Effects of Synchronizing Foot Strike and Cardiac Phase on Exercise Hemodynamics in Patients With Cardiac Resynchronization Therapy: A Within-Subjects Pilot Study to Fine-Tune Cardio-Locomotor Coupling for Heart Failure

Twenty individuals with compensated CHF, who had already undergone CRT for an accepted clinical indication, receiving stable medical therapy for at least 3 months, participated in this study; notably, only 10 participants were included in final analysis after completion of the study protocol (Table 1). Participants were excluded from enrollment if they presented with decompensated heart failure (New York Heart Association class IIIb or IV), hearing loss, unmanageable joint pain, uncontrolled infectious disease, dependent state of inotropes, current smoker, pulmonary disease (eg, idiopathic pulmonary hypertension), other cardiovascular diseases (eg, valve or adult congenital heart disease), musculoskeletal disease, or other existing health conditions that precluded participation in physical activity. Participants were also excluded during screening if they were unable to synchronize their foot strike with an auditory tone representing cardiac cycle phase. During screening, we also performed echocardiography (Philips Epiq 7) during supine rest to characterize left ventricular ejection fraction at the time of study; one participant’s images were not analyzable; thus, echocardiography data are reported from 9 participants. After study completion, 10 individuals were excluded from data analysis, as they were not able to complete at least one pair of systolic-diastolic stepping with a minimum of 50% synchronization for each respective cardiac cycle phase. Participant characteristics are detailed in Table 2. Testing was performed at the Institute for Exercise and Environmental Medicine in Dallas, TX. All participants provided written informed consent, and study approval was granted by the institutional review board of the University of Texas Southwestern (STU-2018-0262).

All participants attended the laboratory on a maximum of 2 occasions to complete screening and testing. After consent, participants were familiarized with walking on a treadmill (Biodex RTM 600) to determine an individualized, appropriate, and comfortable walking speed (equating to ≈30–40% of their maximal oxygen uptake). Walking speed ranged from 1.5 to 3.0 mph at 0% to 4% incline, and all participants completed 3 to 5 bouts of walking for 5 minutes during the study day, with the aim of completing at least one pair of systolic-diastolic stepping. Walking was completed during one control-free walking condition, without any foot strike timing guidance to further familiarize participants with the measurements during exercise and 2 experimental conditions: (1) systolic walking, defined as stepping at 0% or 100±15% of the R-R interval; and (2) diastolic walking, defined as stepping at 45±15% of the R-R interval (Table 1). Free walking was always completed first; the order of systolic-diastolic walking was randomized, and the participants were blinded to the targeted cardiac phase. During these sessions, heart rate and step frequencies, gas exchange, and cardiac output were measured. Heart rate and rhythm were monitored through a 5-lead ECG. Each participant completed all walking at the same speed for all conditions.

All participants wore a chest strap (Zephyr Bioharness, Annapolis, MD) with an attached sensor to detect heart rate (through ECG) and foot strike (through accelerometry). Data were transmitted wirelessly in real time to a tablet (Apple Inc. iPad, Cupertino, CA) running proprietary analysis software (Counterpace; Pulson, Mountain View, CA). Counterpace produced auditory tones corresponding to occur exactly on the R wave of the ECG (ie, ventricular systole) or at 45% of the R-R interval (ie, diastole). Tones were projected to the participant over a loudspeaker.

Valid step counts (%) were determined using Counterpace. The software generated step and R-wave time stamps on the raw quantitative data to assist in calculating the relative phase of each walking bout: %RRI (percentage of R-R interval)=100(t_step–t_R-wave1)/(t_R-wave2–t_R-wave1), where t_R-wave1 and t_R-wave2 refer to the time point of 2 consecutive R waves, and t_step indicates the time point that a step occurred between the 2 R waves. The number of steps that fell in phase (100±15% of the R-wave for systolic walking and 45±15% of the R-R interval for diastolic walking) over the total steps taken was used to calculate the percentage of valid steps. Valid steps were examined in the last minute of each walking bout during which data collection occurred. Bouts that had <50% of valid or “in-phase” steps were excluded. For participants who walked with an uneven gait or had more complex ECG morphology, we completed post hoc analysis of foot strike waveforms collected through pressure sensors (BIOPAC Systems Inc., Goleta, CA) applied to the bottoms of participants’ shoes (n=5). ECG and the foot sensor waveforms were collected in real time through data acquisition software (BIOPAC Systems Inc., Goleta, CA). Heel strike (Figure 1) was marked and used to calculate the foot strike time in relation to the ECG. Foot strike timing was calculated as detailed above to assess valid step timing, expressed as a percentage of the R-R interval.

During standing rest and the latter 2.5 minutes of each walking bout, participants breathed through a mouthpiece attached to a 4-way valve. We measured ventilation (Vacumed, Ventura, CA) and fractional gas concentrations (mass spectrometry, MGA01100 Marquette Gas Analysis, Missouri), and calculated oxygen uptake through indirect calorimetry (Beck Integrative Physiological System; KCBeck). All gas exchange variables were calculated from 3 minutes of rest or 45 to 60 seconds during the last minute of each 5-minute walking bout. Cardiac output was assessed using acetylene rebreathing,^27,28 which involved participants inhaling a mixture of inert gases, including 9% helium, 0.6% acetylene, 45% oxygen, and balanced nitrogen, and rebreathing the same volume for 6 to 8 breaths. Stroke volume was calculated as the quotient of cardiac output and heart rate. During the rebreathing maneuvers, brachial blood pressure was monitored using an electrosphygmomanometer (Suntech Tango M2; Morrisville, NC). We calculated effective arterial elastance (systolic blood pressure×0.9/stroke volume) as an index of arterial resistive, not pulsatile, afterload.²⁹ Blood pressure (finger photoplethysmography; BMEYE, Amsterdam, Netherlands) and pulse oxygen saturation (Nellcor N600x, Pleasanton, CA) were continuously monitored throughout rest and exercise; reported blood pressures are from the electrosphygmomanometer.

After testing for normality (Shapiro-Wilk), we addressed our a priori primary aim by comparing diastolic and systolic walking using paired-samples t tests. Free walking data are shown for reference only and were not compared with the experimental conditions due to the participants’ highly variable gait patterns and consequential irregular step-timing. P<0.05 was considered statistically significant. Data are expressed as mean±SD, and all analyses were conducted using GraphPad Prism (Version 9.5.0).

Ten participants with CHF and previously implanted CRT pacemakers were included for analysis (Table 2). Our participants were 58±17 (range, 23–75) years of age, 40% of whom were female, and 90% were White. Average body mass index was 30.3±9.4 kg/m². At the time of study, left ventricular ejection fraction was 37±10% (n=9).

There was no difference between the percentage of valid footsteps during diastolic or systolic walking (diastolic 65±12% versus systolic 62±11%; P=0.556). Hemodynamic and cardiopulmonary data are displayed in Table 3. Oxygen uptake was similar between diastolic and systolic walking (P=0.298). Likewise, heart rate (paced) was not different between conditions (P=0.300). Compared with systolic walking, stroke volume, and cardiac output (P=0.003 and P=0.004, respectively; Figure 2B and 2C) were higher during diastolic walking. Mean arterial pressure (n=9; P=0.007; Figure 2D), total peripheral resistance (n=9; P=0.008; Figure 2E) and diastolic blood pressure (n=9; P=0.005; Table 3) were also lower during diastolic walking than during systolic walking; respiratory exchange ratio was moderately lower during diastolic walking (P=0.075). Systolic blood pressure, effective arterial elastance, and minute ventilation were not different between systolic and diastolic walking (Table 3).

During whole-body exercise, the skeletal muscle pump (venous return) and cardiac ejection (ie, stroke volume) work to ensure continuous blood flow to skeletal muscle.³⁰ For many years, it has been suggested that the synchronization of skeletal muscle and cardiac contraction may increase cardiovascular efficiency.^31,32 The strategy for synchronizing skeletal muscle and cardiac contraction (cardio-locomotor coupling) is uniquely possible when running or walking, as heart rate and step rate are typically similar at any pace. Indeed, heel strike during running (or walking) can produce a backward returning wave in the circulation that can either increase arterial impedance, if it occurs during systole, or facilitate coronary and skeletal muscle perfusion, if it occurs during diastole.¹⁸ Thus, cardiac output and muscle blood flow are maximized for a given level of exercise when skeletal muscle and cardiac contractions are timed in inverse synchrony.^18,33,34 In line with this theory, we previously showed that synchronizing foot strike to the diastolic portion of the cardiac cycle when running at one predetermined speed resulted in a significantly lower heart rate and higher oxygen pulse compared with the systolic portion.³⁵ Herein, we set to extend these findings to patients with stable CHF during walking by using acetylene rebreathing to assess cardiac output and determine stroke volume, as oxygen pulse can be used as an index of stroke volume in healthy adults²⁶ but not in adults with heart failure.³⁶

We observed that stroke volume and cardiac output were higher and mean arterial pressure was lower during diastolic stepping, which suggests that diastolic cardio-locomotor coupling represents a novel approach to augment oxygen delivery during exercise in patients with CHF. Therefore, our current study extends our previous report on athletes who reported a higher oxygen pulse when foot strike occurred during diastole while running.²⁰ The higher exercise cardiac output during diastolic walking occurred despite no difference in heart rate, which we found to be lower in athletes²⁰; the lack of difference in heart rate, despite the higher stroke volume, likely occurred because the individuals tested here had CRT pacemakers.

CRT^6,7 is well known to affect increases in exercise capacity^14,37 and quality of life and to reduce symptoms in those with CHF.^8–10 Therefore, if built into CRT pacemakers, diastolic cardio-locomotor coupling could further improve exercise capacity and reduce fatigue due to increased oxygen delivery. Including this technology in pacemakers would specifically require reverse engineering the concept used in the current study, and, instead of using the accelerometer to send an auditory tone for individuals to step to, it would be used to guide pacing at a time after the waveform crosses >1× the force of gravity (ie, between steps; Figure 3). Taken together, inducing increases in stroke volume, and thereby oxygen delivery, is likely advantageous for patients with CHF regarding augmenting exercise tolerance. Future studies are required to assess whether diastolic cardio-locomotor coupling is effective in increasing time to fatigue in patients with CHF, thus representing a novel strategy to increase physical activity levels and exercise tolerance in these patients.

It was not possible to obtain assessments of preload (ventricular filling pressure or end-diastolic volume) during this study; however, despite the lack of difference in effective arterial elastance, mean arterial pressure and total peripheral resistance were markedly lower during diastolic walking than during systolic walking. The lower pressure and resistance would suggest that a lower afterload facilitates a higher stroke volume during diastolic stepping. It is notable that our data support the previous suggestions of improved cardiovascular efficiency^31,32 and cardiac output and muscle blood flow^18,33,34 when skeletal muscle and cardiac contractions are timed in inverse synchrony. The effects of this synchronization are likely mediated by the production of a backward returning wave in the circulation that facilitates coronary perfusion, thereby increasing skeletal muscle perfusion in the subsequent cardiac cycle.¹⁸ Arterial pulse wave reflections and their consequent effect on ventricular-arterial coupling are now recognized as critical contributors to both systolic and diastolic dysfunction in patients with CHF³⁸ and are considered a risk factor for the development of heart failure in the general population.³⁹ The effects of exercise, especially in clinical groups, on these pressure-flow relationships remain unclear, but our study may indeed highlight their importance during exercise.

Cardiorespiratory fitness has been recommended to be considered a vital sign,⁴⁰ and exercise intolerance is a primary symptom of patients with CHF. Therefore, novel strategies to improve exercise tolerance in these patients are warranted. We show that by using simple technology, which is already built into current CRT pacemakers (tri-axial accelerometer and cardiac rhythm monitor), it is possible to exert significant increases in cardiac output during exercise at the same walking speed when stepping in diastole. Future studies are required to determine the effectiveness of such technology when developed within a CRT pacemaker on patients’ physical activity levels, exercise capacity (ie, time to onset of fatigue), and sensations of fatigue. Furthermore, future studies could assess the effects of alternative coupling ratios (1 step to 2 heart beats) and respiratory-locomotor coupling. Nevertheless, using diastolic cardio-locomotor coupling to boost exercise cardiac output represents a novel intervention that can be incorporated within current CRT devices to increase exercise capacity and patient quality of life and avoid the detriments of systolic walking.

Several methodological considerations are worthy of discussion. First, this is a small pilot study with usable data in 50% of the individuals consented. Future studies are required to confirm these findings in a larger sample to provide more data regarding the potential benefit of fine-tuning cardio-locomotor coupling in heart failure. Second, synchronization of foot strike timing and cardiac contraction was conducted by having participants step to an audible tone, making our results dependent on a participant’s ability to “step in time.” Related to this issue, we selected the criterion of ≥50% synchronization of foot strike and cardiac cycle phase. It is notable that there were still effects of diastolic stepping, and, if this technology is applied within CRT devices, the synchronization will be automated (participant-independent), providing continuous diastolic stepping that could lead to greater effectiveness. Third, we used 2 methods of assessing foot strike timing due to the 2 different methods of data collection. We used either the raw data output from the Counterpace software or conducted manual analysis of the foot strike waveforms (Figure 1); in individuals with a regular gait pattern, these methods provided similar values of foot strike timing (<5% difference in valid step percentage). Finally, due to the severity of exercise intolerance in those studied here, we were unable to collect reliable data at a higher exercise intensity (ie, fast walking); the appropriate synchronization of foot strike and cardiac cycle phase was much more intermittent, and despite our best attempts, we were unable to coach the participants to complete exercise at faster speeds. Nevertheless, our study had several strengths, especially regarding the reliable measurement of exercise cardiac output through acetylene rebreathing^27,28 and a counterbalanced order of systolic and diastolic walking to offset any order effects.

Timing foot strike when walking to the diastolic phase of the cardiac cycle in patients with CHF achieved a higher stroke volume and cardiac output, as well as a lower mean arterial pressure and total peripheral resistance. The hemodynamic effects of diastolic walking are of benefit in terms of increasing exercise stroke volume, and future studies are required to confirm these findings and then to assess the clinical effects of this technology if built into CRT pacemakers. In patients with CHF, diastolic cardio-locomotor coupling may represent a novel strategy to augment the effectiveness of CRT on exercise tolerance (ie, time to fatigue), as well exercise capacity, together increasing patient quality of life.