Novel Neuromodulation Approach to Improve Left Ventricular Contractility in Heart Failure

Introduction

Approximately 5 million people in the United States suffer from congestive heart failure with over 1 million hospitalizations annually.¹ Heart failure is the leading discharge diagnosis in those over 65 years of age, and it is the major cost driver associated with treating congestive heart failure.¹ These hospitalizations are often due to worsening hemodynamic function and fluid retention which ultimately leads to pulmonary congestion. Symptom relief is the current standard of care in the patient with acutely decompensated heart failure (ADHF), primarily using intravenous diuretics.² Although symptom relief is often achieved before discharge, in-hospital morbidity and mortality in addition to 30-day readmission rates remain high in this population.

It has been suggested that the use of an acute inotrope can help stabilize hemodynamics while pharmacologically reducing both preload and afterload in patients with ADHF. Intravenous inotropic agents are known to increase myocardial contractility via elevation of myocyte calcium concentrations.³ However, the negative side effects of increasing cytosolic calcium concentrations—including increased heart rate, myocardial oxygen debt, cardiac arrhythmias, and end organ ischemia—severely curtail intravenous inotropic use. There does not appear to be any sympathomimetic drug that can be delivered systemically that can increase the inotropic state of the heart without affecting these parameters, as well as repolarization and systemic vascular resistance.

It has been hypothesized that autonomic innervation at the level of the heart is differentiated between chronotropic and inotropic effects.⁴ Animal data substantiates this differential innervation pattern of the cardiac autonomic nerves. In previous studies of normal animals, an epicardial approach for stimulation of distal cardiac nerves caused an increase in stroke work without a substantial effect on heart rate.⁵ However, during coronary artery bypass surgery, epicardial stimulation of visible nerves in humans, although demonstrated to be safe, did not show a consistent pattern of increasing inotropic state: the heart rate increased in 8 subjects and decreased in 8.⁶ Furthermore, in 12 subjects, stimulation increased mean aortic pressure, while decreasing in 8 others. The authors concluded that these variable effects on hemodynamics and heart rate by epicardial cardiac neuronal stimulation could be explained by the effects of stimulating different branches of the autonomic nervous system. Since these variable effects were not observed preclinically when the cardiac autonomic nerves were stimulated using an intravascular approach, we hypothesized that if an intravascular stimulation approach was restudied in humans, the inotropic effect would be more reproducible.

Methods

Because of the sensitive nature of the data collected for this study, requests to access the data set from qualified researchers trained in human subject confidentiality protocols may be sent to Cardionomic at info@cardionomicinc.com.

Subject Population

Subjects undergoing either the implantation of a cardiac resynchronization device for the treatment of heart failure or a prophylactic implantable cardioverter defibrillator were enrolled. Additional inclusion criteria included subjects with a left ventricular ejection fraction ≤35%, currently receiving oral pharmacological management per American College of Cardiology/American Heart Association treatment guidelines for congestive heart failure. Subjects were excluded for (1) atrial arrhythmias at the time of implant; (2) unstable angina or evidence of ST-segment elevation on preprocedure ECG; (3) severe mitral valve regurgitation on preoperative transthoracic echo; or (4) prior thoracic surgery. This study was conducted before implantation and was approved by Homolka Hospital’s Ethics Committee and all subjects signed an informed consent before enrollment.

Protocol

Before device implantation, patients were lightly sedated, and 8F introducers were inserted into the right femoral vein and left femoral artery. Via the left femoral artery, a multipurpose guide catheter was advanced across the aortic valve and two 2.5F Mikro-Cath pressure transducers (Millar, Inc, Houston, TX) were placed in the left ventricle (LV). The guide catheter was retracted into the aortic arch and, under pressure guidance, one of the transducers was placed at the aortic root. Left ventricular and arterial pressure waveform data was acquired using LabChart (Ver. 7) software and PowerLabs data hardware (ADI Inc. Colorado Springs, CO). A 7F Agilis steerable guide catheter (St. Jude Medical, St. Paul, MN) was advanced to the right ventricular outflow tract. Either a 20 pole Lasso catheter (BioSense Webster Inc, Diamond Bar, CA) or a 64 pole Constellation catheter (Boston Scientific Inc, Maple Grove, MN) was advanced into either the left or right pulmonary artery (PA). Standard anatomic mapping was performed using either the Carto (BioSense Webster Inc, Diamond Bar, CA) or EnSite NavX (Abbott Inc, Abbott Park, IL) imaging system. An anatomic map of the PA from the level of the right ventricular outflow tract to the distal branching of both the left and right PA was created (Figure 1). The mapping catheter was then placed in the right PA at the bifurcation and bipolar simulation was performed at 20 Hz, with a pulse width of 4 ms using a Micropace EPS320 cardiac stimulator (Micropace EP Inc, Santa Ana, CA). Stimulation amplitude began at 2 mA and increased to a maximum of 20 mA. If no effect was observed, the catheter was repositioned, or a different electrode pair was chosen, and the stimulation protocol was repeated. A successful effect was deemed to be an increase in LV max positive dP/dt by a minimum of 5% from a nonstimulation baseline. If time allowed stimulation was turned off and once the pressure parameters returned to baseline stimulation was resumed to assess repeatability of the effect.

• Download figure
• Download PowerPoint

Hemodynamic Assessment

Ventricular and aortic pressure waveforms were exported from LabChart at a rate of 2000 samples per second. Heart rate detection as well as beat-to-beat hemodynamic measurements were postprocessed using MATLAB (V2014 and Mathworks, Natick MA). Calculation of heart rate was performed using the ventricular and aortic pressure waveforms. While ECG data was collected, stimulation artifacts often overwhelmed the recordings. In contrast, the pressure recordings were unaffected by stimulation. Heartbeat detection was, therefore, defined as the point when ventricular pressure first exceeded aortic pressure. Systolic and diastolic pressures for both aortic and ventricular pressure waveforms were obtained after the waveform was passed through a 10 and 30 Hz Butterworth lowpass filter. The single maximum and minimum pressures for each beat on each waveform were taken. The mean arterial pressure was calculated from all pressure samples from end diastolic to end diastolic of the subsequent beat. Ventricular dP/dt was calculated using the sample by sample difference of the raw ventricular pressure waveform. For max positive and max negative dP/dt measurements, the raw dP/dt waveform was smoothed first using a 9-sample median filter to remove valve closure artifacts, then using a 30 Hz Butterworth lowpass filter before the maximum and minimum values were taken. After hemodynamic measurements were made for each beat, measurements taken from preventricular contractions, beats following preventricular contractions as well as from other artifacts such as measurements made during catheter repositioning were eliminated using clustering techniques. For each subject, a scatter plot of the beat-to-beat interval and pulse pressure over the entire stimulation period was generated. Clusters of beats to be included in the analysis versus outlier measurements were visually identified.

Baseline and Stimulation Data

Mean and SDs for maximum positive dP/dt (dP/dt_Max), maximum negative dP/dt (dP/dt_Min), LV systolic pressure, left ventricular diastolic pressure, aortic systolic pressure, aortic diastolic pressure, mean arterial pressure, and heart rate were calculated at nonstimulation baseline and peak dP/dt_Max, during stimulation. For each subject, the baseline was considered to be all usable pressure waves 30 seconds before stimulation. Peak stimulation was determined to be the largest change in dP/dt_Max from all usable pressure waves over a 30 second period. The data was then normalized as a percentage change from baseline to account for subject to subject variation in baseline hemodynamics.

Statistical Methods

Baseline and peak stimulation subject mean values were summarized via sample means and SDs. Each subject’s percent change from baseline to peak stimulation was also calculated to facilitate hypothesis testing via nonparametric methods given the small sample size. The Wilcoxon Signed Rank Test⁷ was used to generate 2-sided P values testing indicating if there is evidence that the population median percent change from baseline can be considered significantly different from 0. P values <0.05 were considered evidence of therapeutic effect. The 1-Sample Wilcoxon procedure in Minitab 17 was used to provide a nonparametric estimate of the median percent change from baseline and to generate the tests of significance. Summary statistics were calculated using Microsoft Excel 2013.

Results

A total of 15 subjects signed the informed consent. It was subsequently realized that one subject had prior thoracic surgery and was not studied due to abnormal PA anatomy. Three additional subjects were not studied due to a technical failure of the electro anatomic mapping system malfunctioning (1) or failure of the system to collect the data (2). One subject had insufficient pulse pressure at the time of the study, and therefore, data collection was aborted.

Of the 10 subjects who underwent the full stimulation protocol, the average age was 69.2±6.2 years. Four were women, and all subjects were New York Heart Association Class III at the time of the study with an average left ventricular ejection fraction of 30±5%. All subjects were taking a loop diuretic and an angiotensin converting enzyme-inhibitor or angiotensin antagonist. Nine of the 10 subjects were taking a β-blocker, with 5 on the mixed antagonist carvedilol and the remainder equally split between bisoprolol and metoprolol.

Effectiveness

Of the 10 studied subjects, stimulation began at 2 mA and increased, or the catheter was repositioned until a contractility effect was observed. In all subjects, a response was observed. As seen in Figure 2, an example response observed in one subject, it was typical that a visually observable response with a rapid rise in dP/dt_Max accompanied by a rise in mean arterial pressure occurred within 30 seconds of stimulation onset, but often took a full minute to reach a plateau. As observed in the middle panel, dP/dt_Max increased and reached a plateau within 1 minute, and it was maintained for the duration of the stimulation. It was difficult to hold the electrophysiology catheter in place, and it was not uncommon for it to dislodge multiple times. When dislodgements occurred, the catheter was quickly repositioned as close to the original site as possible.

• Download figure
• Download PowerPoint

Table displays the summary statistics and results of the statistical testing performed, while Figure 3 displays each subject’s change in LV dP/dt_Max from baseline to stimulation response. Cardiac contractility as measured by dP/dt_Max increased from baseline by an average of 22.6% (P=0.006). The rate of relaxation, as measured by dP/dt_Min, also increased 13.6% (P=0.041) from baseline. As would be expected from an increase in contractility, both left ventricular systolic and systemic systolic pressure increased 15.8% (P=0.006) and 11.4% (P=0.011), respectively. There was a significant 13.5% (P=0.006) increase in mean arterial pressure due primarily to the increase in systolic pressure. There was a slight tendency for the heart rate to increase on average but this increase (absolute mean change of 2.6 beats per minute) was neither clinically nor statistically significant (P=0.19). Additionally, there was a nonsignificant increase in LV diastolic pressure. However, the absolute change from baseline (median change in LV diastolic pressure of 0.3 mm Hg) was not clinically significant.

• Download figure
• Download PowerPoint

Safety

Of the 10 studied subjects, 6 adverse events were reported in 6 subjects. There were no adverse events reported in the subjects who were exited before the procedure. None of the adverse events were classified as serious but all were related to the procedure. While searching for the effective site, atrial flutter was reported in one subject, but it resolved without intervention. Cough, while searching for the effective site was also observed in four subjects, although no intervention was required. Cough was not observed at sites of effective stimulation. An arteriovenous fistula in the groin was reported in one subject 1 week after the procedure; however, no treatment was prescribed. No arrhythmias were observed when the stimulation produced an increase in contractility.

Discussion

The cardiovascular system is richly innervated with both sympathetic and parasympathetic fibers. Sympathetic fibers that innervate the heart originate from stellate and thoracic sympathetic ganglia and are responsible for increases in the chronotropic, lusitropic, and inotropic, state of the heart. Studies performed in human cadavers demonstrate that the fibers responsible for the lusitropic and inotropic state of the ventricles pass along the posterior surface of the right and main PA.⁸

Multiple studies have demonstrated that stimulation of the stellate and thoracic sympathetic ganglia increase LV contractility, as measured by an increase in LV dP/dt_Max.^5,9–15 These increases in LV contractile performance were usually accompanied by increases in heart rate, which were greater with right-sided than left-sided ganglia stimulation. Using speckle tracking echocardiography in a porcine model, increases in contractility during left and right stellate ganglia stimulation was accompanied by significantly increased global epicardial and endocardial LV rotation and diastolic untwisting rate; similar to the increase in both max positive and max negative dP/dt in our study.¹⁵

Similar increases in contractility were observed in subjects undergoing left stellate ganglia epivascular stimulation just before sympathectomy for palmar hyperhidrosis.¹⁶ Baseline left ventricular ejection fraction (measured by M-mode echocardiography), heart rate and systolic blood pressure was 54.7%, 65 beats per minute and 115 mm Hg, and increased to 62.8%, 73 beats per minute and 123 mm Hg after 45 seconds of nerve stimulation. In this study, stimulation of the distal branches of these fibers closer to the heart appears to abolish the previously observed increases in heart rate.

We tested a novel therapy by selectively stimulating the nerve fibers on the posterior surface of the right PA from an endovascular approach using standard electrophysiology catheters and stimulators. Increases in the inotropic and lusitropic state to the heart were inferred from changes in LV dP/dt. In the 10 subjects in this feasibility study, LV dP/dt_Max increased by >20% on average, while the rate of relaxation dP/dt_Min increased by 14%. There was no statistically or clinically significant change in heart rate during the time that contractility increased. These results suggest that the nerves passing along the posterior surface of the PA are predominately nerve fibers with little or no innervation to the sinoatrial node. The increase in both the inotropic and lusitropic state of the LV would also suggest that the effect on the ventricle was predominately due to an increase in sympathetic outflow.

Importantly, all subjects in this study had heart failure with reduced ejection fraction. Most were receiving β-blockers which did not seem to inhibit the inotropic effect. Our assumption was that stimulation caused sufficient release of norepinephrine to activate β-receptors due to mass effect, similar to that observed with dobutamine infusion in patients with ADHF. There was a significant increase in mean arterial pressure predominately due to the increase in systolic pressure. The increase in mean arterial pressure due to the increase in systolic pressure would suggest that there was an increase in systemic blood flow, and there was no effect on systemic vascular resistance. Of course, future studies measuring cardiac output and systemic vascular resistance will be needed to verify this assumption.

As demonstrated in Figure 2, the increase in contractility remained relatively constant during the stimulation period—provided that the electrophysiology stimulation catheter remained physically stable. Twelve-hour stimulation of the stellate ganglia in normal animals demonstrated that the initial increase in contractility could be maintained for the duration of the experiment (not shown). In our clinical study, the procedure was limited to a total of 90 minutes including catheter insertion, anatomic mapping, and stimulation. Accordingly, future clinical studies will be necessary to better determine the sustainability of the response over multiple days.

Because the catheters used for stimulation in this study were standard off-the-shelf electrophysiology catheters, it is not particularly surprising that periodic catheter dislodgements occurred. There is little doubt that for this approach to become clinically practical for the treatment of patients with ADHF, catheters will need to be designed to specifically maintain position at the proximal right PA.

Limitations

The small number of subjects and non-normality of the data makes generalization of these results difficult. While there is evidence to suggest a therapeutic effect on several of the cardiac parameters included, such findings are considered hypothesis-generating and should be validated in larger, powered clinical studies. Indeed a number of issues need to be clarified before this neuromodulation therapy’s use in patients who present with ADHF: (1) can a device be created that can deliver stable nerve stimulation for extended periods of time (days); (2) will awake, nonsedated patients tolerate this stimulation therapy; and (3) will contractility remain elevated if this therapy is applied for multiple days after admission for ADHF.

Clinical Perspectives

Competencies in Medical Knowledge

In the setting of heart failure with reduced ejection fraction, treatment of acute decompensated heart failure is largely limited to pharmacological reduction in preload and afterload; use of sympathomimetics is limited by the attendant increase in heart rate and myocardial oxygen debt. In this study, we demonstrate that in patients with reduced ejection fraction, temporary stimulation of the distal sympathetic fibers at the level of the pulmonary artery can increase ventricular contractility and arterial blood pressure without a substantial increase in the heart rate.

Translational Outlook

If a dedicated device can be developed to deliver stable stimulation of this cardiac nerve for extended periods of time, this novel neuromodulation therapy may play an important role in the management of patients who present with acute decompensated heart failure.

Footnotes