Abstract— Rate-responsive pacemakers (PMs) aim at

having pacing rates as similar to physiological cardiac rhythms as possible. The pacemaker INOS2+-CLS (Biotronik, Germany) implements a closed loop strategy (CLS) based on indirect measures of right ventricle contractility using intracardiac impedance signal. The contractility is, in turn, related to the autonomic nervous system control to the heart. Aim of this study was to evaluate the 24h beat-to-beat heart rate and blood pressure profiles in patients implanted with CLS rate adaptive PM. 24h ECG and arterial pressure waveform acquisition were performed by a digital HOLTER system by the Portapres equipment, respectively. A proper-designed algorithm was developed to classify PM pacing modalities. For each beat we estimated the heart rate (HR), and the systolic and diastolic pressure values (SP, DP). So far, 6 patients have been studied: 4 patients have been analyzed both with and without rate responsive modalities (DDD-R and DDD, respectively); 2 patients have been studied only with rate-responsive modality. Results obtained in 6 patients show that this rate adaptive PM accurately preserve the heart rate and blood pressure variability throughout the 24h. In particular, the rate adaptation of PM based on impedance measurements succeeds in maintaining the spontaneous HR, SP and DP on a beat-to-beat basis.

Keywords—Cardiovascular signal processing, heart rate

variability, rate responsive pacemakers

I. INTRODUCTION Rate-responsive cardiac pacemakers (PMs) aim at having pacing rates as similar to physiological cardiac rhythms as possible. In the last two decades different sensors and algorithms have been developed to modulate pacing rate according to physiological demand [1][2]. Latest generation PMs use signals derived from sensors or algorithms to modify the paced rate, but present several limits [3]: signals derived from sensors can not account for non-movement stress, while algorithm-based pacing strategies exhibit long reaction times. In addition, sensors and algorithms can have different PM parameters to be programmed individually. Dual sensor systems may find difficulties in their coordination. From a physiological point of view, the autonomic tone contains the information to adapt the pacing rate to different kinds of

haemodynamic challenges, thus overcoming the limitations of previous technologies [4]. The pacemaker INOS2+-CLS (Biotronik, Germany) implements a closed loop strategy (CLS) based on indirect measures of right ventricle contractility using intracardiac impedance signal [5][6]. The contractility is, in turn, related to the autonomic nervous system control to the heart. This PM has demonstrated to measure neurovegetative tone and to control paced heart rate better than any other sensors [3][7][8]. However, the hemodynamic profiles of this rate responsive pacing have not been analyzed yet. Aim of this study was to analyze the 24h cardiovascular profiles of patients implanted with CLS rate adaptive PM.

II. METHODOLOGY A. Experimental protocol In order to evaluate the hemodynamic responses to all types of PM pacing (atrial, ventricular or both) compared to the spontaneous beats, we simultaneously measured the ECG and the arterial blood pressure waveform. 24h ECG acquisition was performed by a digital HOLTER system (PRIMA Holter, Remco Italia CARDIOLINE®, Milano, Italy). Arterial pressure waveform (APW) was non-invasively acquired for 24h at the fingers by the Portapres (TNO Biomedical Instrumentation, Amsterdam, The Netherlands). So far, we analyzed 6 patients (4 males; age: 70.3±9.9 yr): 4 patients have been analyzed with the PM programmed both as rate-adaptive pacing (DDD-R) and then as fixed rate modality (DDD); 2 patients have been studied only in DDD-R, not being yet recorded in DDD. B. PM description and programming

The INOS2+-CLS modifies pacing rate by indirect

measures of cardiac contractility delivering a monopolar under-threshold pulse train (pulse amplitude=600mA; pulse duration=30.5msec; frequency=128Hz) for near 300ms, synchronized to each ventricular stimulation. The PM does not require individual programming. We used the following common settings: Lower Rate = 60 -70 bpm; Upper CLS Rate (i.e. maximum allowed increase of paced rate) = 120-130 bpm.

Heart Rate and Blood Pressure Variability in Patients Implanted with Rate-Responsive Pacemaker

F. Censi1, G. Calcagnini1, M. D’Alessandro1, M. Malavasi2, R. Quaglione2, G. Critelli2, P. Bartolini1,

V. Barbaro1 1Department of Technology and Health, Istituto Superiore di Sanità, Roma, Italy

2 Policlinico Umberto I, Università degli Studi di Roma ‘La Sapienza’, Roma, Italy

C. Signal processing A proper-designed algorithm was developed to classify

PM pacing modalities (figure 1): sensed atrium and paced ventricle (AsVp), paced atrium and paced ventricle (ApVp).

First, the PM spikes were filtered out by using a non-linear procedure.

Second, the R wave was recognized in the spikes-free signal according to a modified version of the algorithm of Pan and Tompkins [9].

Then, for each detected beat, sensed or paced atrial beats were classified according to the presence of the PM spike. For each beat we estimated the heart rate (HR) and the systolic and diastolic pressure values (SP and DP, respectively).

Fig. 1. Example of pacing modalities: sensed atrium and paced ventricle

(AsVp), paced atrium and paced ventricle (ApVp). In order to analyze the long term heart rate and blood pressure variability due to rate-adaptive pacing (DDD-R) and to compare it to the conventional fixed rate modality (DDD), we estimated the mean RR interval and the mean SP and DP values, within 5-minute long non-overlapping windows, throughout the 24h recordings. Over 5 minutes, both paced and spontaneous beats occur, and thus it is possible to obtain two series of averaged data to be compared. Moreover 5-minute long windows preserve the long term HR and blood pressure variability information [10]. To assess and quantify the similarity between paced and spontaneous HR, SP and DP trends, we estimated the correlation coefficients between the 5-minute mean values either during rate-responsive pacing and at fixed rate.

III. RESULTS

Table I shows the average HR over the 24h recording for the spontaneous (AsVp) and the paced (ApVp) beats. For AsVp beats, the mean values are averaged over all the spontaneous beats in the two recordings (DDD-R and DDD). As for the paced beats, the mean RR interval in DDD modality is not reported, being the heart rate always equal to the PM lower rate (60-70 bpm).

Tables II and III show the SP and the DP, respectively, averaged over the 24h recording for the spontaneous (AsVp) and the paced (ApVp) beats.

For AsVp beats, the mean values are averaged over all the spontaneous beats in the two recordings (DDD-R and DDD). Patients 5 and 6 did not undergo the 24h recording in DDD modality yet.

TABLE I 24H MEAN RR INTERVAL FOR ALL THE PATIENTS.

RR interval [s]

Patient AsVp

DDD&DDD-R ApVp

DDD-R 1 0.81±0.12 0.89±0.10 2 0.80±0.07 0.80±0.09 3 0.84±0.09 0.93±0.08 4 0.87±0.17 0.88±0.09 5 0.76±0.13 0.93±0.09 6 0.75±0.09 0.78±0.12

TABLE II 24H MEAN SYSTOLIC PRESSURE (SP) FOR ALL THE PATIENTS.

SP [mmHg]

Patient AsVp

DDD&DDD-R ApVp

DDD-R ApVp DDD

1 152.7±0.4 165.9±0.3 124.7±0.3 2 110.9±0.2 105.5±0.2 101.2±0.1 3 116.2±0.2 100.9±0.1 135.7±0.3 4 107.8±0.2 101.1±0.2 105.2±0.2 5 133.3±0.3 124.1±0.2 6 113.3±0.2 112.2±0.1

TABLE III 24H MEAN DIASTOLIC PRESSURE (DP) FOR ALL THE PATIENTS.

DP [mmHg]

Patient AsVp

DDD-R&DDD ApVp

DDD-R ApVp DDD

1 70.8±0.2 75.6±0.2 57.0±0.1 2 60.5±0.1 62.7±0.1 48.8±0.1 3 64.2±0.1 58.5±0.1 90.2±0.2 4 50.1±0.1 50.3±0.1 48.5±0.1 5 78.5±0.1 76.8±0.1 6 58.3±0.1 58.8±0.1

AsVp

ApVp

Ventricular spike

Spontaneous P-wave

Ventricular spike

Atrial spike

Note that, although a certain inter-patient variability does exist, the mean values of HR, SP and DP obtained for paced beats are comparable to those obtained for spontaneous beats. Figure 2 shows the correlation coefficients (ρ) between the values of RR intervals, SP and DP averaged over 5-minute long windows obtained for paced beats and for spontaneous beats in the two pacing modalities. Correlation coefficients obtained in DDD modality are lower than those obtained in DDD-R for RR intervals, SP and DP. Figure 3 shows the mean heart rate (in beats per minute, bpm) within 5-minute long non-overlapping windows, throughout the 24h recording for both pacing modalities, for one patient. Similar results have been obtained for the other patients. Note how the PM programmed as DDD-R (upper panel) succeeds in accurately following the physiological heart rate variability at any day time.

IV. DISCUSSION AND CONCLUSION

The characterization of PM which use the estimation of the autonomic tone to adapt the pacing rates requires specific methods and procedures. The short-term beat-to-beat changes in heart rate and systolic pressure in patients implanted with the PM INOS2+-CLS have been evaluated in response to a set of tests consisting in both physical and emotional inputs to the autonomic nervous system [8].

Correlation coefficients

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0

0.0 0.2 0.4 0.6 0.8 1.0

DDD DDD-R 0.0 0.2 0.4 0.6 0.8 1.0

RR interval

SP

DP

Fig. 2. Correlation coefficients between the mean values of RR intervals, SP and DP obtained for paced and spontaneous

beats in the two modalities.

Fig. 3. Mean heart rate in beats per minute (bpm) within 5-minute long non-overlapping windows, throughout the 24h recording for both pacing modalities, for one patient.

9:31:43 day 1

9:20:04 day 2

9:36:05 day 1

9:05:28day 2

Heart rate - DDD-R

Heart rate - DDD

AsVp ApVp

55

105

55

135

[bpm

] [b

pm]

day

day night

night

This study aimed to assess the long-term beat-to-beat cardiovascular variability in patients implanted with rate adaptive PM. We selected patients not being completely dependent by the PM, so as to have a certain number of spontaneous beats to compare with the artificial rhythm imposed by the PM. Results obtained in 6 patients show that this rate adaptive PM accurately preserve the heart rate and blood pressure variability throughout the 24h. In particular, the rate adaptation of PM based on impedance measurements succeeds in maintaining the spontaneous HR, SP and DP on a beat-to-beat basis.

As expected, paced HR during DDD showed poor correlation to the spontaneous one, while a good correlation is found when PM is switched to the rate-responsive modality.

Systolic and diastolic pressure showed a similar correlation structure, but the difference of the correlation coefficients between fixed and rate-responsive pacing was smaller.

In other words, in fixed rate pacing modality, the blood pressure obtained for paced beats still shows a physiological variability, probably expression of other blood pressure control mechanisms, and to a minor extent depending by the HR.

From our data it emerges that the closed loop algorithm based on ventricular impedance signal responds to the changes in the autonomic drive not only on short-term basis, as previously demonstrated [8], but also over longer periods. In conclusion, the efficacy of rate adaptive PM based on CLS have been demonstrated only on a short-term basis (few minutes), by means of response of the paced HR to short physical and/or mental stressors. By using 24h HR and blood pressure monitoring instrumentation, we evaluated the efficacy of the CLS rate

adaptive pacing during the patient normal life, by analyzing the HR and blood pressure changes due to the several situations a patient commonly faces during his/her day activity.

REFERENCES

[1] M. Schaldach, “Compensation of chronotropic incompetence with temperature-controlled rate adaptive pacing”, Biomed Tech, vol. 33, pp. 286-94, 1988. [2] K. Stangl, A. Wirtzfeld, R. Heinze, M. Laule, “A new multisensor pacing system using stroke volume, respiratory rate, mixed venous oxygen saturation, and temperature, right atrial pressure, right ventricular pressure, and dP/dt”, Pacing Clin Electrophysiol, vol. 11, pp. 712-24, 1988. [3] K. Malinowski, “Interindividual comparison of Different Sensors for Rate-adaptive pacing”, Pacing Clin Electrophysiol, vol. 21, pp. 2209-2213, 1998. [4] A.M. Pichlmaier, Braile D, Ebner E, Greco OT, Hutten H, von Knorre GH, Niederlag W, Rentsch W, Volkmann H, Weber D, et al., “Autonomic nervous system controlled closed loop cardiac pacing”, Pacing Clin Electrophysiol,, vol. 15, pp. 1787-91, 1992. [5] M. Schaldach, “Automatic adjustment of pacing parameters based on intracardiac impedance measurements”, Pacing Clin Electrophysiol, vol. 13, pp. 1702-10, 1990. [6] M. Schaldach, H. Hutten, “Intracardiac impedance to determine sympathetic activity in rate responsive pacing”, Pacing Clin Electrophysiol, vol. 15, pp. 1778-86, 1992. [7] A.S. Menezes Jr, Freitas Jr AF, Melo CSN et al., “Cardiac stimulation in closed loop controlled by ANS versus accelerometer - comparative analysis", Progress in Biomedical Research, pp.190-193, 1999. [8] M. Santini, Ricci R, Pignalberi C, Biancalana G, Censi F, Calcagnini G, Bartolini P, Barbaro V, “Effect of autonomic stressors on rate control in pacemakers using ventricular impedance signal”, Pacing Clin Electrophysiol, vol. 27(1), pp: 24-32, 2004. [9] Pan J, Tompkins WJ, “A real-time QRS detection algorithm”, IEEE Trans Biomed Eng, vol. 32, pp. 230-6, 1985. [10] Task Force, Heart Rate Variability, “Standard of measurement, physiological interpretation and clinical use”, Circulation, vol. 93, pp:1043-1065, 1996.