Impedance Cardiography Accurately Measures Cardiac Output During Exercise in Children With Cystic Fibrosis (Materials and Methods)
This study received approval from the University of Manitoba Faculty of Medicine Committee for Use of Human Subjects in Research and informed, written consent was obtained from parents and patients alike. The protocol was identical to our study done in healthy children, except that children with CF from the Children’s Hospital of Winnipeg CF Clinic were recruited for the present study. Routine spirometry was performed on a portable spirometer (model AT-6; Schiller AG; Baar, Switzerland) and expressed as percent of predicted. Exercise was performed on an electrically braked cycle ergometer (Excalibur; Quinton Instruments; Seattle) at two work rates: 0.5 and 1.5 W/kg. Subjects pedaled for 4 min before gas exchange measurements commenced, starting with blood gas sampling, followed by mixed expired gas collection for 2 min, and ending with duplicate RB maneuvers. Total duration of work at each load was approximately 8 min.
Expired ventilation was measured on a spirometer (Transferscreen II; Erich Jaeger Co; Wurzburg, Germany) equipped with a heated pneumotachograph, calibrated before and after tests using a volumetric syringe. Mixed expired gases were collected in gas collection bags and analyzed with zirconium oxide 02 and infrared C02 analyzers (S-3A/1, CD-3A, respectively; Ametek Inc; Pittsburgh). Analyzers were calibrated with reference gases (room air; 5% C02, 15% 02, balance N2; and 13% C02, balance 02) before and after testing. Arterialized capillary blood was obtained from a finger warmed by wrapping in a heating pad, stored on ice, and analyzed within 20 min on a blood gas analyzer (Radiometer ABL500; Copenhagen) to measure pH and PaC02. Hemoglobin (g/L) was also measured from these samples (on a Radiometer Hemoximeter OSM 3). The rebreathing bag was filled with an appropriate mixture of C02 (10 to 15%) in 02, with the volume of the rebreathing mixture approximating the subject’s vital capacity. At end-expiration, the subject was switched into the rebreathing bag for a period of approximately 15 s. This maneuver was attempted in duplicate with at least 1 min between each, and if both trials resulted in an equilibrium plateau, the mean mixed venous Pco2 was used, incorporating the “downstream” corrected C02 partial pressure. Maneuvers that did not achieve a plateau in the Pco2 tracing were discarded. Signals were recorded in real time on a thermal chart recorder (Gould TA2000; Cleveland).
The ICG-M401 employs a tetrapolar lead system with paired inner electrodes placed on either side in the supraclavicular fossa just above the level of the suprasternal notch and along the midaxillary line at the level of the xiphoid. The outer electrodes are placed 6 cm cephalad and caudad, respectively. The phonocardiogram was recorded by a proprietary microphone attached with double-sided adhesive and placed along the left sternal border at a site where the second heart sound was loudest. The software program used a modification of the Kubicek equation based on empirical corrections for body habitus.
where SV=stroke volume (mL), r—blood resistivity (assumed constant 135 ll/cm), L=distance between inner electrodes (measured), Z0=baseline thoracic impedance (II), VET=ventricular ejection time (s), and dZ/dt=maximum value of first time derivative of impedance. The three types of body habitus were ectomorph, mesomorph, and endomorph, which we defined for use in children as percent ideal weight for height <85, 85 to 120, and >120, respectively. Thoracic length was measured while the subject was seated on a stool after electrode placement, and verified later with the subject seated on the ergometer. The mean of these (nearest centimeter) was entered as “L” into the computations. Subjects also underwent duplicate measurement of anteroposterior and lateral chest diameters at the level of the xiphoid. The ratio of these two diameters (thoracic index) was also calculated.
A single measurement of impedance cardiac output was made once each minute from all cardiac cycles over an 8-s interval, by computing beat by beat stroke volume and multiplying by heart rate. Unlike the experience in healthy children, obtaining good analog signals for impedance and phonoeardiogram was difficult in some CF patients. The ICG-M401 saves all raw signals for subsequent review, permitting manual elimination of beats in which the computer algorithm had erroneously determined dZ/dtmax because the signal contained too much motion artifact (Fig 1). In practice, the phonocardiographic signal did not cause any beats to be discarded, but in six patients up to seven cardiac cycles were omitted from stroke volume calculations, leaving no fewer than 10 beats from which to compute stroke volume during any given minute. Impedance measurements from the fourth or fifth to eighth minutes of work were averaged to provide a single value for heart rate and Q.
Values of Q obtained by both methods were regressed against oxygen uptake (Vo2). Measurements of Q by the two methods were compared by the following: regression of impedance cardiac output (Qicg) on rebreathing cardiac output (Qrb); by the method of Bland and Altman; and by paired t test of the difference between results (Qrb —Qicg). The ratio of Qicg/ Qrb was regressed against thoracic index to test the appropriateness of Kubicek’s equation in patients with varying degrees of barrel-chest deformity.
Figure 1. Graphic output of waveforms processed by the ICG M401. Upper trace is ECG; second from the top is dZ/dt; third trace is the phonocardiogram (PCG); and lower trace is the thoracic electrical bioimpedance (ICG). The calculated numeric value of dZ/dt is shown for three individual heartbeats, one of which (No. 3) is distorted by artifact. Below is a list of beat-by-beat computations of stroke volume for these beats, in which one can readily appreciate the error inherent in including such a beat (eg, beat 3) in the stroke volume calculation for that sampling interval.