Reverse re-modelling chronic heart failure by reinstating heart rate variability

Abstract

Heart rate variability (HRV) is a crucial indicator of cardiovascular health. Low HRV is correlated with disease severity and mortality in heart failure. Heart rate increases and decreases with each breath in normal physiology termed respiratory sinus arrhythmia (RSA). RSA is highly evolutionarily conserved, most prominent in the young and athletic and is lost in cardiovascular disease. Despite this, current pacemakers either pace the heart in a metronomic fashion or sense activity in the sinus node. If RSA has been lost in cardiovascular disease current pacemakers cannot restore it. We hypothesized that restoration of RSA in heart failure would improve cardiac function. Restoration of RSA in heart failure was assessed in an ovine model of heart failure with reduced ejection fraction. Conscious 24 h recordings were made from three groups, RSA paced (n = 6), monotonically paced (n = 6) and heart failure time control (n = 5). Real-time blood pressure, cardiac output, heart rate and diaphragmatic EMG were recorded in all animals. Respiratory modulated pacing was generated by a proprietary device (Ceryx Medical) to pace the heart with real-time respiratory modulation. RSA pacing substantially increased cardiac output by 1.4 L/min (20%) compared to contemporary (monotonic) pacing. This increase in cardiac output led to a significant decrease in apnoeas associated with heart failure, reversed cardiomyocyte hypertrophy, and restored the T-tubule structure that is essential for force generation. Re-instating RSA in heart failure improves cardiac function through mechanisms of reverse re-modelling; the improvement observed is far greater than that seen with current contemporary therapies. These findings support the concept of re-instating RSA as a regime for patients who require a pacemaker.

Keywords: Cardiac output; Heart failure; Pacemaker; Respiratory sinus arrhythmia.

Fig. 1

A. Schematic representation of the experimental set-up. Fully instrumented conscious 3–5 year old, outbred, female Romney sheep were paced via the left atrium, 1.5–2.5 V, using a neuron-based pace making device that modulates heart rate in phase with respiration on a breath by breath basis mimicking respiratory sinus arrhythmia (RSA). The device receives diaphragmatic EMG (electromyogram) input to define the phase of inspiration. RSA increases heart rate during inspiration as indicated by the highlighted regions in panels A, C–F Cardiac output was measured on a beat by beat basis using an implantable flow probe attached to the ascending aorta. B Time-line of the six-month experimental protocol. Heart failure with reduced ejection fraction was induced by microembolization and followed by an 8–10 week holding period for chronic heart failure to develop prior to pacing. Following instrumentation and stabilization of hemodynamics parameters, animals were divided into three groups: two were paced (RSA or monotonically) and one not paced (time control) for 4 weeks. This was followed by 1 week off pacing before animals were anaesthetized for baroreflex testing and cardiac tissue collected for post hoc analysis. C Representative raw data traces from a healthy sheep showing native RSA. D A heart failure sheep showing loss of RSA. E A heart failure sheep paced monotonically, showing a stable heart rate. F A heart failure sheep being RSA paced. Note that there was no difference between the mean heart rate of monotonically and RSA paced sheep. Acclim acclimatization to the laboratory; Avg. average; bpm, beats per minute; DEMG diaphragmatic electromyogram; HR heart rate; mV milli Volts

Fig. 2

Data were collected for 24 h a day for 6 weeks (1 week baseline or pre-pacing, 4 weeks pacing; 1 week pace off post-pacing) from chronically instrumented, conscious sheep. A Cardiac output was measured directly on a beat by beat basis and each data point represents the 24 h average change in cardiac output. Monotonically paced (Mono: magenta, n = 6 paced to day 14, n = 5 paced for 4 weeks) and respiratory sinus arrhythmia paced (RSA: green, n = 6 paced to day 14, n = 5 paced for 4 weeks) were paced for 28 days. A time control group (TC: blue, n = 5 to day 32, n = 3–4 weeks) was not paced. B Cardiac output change from baseline on expanded time scale over first week of pacing. Each data point represents 2 h average cardiac output. Mono + RSA (n = 6 each) animals only. C Mean heart rate over time; each data point is a 24 h average (Mono: n = 6 paced to day 14, n = 5 paced for 4 weeks. RSA: n = 6 paced to day 14, n = 5 paced for 4 weeks. TC: n = 5 to day 32, n = 3–4 weeks). D Ejection fraction measured using echocardiography from conscious sheep (TC, n = 2–5, Mono n = 5, RSA n = 5). E Mean arterial pressure (all groups, n = 4). F Systemic vascular resistance (all groups, n = 4). Note that the Posthoc test significance for the RSA group week 4 vs. baseline was P = 0.057 For D–F, all data are presented as a 24 h average. 2-way ANOVA interaction effect; #P < 0.05. Posthoc analysis (Dunnetts) to compare the time points to baseline;*P < 0.05

Fig. 3

A Renal sympathetic nerve activity (RSNA) was recorded under general anesthesia at the end of the sixth week. Representative raw data traces of RSNA and blood pressure (BP) in response to a baroreflex challenge with phenylephrine (PE) and sodium nitroprusside (SNP). Time control (blue), Mono (magenta) and RSA paced (green). B RSNA baroreflex function curves for the three groups of sheep. C Baroreflex saturation is the systolic blood pressure when RSNA reaches zero percent of maximum. D Threshold is the point of the baroreflex curve when RSNA reaches a maximum at low-pressure levels (TC n = 3; Mono n = 3; RSA n = 4)

Fig. 4

A A representative example of the diaphragmatic electromyographic activity (DEMG), which indicated the inspiratory phases, of a conscious sheep in heart failure. Note the instability in DEMG burst frequency including periods of apnoea. B Breath rate was calculated to form the burst frequency of the diaphragmatic electromyography activity (TC, n = 3–4, Mono n = 3, RSA n = 5). C, D The number of apneas were totaled over 24 h periods for each week of the 6 week protocol. Apnoea incidence of different length was calculated and represented as apnoeas > 3, 4, 5 and 6 s for RSA paced animals (D; n = 5) and monotonically paced sheep (C; Mono; n = 3). DEMG diaphragmatic electromyogram. 1-way ANOVA; *P < 0.05. Pace-off data not included in statistical analysis

Fig. 5

T-tubule morphology and myocyte size were measured from cells labelled with the cell membrane stain wheatgerm agglutin (WGA). A–D Representative images of the lateral wall of the left ventricle from a healthy (control), heart failure (time control), Mono and RSA paced heart failure sheep, respectively. For each sheep group, T-tubule fluorescence intensity, a Fourier Transformation (FFT), and power spectrum of the FFT were plotted from 20 µm segments of heart cells (examples indicated with an *) are shown. T-tubule structures display a 0.5 Hz frequency (FFT power spectrum), the amplitude of the 0.5 Hz peak demonstrates their presence and uniformity. E T-tubule uniformity is reduced in both time control and Mono paced sheep but T-tubules are re-instated after RSA pacing and not different to healthy controls. Number of animals and cells were: (healthy control, n = 5 sheep, 176 cells; Time control, n = 5 sheep, 128 cells; Mono n = 4 sheep, 153 cells; RSA n = 3 sheep, 93 cells). F RSA pacing reduced cardiomyocyte cell size relative to time control and Mono paced sheep. Cell size was calculated as cross-sectional area. Data analysed included multiple animals and cell numbers: healthy controls (n = 5 sheep, 164 cells), time controls (n = 5, 174 cells), Mono (n = 4, 229 cells), RSA (n = 3, 158 cells)

Fig. 6

Three sheep in heart failure were paced with a sinusoidal pacing regimen that was not phase locked to respiration. A Representative raw data trace showing sine-wave pacing, which was generated using a grass-stimulator. Note the free-running cyclic changes in heart rate (HR); the amplitude of these oscillations was approximately 12 beats per minute and matched the RSA pacing peak-to-trough amplitude. Cardiac output (CO) and blood pressure (BP) in addition to HR are shown. B–F Cardiac output, mean arterial pressure and systemic vascular resistance, respectively, where data points represent 24 h average changes from pre-pacing baseline. Note the severe decline in cardiac output