Comparison of cardiac time intervals between echocardiography and impedance cardiography at various heart rates

When measuring the electrical impedance of the human thorax, a distinct variation can be observed, synchronous with the activity of the heart. The time derivative of this variation is called the impedance cardiogram (ICG) and was introduced by Kubicek and Patterson in 1966 in an effort to determine cardiac stroke volume (SV) [1]. Their paper has initiated numerous studies that aimed to use the amplitude of the minimum value (C-point or [dZ/dt]_min) of the declining wave (C-wave) in the ICG during the systolic phase to assess SV. To this end, various physiological models and algorithms were used, e.g. by Sramek et al. and Bernstein [2, 3, 4]. Most of the estimates of SV were essentially based on the assumptions that the C-wave originates from small fluctuations in a simply distributed electrical field caused by blood volume changes or velocity changes in the aorta. Meanwhile, several studies have shown that these assumptions are too simplistic [5, 6, 7, 8, 9]. Therefore, it appears to be futile to make an interpretation of the amplitude of the ICG-signal based on simplistic models.

A relevant property of the ICG-signal, however, can be found in the time relationships, especially when compared to specific marker points in the electrocardiogram (ECG). Regardless of the multiple sources of the signal, the ICG reflects the mechanical/hydrodynamical aspects of the cardiac cycle, while the ECG reflects the electrical aspect of the cardiac cycle [10, 11]. The variation in the ICG is believed to be the result of two physiological effects [10]. Firstly, the diameters of the aorta and other contributing blood vessels vary within a heartbeat. Since blood plasma is an electrical conductor, the impedance of the thorax varies with varying blood volume in these vessels. This is called the volume effect. Secondly, the orientation of erythrocytes changes the electrical conductivity of the blood in the major vessels. The shear stress, present in flowing blood plasma, causes the disk-shaped erythrocytes to change their orientation parallel to the velocity vectors of the blood. This causes the resistivity of the blood to decrease in the direction of the flow, which may result in a pulsating decrease in thoracic impedance. This is called the orientation effect. The relative contributions of both effects to the ICG have been reported to be comparable in magnitude by some investigators [10, 12].

A parameter that links electrical and mechanical cardiac activity in time is the widely used pre-ejection period (PEP) [13, 14]. The PEP is an index of contractility, which is influenced by sympathetic but not by parasympathetic activity in humans. PEP is defined as the interval from the onset of left ventricular depolarization, reflected by the Q-wave onset in the ECG, to the opening of the aortic valves, reflected by the B-point in the ICG signal [14, 15, 16]. However, despite efforts to improve the quality of the signal by ensemble averaging over multiple beats, time locked to the R-peak, substantial uncertainties in positioning of the Q-wave onset and of the B-point remain [16, 17, 18, 19]. It has also been reported that the B-point in the ICG does not reflect a well-defined point but rather an intersection between two waves, one of which is associated with the left ventricular ejection and the other is of unknown origin [20, 21]. The two waves become apparent in elderly, under pathologic cardiac conditions, and at higher heart rates in young, healthy volunteers during an exercise test. This phenomenon could reflect an asynchrony between the right and left part of the heart.

To find a solution to these problems, the initial systolic time interval (ISTI) was introduced as an indicator for the time delay between electrical and mechanical pumping activity of the heart [17, 19]. This parameter is defined as the time difference between the C-point in the ICG signal, also known as [dZ/dt]_min – a minimum of the dZ/dt-signal during systole – and the preceding R-peak in the simultaneously measured ECG (see Figure 1). Both begin and end points of the ISTI can be detected robustly. Several studies reported significant trends in ISTI during fluid administration to hypovolaemic patients at the intensive care unit after cardiac surgery [22] and during excess fluid withdrawal in patients during haemodialysis [23]. Respiratory arrhythmic deviations in ISTI were observed in patients suffering from Parkinson’s disease [24]. Hoekstra et al. investigated the dependence of ISTI on heart rate and physical training [25]. These studies demonstrated a dependence of ISTI on circulating blood volume, central nervous control and on physical training. However, despite the clinically and physiologically relevant variations in ISTI, no studies are known that position the moment of occurrence of the C-point, the end marker of the ISTI, in the cardiac cycle in healthy humans and over a range of heart rates. A limited number of studies report observations during surgery, in animals, in an isolated test, or at single unknown heart rates [4, 9, 11, 26, 27, 28]. Therefore, the aim of the present study was to establish the timing of the C-point in healthy humans and at various heart rates. The moment of the C-point was compared with the timing of four other cardiac cycle markers obtained from echocardiography, in order to facilitate an interpretation of the timing of the ISTI within the cardiac cycle. To this end, the moments of opening and closing of the aortic valves, the moment of maximum diameter of the aortic arch, and the moment of maximum diameter of the descending aorta, just below the sternum, were determined simultaneously with the moment of the C-point in the ICG in 16 healthy subjects. The heart rates of the subjects were varied using an exercise stimulus on a stationary bicycle. All moments were synchronized with respect to the R-peak of the ECG. Additionally, the timing of the C-point with respect to the opening of the aortic valves was also established at various heart rates from the same recordings.

Fig.1

The measurements resulted in a dataset of heart cycle marker moments at varying heart rates for each subject. The relationship of each marker moment with the RR-interval was established by means of linear regression. All linear relationships were found to be significant (all R > 0.86, all p < 0.005). A typical example of these relationships in one of the subjects is shown in Figure 4, in which the moment of occurrence after the R-peak of the markers is presented as a function of the RR-interval. This figure shows that all markers, including the C-point, occurred earlier after the R-peak for smaller RR-intervals. This was observed in all subjects. The C-point occurred at about the same moment in the cardiac cycle as the moment of maximum diameter of the aortic arch. Furthermore, the slopes of the regression lines for the C-point and the opening of the aortic valves (AV open) were indistinguishable.

Figure 4

Table 2 presents a comparison of the moment of the C-point with the other heart cycle markers, both in rest and after exercise at the two chosen heart rates. It shows that the moment of occurrence of the C-point differed significantly from all other markers both in rest and after exercise, except the maximum diameter of the aortic arch.

In order to evaluate the moments of occurrence of the markers in the ejection phase, the regression line of the moments of opening of the aortic valves was subtracted from the other lines for each subject. An example of these relationships, originating from the same subject, is presented in Figure 5. Every line represents the time interval between the start of the ejection phase and the corresponding heart cycle marker.

Figure 5

Figure 5 illustrates that the interval between the ultrasound markers and the opening of the aortic valves decreased for smaller RR-intervals, while the time interval between the C-point and opening of the aortic valves was constant and independent of the RR-interval.

The t-tests were applied to the slopes of the regression lines, shown in Figure 5, to test whether they differ from zero. As such, all markers were tested for having a constant time interval between the opening of the aortic valves and the marker. This procedure was followed for each subject. The results are shown in Table 3.

Table 3 shows that the C-points occurred at a constant time interval after the opening of the aortic valves, independently of the RR-interval. The slopes of the regression lines for the other heart cycle markers were significantly different from zero. Thus, the time interval of these other markers after the aortic valves open decreased for decreasing RR-intervals.

In this study a comparison was made between the timing of the C-point in the impedance cardiogram and the timing of four different cardiac cycle markers obtained from echocardiography in healthy, young volunteers at various heart rates. The study was not intended to identify the sources of the C-wave in the ICG, which may be complex, but was designed to position the moment of the C-point (minimum of the C-wave or [dZ/dt]_min) in the cardiac cycle. This provides an interpretation of the ISTI, the time difference between the R-point in the ECG and the C-point in the ICG.

It was found that the C-point occurred around the same moment in time as the maximum diameter of the aortic arch for all sixteen subjects and at all heart rates. No significant difference was found between the timing of the C-point and the moment of the maximum diameter of the aortic arch in the cardiac cycle within the time resolution of the echocardiographic recordings. This implies that the C-point occurs around the moment that the blood pressure in the aortic arch reaches its maximum.

Consistent with the observations reported by Hoekstra et al. [25], the C-point occurred earlier in the cardiac cycle for higher heart rates (shorter RR-intervals), which means that the ISTI shortened accordingly. However, the present study shows that this is entirely the consequence of a shortening of the period between the R-peak and the opening of the aortic valves, because the time interval between the opening of the aortic valves and the C-point remained constant. Thus, the latter period was independent of the heart rate in this study. This means that the regression line of the C-point with the RR-interval followed a different trend than the moment of maximum diameter of the aortic arch. A possible explanation for this observation is the erythrocyte orientation effect [5, 12], which causes thorax conductivity to decrease as a consequence of the velocity change of the blood. Aortic velocity flow patterns have recently been visualized by Tanaka et al. [29]. They used echo-dynamography to show that the blood in the aortic arch spirals during the systole; the rotation cycle of this flow is inversely correlated with the maximum velocity. These complex flow profiles may well influence the orientation of the erythrocytes for subjects in exercise. Consequently, the C-point might also coincide with the maximum velocity of aortic arch flow.

This observation also means that the shortening of the ISTI with increasing heart rate in response to an exercise stimulus can be attributed to a shortening of the pre-ejection period (PEP) in every individual subject in this study. This is not contradictory to the report of van Lien et al. [18] that concluded that ISTI and PEP were not interchangeable by one single regression equation, pooled for multiple individuals. Whether the intra-individual relationship between PEP and ISTI is fixed over a wide range of physiological and clinical conditions has to be established in future studies.

In line with the method used by Hoekstra et al. [25], the relationships between ISTI and RR-interval were described using linear regression analysis. However, despite the execution of a strict measurement protocol, it is possible that the moment of the C-point in the cardiac cycle was influenced by variables other than heart rate alone. For example, vascular tonus may be different in rest than during exercise. However, such effects on the timing of the C-point are believed to have been of minor influence on the observed relationships, since all linear fits were highly significant.

The moment of occurrence of the C-point in the cardiac cycle was found to be indistinguishable from the moment of the maximum diameter of the aortic arch. This means that the non-invasively measured ISTI, obtained by impedance cardiography, can be interpreted directly in terms of known events in the cardiac cycle. Therefore, the ISTI can be used as a clinical and physiological parameter indicating the time difference between the electrical activity and the pumping activity of the heart. The ISTI can be used both in intra- and extramural settings.

The time interval between the opening of the aortic valves (start of the ejection phase) and the C-point was found to be independent of the heart rate and constant within subjects, while the time interval between the opening of the aortic valves and the moment of maximum diameter of the aortic arch shortened at increased heart rates. This means that the shortening of the ISTI with increasing heart rate in response to an exercise stimulus was caused by a shortening of the pre-ejection period (PEP) in this study.