Journal of Cardiac Failure Vol. 21 No. 1 2015
Heart RateeDependent Left Ventricular Diastolic Function in Patients With and Without Heart Failure SAM ESFANDIARI, MSc,1,2 FELIPE FUCHS, MD,3 RODRIGO V. WAINSTEIN, MD,4 ANJALA CHELVANATHAN, MD,1 PETER MITOFF, MD,1 ZION SASSON, MD,1 AND SUSANNA MAK, MD, PhD1 Toronto, Ontario, Canada; and Porto Alegre, Brazil
ABSTRACT Background: Chronic heart rate (HR) reduction in the treatment of heart failure (HF) with systolic dysfunction is beneficial, but the immediate mechanical advantages or disadvantages of altering HR are incompletely understood. We examined the effects of increasing HR on early and late diastole in humans with and without HF. Methods and Results: We studied force-interval relationships of the left ventricle (LV) in 11 HF patients and 14 control subjects. HR was controlled by right atrial pacing, and LV pressure was recorded by a micromanometer-tipped catheter. The time constant of isovolumic relaxation (tau) was calculated, and simultaneous sonographic images were analyzed for LV volumes. The end-diastolic pressure-volume relationship (EDPVR) was analyzed with the use of a single-beat method. Tau was shortened in response to increasing HR in both groups; the slope of this relationship was steeper in HF than in control subjects. The predicted volume at a theoretic pressure of 0 mm Hg (V30) increased at higher HRs compared with baseline, shifting the predicted EDPVR compliance curve to the right in HF patients but not in control subjects. Conclusions: In HF, changes in HR affect early relaxation and diastolic compliance to a greater extent than in control subjects. Our study reinforces current recommendations for HR-lowering drug treatment in HF. (J Cardiac Fail 2015;21:68e75) Key Words: Heart failure, diastole, left ventricle, heart rate.
HF hospitalizations by 3% and 16%, respectively.1 The efficacy of cornerstone pharmacotherapeutic interventions, such as adrenergic antagonists, can be evaluated by reductions in HR. More recently, ivabradine, a nonselective inhibitor of the If channels in the sinoatrial node with purely negative chronotropic effects, has been clearly demonstrated to confer a survival advantage to HF patients.2 As such, current evidence-based ambulatory care for chronic HF mandates the acute manipulation of HR in these patients. Although the long-term benefits are clear, the immediate mechanical advantages or disadvantages of altering HR in HF patients are incompletely understood. It is well established that the left ventricular (LV) inotropic response to increases in HR (ie, force-frequency relationship) is blunted in the failing human myocardium.3 However, less is documented regarding the acute effects of changes in HR on the LV diastolic response (ie, relaxationfrequency) in vivo in humans with and without HF. Such data arise from older studies that included relatively decompensated HF patients. There are few data assessing the
High resting heart rate (HR) is a known marker of cardiovascular outcomes in patients with heart failure (HF) due to systolic dysfunction.1 One-beat and 5-beat increases of resting HR increase the risk of cardiovascular death and
From the 1Division of Cardiology, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada; 2Institute of Medical Science, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada; 3Division of Cardiology, Mount Sinai Hospital/University Health Network, Toronto, Ontario, Canada and 4Division of Cardiology, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil. Manuscript received June 6, 2014; revised manuscript received October 9, 2014; revised manuscript accepted October 20, 2014. Reprint requests: Susanna Mak, MD, PhD, Director, Assistant Professor, Division of Cardiology, Department of Medicine, Mount Sinai Hospital, University of Toronto, 600 University Avenue, Murray Side Rm 18-365, Toronto, Ontario M5G 1X5. Tel: þ1 416-586-4800 ext 5554; Fax: þ1 416-485-4775. E-mail: [email protected] Funding: Heart and Stroke Foundation of Ontario (grant-in aid T-7336). See page 74 for disclosure information. 1071-9164/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2014.10.013
68
Relaxation-frequency Relationship in HF
relaxation-frequency response in contemporary cohorts of HF patients. Because diastolic dysfunction is a common feature of HF, investigating the mechanical effects of changes in HR on LV diastolic performance is of clinical relevance to the administration of therapies available to HF patients. Therefore, our purpose was to examine the effects of increasing HR on early and late diastole in a cohort of humans with and without HF.
Methods Study Participants The study population consisted of patients with normal systolic LV function (NLV) and patients with systolic HF who underwent elective diagnostic cardiac catheterization at the Cardiac Catheterization Research Laboratory of the Mount Sinai Hospital. Patients agreed to participate in the research experiment after the diagnostic procedure was completed. The study was approved by the Mount Sinai Ethics Review Board for experimentation involving human subjects, and every patient gave written informed consent. Patients with NLV enrolled in this protocol were referred from outpatient clinics for assessment of chronic chest pain syndrome. All patients underwent stress nuclear perfusion imaging or stress echocardiography, and those with ischemia at low or moderate stress were excluded. Briefly, inclusion criteria for the NLV group were LV ejection fraction (LVEF) O55%, normal LV mass index documented by 2-dimensional (2D) echocardiography, no history or symptoms of HF, and LV end-diastolic pressure (LVEDP) #15 mm Hg at the time of catheterization. Patients in the HF group were recruited from a subspecialty HF clinic and were referred for right and left heart catheterization as part of the clinical evaluation. Similarly to the NLV patients, results of stress perfusion imaging or stress echocardiography were reviewed to exclude patients demonstrating ischemia at low or moderate stress. The etiology of patients with HF was nonischemic in all but 2 patients. Inclusion criteria for the HF group were LVEF !35% according to 2D echocardiography (confirmed on the day of the study) and stable symptomatic HF (New York Heart Association functional class II or III) of $3 months’ duration. General exclusion criteria were acute coronary syndromes or revascularization procedures within the preceding 6 months, ventricular pace-dependency, any rhythm other than sinus, QRS O110 ms, uncontrolled hypertension, and inability to give informed consent.
Esfandiari et al
69
condition was recorded. Two separate pacing protocols were then performed.
Pacing Protocol 1: Frequency-Dependent LV Pressure. Right atrial (RA) pacing was initiated $4e15 beats/min above the intrinsic rate to a multiple of 10 beats/min. Thereafter, HR was increased by increments of 10 beats/min every 3 minutes to a maximum of 120 beats/min or until Mobitz type 1 or 2:1 AV block occurred. If AV conduction fell below 1:1, the pacing was reduced by increments of 5 beats/min until 1:1 conduction returned. After this pacing protocol was completed, a rest period was allotted to allow LV pressure and hemodynamics to return to control conditions.
Pacing Protocol 2: Combined Sonographic and Hemodynamic Study. In this pacing run, LV pressure was recorded at the same time that echocardiography images were acquired by a research sonographer, as previously described.4 Four conditions were acquired: control and RA pacing at 80 beats/min, 100 beats/min, and 120 beats/min. Chronometers for the imaging system and hemodynamic recording system were synchronized to enable offline beat-to-beat simultaneous analysis of LV pressure and the time-stamped echocardiographic image frames. 2D and Doppler echocardiography imaging (GE Vivid 7 Imaging System, version BT03e5; GE Healthcare, Canada) and analyses were performed in accordance with the American Society of Echocardiography guidelines with the use of an M4S probe using optimized windows, and were analyzed offline with the use of a proprietary workstation (Echopac, version 7, GE Healthcare). Depending on the HR condition, images were acquired at 70e120 frames/s, with higher frame rates at higher HR conditions. Hemodynamic Analysis Electrocardiography (ECG) and LV pressure were continuously digitally recorded (1,000 Hz) online. For each study condition, LV peak positive dP/dt (LV þdP/dtmax), LV peak negative dP/dt (LV dP/dtmin), LV end-diastolic pressure (LVEDP), minimum LV pressure (LVmin), and LV systolic pressure (LVSP) were calculated offline with the use of a customized software program (Labview version 5.0; National Instruments Corp). The time constant of isovolumic relaxation (t) was also calculated by means of 2 separate methods: the monoexponential method with the use of a nonzero asymptote (tM),5 and the pressure half-time (t1/2) method.6 Arterial pressure was acquired continuously from the sidearm of the femoral sheath. These analyses were performed by a research technician blinded to the clinical status of the patient and the purpose of the study. Analysis of Frequency-Dependent LV Characteristics
Combined Cardiac Catheterization and Echocardiography Study Procedures Our study procedure has been described previously.4 All patients underwent selective coronary angiography, and HF patients underwent standard right heart catheterization from the femoral approach. Treatment with all oral medications was withheld on the morning of the investigation. Pharmacologic agents known to inhibit atrioventricular (AV) conduction were withheld 48 hours before the study day. To control HR, a 6-Fr bipolar pacing catheter was advanced to the high right atrium and aligned with a programmable stimulator (Prucka GE Cardiolab). A 7-Fr micromanometer-tipped catheter (Micro-tip Catheter; Millar Industries) was positioned in the LV for pressure measurement. A rest period was allotted after instrumentation, and a control
Protocol 1. For each HR condition, 50 consecutive beats from the 2nd minute of RA pacing were selected. If the preceding R-R interval was not within 2% of the planned pacing cycle length, the specific beat and the subsequent 5 beats were discarded. Per beat, t, LV þdP/dtmax, LV dP/dtmin, LVEDP, and LVmin were measured. The mean of these 50 cardiac cycles are reported. The relationship between HR and LV þdP/dtmax (the force-frequency relationship) as well as t (the relaxationfrequency relationship) were plotted. Protocol 2. At each HR condition, we used LV pressure recordings and sonographic images to estimate the end-diastolic pressure volume relationship (EDPVR) as described by Klotz et al.7 The calculation of EDPVR is dependent on estimates of cardiac volumes by transthoracic echocardiography. We did not
70 Journal of Cardiac Failure Vol. 21 No. 1 January 2015 calculate the EDPVR if the image quality was inadequate for these measurements. Three points of the EDPVR were estimated by means of equations listed in the Appendix. LV end-diastolic volume (LVEDV; Vm) and simultaneous LVEDP (Pm) were measured by matching the time stamp from selected sonographic images and the continuous LV pressure recording. With the use of Vm and Pm, 2 other points on the EDPVR can be estimated: V0 and V30, the theoretical volumes at pressure of 0 mm Hg and 30 mm Hg, respectively.
Statistical Analysis Data are presented as mean 6 SD. Between-group comparisons of baseline variables were made with the use of Student t test for continuous variables and Fisher exact test for categoric variables. Within-group and between-group comparisons of the effect of HR on LV hemodynamic variables within the HF and the NLV groups were assessed with the use of analysis of variance for repeated measures. To account for differences in resting HR, LV hemodynamic variables were analyzed in relation to incremental changes in HR above resting. To examine the effect of group-based differences between patients with and without HF, the relationship between HR and t was analyzed by means of an analysis of covariance with HR as a covariate. Simple linear regression analysis was used to describe the relationship between LV þdP/dtmax and t. All comparisons were based on a 95% confidence limit (P ! .05). All data were analyzed with the use of the statistical software PASW (Predictive Analytics Software) Statistics, version 19.0.
Study Population
Eleven HF patients and 14 NLV subjects without HF underwent the study protocol. Baseline characteristics are documented in Table 1. LV hemodynamic measurements are documented in Table 2, including the results of right heart catheterization in the HF population, who were relatively well compensated. Table 1. Clinical Characteristics
Age (y) Female (n) Body surface area (m2) Hemoglobin (g/L) Estimated glomerular filtration rate (mL/min) Hypertension (n) Diabetes mellitus Coronary artery disease Beta-blockers ACE inhibitors/ARBs Calcium channel blockers Diuretics Digoxin
NLV (n 5 14)
Variable Heart rate (beats/min) SAP (mm Hg) DAP (mm Hg) MAP (mm Hg) LVSP (mm Hg) LVEDP (mm Hg) LVmin (mm Hg) LV þdP/dtmax (mm Hg/s) LV dp/dtmin (mm Hg/s) tM (ms) t1/2 (ms) Mean right atrial pressure (mm Hg) Mean pulmonary arterial pressure (mm Hg) Mean capillary wedge pressure (mm Hg) Fick cardiac output (L/min)
65 6 123 6 60 6 85 6 112 6 10 6 3.9 6 1348 6 1466 6 42 6 30 6 d
HF (n 5 11)
8 21 10 13 20 5 3.5 285 304 7 6
d
72 134 73 94 115 14 6.5 939 1015 54 43 3
6 6 6 6 6 6 6 6 6 6 6 6
P Value
10 20 13 13 23 9 6.1 239 340 14 10 5
.08 .23 !.01 .10 .70 .19 .23 !.01 !.01 .02 !.01 d
18 6 7
d
d
9 6 6.5
d
d
4.7 6 1.3
d
DAP, diastolic arterial pressure; LV þdP/dtmax, left ventricular peak positive dP/dt; LV dP/dtmin, left ventricular peak negative dP/dt; LVEDP, left ventricular end-diastolic pressure; LVmin, minimum left ventricular pressure; LVSP, left ventricular systolic pressure; MAP, mean arterial pressure; SAP, systolic arterial pressure; tM, t calculated by means of the monoexponential method with the use of a nonzero asymptote; t1/2, t calculated by means of the pressure half-time method; other abbreviations as in Table 1. Data are presented as mean 6 SD.
Heart RateeDependent Effects on the LV Pressure Waveform
All patients achieved $4 RA-paced conditions above their resting HR (Table 3). For both groups, increases in
Results
Variable
Table 2. Right Heart and Left Heart Hemodynamics
NLV (n 5 14)
HF (n 5 11)
P Value
59 6 7 6 1.93 6 0.19 140 6 18 80.9 6 19.7
54 6 12 4 1.96 6 0.26 142 6 19 84.5 6 25.5
.21 1.00 .71 .71 .70
6 3 2 6 6 3 4 0
4 1 3 10 9 0 6 2
1.00 .60 .62 .03 .10 .23 .24 .18
ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; HF, heart failure; NLV, normal left ventricular function. Data are presented as mean 6 SD or n.
Table 3. Heart RateeDependent Effects on the Left Ventricular Pressure Waveform Variable
þ10
þ20
SAP (mm Hg)* NLV 124 6 26 125 HF 135 6 19 133 DAP* NLV 65 6 11 68 HF 81 6 21 83 MAP* NLV 89 6 16 92 HF 97 6 13 99 LVEDP* NLV 965 7 HF 15 6 11 12 LV þdp/dtmax (mm Hg/s)*,y NLV 1,453 6 307 1,537 HF 993 6 244 1,030 ,y tM (ms)* NLV 40 6 7 39 HF 53 6 15 50 t1/2 (ms)*,y NLV 29 6 5 28 HF 41 6 10 40
þ30
þ40
6 25 6 20
124 6 27 132 6 18
121 6 27 129 6 18
6 12 6 20
70 6 13 86 6 20
72 6 13 87 6 19
6 17 6 14
94 6 18 101 6 14
94 6 17 100 6 15
64 6 10
664 11 6 9
564 968
6 333 6 239
1,638 6 362 1,072 6 246
1,717 6 385 1,083 6 241
67 6 15
37 6 6 48 6 15
38 6 8 46 6 13
65 6 10
28 6 4 39 6 10
28 6 4 37 6 10
Abbreviations as in Tables 1 and 2. Data are presented as mean 6 SD. *Effect of HR. y Interaction of group and HR.
Relaxation-frequency Relationship in HF
Esfandiari et al
71
HR were associated with reductions in systolic arterial pressure and LVEDP, whereas diastolic arterial pressure and mean arterial pressure increased. LVSP decreased at the highest pacing rates. LVmin did not change with RA pacing, but was significantly higher in HF patients at each pacing rate. The force-frequency relationship was assessed by inscribing the effect of pacing-controlled HR on LV þdP/ dtmax. As expected, the slope of this linear relationship was significantly steeper in the NLV group than in the HF patients, who demonstrated a blunted LV þdP/dtmax response to increased HR. The interaction between HR and the studied group with a steeper HR-related rise in NLV group was significant (Fig. 1). The relaxation-frequency relationship was assessed by examining the effect of pacing-controlled increases in HR on t calculated by 2 separate methods (tM and t1/2; Fig. 2). We observed significant shortening of t in response to increasing HR in both groups, In contrast to the forcefrequency relationship, we observed that the slope of this relationship was significantly steeper in HF patients than in the NLV group. Contractility-Relaxation Coupling
The contrasting effects of HR on LV þdP/dtmax and t suggests that coupling between contractility and early diastolic relaxation was also different between the NLV and HF groups. We therefore examined the relationship between LV þdP/dtmax and t measured by 2 methods across the range of HRs. A linear relationship between LV þdP/ dtmax and both tM and t1/2 was observed in both groups. However, the slope of this relationship was significantly different between the NLV and HF groups. In the HF group, the inverse relationship was significantly steeper, such that relatively small positive differences in contractile state were
Fig. 2. The relaxation-frequency relationship in subjects with and without heart failure. t was significantly shortened in response to increasing HR in both groups. However, the slope of this relationship was significantly steeper in HF compared to NLV. tM, t calculated by means of the monoexponential method with the use of a nonzero asymptote; t1/2, t calculated by means of the pressure half-time method; other abbreviations as in Figure 1.
associated with incrementally greater shortening of both tM and t1/2 (Fig. 3). HR RateeDependent Effects on the Events of Late Diastole
Fig. 1. The force-frequency relationship in subjects with and without heart failure (HF). The slope of this linear relationship was significantly steeper in subjects with normal systolic left ventricular function (NLV) compared with HF, who showed a blunted left ventricular peak positive dP/dt (LV þdP/dtmax) response to increased heart rate (HR). The interaction between HR and the steeper HR-related rise in NLV was significant. RA, right atrial.
We examined the EDPVR in a subset of 18 patients who had adequate sonographic images: 9 in the NLV group (age 59 6 7 y, LVEF 58 6 3%) and 9 in the HF group (age 56 6 11 y, LVEF 28 6 7%). There were 4 women in the NLV group and 2 in the HF group. The single-beat method of calculating EDPVR allows for estimation of LVEDVs at a theoretic end-diastolic pressure of 0 mm Hg (V0) and 30 mm Hg (V30; Table 4). As expected, HF patients had higher measured LVEDP (Pm) and LVEDV (Vm) than NLV patients. Estimated V0 and V30 were also increased in the HF group compared with the NLV group. In the NLV group, increases in HR were associated with a fall in LVEDP but no change in V0, Vm, or V30. In the HF group, increases in HR were also associated with a fall in LVEDP. However, in contrast to the NLV group, estimated V30 increased significantly in the HF patients. These results predict a rightward shift in the EDPVR associated with increasing HR in patients with impaired systolic LV function, a behavior we did not observe in the control group (Fig. 4).
72 Journal of Cardiac Failure Vol. 21 No. 1 January 2015
Fig. 3. The relationship between contractility and early diastolic relaxation in subjects with and without HF. A linear relationship between LV þdP/dtmax and both tM and t1/2 was observed in both groups, yet the slope of this relationship was significantly different between the NLV and HF groups. Abbreviations as in Figures 1 and 2.
Discussion This study examined the acute in vivo effects of HR manipulation on LV chamber diastolic function, comparing a contemporary cohort of HF patients with systolic impairment and a control group of patients with normal LV function. The immediate mechanical effects of changes in HR on LV chamber performance are complex. Importantly, our findings illustrate that prediction of diastolic behavior can not be derived from knowledge of systolic function alone. This knowledge is clinically relevant to the administration of pharmacologic and device therapies now available for the care of the contemporary HF patient. Early Diastolic Relaxation
The attenuated effect of HR on isovolumic contractility has been consistently documented in HF patients compared with control subjects.3,4,8,9 The influence of HR on isovolumic relaxation in humans with HF in vivo has been reported
less frequently.9,10 In healthy humans, a trend or modest effect of atrial-pacing tachycardia to shorten t has been observed.9,11 In a study of a novel inotrope, Givertz et al10 examined both the force-frequency and relaxationfrequency relationships in 10 patients with HF. Consistently with our observations, increasing HR was associated with shortening of t while the LV þdP/dtmax response appeared blunted. As noted by those investigators, it was not clear whether their findings indicated preservation of the relaxation-frequency relationship, owing to absence of a control group. Our findings address this issue, demonstrating not only preservation, but apparent augmentation of the relaxation-frequency relationship in the failing LV compared with the nonfailing LV. In the intact LV chamber, inotropic and lusitropic responses are coupled by a number of mechanisms. Increased inotropy increases the elastic recoil and restorative forces of the LV chamber,12 and both isovolumic contractility and relaxation are similarly dependent on myocardial
Table 4. Hemodynamic and Sonographic Measurements at Baseline and Peak Heart Rate in NLV and HF Patients NLV (n 5 9) Baseline ESV (mL) EDV (mL) SV (mL) SVI (L/m2) CO (L/min) CI (L/min/m2) EDP (mm Hg) V0 (mL) V30 (mL)
14 84 70 38 4.7 2.5 12 45 101
6 6 6 6 6 6 6 6 6
27 22 13 6 1 0.5 4 12 27
HF (n 5 9)
Peak 20 72 51 28 5.6 3.0 7 40 97
6 6 6 6 6 6 6 6 6
24 15 15y 9 1.5 0.9 5y 9 25
Baseline 144 192 48 24 3.1 1.9 18 94 212
6 6 6 6 6 6 6 6 6
49* 53* 29 11* 2.3 1 7* 26* 56*
Peak 158 194 36 19 4.0 2.0 11 102 248
6 6 6 6 6 6 6 6 6
77* 73* 15 7* 1.7* 1* 8y 36* 90*,y
ESV, end-systolic volume; EDV, end-diastolic volume; SV, stroke volume; SVI, stroke volume index; CO, cardiac output; CI, cardiac index; EDP, end-systolic volume; V0, estimation of end-diastolic ventricular volume at a theoretic end-diastolic pressure of 0 mm Hg by means of a single-beat method; V30, estimation of end-diastolic ventricular volume at a theoretic end-diastolic pressure of 30 mm Hg by means of a single beat method. Data are presented as mean 6 SD. *P ! .05 vs control. y P ! .05 vs baseline.
Fig. 4. Prediction of the end-diastolic pressure-volume relationship (EDPVR) from baseline to peak HR in subjects with and without HF by means of the single-beat method. Estimated V0 and V30 were increased in HF compared with NLV. In NLV, increases in HR were associated with a fall in left ventricular enddiastolic pressure (LVEDP) but no change in V0, Vm, or V30. In HF, increases in HR were also associated with a fall in LVEDP. In contrast to NLV, estimated V30 increased significantly in the HF patients, predicting a rightward shift in the EDPVR associated with increasing HR in HF. Abbreviations as in Figures 1 and 2.
Relaxation-frequency Relationship in HF
calcium-handling processes.13 In the present study, contractility-relaxation coupling was clearly observed, but the relative scale of this relationship was different between groups. In the HF group, HR-mediated change in contractility was accompanied by an incrementally greater acceleration of relaxation. Our results are consistent with those of Parker et al,14 who observed a similar pattern of blunted inotropic yet preserved lusitropic responses to dobutamine stimulation in humans with HF, again different from those in control subjects. Although human HF data are sparse, they suggest that impairment of Ca2þ-handling processes may affect contractility and relaxation to different extents. End-Diastolic Pressure-Volume Relationship
The EDPVR is a quantitative description of the passive properties of the ventricular chamber as it fills. We estimated points on the EDPVR with the use of a single-beat method reported by Klotz et al.7 The method is based on the observation that the EDPVR across mammalian species and disease states can be described by a common nonlinear analytic expression. The validity of this concept has been tested in vivo in a limited number of patients,7,15 correlating well with the acquisition of multiple pressure-volume loops during manipulation of load. It has been demonstrated that the single-beat estimate of EDPVR was sufficient to discriminate the improvement of LV chamber diastolic performance observed after alcohol septal ablation in patients with hypertrophic obstructive cardiomyopathy.16 We observed that increases in HR would predict a rightward shift of the EDPVR in HF patients, which was not observed in patients with normal LV function. In other words, in patients with systolic dysfunction, higher HR was associated with a state of relative higher compliance of the LV. Similarly to our findings, Feldman et al9 observed that modest shortening of the RR interval in HF lowered LVEDP without lowering LVEDV, whereas in control subjects both LV end-diastolic dimension and LVEDP decreased. Those investigators concluded that diastolic reserve in HF may be impaired, based on anticipated limitation of gains in LV volume with longer filling period. Although the present and earlier work have tested manipulation of HR in only 1 direction (ie, increased, not decreased, from resting HR), these mechanistic data suggest that prolonging the diastolic filling period does not further optimize diastolic performance in HF. We speculate that LV compliance increases as HR increases because shortening the diastolic filling period may attenuate abnormal diastolic ventricular interactions.17 Clinical Relevance: The Balance of Heart Rate and Cardiac Function in Heart Failure
It is well understood that increases in HR offer very little advantage in terms of systolic performance in the failing LV. As our present and previous work revisit,4 there is no increase in contractility and minimal increases in cardiac output. Moreover, tachycardia is likely to exacerbate the impairment of Ca2þ-handling processes and provoke
Esfandiari et al
73
ischemia with adverse effects on diastolic performance. Our observations of the effect of HR on isovolumic relaxation and EDPVR predict that tolerance of bradycardia may also be quite limited, even in this relatively well compensated HF cohort. No gains in diastolic filling and higher filling pressures would be anticipated if HR is lowered acutely. The notion that acute reductions in HR may be poorly tolerated is supported by the observation that sick sinus syndrome can itself precipitate overt HF decompensation.18,19 This concept should be separated from the observation that chronically lower HR is associated with improved survival.20 Clearly, evidence-based HF therapies, such as beta-blockade, lower HR.21,22 However, the mechanism of benefit is attributable to effective neurohumoral blockade rather than the direct effect of changing HR per se.23 We also made our observations in a contemporary cohort of ambulatory patients in which the use of evidence-based pharmacotherapies, including angiotensin and adrenergic system antagonists, was O80%. Right heart catheterization confirmed the HF cohort to be well compensated. The HF group was clearly myopathic on the basis of significantly impaired contractility and prolonged relaxation, but well matched to the control group in terms of LV systolic and diastolic filling pressures. As such, the experimental intervention was applied at the same time that loading conditions were similar between the groups, unlike older experimental reports in which HF subjects were often much more decompensated.9,10,14 Our findings lend mechanistic support to best-practice recommendations that beta-blockers be introduced cautiously in HF patients who continue to demonstrate evidence of congestion or if resting HR is !60 beats/min.24 Similar precautions are recommended for ivabradine, which has become available to clinicians for the treatment of HF.25,26 Although the HR-lowering effect of ivabradine is attenuated as HR decreases, our findings substantiate the recommendation that ivabradine can be prescribed if resting HR is O70 beats/min.26 Such cautions allow eligible HF patients to safely receive the long-term benefits of these evidence-based therapies. Study Limitations
Estimation of EDPVR by the single-beat method is based on several assumptions. Nonetheless, this technique has been validated in humans ex vivo and in vivo and in other species.7,15 The accuracy of assessing ventricular volumes by 2D echocardiography is limited, although our main interest was relative changes in LVEDV from baseline to peak HR. Our primary goal was to evaluate diastolic LV chamber performance, and although our work has clinical implications, we did not assess clinical outcomes. We examined indices that reflect separate elements of diastole, early LV pressure decay, and the ability of the LV chamber to accommodate filling late in diastole. Although changes in these metrics were concordant, we do not suggest that these observations are interdependent. In subsequent
74 Journal of Cardiac Failure Vol. 21 No. 1 January 2015 experiments, we would propose to further evaluate the mechanisms by which changing HR may acutely affect LV chamber compliance. We hypothesize that prolongation of diastolic filling in the setting of an enlarged failing LV results in distention to the limits of pericardial constraint,27 which would explain bradycardia-mediated decline in compliance. The effects of prolonging diastolic filling period may be equally disadvantageous to the RV, particularly in patients with biventricular dysfunction, and such individuals would be most susceptible to adverse diastolic ventricular interaction. Conclusion In summary, increasing HR affects early LV relaxation and late diastolic LV performance in vivo in humans with and without HF. Our observations suggest that in HF, acute increases in HR may (1) accelerate early relaxation to a comparatively greater extent than in control subjects and (2) shift the LV EDPVR to the right. Conversely, these results also imply that abrupt HR change can be mechanically disadvantageous to the failing LV, particularly if filling pressures are high. Our findings reinforce contemporary best-practice recommendations for avoidance of bradycardia when administering negative chronotropic therapy for HF patients. Disclosures None.
References 1. Bohm M, Swedberg K, Komajda M, Borer JS, Ford I, DubostBrama A, et al. Heart rate as a risk factor in chronic heart failure (SHIFT): the association between heart rate and outcomes in a randomised placebo-controlled trial. Lancet 2010;376:886e94. 2. Savelieva I, Camm AJ. If inhibition with ivabradine: electrophysiological effects and safety. Drug Saf 2008;31:95e107. 3. Hasenfuss G, Holubarsch C, Hermann HP, Astheimer K, Pieske B, Just H. Influence of the force-frequency relationship on haemodynamics and left ventricular function in patients with non-failing hearts and in patients with dilated cardiomyopathy. Eur Heart J 1994;15:164e70. 4. Mak S, Van Spall HG, Wainstein RV, Sasson Z. Strain, strain rate, and the force frequency relationship in patients with and without heart failure. J Am Soc Echocardiogr 2012;25:341e8. 5. Weiss JL, Frederiksen JW, Weisfeldt ML. Hemodynamic determinants of the time-course of fall in canine left ventricular pressure. J Clin Invest 1976;58:751e60. 6. Mirsky I. Assessment of diastolic function: suggested methods and future considerations. Circulation 1984;69:836e41. 7. Klotz S, Hay I, Dickstein ML, Yi GH, Wang J, Maurer MS, et al. Single-beat estimation of end-diastolic pressure-volume relationship: a novel method with potential for noninvasive application. Am J Physiol Heart Circ Physiol 2006;291:H403e12. 8. Cotton JM, Kearney MT, MacCarthy PA, Grocott-Mason RM, McClean DR, Heymes C, et al. Effects of nitric oxide synthase inhibition on Basal function and the force-frequency relationship in the normal and failing human heart in vivo. Circulation 2001;104: 2318e23.
9. Feldman MD, Alderman JD, Aroesty JM, Royal HD, Ferguson JJ, Owen RM, et al. Depression of systolic and diastolic myocardial reserve during atrial pacing tachycardia in patients with dilated cardiomyopathy. J Clin Invest 1988;82:1661e9. 10. Givertz MM, Andreou C, Conrad CH, Colucci WS. Direct myocardial effects of levosimendan in humans with left ventricular dysfunction: alteration of force-frequency and relaxation-frequency relationships. Circulation 2007;115:1218e24. 11. Izawa H, Yokota M, Takeichi Y, Inagaki M, Nagata K, Iwase M, et al. Adrenergic control of the force-frequency and relaxation-frequency relations in patients with hypertrophic cardiomyopathy. Circulation 1997;96:2959e68. 12. Firstenberg MS, Smedira NG, Greenberg NL, Prior DL, McCarthy PM, Garcia MJ, et al. Relationship between early diastolic intraventricular pressure gradients, an index of elastic recoil, and improvements in systolic and diastolic function. Circulation 2001;104(12 Suppl 1):I330e5. 13. Pieske B, Maier LS, Bers DM, Hasenfuss G. Ca2þ handling and sarcoplasmic reticulum Ca2þ content in isolated failing and nonfailing human myocardium. Circ Res 1999;85:38e46. 14. Parker JD, Landzberg JS, Bittl JA, Mirsky I, Colucci WS. Effects of betaadrenergic stimulation with dobutamine on isovolumic relaxation in the normal and failing human left ventricle. Circulation 1991;84:1040e8. 15. Ten Brinke EA, Burkhoff D, Klautz RJ, Tschope C, Schalij MJ, Bax JJ, et al. Single-beat estimation of the left ventricular end-diastolic pressurevolume relationship in patients with heart failure. Heart 2010;96:213e9. 16. Meliga E, Steendijk P, Valgimigli M, ten Cate FJ, Serruys PW. Effects of percutaneous transluminal septal myocardial ablation for obstructive hypertrophic cardiomyopathy on systolic and diastolic left ventricular function assessed by pressure-volume loops. Am J Cardiol 2008;101:1179e84. 17. Atherton JJ, Moore TD, Lele SS, Thomson HL, Galbraith AJ, Belenkie I, et al. Diastolic ventricular interaction in chronic heart failure. Lancet 1997;349:1720e4. 18. Alboni P, Brignole M, Menozzi C, Scarfo S. Is sinus bradycardia a factor facilitating overt heart failure? Eur Heart J 1999;20:252e5. 19. Alboni P, Menozzi C, Brignole M, Paparella N, Gaggioli G, Lolli G, et al. Effects of permanent pacemaker and oral theophylline in sick sinus syndrome the THEOPACE study: a randomized controlled trial. Circulation 1997;96:260e6. 20. Greene SJ, Vaduganathan M, Lupi L, Ambrosy AP, Mentz RJ, Konstam MA, et al. Prognostic significance of serum total cholesterol and triglyceride levels in patients hospitalized for heart failure with reduced ejection fraction (from the EVEREST trial). Am J Cardiol 2013;111:574e81. 21. CIBIS Investigators and Committees. A randomized trial of betablockade in heart failure. The Cardiac Insufficiency Bisoprolol Study (CIBIS). Circulation 1994;90:1765e73. 22. Hjalmarson A, Goldstein S, Fagerberg B, Wedel H, Waagstein F, Kjekshus J, et al. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet 1999;353:2001e7. 23. Packer M. Beta-blockade in the management of chronic heart failure. Another step in the conceptual evolution of a neurohormonal model of the disease. Eur Heart J 1996;17(Suppl B):21e3. 24. McMurray J, Cohen-Solal A, Dietz R, Eichhorn E, Erhardt L, Hobbs FD, et al. Practical recommendations for the use of ACE inhibitors, beta-blockers, aldosterone antagonists and angiotensin receptor blockers in heart failure: putting guidelines into practice. Eur J Heart Fail 2005;7:710e21. 25. Fox K, Ford I, Steg PG, Tendera M, Ferrari R, Investigators B. Ivabradine for patients with stable coronary artery disease and leftventricular systolic dysfunction (BEAUTIFUL): a randomised, double-blind, placebo-controlled trial. Lancet 2008;372:807e16. 26. Swedberg K, Komajda M, Bohm M, Borer JS, Ford I, Dubost-Brama A, et al. Ivabradine and outcomes in chronic heart failure (SHIFT): a randomised placebo-controlled study. Lancet 2010;376:875e85. 27. Williams L, Frenneaux M. Diastolic ventricular interaction: from physiology to clinical practice. Nat Clin Pract Cardiovasc Med 2006;3:368e76.
Relaxation-frequency Relationship in HF
(ii) Calculate a from a5
EDPVR Equations: EDP 5 An EDVBn, with An 5 28.2 mm Hg and Bn 5 2.79. V05 Vm(0.6 0.006$Pm)
30 b V30
(B) Determining EDPVR for measured Pm O22 mm Hg: (i) Calculate V15 from V15 50:8ðV30 V0 Þ þ V0
V30 5 V0 þ (Vm,n Vo)/(Pm/An)(1/Bn) (A) Determining EDPVR for measured Pm #22 mm Hg:
Pm Log 30 (i) Calculate b from b5 Vm Log V30
75
#
"
Appendix
Esfandiari et al
Pm Log 15 (ii) Calculate b from b5 Vm Log V15 Pm (iii) Calculate a from a5 b Vm Use the estimates of a and b to specify the entire EDPVR by EDP5aEDV b.
Heart RateeDependent Left Ventricular Diastolic Function in Patients With and Without Heart Failure SAM ESFANDIARI, MSc,1,2 FELIPE FUCHS, MD,3 RODRIGO V. WAINSTEIN, MD,4 ANJALA CHELVANATHAN, MD,1 PETER MITOFF, MD,1 ZION SASSON, MD,1 AND SUSANNA MAK, MD, PhD1 Toronto, Ontario, Canada; and Porto Alegre, Brazil
ABSTRACT Background: Chronic heart rate (HR) reduction in the treatment of heart failure (HF) with systolic dysfunction is beneficial, but the immediate mechanical advantages or disadvantages of altering HR are incompletely understood. We examined the effects of increasing HR on early and late diastole in humans with and without HF. Methods and Results: We studied force-interval relationships of the left ventricle (LV) in 11 HF patients and 14 control subjects. HR was controlled by right atrial pacing, and LV pressure was recorded by a micromanometer-tipped catheter. The time constant of isovolumic relaxation (tau) was calculated, and simultaneous sonographic images were analyzed for LV volumes. The end-diastolic pressure-volume relationship (EDPVR) was analyzed with the use of a single-beat method. Tau was shortened in response to increasing HR in both groups; the slope of this relationship was steeper in HF than in control subjects. The predicted volume at a theoretic pressure of 0 mm Hg (V30) increased at higher HRs compared with baseline, shifting the predicted EDPVR compliance curve to the right in HF patients but not in control subjects. Conclusions: In HF, changes in HR affect early relaxation and diastolic compliance to a greater extent than in control subjects. Our study reinforces current recommendations for HR-lowering drug treatment in HF. (J Cardiac Fail 2015;21:68e75) Key Words: Heart failure, diastole, left ventricle, heart rate.
HF hospitalizations by 3% and 16%, respectively.1 The efficacy of cornerstone pharmacotherapeutic interventions, such as adrenergic antagonists, can be evaluated by reductions in HR. More recently, ivabradine, a nonselective inhibitor of the If channels in the sinoatrial node with purely negative chronotropic effects, has been clearly demonstrated to confer a survival advantage to HF patients.2 As such, current evidence-based ambulatory care for chronic HF mandates the acute manipulation of HR in these patients. Although the long-term benefits are clear, the immediate mechanical advantages or disadvantages of altering HR in HF patients are incompletely understood. It is well established that the left ventricular (LV) inotropic response to increases in HR (ie, force-frequency relationship) is blunted in the failing human myocardium.3 However, less is documented regarding the acute effects of changes in HR on the LV diastolic response (ie, relaxationfrequency) in vivo in humans with and without HF. Such data arise from older studies that included relatively decompensated HF patients. There are few data assessing the
High resting heart rate (HR) is a known marker of cardiovascular outcomes in patients with heart failure (HF) due to systolic dysfunction.1 One-beat and 5-beat increases of resting HR increase the risk of cardiovascular death and
From the 1Division of Cardiology, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada; 2Institute of Medical Science, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada; 3Division of Cardiology, Mount Sinai Hospital/University Health Network, Toronto, Ontario, Canada and 4Division of Cardiology, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil. Manuscript received June 6, 2014; revised manuscript received October 9, 2014; revised manuscript accepted October 20, 2014. Reprint requests: Susanna Mak, MD, PhD, Director, Assistant Professor, Division of Cardiology, Department of Medicine, Mount Sinai Hospital, University of Toronto, 600 University Avenue, Murray Side Rm 18-365, Toronto, Ontario M5G 1X5. Tel: þ1 416-586-4800 ext 5554; Fax: þ1 416-485-4775. E-mail: [email protected] Funding: Heart and Stroke Foundation of Ontario (grant-in aid T-7336). See page 74 for disclosure information. 1071-9164/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2014.10.013
68
Relaxation-frequency Relationship in HF
relaxation-frequency response in contemporary cohorts of HF patients. Because diastolic dysfunction is a common feature of HF, investigating the mechanical effects of changes in HR on LV diastolic performance is of clinical relevance to the administration of therapies available to HF patients. Therefore, our purpose was to examine the effects of increasing HR on early and late diastole in a cohort of humans with and without HF.
Methods Study Participants The study population consisted of patients with normal systolic LV function (NLV) and patients with systolic HF who underwent elective diagnostic cardiac catheterization at the Cardiac Catheterization Research Laboratory of the Mount Sinai Hospital. Patients agreed to participate in the research experiment after the diagnostic procedure was completed. The study was approved by the Mount Sinai Ethics Review Board for experimentation involving human subjects, and every patient gave written informed consent. Patients with NLV enrolled in this protocol were referred from outpatient clinics for assessment of chronic chest pain syndrome. All patients underwent stress nuclear perfusion imaging or stress echocardiography, and those with ischemia at low or moderate stress were excluded. Briefly, inclusion criteria for the NLV group were LV ejection fraction (LVEF) O55%, normal LV mass index documented by 2-dimensional (2D) echocardiography, no history or symptoms of HF, and LV end-diastolic pressure (LVEDP) #15 mm Hg at the time of catheterization. Patients in the HF group were recruited from a subspecialty HF clinic and were referred for right and left heart catheterization as part of the clinical evaluation. Similarly to the NLV patients, results of stress perfusion imaging or stress echocardiography were reviewed to exclude patients demonstrating ischemia at low or moderate stress. The etiology of patients with HF was nonischemic in all but 2 patients. Inclusion criteria for the HF group were LVEF !35% according to 2D echocardiography (confirmed on the day of the study) and stable symptomatic HF (New York Heart Association functional class II or III) of $3 months’ duration. General exclusion criteria were acute coronary syndromes or revascularization procedures within the preceding 6 months, ventricular pace-dependency, any rhythm other than sinus, QRS O110 ms, uncontrolled hypertension, and inability to give informed consent.
Esfandiari et al
69
condition was recorded. Two separate pacing protocols were then performed.
Pacing Protocol 1: Frequency-Dependent LV Pressure. Right atrial (RA) pacing was initiated $4e15 beats/min above the intrinsic rate to a multiple of 10 beats/min. Thereafter, HR was increased by increments of 10 beats/min every 3 minutes to a maximum of 120 beats/min or until Mobitz type 1 or 2:1 AV block occurred. If AV conduction fell below 1:1, the pacing was reduced by increments of 5 beats/min until 1:1 conduction returned. After this pacing protocol was completed, a rest period was allotted to allow LV pressure and hemodynamics to return to control conditions.
Pacing Protocol 2: Combined Sonographic and Hemodynamic Study. In this pacing run, LV pressure was recorded at the same time that echocardiography images were acquired by a research sonographer, as previously described.4 Four conditions were acquired: control and RA pacing at 80 beats/min, 100 beats/min, and 120 beats/min. Chronometers for the imaging system and hemodynamic recording system were synchronized to enable offline beat-to-beat simultaneous analysis of LV pressure and the time-stamped echocardiographic image frames. 2D and Doppler echocardiography imaging (GE Vivid 7 Imaging System, version BT03e5; GE Healthcare, Canada) and analyses were performed in accordance with the American Society of Echocardiography guidelines with the use of an M4S probe using optimized windows, and were analyzed offline with the use of a proprietary workstation (Echopac, version 7, GE Healthcare). Depending on the HR condition, images were acquired at 70e120 frames/s, with higher frame rates at higher HR conditions. Hemodynamic Analysis Electrocardiography (ECG) and LV pressure were continuously digitally recorded (1,000 Hz) online. For each study condition, LV peak positive dP/dt (LV þdP/dtmax), LV peak negative dP/dt (LV dP/dtmin), LV end-diastolic pressure (LVEDP), minimum LV pressure (LVmin), and LV systolic pressure (LVSP) were calculated offline with the use of a customized software program (Labview version 5.0; National Instruments Corp). The time constant of isovolumic relaxation (t) was also calculated by means of 2 separate methods: the monoexponential method with the use of a nonzero asymptote (tM),5 and the pressure half-time (t1/2) method.6 Arterial pressure was acquired continuously from the sidearm of the femoral sheath. These analyses were performed by a research technician blinded to the clinical status of the patient and the purpose of the study. Analysis of Frequency-Dependent LV Characteristics
Combined Cardiac Catheterization and Echocardiography Study Procedures Our study procedure has been described previously.4 All patients underwent selective coronary angiography, and HF patients underwent standard right heart catheterization from the femoral approach. Treatment with all oral medications was withheld on the morning of the investigation. Pharmacologic agents known to inhibit atrioventricular (AV) conduction were withheld 48 hours before the study day. To control HR, a 6-Fr bipolar pacing catheter was advanced to the high right atrium and aligned with a programmable stimulator (Prucka GE Cardiolab). A 7-Fr micromanometer-tipped catheter (Micro-tip Catheter; Millar Industries) was positioned in the LV for pressure measurement. A rest period was allotted after instrumentation, and a control
Protocol 1. For each HR condition, 50 consecutive beats from the 2nd minute of RA pacing were selected. If the preceding R-R interval was not within 2% of the planned pacing cycle length, the specific beat and the subsequent 5 beats were discarded. Per beat, t, LV þdP/dtmax, LV dP/dtmin, LVEDP, and LVmin were measured. The mean of these 50 cardiac cycles are reported. The relationship between HR and LV þdP/dtmax (the force-frequency relationship) as well as t (the relaxationfrequency relationship) were plotted. Protocol 2. At each HR condition, we used LV pressure recordings and sonographic images to estimate the end-diastolic pressure volume relationship (EDPVR) as described by Klotz et al.7 The calculation of EDPVR is dependent on estimates of cardiac volumes by transthoracic echocardiography. We did not
70 Journal of Cardiac Failure Vol. 21 No. 1 January 2015 calculate the EDPVR if the image quality was inadequate for these measurements. Three points of the EDPVR were estimated by means of equations listed in the Appendix. LV end-diastolic volume (LVEDV; Vm) and simultaneous LVEDP (Pm) were measured by matching the time stamp from selected sonographic images and the continuous LV pressure recording. With the use of Vm and Pm, 2 other points on the EDPVR can be estimated: V0 and V30, the theoretical volumes at pressure of 0 mm Hg and 30 mm Hg, respectively.
Statistical Analysis Data are presented as mean 6 SD. Between-group comparisons of baseline variables were made with the use of Student t test for continuous variables and Fisher exact test for categoric variables. Within-group and between-group comparisons of the effect of HR on LV hemodynamic variables within the HF and the NLV groups were assessed with the use of analysis of variance for repeated measures. To account for differences in resting HR, LV hemodynamic variables were analyzed in relation to incremental changes in HR above resting. To examine the effect of group-based differences between patients with and without HF, the relationship between HR and t was analyzed by means of an analysis of covariance with HR as a covariate. Simple linear regression analysis was used to describe the relationship between LV þdP/dtmax and t. All comparisons were based on a 95% confidence limit (P ! .05). All data were analyzed with the use of the statistical software PASW (Predictive Analytics Software) Statistics, version 19.0.
Study Population
Eleven HF patients and 14 NLV subjects without HF underwent the study protocol. Baseline characteristics are documented in Table 1. LV hemodynamic measurements are documented in Table 2, including the results of right heart catheterization in the HF population, who were relatively well compensated. Table 1. Clinical Characteristics
Age (y) Female (n) Body surface area (m2) Hemoglobin (g/L) Estimated glomerular filtration rate (mL/min) Hypertension (n) Diabetes mellitus Coronary artery disease Beta-blockers ACE inhibitors/ARBs Calcium channel blockers Diuretics Digoxin
NLV (n 5 14)
Variable Heart rate (beats/min) SAP (mm Hg) DAP (mm Hg) MAP (mm Hg) LVSP (mm Hg) LVEDP (mm Hg) LVmin (mm Hg) LV þdP/dtmax (mm Hg/s) LV dp/dtmin (mm Hg/s) tM (ms) t1/2 (ms) Mean right atrial pressure (mm Hg) Mean pulmonary arterial pressure (mm Hg) Mean capillary wedge pressure (mm Hg) Fick cardiac output (L/min)
65 6 123 6 60 6 85 6 112 6 10 6 3.9 6 1348 6 1466 6 42 6 30 6 d
HF (n 5 11)
8 21 10 13 20 5 3.5 285 304 7 6
d
72 134 73 94 115 14 6.5 939 1015 54 43 3
6 6 6 6 6 6 6 6 6 6 6 6
P Value
10 20 13 13 23 9 6.1 239 340 14 10 5
.08 .23 !.01 .10 .70 .19 .23 !.01 !.01 .02 !.01 d
18 6 7
d
d
9 6 6.5
d
d
4.7 6 1.3
d
DAP, diastolic arterial pressure; LV þdP/dtmax, left ventricular peak positive dP/dt; LV dP/dtmin, left ventricular peak negative dP/dt; LVEDP, left ventricular end-diastolic pressure; LVmin, minimum left ventricular pressure; LVSP, left ventricular systolic pressure; MAP, mean arterial pressure; SAP, systolic arterial pressure; tM, t calculated by means of the monoexponential method with the use of a nonzero asymptote; t1/2, t calculated by means of the pressure half-time method; other abbreviations as in Table 1. Data are presented as mean 6 SD.
Heart RateeDependent Effects on the LV Pressure Waveform
All patients achieved $4 RA-paced conditions above their resting HR (Table 3). For both groups, increases in
Results
Variable
Table 2. Right Heart and Left Heart Hemodynamics
NLV (n 5 14)
HF (n 5 11)
P Value
59 6 7 6 1.93 6 0.19 140 6 18 80.9 6 19.7
54 6 12 4 1.96 6 0.26 142 6 19 84.5 6 25.5
.21 1.00 .71 .71 .70
6 3 2 6 6 3 4 0
4 1 3 10 9 0 6 2
1.00 .60 .62 .03 .10 .23 .24 .18
ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; HF, heart failure; NLV, normal left ventricular function. Data are presented as mean 6 SD or n.
Table 3. Heart RateeDependent Effects on the Left Ventricular Pressure Waveform Variable
þ10
þ20
SAP (mm Hg)* NLV 124 6 26 125 HF 135 6 19 133 DAP* NLV 65 6 11 68 HF 81 6 21 83 MAP* NLV 89 6 16 92 HF 97 6 13 99 LVEDP* NLV 965 7 HF 15 6 11 12 LV þdp/dtmax (mm Hg/s)*,y NLV 1,453 6 307 1,537 HF 993 6 244 1,030 ,y tM (ms)* NLV 40 6 7 39 HF 53 6 15 50 t1/2 (ms)*,y NLV 29 6 5 28 HF 41 6 10 40
þ30
þ40
6 25 6 20
124 6 27 132 6 18
121 6 27 129 6 18
6 12 6 20
70 6 13 86 6 20
72 6 13 87 6 19
6 17 6 14
94 6 18 101 6 14
94 6 17 100 6 15
64 6 10
664 11 6 9
564 968
6 333 6 239
1,638 6 362 1,072 6 246
1,717 6 385 1,083 6 241
67 6 15
37 6 6 48 6 15
38 6 8 46 6 13
65 6 10
28 6 4 39 6 10
28 6 4 37 6 10
Abbreviations as in Tables 1 and 2. Data are presented as mean 6 SD. *Effect of HR. y Interaction of group and HR.
Relaxation-frequency Relationship in HF
Esfandiari et al
71
HR were associated with reductions in systolic arterial pressure and LVEDP, whereas diastolic arterial pressure and mean arterial pressure increased. LVSP decreased at the highest pacing rates. LVmin did not change with RA pacing, but was significantly higher in HF patients at each pacing rate. The force-frequency relationship was assessed by inscribing the effect of pacing-controlled HR on LV þdP/ dtmax. As expected, the slope of this linear relationship was significantly steeper in the NLV group than in the HF patients, who demonstrated a blunted LV þdP/dtmax response to increased HR. The interaction between HR and the studied group with a steeper HR-related rise in NLV group was significant (Fig. 1). The relaxation-frequency relationship was assessed by examining the effect of pacing-controlled increases in HR on t calculated by 2 separate methods (tM and t1/2; Fig. 2). We observed significant shortening of t in response to increasing HR in both groups, In contrast to the forcefrequency relationship, we observed that the slope of this relationship was significantly steeper in HF patients than in the NLV group. Contractility-Relaxation Coupling
The contrasting effects of HR on LV þdP/dtmax and t suggests that coupling between contractility and early diastolic relaxation was also different between the NLV and HF groups. We therefore examined the relationship between LV þdP/dtmax and t measured by 2 methods across the range of HRs. A linear relationship between LV þdP/ dtmax and both tM and t1/2 was observed in both groups. However, the slope of this relationship was significantly different between the NLV and HF groups. In the HF group, the inverse relationship was significantly steeper, such that relatively small positive differences in contractile state were
Fig. 2. The relaxation-frequency relationship in subjects with and without heart failure. t was significantly shortened in response to increasing HR in both groups. However, the slope of this relationship was significantly steeper in HF compared to NLV. tM, t calculated by means of the monoexponential method with the use of a nonzero asymptote; t1/2, t calculated by means of the pressure half-time method; other abbreviations as in Figure 1.
associated with incrementally greater shortening of both tM and t1/2 (Fig. 3). HR RateeDependent Effects on the Events of Late Diastole
Fig. 1. The force-frequency relationship in subjects with and without heart failure (HF). The slope of this linear relationship was significantly steeper in subjects with normal systolic left ventricular function (NLV) compared with HF, who showed a blunted left ventricular peak positive dP/dt (LV þdP/dtmax) response to increased heart rate (HR). The interaction between HR and the steeper HR-related rise in NLV was significant. RA, right atrial.
We examined the EDPVR in a subset of 18 patients who had adequate sonographic images: 9 in the NLV group (age 59 6 7 y, LVEF 58 6 3%) and 9 in the HF group (age 56 6 11 y, LVEF 28 6 7%). There were 4 women in the NLV group and 2 in the HF group. The single-beat method of calculating EDPVR allows for estimation of LVEDVs at a theoretic end-diastolic pressure of 0 mm Hg (V0) and 30 mm Hg (V30; Table 4). As expected, HF patients had higher measured LVEDP (Pm) and LVEDV (Vm) than NLV patients. Estimated V0 and V30 were also increased in the HF group compared with the NLV group. In the NLV group, increases in HR were associated with a fall in LVEDP but no change in V0, Vm, or V30. In the HF group, increases in HR were also associated with a fall in LVEDP. However, in contrast to the NLV group, estimated V30 increased significantly in the HF patients. These results predict a rightward shift in the EDPVR associated with increasing HR in patients with impaired systolic LV function, a behavior we did not observe in the control group (Fig. 4).
72 Journal of Cardiac Failure Vol. 21 No. 1 January 2015
Fig. 3. The relationship between contractility and early diastolic relaxation in subjects with and without HF. A linear relationship between LV þdP/dtmax and both tM and t1/2 was observed in both groups, yet the slope of this relationship was significantly different between the NLV and HF groups. Abbreviations as in Figures 1 and 2.
Discussion This study examined the acute in vivo effects of HR manipulation on LV chamber diastolic function, comparing a contemporary cohort of HF patients with systolic impairment and a control group of patients with normal LV function. The immediate mechanical effects of changes in HR on LV chamber performance are complex. Importantly, our findings illustrate that prediction of diastolic behavior can not be derived from knowledge of systolic function alone. This knowledge is clinically relevant to the administration of pharmacologic and device therapies now available for the care of the contemporary HF patient. Early Diastolic Relaxation
The attenuated effect of HR on isovolumic contractility has been consistently documented in HF patients compared with control subjects.3,4,8,9 The influence of HR on isovolumic relaxation in humans with HF in vivo has been reported
less frequently.9,10 In healthy humans, a trend or modest effect of atrial-pacing tachycardia to shorten t has been observed.9,11 In a study of a novel inotrope, Givertz et al10 examined both the force-frequency and relaxationfrequency relationships in 10 patients with HF. Consistently with our observations, increasing HR was associated with shortening of t while the LV þdP/dtmax response appeared blunted. As noted by those investigators, it was not clear whether their findings indicated preservation of the relaxation-frequency relationship, owing to absence of a control group. Our findings address this issue, demonstrating not only preservation, but apparent augmentation of the relaxation-frequency relationship in the failing LV compared with the nonfailing LV. In the intact LV chamber, inotropic and lusitropic responses are coupled by a number of mechanisms. Increased inotropy increases the elastic recoil and restorative forces of the LV chamber,12 and both isovolumic contractility and relaxation are similarly dependent on myocardial
Table 4. Hemodynamic and Sonographic Measurements at Baseline and Peak Heart Rate in NLV and HF Patients NLV (n 5 9) Baseline ESV (mL) EDV (mL) SV (mL) SVI (L/m2) CO (L/min) CI (L/min/m2) EDP (mm Hg) V0 (mL) V30 (mL)
14 84 70 38 4.7 2.5 12 45 101
6 6 6 6 6 6 6 6 6
27 22 13 6 1 0.5 4 12 27
HF (n 5 9)
Peak 20 72 51 28 5.6 3.0 7 40 97
6 6 6 6 6 6 6 6 6
24 15 15y 9 1.5 0.9 5y 9 25
Baseline 144 192 48 24 3.1 1.9 18 94 212
6 6 6 6 6 6 6 6 6
49* 53* 29 11* 2.3 1 7* 26* 56*
Peak 158 194 36 19 4.0 2.0 11 102 248
6 6 6 6 6 6 6 6 6
77* 73* 15 7* 1.7* 1* 8y 36* 90*,y
ESV, end-systolic volume; EDV, end-diastolic volume; SV, stroke volume; SVI, stroke volume index; CO, cardiac output; CI, cardiac index; EDP, end-systolic volume; V0, estimation of end-diastolic ventricular volume at a theoretic end-diastolic pressure of 0 mm Hg by means of a single-beat method; V30, estimation of end-diastolic ventricular volume at a theoretic end-diastolic pressure of 30 mm Hg by means of a single beat method. Data are presented as mean 6 SD. *P ! .05 vs control. y P ! .05 vs baseline.
Fig. 4. Prediction of the end-diastolic pressure-volume relationship (EDPVR) from baseline to peak HR in subjects with and without HF by means of the single-beat method. Estimated V0 and V30 were increased in HF compared with NLV. In NLV, increases in HR were associated with a fall in left ventricular enddiastolic pressure (LVEDP) but no change in V0, Vm, or V30. In HF, increases in HR were also associated with a fall in LVEDP. In contrast to NLV, estimated V30 increased significantly in the HF patients, predicting a rightward shift in the EDPVR associated with increasing HR in HF. Abbreviations as in Figures 1 and 2.
Relaxation-frequency Relationship in HF
calcium-handling processes.13 In the present study, contractility-relaxation coupling was clearly observed, but the relative scale of this relationship was different between groups. In the HF group, HR-mediated change in contractility was accompanied by an incrementally greater acceleration of relaxation. Our results are consistent with those of Parker et al,14 who observed a similar pattern of blunted inotropic yet preserved lusitropic responses to dobutamine stimulation in humans with HF, again different from those in control subjects. Although human HF data are sparse, they suggest that impairment of Ca2þ-handling processes may affect contractility and relaxation to different extents. End-Diastolic Pressure-Volume Relationship
The EDPVR is a quantitative description of the passive properties of the ventricular chamber as it fills. We estimated points on the EDPVR with the use of a single-beat method reported by Klotz et al.7 The method is based on the observation that the EDPVR across mammalian species and disease states can be described by a common nonlinear analytic expression. The validity of this concept has been tested in vivo in a limited number of patients,7,15 correlating well with the acquisition of multiple pressure-volume loops during manipulation of load. It has been demonstrated that the single-beat estimate of EDPVR was sufficient to discriminate the improvement of LV chamber diastolic performance observed after alcohol septal ablation in patients with hypertrophic obstructive cardiomyopathy.16 We observed that increases in HR would predict a rightward shift of the EDPVR in HF patients, which was not observed in patients with normal LV function. In other words, in patients with systolic dysfunction, higher HR was associated with a state of relative higher compliance of the LV. Similarly to our findings, Feldman et al9 observed that modest shortening of the RR interval in HF lowered LVEDP without lowering LVEDV, whereas in control subjects both LV end-diastolic dimension and LVEDP decreased. Those investigators concluded that diastolic reserve in HF may be impaired, based on anticipated limitation of gains in LV volume with longer filling period. Although the present and earlier work have tested manipulation of HR in only 1 direction (ie, increased, not decreased, from resting HR), these mechanistic data suggest that prolonging the diastolic filling period does not further optimize diastolic performance in HF. We speculate that LV compliance increases as HR increases because shortening the diastolic filling period may attenuate abnormal diastolic ventricular interactions.17 Clinical Relevance: The Balance of Heart Rate and Cardiac Function in Heart Failure
It is well understood that increases in HR offer very little advantage in terms of systolic performance in the failing LV. As our present and previous work revisit,4 there is no increase in contractility and minimal increases in cardiac output. Moreover, tachycardia is likely to exacerbate the impairment of Ca2þ-handling processes and provoke
Esfandiari et al
73
ischemia with adverse effects on diastolic performance. Our observations of the effect of HR on isovolumic relaxation and EDPVR predict that tolerance of bradycardia may also be quite limited, even in this relatively well compensated HF cohort. No gains in diastolic filling and higher filling pressures would be anticipated if HR is lowered acutely. The notion that acute reductions in HR may be poorly tolerated is supported by the observation that sick sinus syndrome can itself precipitate overt HF decompensation.18,19 This concept should be separated from the observation that chronically lower HR is associated with improved survival.20 Clearly, evidence-based HF therapies, such as beta-blockade, lower HR.21,22 However, the mechanism of benefit is attributable to effective neurohumoral blockade rather than the direct effect of changing HR per se.23 We also made our observations in a contemporary cohort of ambulatory patients in which the use of evidence-based pharmacotherapies, including angiotensin and adrenergic system antagonists, was O80%. Right heart catheterization confirmed the HF cohort to be well compensated. The HF group was clearly myopathic on the basis of significantly impaired contractility and prolonged relaxation, but well matched to the control group in terms of LV systolic and diastolic filling pressures. As such, the experimental intervention was applied at the same time that loading conditions were similar between the groups, unlike older experimental reports in which HF subjects were often much more decompensated.9,10,14 Our findings lend mechanistic support to best-practice recommendations that beta-blockers be introduced cautiously in HF patients who continue to demonstrate evidence of congestion or if resting HR is !60 beats/min.24 Similar precautions are recommended for ivabradine, which has become available to clinicians for the treatment of HF.25,26 Although the HR-lowering effect of ivabradine is attenuated as HR decreases, our findings substantiate the recommendation that ivabradine can be prescribed if resting HR is O70 beats/min.26 Such cautions allow eligible HF patients to safely receive the long-term benefits of these evidence-based therapies. Study Limitations
Estimation of EDPVR by the single-beat method is based on several assumptions. Nonetheless, this technique has been validated in humans ex vivo and in vivo and in other species.7,15 The accuracy of assessing ventricular volumes by 2D echocardiography is limited, although our main interest was relative changes in LVEDV from baseline to peak HR. Our primary goal was to evaluate diastolic LV chamber performance, and although our work has clinical implications, we did not assess clinical outcomes. We examined indices that reflect separate elements of diastole, early LV pressure decay, and the ability of the LV chamber to accommodate filling late in diastole. Although changes in these metrics were concordant, we do not suggest that these observations are interdependent. In subsequent
74 Journal of Cardiac Failure Vol. 21 No. 1 January 2015 experiments, we would propose to further evaluate the mechanisms by which changing HR may acutely affect LV chamber compliance. We hypothesize that prolongation of diastolic filling in the setting of an enlarged failing LV results in distention to the limits of pericardial constraint,27 which would explain bradycardia-mediated decline in compliance. The effects of prolonging diastolic filling period may be equally disadvantageous to the RV, particularly in patients with biventricular dysfunction, and such individuals would be most susceptible to adverse diastolic ventricular interaction. Conclusion In summary, increasing HR affects early LV relaxation and late diastolic LV performance in vivo in humans with and without HF. Our observations suggest that in HF, acute increases in HR may (1) accelerate early relaxation to a comparatively greater extent than in control subjects and (2) shift the LV EDPVR to the right. Conversely, these results also imply that abrupt HR change can be mechanically disadvantageous to the failing LV, particularly if filling pressures are high. Our findings reinforce contemporary best-practice recommendations for avoidance of bradycardia when administering negative chronotropic therapy for HF patients. Disclosures None.
References 1. Bohm M, Swedberg K, Komajda M, Borer JS, Ford I, DubostBrama A, et al. Heart rate as a risk factor in chronic heart failure (SHIFT): the association between heart rate and outcomes in a randomised placebo-controlled trial. Lancet 2010;376:886e94. 2. Savelieva I, Camm AJ. If inhibition with ivabradine: electrophysiological effects and safety. Drug Saf 2008;31:95e107. 3. Hasenfuss G, Holubarsch C, Hermann HP, Astheimer K, Pieske B, Just H. Influence of the force-frequency relationship on haemodynamics and left ventricular function in patients with non-failing hearts and in patients with dilated cardiomyopathy. Eur Heart J 1994;15:164e70. 4. Mak S, Van Spall HG, Wainstein RV, Sasson Z. Strain, strain rate, and the force frequency relationship in patients with and without heart failure. J Am Soc Echocardiogr 2012;25:341e8. 5. Weiss JL, Frederiksen JW, Weisfeldt ML. Hemodynamic determinants of the time-course of fall in canine left ventricular pressure. J Clin Invest 1976;58:751e60. 6. Mirsky I. Assessment of diastolic function: suggested methods and future considerations. Circulation 1984;69:836e41. 7. Klotz S, Hay I, Dickstein ML, Yi GH, Wang J, Maurer MS, et al. Single-beat estimation of end-diastolic pressure-volume relationship: a novel method with potential for noninvasive application. Am J Physiol Heart Circ Physiol 2006;291:H403e12. 8. Cotton JM, Kearney MT, MacCarthy PA, Grocott-Mason RM, McClean DR, Heymes C, et al. Effects of nitric oxide synthase inhibition on Basal function and the force-frequency relationship in the normal and failing human heart in vivo. Circulation 2001;104: 2318e23.
9. Feldman MD, Alderman JD, Aroesty JM, Royal HD, Ferguson JJ, Owen RM, et al. Depression of systolic and diastolic myocardial reserve during atrial pacing tachycardia in patients with dilated cardiomyopathy. J Clin Invest 1988;82:1661e9. 10. Givertz MM, Andreou C, Conrad CH, Colucci WS. Direct myocardial effects of levosimendan in humans with left ventricular dysfunction: alteration of force-frequency and relaxation-frequency relationships. Circulation 2007;115:1218e24. 11. Izawa H, Yokota M, Takeichi Y, Inagaki M, Nagata K, Iwase M, et al. Adrenergic control of the force-frequency and relaxation-frequency relations in patients with hypertrophic cardiomyopathy. Circulation 1997;96:2959e68. 12. Firstenberg MS, Smedira NG, Greenberg NL, Prior DL, McCarthy PM, Garcia MJ, et al. Relationship between early diastolic intraventricular pressure gradients, an index of elastic recoil, and improvements in systolic and diastolic function. Circulation 2001;104(12 Suppl 1):I330e5. 13. Pieske B, Maier LS, Bers DM, Hasenfuss G. Ca2þ handling and sarcoplasmic reticulum Ca2þ content in isolated failing and nonfailing human myocardium. Circ Res 1999;85:38e46. 14. Parker JD, Landzberg JS, Bittl JA, Mirsky I, Colucci WS. Effects of betaadrenergic stimulation with dobutamine on isovolumic relaxation in the normal and failing human left ventricle. Circulation 1991;84:1040e8. 15. Ten Brinke EA, Burkhoff D, Klautz RJ, Tschope C, Schalij MJ, Bax JJ, et al. Single-beat estimation of the left ventricular end-diastolic pressurevolume relationship in patients with heart failure. Heart 2010;96:213e9. 16. Meliga E, Steendijk P, Valgimigli M, ten Cate FJ, Serruys PW. Effects of percutaneous transluminal septal myocardial ablation for obstructive hypertrophic cardiomyopathy on systolic and diastolic left ventricular function assessed by pressure-volume loops. Am J Cardiol 2008;101:1179e84. 17. Atherton JJ, Moore TD, Lele SS, Thomson HL, Galbraith AJ, Belenkie I, et al. Diastolic ventricular interaction in chronic heart failure. Lancet 1997;349:1720e4. 18. Alboni P, Brignole M, Menozzi C, Scarfo S. Is sinus bradycardia a factor facilitating overt heart failure? Eur Heart J 1999;20:252e5. 19. Alboni P, Menozzi C, Brignole M, Paparella N, Gaggioli G, Lolli G, et al. Effects of permanent pacemaker and oral theophylline in sick sinus syndrome the THEOPACE study: a randomized controlled trial. Circulation 1997;96:260e6. 20. Greene SJ, Vaduganathan M, Lupi L, Ambrosy AP, Mentz RJ, Konstam MA, et al. Prognostic significance of serum total cholesterol and triglyceride levels in patients hospitalized for heart failure with reduced ejection fraction (from the EVEREST trial). Am J Cardiol 2013;111:574e81. 21. CIBIS Investigators and Committees. A randomized trial of betablockade in heart failure. The Cardiac Insufficiency Bisoprolol Study (CIBIS). Circulation 1994;90:1765e73. 22. Hjalmarson A, Goldstein S, Fagerberg B, Wedel H, Waagstein F, Kjekshus J, et al. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet 1999;353:2001e7. 23. Packer M. Beta-blockade in the management of chronic heart failure. Another step in the conceptual evolution of a neurohormonal model of the disease. Eur Heart J 1996;17(Suppl B):21e3. 24. McMurray J, Cohen-Solal A, Dietz R, Eichhorn E, Erhardt L, Hobbs FD, et al. Practical recommendations for the use of ACE inhibitors, beta-blockers, aldosterone antagonists and angiotensin receptor blockers in heart failure: putting guidelines into practice. Eur J Heart Fail 2005;7:710e21. 25. Fox K, Ford I, Steg PG, Tendera M, Ferrari R, Investigators B. Ivabradine for patients with stable coronary artery disease and leftventricular systolic dysfunction (BEAUTIFUL): a randomised, double-blind, placebo-controlled trial. Lancet 2008;372:807e16. 26. Swedberg K, Komajda M, Bohm M, Borer JS, Ford I, Dubost-Brama A, et al. Ivabradine and outcomes in chronic heart failure (SHIFT): a randomised placebo-controlled study. Lancet 2010;376:875e85. 27. Williams L, Frenneaux M. Diastolic ventricular interaction: from physiology to clinical practice. Nat Clin Pract Cardiovasc Med 2006;3:368e76.
Relaxation-frequency Relationship in HF
(ii) Calculate a from a5
EDPVR Equations: EDP 5 An EDVBn, with An 5 28.2 mm Hg and Bn 5 2.79. V05 Vm(0.6 0.006$Pm)
30 b V30
(B) Determining EDPVR for measured Pm O22 mm Hg: (i) Calculate V15 from V15 50:8ðV30 V0 Þ þ V0
V30 5 V0 þ (Vm,n Vo)/(Pm/An)(1/Bn) (A) Determining EDPVR for measured Pm #22 mm Hg:
Pm Log 30 (i) Calculate b from b5 Vm Log V30
75
#
"
Appendix
Esfandiari et al
Pm Log 15 (ii) Calculate b from b5 Vm Log V15 Pm (iii) Calculate a from a5 b Vm Use the estimates of a and b to specify the entire EDPVR by EDP5aEDV b.
