate variability in patients receiving an angiotensinconverting enzyme inhibitor (ACE INHIBITION) or placebo. A
50% increase (p = 0.09) in total heart rate variability (power) is noted after 12 weeks of treatment (hatched bars) with the angioten
sin-convortiq enzyme inhibitor. In contrast, there is no change in
total heart rale variability in the placebo contml group. bpm2 = (beatsimin)‘.
ACE tNHlBlTtON
PLACEBO
mtdiatcd high
frequency hean rare variability in patients refeiviagan angiotensinor placebo. A converting enzyme inhibitor [ACE INHIBtTION) significant (p = 0.03) twofold increase in high frequency area is noted after I2 weeks of treatment (hatched bars) with the angio’ensin-convertina enzyme inhibitor. indicatina an increase in oarasvmpathetic tone. No-change in high frequ&y area is notid in-the p!Zebo group, indicating conrinucd atlenuation of parasympathelic typical of palients with congestive heart failure. bpm2 = WatslminY.
II%
I
witi an r&kHz) heart rate variability consistent tion of parasympathetic tone (33,461. After the treatment period, there are striking differences in the power density spectrum, with a marked increase in variability in the band >O.i Hz. In this particular example, there is also a reduction in the sympathetically intIuenced frequency band in the region ~0. I Hz. In contrast, the power density spectrum for the patient in the placebo group (Fig. 2) shows a consistent configuraticn, with no change between baseline and the I’Lwerk treatment period. Clhtbl response. None of the patients in either the placebo group or the group treated with the angiotensinconvening enzyme inhibhor zofenopril required in-hospital treatment of decompensated consstive heart failure. As outlined in the study protocol, all patients were successfully managed with stable doses of diuretic drugs and a digitalis preparation as established in the run-in period. Mean sys-
temic blood pressure did not change from the baseline value of 87 t 9 mm Hg in the placebo group and 88 + 10 mm Hg in the zofenopril treatment group. Mean heart rake did cot change in the placebo group (81 t tcatsimin al baseline
intravenous infusion of angiotensin II. Considering the elevation of angictensin 11 in patiems with congestive heart failure. the characteristic blunting of ;he arterial baroreflex rcsponsc sod the attenuation of parasympathetic tone ob=
and 81 c 6 beatsimin at the end of the placebo period). A
perved in thete patients may derive in part from angiotennin
numeric decrease in mean heart rate thal did nor attain
ing to 85 L IO beatslmin. Respiratory rate measi;red at rest immediately before the ECG recording petiod did not change h iithergroup, measuring20 2 2 breathslmin at baseline and
It-mediated central suppression of vagal tone. Accordingly, a reduction of angiotensin II concentration. as occurs with angiorensin-converting enzyme inhibition, would reverse central suppression of vagal outflow, resulting in an augmentation of parasympathetic tone. The enhanced parasympathetic tone with angioteedn.convcrting enzyme jnhibition
21 i 2 breath&in
observed in the torrent
II
statistical significance was noted in the zofenoprii treatment group with the baseline value of 89 T IO beats/mm decreas-
after 12 weeks in the zofenopril rreatmenr
group. In the placebo group, respiratory breathslmin
rate ~3s ?I * 4
at baseline and 19 + 2 brearhslmin
weeks (bath p = NS).
New York
afler
I2
Heart Association func-
tional class significantly (p = 0.007) improved in the zofeno-
pril treatment group, decreasing from 2.4 + C.4 to I .8 ? 0.8. No change in functional class was noted in the placebo @‘oUp, with a mean vatue of 2.3 z 0.4 measored at the end of the study as compared with 2.4 * 0.4 recorded at baseline.
investigation
woilld therefore
he
predicted on the basis of these fundamental mech:aisns. The current investigation cannot verify that the observed augmentation
of parasympathetic
tone does indeed derive
from the angiotensin II-related mechanisms discussed. It is conceivable that some propeny of the anpiolensinconverting cnryme inhibitor unrelated to aogiotensin IImedialcd mechanisms may have produced the notedchanges in parasympathetic tone. Accordingly. it is not possible from the current dats to verify whether this vagotonic influence is
Discussion The current investigation demonstrate, that the autonomic imbalance in patients with congestive heart failure can be significantly altered by long-term rhenpy with an angiotensin-converting enzyme inhibitor. Autonomic tone was assessed with the technique of spectral analysis ofheart rate variability,
which has been shown to provide a nonin-
vasive in viva assessment of autonomic activity given that
I Hz)
variations in heart rate are exclusive& under the control of the parasympathetic nervous system and that low frequency heart rate variability (co.1 is influenced to a significant extent by sympathetic drive (39,40,47). Speciiically, with use of a randomized double-blind placebo-controlled stttdy design, it was shown high frequency 00.
Hz)
that 12 weeks of therapy with the aogiotensin-converting enzyme inhibitor zofeoopril produced an increase in total
(Fig. 3). This increase was due to a significant twofold increase in high frequency heart rate variability (Fig. 4), which is specifically a direct manifestation of a significant augmentation of parasympathetic tone. These findings contrast markedly with those in the placebo control group, which showed no change in variables of heart rate variability
autonomic tone after the la-week
treatment period (Fig. 3
a class effect of all angiotensin-converting enzyme inhibitors or is specific to zofenopril. However. if interruption of angiorensin [I-mediated suppression of vagal tone is indeed the operarive mechanism for the increased parasympathetic tone noted in lhis investigation. it would h predicted that any angiotensin-convening enzyme inhibitor tha1 reduces central nervous system concentntions of angiotenain II would produce a similar augmentation
of parasympathetic
heart failure. Clinical implications. Heart rate variability has been shown to be perhaps the most powerful independent predictorof malignant ventricular arrhythmias and sudden death in patients with ischemic heart disease (50-56). These clinical observations confam basic investigations in animal models thal have shown that the absence ofvagal tone as manifested by diminished high frequency heart rate variability is associated with a reduction of ventricular fibrillation threshold (53) and that restoration of vagai tone reduces this propensity to ventricular arrhythmias (57). if these principles pertain to patients with dilated cardiomyopathy, then the attenuaibn of parasympathetic tone thal is evident by the signiticant decrease in high frequency heart rate variability tone when administered
to patients with congestive
may contribute to the pmarrhythmic
environment that pre-
disposes this patient group to sudden cardiac death. There-
and 4). Proposed mechanisms for vagolonic effectsof angidensinconverting enzyme inhibition in coageslive heart f&w. The mechanism for the observed effects of angiotensinconverting enzyme inhibition are speculative at this time but are suggested by prior observations regarding the influence of angiotensin 11 on autonomic tone (16-24). Angiotensin
fore. it may be speculated that the augmentation of parasympathetic tone reported in this investigation may contribute to the improved survival associated with angioteasinconverting enzyme inhibitors in the setting of congestive
mediates a central suppression ofvagal o&low
strates that the technique of spectral analysis of heart rate variability may he successfully used for the prospective
when admin-
istered in a variety of animal models (16.23). Lumbers et al. (16) demonstrated a significant reduction of parasympathetic
activity recorded directly from vagal e&rent
fibers during
heart faiture (6,8). Methadolagic nmsiderations.
This investigation
demon-
serial assessment of changes in autonomic tone associated with long-term treatment of congestive heart Failure. The
660
RINKLEY ET AL. AuGMENT*TwN OF FARASYMPATHETIC
JACC Vol. 21. No. 3 hhrcb 1. lYw655-61
TONE
consistency of the measures obtained in the placebo contml group at baseline study and after a 12-week period in terms of total heart rate variability and its low and high frequency components demonstrates the reproducibility of this technique even over an extended period of time (Fig. 2 to 4). A consistent change in low frequency heart rate variability was not observed in the angiotensin-converting enzyme inhibitor trcatmeni group and, as a consequence, the mean value for this frequency band did not change after the treatment period. Because angiotensin II has been shown in some models to stimulate central sympathetic outflow (17), it may be anticipated that angiotensin-converting enzyme inhibition would reduce sympathetic tone and thus the low frequency band of hearl rate vaizbility. However. low frequency heart rate variability is not devoid of some degree of influence by the parasympathetic nervous system (40,47). Therefore, an increase in parasympathetic tone may result in some increase in low ficqtiency heart rate variability that would offset decreases in this frequency band caused by reduced sympathetic drive. The net result would be no change in the low frequency band as observed in the current investigation. Respiratory rate can influence the magnitude and frequency distribution of heart rate variability and thus produce changes in the frequency bands of interest independent of any primary change in autonomic tone ($59). However, in the current investigation, there was no change in respiratory rate in either group after the treatment period, a finding that further substantiates the conclusion that the observed changes in high frequency heart rate variability were indeed due to an augmentation of parasympathetic tone in the zofenopril treatment group. Conclusions. In this randomized double-blind placebocontrolled trial, a significant augmentation of parasympathetic tone was associated with long-term administration of the angiotensin-converting enzyme inhibitor wfenapril in patients with congestive heart failure. Whether in patients with ventricular failure such augmentation of vagal tone confers a protective effect against malignant ventricular arrhythmias and sudden death, as is seen in patients with ischemia, requires further study. Likewise, whether the changes in autonomic tone observed in this report are a nonspecific response to vasodilation or are specific to angiotensin-converting enzyme inhibition provides the foundation for future investigations. We ‘bank Trichia Zavilla for expen nssislance in Ihe preparation of the manl!script. .-
References I. Capbwil Multicenter Research Group. A placebcwntmlled lrial of caploprjl irarefractory chmnic congestive heart failure. I Am Colt Cardiat 1983;2:755-63. 2. Crcager MA, Massie BM. Faxon DP. et al. Acue and long-term effeleclsnf enaiapril un the cardmvascular response lo exercise and exercise loler-
3. Levine TB. Olivsri Ml. Garberg V. Sbarkey SW. Cohn JN. Hemodynamic and clinical response Lo enalapril. P longs&g convening-enzyme InhIbitor. in patients with congeslive hcarl fadurc. Circulalion 1984~69. 548-53. 4. Packer ht, Medina N. Yushak M. Comparative hrmodynamz and Clinical effects or long-term tratment with prazosin and captopnl for severe chronic congestive heart failure secondary to coronary a&y disease or idiopathic ddafed cardiomyopathy. Am 5 Cardiol 1986:5?: t32&7. 5. Cody RI. Cowl AB. Schacr CL. La@ JH. Evaluatmn of a long-acting converlinp enzyme inhibitor lenalapril) for Ihe freatmenl of chronic congeslive heart failure. I Am Co8 Cardiol 1983:1:1154-9. 6. The CONSENSUSTrial Study Oruup. EDctsofenalaprit on monalhy in severe congcslive heari f&x: rcrukr of Ihe Cooperative N*dh Scandinavian Enalapril Survival Sludy ICONSENSUSL N Engl J Mcd 1987: 316:1429-x 7. TheSOLVD lnvesli@ors. EIlecrof enalapril on survival in patiems whh reduced left ventricular eJection fractions and conge~Iwe hrart fadurc. N En@ J Mcd 1991:325:193-302. 8. fohn IN, Johnson G. tieschc S. et al. A compansan of enalapnl with hydratazine-irorbide dmrlrale in the rramentofchmn~c congerrive hean failure. N Engl J bled 1991:325:30~10. 9. tiollr J. Busse R, Sommer 0,Basrenge E. Dilatation ofepicardial arteries in consciaus dags induced by angiotenrin-convening enzyme Inhibition with enalaarilat. J Cardiovasc F’harmacal 1987:9:348-25. IO. Lindpaintner K. Ganten D. The cardiac reniwangiotensin ayeem: w rwraisal of Present extwimenlal and clinical evidence. Circ Res 1991: 68:%5-21. It. Pfeffer JM. PIa& MA. Mirsky I. Braunwald E. Regression of ten venrricular hypcrtrophy andprevention of kfr venlricular dy&r&n by caploprit in the sponlaneourly hypatensive rat. Proc Nad Acad ki USA 1982;?9:3310-4. 12. Tarazi RC. Fouad Phi. Ceimo JK. Brava EL. Renin. Pdorterane and cardiac decomperwtion: studies with an oral convertingenryme inhibitor in heart failure. Am J Cardtol 1979:44X113-8. 13. Cody RI, Franklin KW. Kluger 1. Laragb JH. Sympathetic rerponra and plasma norepinepbrine during therapy of chronic congestive heti failure with captopril. Am J Med 1%2;72:791-7. 14. Cody RJ, Bravo EL, Fouad FM. Tarui RC. Cardiovascular r&exer durine chronic convenina enzvme inhibition and sodium de&&m: the rest&e lo lilt in hyprt&& palientr. Am I Med 1981;71:~22-6. IS. Cody RI. Franklin KW. Klugerl. La& JH. Mechanismsgoverning the partural response and barnreceptor abnormalities in chroeic ccmgsstivr heanfaiture: eects of acule and tong-term convtrting~nzyme inhibibon. Clnulation 1982.66: 1X-42. 16. Lumbers ER. McClorkcy DI. Potter EK. tnlmbition by angiolenir, It OF baroreceptor-evoked aclivity in cardiac vagrd &snl nerves in the dog. I Phvriol 1979:194:#-80. I?. B&non RK. Buckley JP. Evidence for a centnt mechanism in an& leusin induced hypertension. Proc SOF Exp Biol Med 1%1.106:834-86. 18. Ismay MJA. Lumbers ER. Stevens AD. The action ofangiotrosin II on the baroreilex response of the conscious we and the wuscious foetus. J Physiol 1979:288:467-81. t9. Scrap GC. I owe RD. Efkrent pathways of the cardiovarculaF respome 10 vertebral infusions of smgiclcnsin in the dog. Clin Sci l%v37: M1!-19. 20. Sweel CS. GXIY MJ. Central inhibition of reRex vascdilaetion by angiolcnsin atid reduced renal treasure. Am J Phvriol l970:219:1751-8. 21. R&r GP. Raw JH. Aogisienrin II potential& the vasownstrictive e&t of norrpinephrine in normotemive rxrd hypertenrive man. J Cbn Hyperrcns t987:3:610-6. 22. Guo 08. Abboud FM. Angiotensin II ~ltenulltes barar&x conlrol of Am J Phvsiol 1984:24&H&9.
2‘tery
37. Horn EM. Cowm SJ. Sler,be~?SF. et al Reducedlymphoryle ~I~muktory gupniw nucleotide reg.ula~oryprotein and bera-adrenelgic receplarr in congestive heart failure and revenal wlh ;mgiolensin converllng enzyme inhibitor therapy. Circuletiun 1988.78:1371-Y. 38. Appel ML. Belger RD. Saul I?. Smith JM. Cohen RJ. Brat to bea variability in cardiovascularvariables: noise or music? f Am Coil Csrdlol 191L9.1d.11w9-lR 39. Pomemnz 8. Macauley RJB. Caudill MA. et al Asaessmenlafsutwomlc ;;;p; in humansby hean nle spectral a.dIws. Am J Physiol IYSN:
._“_ ,....._.
40. Aksebod S. Go&o D. Shannon DC. Bargcr AC. Cohci, RJ Puwcr sger~~m analysis of hean rate Ructualian: a quantitative probe of bat-to-ba cardwascubr control. Science 1981:?13:230-? 41. Kamna PG. Jh F. Respiratoryvnuranhylhmw: noninvaiivc mrawre of pwarym~helic cardiac contml. J Appl Phyrml 197S:;Y1801-S. 42 Abrrkod S, Gordon D. MadwedJR. Sn,dman NC, ShannonDC. Cohen RJ. Hemcdynamic rwlalion. inverdgarion by w:ml analyrir Am J ^.
.^^IX.. __-r
mgordrdlal mfuctmn m consciousdogs at hgh and I& nrk at sudden dcalh J ,AmCell Cadiol 1wO:16.97S-S~. ?5 Kleipr RE. Mdlcr JP. B~pgerJT. Moss.AJ, for the Muhicenter PostInhrwon Group. Cwrearcd hcxl rate vtiabildy and ds awxiallon with increased monalily afler acute myoadial infarclian. Am J Cardid ,987.59:2.i6-62. 56. B~wcrJT. Klcipr RE. FleibsIL, Rolnhrky LM. SleinmanRC, MillerJP. far the hluft~cenler Ponl.lniarclion Gwup. Components 01 hean raIe vorwbiliry mcrrureddurinp healtngofacute m)_aardml mfarclion. Am J Cardiol IRR;61:2MI-IS. 57. Billman GE. ERect of carbachol and cyclic GMP on sussepl!bilily to vrntncul.r Cbnllation. FASEB J 1990:1:1668-73. 58. Hlmh JA. Bishop B. Respiratory sinus anhythmia in humans: how hrcalhiq palrun modulatesheart rale. Am J Physic41981:?JI:H62U-9. 19. Pwni M. Lvmbardi F. CuEelli S. er sl. Powerrp~rnl analysis.ofhrar!