Accuracy of electrical impedance cardiography for measuring cardiac output in children with congenital heart defects

Accuracy of Electrical ImpedanceCardiography for Measuring Cardiac Output in Children with Congenital Heart Defects DANIEL S. MILES, PhD, ROBERT W. GOTSHALL, PhD, JANE C. GOLDEN, MS, DWIGHT T. TUURI, MD, ROBERT H. BEEKMAN Ill, MD, and TERRENCE DILLON, MD

Thii study determined whether noninvasive electrical impedance cardiography accurately measures systemic blood ftow (cardiac output) in children with congenital heart defects. A total of 37 patients ranging in age from 2 to 171 months underwent complete right- and left-sided heart catheterizations that included simultaneous Fick and impedance measurement of cardiac output. Based on the diagnosis, 4 groups were formed consisting of a control group (n = 11) with no shunts, a group with intracardiac left-to-right shunting and an atrial septal defect (n = 7), another with a ventricular septal defect (n = 12) and an extracardiac left-to-right shunting with patent ductus arteriosus group (n = 7). Impedance values for systemic blood flow were compared wlth systemic and pulmonary blood flow obtained by the direct Fick method with measured oxygen consumption. The difference between im-

pedance and Fick systemic blood flow was 55% in each of the 4 groups. The hghest correlation between impedance and Fick systemic blood flow was with the atrial septal defect group (r = 0.89) and lowest with the ventricular septal defect and control (r = 0.69) groups. Fick pulmonary blood flow was significantly greater than impedance or Fick systemic flow in all 3 shunt groups. Impedance cardiography accurately measured systemic blood flow in children without shunts or valvular insufficiency. Likewise, systemic blood flow was accurately measured by impedance in the presence of intracardiac left-to-right shunts (atrial and ventricular septal defects) and extracardiac left-to-right shunts (patent ductus arteriosus). Although such results are promising, further studies are needed to evaluate the role of impedance cardiography in pediatric medicine. (Am J Cardiol 1988;61:612-616)

I

mpedance cardiography is a noninvasive technique used to measure stroke volume on a beat-by-beat basis.1-4 Most validation studies’x5,fi*7 and clinical tria1.G” have used adults or animal models almost exclusivelv. The application of this technique in a neonatal or ped”i-

atric population has received very little attention.l”-‘” Due to its ease of application, cost effectiveness and atraumatic nature, impedance cardiography may provide an effective adjunct for assessing a child’s cardiovascular status or response to medical or surgical therapy. Our preliminary work with canine pups” and neonatesl” suggested that the impedance technique may prove useful in measuring cardiac output in premature newborns, neonates and children. We were particularly interested in evaluating whether the impedance technique could measure cardiac output in children with a congenital heart defect. The only previousstudy to evaluate children with congenital heart defects reported that the impedance technique overestimated systemic flow but correlated well with pulmonary flow.14 Unfortunately, because oxygen consumption was estimated in that study,14 the Fick blood flow measurements may have been questionable. The present study determines whether impedance cardiography accurately measures cardiac outpllt in the presence of

From the Department of Physiology and Biophysics, School of Medicine and College of Science and Mathematics, Wright State University, and the Department of Pediatrics, The Children’s Medical Center, Dayton, Ohio. This work was supported by a grant from the Miami Valley Chapter of the American Ilcart Association, Dayton, Ohio. Manuscript received August 10. 1987; revised manuscript received and accepted November 2.1987. Dr. Beekman’s present address: C.S. Mott Children’s I lospital, University of Michigan, F1123 Box66, Ann Arbor, Michigan. Ms. Golden’s present address: Physical Therapy Education, 2600 Stcindler Building, University of Iowa, Iowa City, Iowa. Address for reprints: Daniel S. Miles, PhD, Department of Physiology and Biophysics, School of Mcdicinc:, Wright State University, Dayton, Ohio 45401-0927. 612

March

either intracardiac or extracardiac shunting dren with dongenital heart defects.

1. 1988

THE AMERICAN

Pt

Patients: All participants underwent complete right- and left-sided heart catheterizations to confirm the diagnosis, evaluate the extent of a congenital defect or determine the success of corrective surgery. All catheterizations were completed at The Children’s Medical Center. Approval for this study was obtained from the Wright State University’s Human Subjects Institutional Review Board and the hospital’s review board. A total of 37 participants were divided into 4 groups based on the diagnosis. No shunts were present in a control group of patients (n = 11) presenting with a variety of disorders, such as aortic or pulmonary stenosis or aortic coarctation, as well as successfully repaired shunts. Intracardiac shunts were present in the secundum atria1 septal defect (ASD) group (n = 7) and in the ventricular septal defect (VSD) group (n = 12). Extracardiac shunts occurred with patent ductus arteriosus (PDA) patients (n = 7). Descriptive characteristics are listed in Table I. Catheterization: The patients were sedated preoperatively with a variety of medications such as morphine, vistaril, chlorohydrate or ketamine. The right femoral vein was used for right-heart catheterization. The inguinal region was infiltrated with 1 to 2% xyloCaine or lidocaine before percutaneous insertion of a Berman or Edwards angiographic catheter (4 to 8Fr) and passed with the aid of fluoroscopy. Left-heart catheterization was performed by inserting a regular Cook pigtail catheter (3 to 6 Fr) into the femoral artery. Pressures were measured with a fluid-filled transducer and recorded on a 4ichannel oscillographic recorder (Electronics for Medicine) Cardiac output: Systemic and pulmonary blood flows were measured by the direct Fick procedure. The direction and magnitude of shunting were also calculated using this procedure, with the superior vena cava providing the mixed venous sample. Oxygen saturation was measured (American Optical Unistat Oximeter) using 0.25 ml aliquots of blood. Blood oxygen carrying capacity was calculated as hemoglobin (g/dl) X 1.36. Oxygen consumption was measured with a Waters instruments continuous flow-through

II

Summary

(Mean

HR Control ASD VSD PDA

102 f (81-146) 100 f (75-128) 122 i (86-152) 102 f (79-135)

f Standard GO*

5.3 6.6 6.8 6.9

HR = heart rate (beatslmin); cardiac index; RLS = right-to-left oxygen consumption (cclminlm2).

156 f 8.1 (130-212) 119 f 7.0 (98-147) 165 f 9.4 (125-225) 151 f 8.8 (125-196)

Error

OF CARDIOLOGY

TABLE I Descriptive Characteristics and Range) of the Patients

in chil-

Methods

TABLE

JOURNAL

Age 0-m) Height

(cm)

Weight

(kg)

ASD = atrial arteriosus; VSD

Volume

(Mean

Control (n = 11)

ASD (n = 7)

58f 12 (6-130) 100 f 7 (63-140) 16.7 f 3 (5-35)

72 f 22 (6-171) 109 f 14 (57-164) 25.7 f 7 (8-60)

f Standard

VSD (n = 12) 32 f (2-90) 86 f (56-120) 11.8f2 (4-20)

septal defect: control = no shunts: = ventricular septal defect.

PDA

613

61

Error

PDA (n = 7) 9 6

39 f 12 (9-88) 89 f 7 (68-107) 12.7 f 1.8 (6-20) = patent

ductus

head-box system (MRMB - Oa Consumption Monitor). An average of three l-minute samples was used in the calculations. Cardiac output was measured by impedance at the time of the oxygen consumption measurements.17 Mylar tape electrode strips were positioned on the forehead, circumferentially around the neck and on a midaxillary to midaxillary line at the level of the xiphisternal notch and lower abdomen (Figure 1). The Minnesota Impedance Cardiograph Model 304B was used to introduce a constant sinusoidal alternating current of 4 mA at 100 kHz between the outer 2 electrodes (the inner electrodes are sensing electrodes and detect changes in voltage). Knowing the current applied and voltage produced, the cardiograph calculates the impedance based on Ohm’s law. Heart sounds were measured at the lower sternal border using a Hewlett Packard microphone (21050 A) and the electrocardiogram was obtained from an Electronics for Medicine ECG amplifier. The electrical impedance technique provides realtime values of stroke volume which are used with the heart rate to determine cardiac output.Z*4 The stroke volume is derived from analysis of the change in impedance with respect to time (dZ/dt) during the cardiac cycle. The electrocardiogram, phonocardiogram and dZ/dt were recorded at a paper speed of 100 mm/s with a Gould 4-channel strip-chart recorder (Figure 1). The equation used to calculate stroke volume was SV = pa (L/Z@ - dZ/dt - T and expressed as follows: SV = (D -cm) (cmZ/?2) (Q/s) (s) = cm3 (ml), where SV = stroke

and Range)

of the Catheterization

Procedure

GPl/&l

LRS

RLS

RPIRS

1.0

0

0

2.60 f 0.36 (1.3-4.5) 2.30 f 0.38 (1.3-6.0) 1.86 f 0.15 (1.3-2.3)

4.68 f 0.66 (1.4-7.0) 5.55 f 1.21 (1.2-16.7) 3.84 f 0.59 (1.4-6.0)

0.02 i 0.02 (O-O. 1) 0

0.13 f 0.02 (0.06-0.30) 0.04 f 0.008 (0.02-0.07) 0.09 f 0.02 (0.04-0.21) 0.09 f 0.03 (0.05-0.26)

LRS = left-to-right shunt (liters/min/m*); shunt (liters/min/m2); RPlRS = ratio Other abbreviations as in Table I.

0

GPl/&l = ratio of pulmonary to systemic of pulmonary to systemic resistance; VOn =

614

IMPEDANCE

CARDIOGRAPHY

FOR CARDIAC

TABLE III Summary (Mean f Standard Direct Fick and impedance Blood Flow

Error

4.S Control

VSD PDA

MEASUREMENT

and Range)

of

i;)P

3109 f 0.45 (0.96-5.17) 2.72 f 0.49 (0.36-4. IO) 2.59 f 0.36 (0.87-4.51) 2.48 f 0.35 (1.34-3.75)

ASD

OUTPUT

3.22 f 0.59 (0.41-6.31) 2.86 f 0.71 (0.41-5.62) 2.52 f 0.48 (0.26495) 2.49 f 0.44 (0.90-3.76)

: p 6s or bz. QP = Fick.pulmonary blood flow (liters/min); (literslmin); QZ = impedance systemic blood tions as in Table I.

3.09 f (0.96-5.17) 6.56” f (1.63-10.96) 5.02” f (1.97-7.33) 4.51’ f (2.46-6.51)

0.45 1.30 0.47 0.62

6s = Fick systemic blood flow flow (liters/min). Other abbrevia-

volume (ml]; p = resistivity (fi . cm] of blood and assumed to be a constant18 of 135; L = anterior midline distance between the inner electrodes (cm); 20 = total thoracic impedance (n] between the sensing electrodes (2 and 3); dZ/dt = the maximum rate of change of the thoracic impedance (Ws) during each heart beat

ECG (Spot Electrodes)

measured from the beginning of the rapid upstroke independent of the calibration baseline as previously described in our laboratorylg; T = left ventricular ejection time (s). Heart rate was determined on a beat-by-beat basis from the Q-Q interval of the electrocardiqgram and multiplied by stroke volume to obtain impedance systemic blood flow. The mean impedance blood flow was obtained by averaging 10 heartbeats. Impedance data were hand-digitized on a Houston Instruments True Grid 1017 digitizing pad interfaced with an IBMPC XT computer. Statistical analyses: Means, standard errors, regression analysis, correlations and Student-dependent t test were run using the Crunch,Interactive Statistical Package (CRISP]. All null hypotheses were rejected at the 5% level.

Results Heart rates for the 4 groups were similar during catheterization (Table II). The greater Fick pulmonary blood flow in the ASD, VSD and PDA groups (compared with Fick systemic blood flow] could be attributed to the left-to-right shunting. The ratio of pulmonary resistance to systemic resistance was similar in all groups. The systemic blood flow measured by impedance was similar (r = 0.84, p <0.05) to Fick measurements (Table III, Figure 2). The difference between the blood flow values for the 2 techniques was 15% in each of the 4 groups (Figure 3). Blood flow correlations between the 2 techniques ranged from 0.69 to 0.89, with the highest in the ASD group and lowest in the VSD

Comparison #I Impedance

x2

of Fick and Impedance

#3 %4 Leads

(ZCG)

PCG 0 Control

2

3

4

5

6

7

Fick 6s (timin) FIGURE 1. Schematic representation of patient instrumentation records of the electrocardiogram (ECG), phonocardiogram and impedance cardiogram (ZCG).

and (PCG)

FIGURE 2. Correlation and regression between the measurement of Fick (4s) and impedance (&) systemic blood flow for all subjecis including atrial septal defects. (ASD), ventricular septal defects (VSD), patent ductus arteriosus (PDA) and no shunts (control).

March

1. 1988

Repeated tests can be easily administered using the impedance technique for assessment of cardiac output without any known risk or discomfort to the patient. It may prove to be particularly useful in providing a means to serially evaluate patients after surgery or over the course of medical or surgical management. It must be convincingly demonstrtited, however, that an accurate estimate of cardiac output is being measured. The present patient population offered an opportunity to evaluate the impedance technique using the established direct Fick procedure as the standard. Nevertheless, it should be noted that even blood flow measured by the direct Fick procedure is not without inherent errors. Very few studies evaluating neonates or children with impedance cardiography have been published. We have previously reported that healthy premature infants weighing 0.8 to 3.8 kg could be evaluated accurately.lG Premature infants with patent ductus arterioSUS,~~~~*healthy newborn infants11J2J2 and childrenz3 also have been studied. In general, these studies hatie reported cardiac output values consistent with the population under consideration but lacked a comparison with an invasive technique. The work of Lababidi et all4 is the only other study published where a comparison between direct Fick values of cardiac output and impedance has been evaluated in children with and without cardiac shunts. Our results were similar to Lababidi et al in demonstrating that cardiac outputs measured by either the Fick or impedance techniques were similar (r = 0.69) in children without shunts or valvular insufficiency. Our results were at variance, however, with their values reported for children with intracardiac left-toright shunting. When they evaluated 21 patients with left-to-right shunts, the impedance cardiac output overestimated systemic flow and correlated well with pulmonary blood flow (r = 0.92) rather than systemic flow [r = 0.21). In sharp contrast, our findings suggested close agreement between the Fick and impedance systemic blood flow measurements in either the ASD (r = 0.89) or VSD (r = 0.691 groups. The reason for the discrepancy between ihe 2 studies of patients with intracardiac shunts is not readily apparent. Oxygen consumption was directly measured in the present study whereas in the Labibidi et al study it was estimated. It has been demonstrated that oxygen consumption in sedated children varies during catherization proceduresz4 and that oxygen consumption is higher in children with cardiovascular disease.25 If oxygen consumption was underestimated by Lababidi et al, Fick estimates of pulmonary and systemic flow would also be underestimated. Perhaps this may explain some of the observed differences. Furthermore, we used a constant of 135 D - cm for the blood resistivity factor in our calculation of stroke volume,l* where-

JOURNAL

OF CARDIOLOGY

Control No Shunts

and control groups. Fick pulmonary blood flow was significantly greater than systemic flow as determined by either Fick or impedance in the ASD, VSD and PDA groups (Table III, Figure 3).

Discussion

THE AMERICAN

Volume

615

61

ASD L-R Intracardiac

Shunt Fick 6P

‘-A.

l

,’

5? E =

4-

0

1

2

3

4

5

6

0

Flck 6s (Imm)

VSD L-R Intracardiac

PDA L-R Extracardiac

Shunt

4

2 Fick &i

4

6

6

10

12

Ffck 6s (I’min)

8r

0

2

6

D.

Shunt ,’

8

(I min)

FIGURE 3. Individual systemic (&) and pulmonary (&, open cjrcles) blood flow determined by the Fick procedure, and systemic blood flow determined by impedance (42) for (A) control, (B) atrial septal defect (ASD), (C) ventricular septal defect (VSD) and (D) patent ductus arteriosus (PDA) groups.

as Lababidi et al based bldod resistivity calculations on the hematocrit. The total number of patients evaluated with intracardiac shunts was only 19 in, the present study and 21 in Lababidi’s. Clearly, further data are needed before a definitive conclusion can be drawn. In the patients with PDA (extracardiac left-to-right shunts), impedance measurements reflected systemic blood flow (r = 0.74). Pulmonary blood flow measured by the Fick technique was consistently higher than either Fick or impedance systemic flow in all patients. These results were again at variance with the only 2 previously published studies of patent ductus arteriosus patients.20J1 Cotton et alzO reported that the magnitude of the dZ signal using a tetrapolar electrode configuration was indicative of pulmonary blood flow in infants with symptomatic PDA. Spontaneous or pharmacologic closure of the ductus was coincident with an amplitude decrease in the AZ signal. Cardiac output, however, was not measured. Sexson et alzl demonstrated in a preliminary report that indotiet’hacin significantly reduced cardiac output within 24 hours of therapy for treatment of PDA. Neither the study by Cotton et alzO nor Sexson et alzl used an alternative technique for corroboration df the impedance values.

References 1. Bernstein DP. Continuous noninvasive real-time monitoring ume ond cardiac output by thorocic electrical bioimpedance.

of stroke volCrit Care Med

616

IMPEDANCE

CARDIOGRAPHY

FOR CARDIAC

OUTPUT

MEASUREMENT

1986;14:898-901. 2. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA, Mattson RH. Development and evaluation of an impedance cardiac output system. Aerospace Med 1966;37:1208-1212. 3. Miles DS, Sawka MN, Hanpeter DE, Foster JE Jr, Doerr BM, Frey MAB. Central hemodynamics during progressive upper- and lower-body exercise and recovery. 1 Appl Physiol 1984;57:366-370. 4. Ebert TJ, Eckberg DL, Vetrovec GM, Cowley MJ. Impedance cardiograms reliably estimate beat-by-beat changes of left ventricular stroke volume in humans. Cardiovasc Res 1984;18:354-360. 5. Gotshall RW, Breay-Pilcher JC, Boelcskevy BD. Cardiac output in adult and neonatal rats. Am J Physiol 1987;253:H1298-Hl304. 6. Miles DS, Gotshall RW, Sexson WR. Evaluation of impedance cardiography in the canine pup. J Appl Physiol 1986;60:260-265. 7. Miles DS, Sawka MN, Wilde SW, Doerr BM, Frey MAB, Glaser RM. Estimation of cardiac oufput by electrical impedance during arm exercise in women. J Appl Physiol 1981;51:1488-1492, 6. Goldstein DS, Cannon RO III, Zimlichman R, Keiser HR. Clinical evaluation of impedance cardiography. Clin Physiol 1986:6:235-251. 9. Hubbard WN, Fish DR, McBrien DJ. The use of impedance cardiography in heart failure. Int r Cardiol 1986;12:71-79. 10. Barbacki M, Gluck A, Sandhage K. Estimation of the correlation between the transcutaneous aortic flow velocity curve and impedance cardiogram in normal children. Cor Vasa 1981;23:291-298. 11. Costeloe K. Stocks J, Godfrey S. Cardiac output in the neonatal period using impedance cardiography. Pediatr Res 1977;11:1171-1177. 12. Freyschuss U, Noack G, Zett&strom R. Serial measurements of thoracic impedance and cardiac output in healthy neonates after normal delivery and caesarean section. Acta Paediatr Stand 1979;68:357-362. 13. Halpern 8, Mannino F. Right-to-left cardiac shunting in neonates determined by tetrapolar cardiac thoracic impedance measurements (abstr]. Ciin Res 1979;27:125A. 14. Lababidi Z, Ehmke DA, Durnin RE, Leaverton PE, Lauer RM. Evaluation of impedance cardiac output in children. Pediatrics 1971;47:870-879. 15. McKinley DF. Pollack MM. A comparison of thoracic bioimpedance to

thermodilution cardiac output in critically ill children (abstr). Crit Care Med 1987;15:358. 16. Sexson WR, Gotshall RW, Miles DS. Serial assessment of central hemodynamics in growing premature infants using impedance cardiography (abstr]. Physiologist 1984;27:210. 17. Kubicek WG, Witsoe DA, Patterson RP, Moshavrata MA, Karnegis JN, From AHL. Development and Evaluation of an Impedance Cardiogiaphic System to Measure Cardiac Output and Development of an Oxygen Consumption Rote Computing System Utilizing a Quadrapole Mass Spectrometer. Houston. Texas: Nationai Aeronautics and Suace Administration. 1969: l-151. 16. Quail AW, Traugott FM, Porges WL, White SW. Thoracic resistivity for stroke volume calculation in impedance cardiography. r Appl Physiol 1981;50:191-195. 19. Doerr BM, Miles DS, Frey MAB. Influence of respiration on stroke volume determined by impedance cardiography. Aviat Space Environ Med 1981;52:394-398. 20. Cotton RB, Lindstrom DP, Olsson T, Riha M, Graham TP, Selstam U, Catterton WZ. Impedance cardiographic assessment of symptomatic patent ductus arteriosus. J Pediatrics 1988:96:711-715. 21. Sexson WR, Holcomb JW, Gotshall RW, Critz AD. Impedance hemodynamic evaluation of neonates with patent ductus arteriosus during indomethatin therapy (abstr). Pediatr Res 1987;21:388A. 22. Noack G. Freyschuss U. The early detection of pneumothorax with transthoracic impedance in newborn infants. Acta Paediatr Stand 1977;66: 677-680. 23. Edmunds AT, Godfrey S, Tooley M. Cardiac output measured by tronsthoracic impedance cardiography at rest, during exercise and at various lung volumes. Clin Sci 1982;63:107-113, 24. Baum D, Brown AC, Church SC. Effect of sedation on oxygen consumption of children undergoing cardiac catheterization. Pediatrics 1967;39:891895. 25. Lees MH, Bristow JD, Griswold HE, Olmstead RW. Relative hypermetabolism in infants with congenital heart disease and undernutrition. Pediatrics 1965;36:183-185.