METHODS

Noninvasive Determination of Systolic, Diastolic and End-Systolic Blood Pressure in Neonates, Infants and

Young Children: Comparison with Central Aortic Pressure Measurements

STEVEN D. COLAN, MD, ALAN FUJII, MD, KENNETH M. BOROW, MD,

DEBBIE MacPHERSON, BS, and STEPHEN P. SANDERS, MD

Noninvasive determinations of systolic and diastolic blood pressure using the osciiiometric method for pressure measurement were combined with exter- nally recorded axiiiary pulse tracings to estimate end-systolic pressure in 32 neonates, infants and young chikiren. Resuits were compared with central aortic pressure measurements made at the time of central aortic catheter placement. Studies were performed in patients aged 1 day to 48 months who weighed 0.9 to 18.1 kg. A wide range of systolic (41 to 141 mm Hg), diastolic (22 to 73 mm Hg) and end-systolic (30 to 111 mm Hg) pressure values were found. The mean absolute pressure differences and percent errors (pressure difference divided by

central aortic pressure) were 1.8 mm Hg and 2.5 % for systolic, 0.8 mm Hg and 0.8% for diastolic and 1.4 mm Hg and 2.1% for end-systolic pressure. No correlation was noted between percent error and age, weight, heart rate, cardiac index or systemic vascular resistance. The ability to perform reliable noninvasive pressure measurements should prove invaluable for clinical and research purposes. in addition, this method of end-systolic blood pressure determination enables sensitive indexes of left ventricular contractility to be measured noninva- siveiy in small children.

(Am J Cardioi 1983;52:867-870)

Assessment of left ventricular (LV) contractility in small children is a difficult problem. Recent studies performed on older children and adults suggest that the LV end-systolic pressure-dimension and wall stress- shortening relations are highly sensitive indexes of LV contractile state that can be determined using nonin- vasive methods.1-4 These indexes are particularly at- tractive for neonates and infants because they overcome the preload and afterload dependency problems asso- ciated with ejection phase indexes. This is done by quantifying the forces (i.e., pressure or wall stress) that resist myocardial fiber shortening during systole and

From the Department of Cardiology, Children’s Hospital, the Cardio- vascular Division, Brigham and Women’s Hospital, and the Departments of Medicine and Pediatrics, Harvard Medical School, Boston, Massa- chusetts. This study was supported in part by Grant HL07193 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland. Manuscript received April 1, 1983; revised man- uscript received June 2 1, 1983, accepted June 23, 1983.

Address for reprints: Steven D. Cola% MD, Departmsnt of Cardiilogy, Children’s Hospital, 300 Longwood Avenue, Boston, Massachusetts 02115.

necessitates accurate methods for noninvasive blood pressure (BP) determinations.5y6 However, standard methods for BP measurements in neonates and infants are unreliable 7-13

This study evaluates the accuracy of an automated noninvasive BP device that uses the oscillometric principle to measure brachial artery systolic and dia- stolic pressures. These values can be used for cali- bration of externally recorded arterial pulse tracings and calculation of the end-systolic pressure.14 Values for systolic, diastolic and end-systolic pressures determined by this noninvasive technique are compared with si- multaneous central aortic pressure values in neonates, infants and young children.

Methods

Patients: Our study population consisted of 32 infants and children who were undergoing aortic catheter placement. The patients were 1 day to 48 months old (median 6 months) and weighed 0.9 to 18.1 kg (median 5.0). Twenty-three had con- genital cardiac lesions and 9 were neonates free of congenital cardiac anomalies but had significant respiratory disease.

“.._

068 NONINVASIVE BLOOD PRESSURE MEASUREMENTS IN INFANTS AND CHILDREN

i

PEAK SYSTOLIC PRESSURE

tND-SYSTOLIC PRESSURE

V

DIASTOLIC PRESSURE

END-SYSTOLIC = + PULSE PRESSURE

PRESSURE 1 I[ 3 II + DIASTOLIL PRESSU'IL J

FIGURE 1. End-systolic pressure calculation. The carotid tracing is calibrated by assignment of systolic pressure to the peak and diastolic pressure to the nadir of the tracing. End-systolic pressure is then de- termined by linear interpolation.

Blood pressure determinations: The noninvasive BP measurements in this study were made using either the Di- namap@ 847 neonatal or the Dinamap@ 845 vital signs monitor. This microprocessor-controlled device uses the oscillometric method to estimate arterial pressure and has been described in detail.5,6 The principle of operation is the detection of ar- terial pressure oscillations transmitted to a pneumatic cuff encircling the upper arm. Central aortic BP was measured using Hewlett-Packard series 1280 physiological pressure transducers attached to either No. 3.2Fr or 4.OFr pigtail catheters or No. 3.5Fr or 5.OFr umbilical artery catheters. External axillary pulse tracings were recorded with a standard, funnel-shaped pick-up connected to an Irex 120-13202 strain gauge. Simultaneous axillary and central pulse tracings were recorded on an Irex Continutrace model 101 at a paper speed of 100 mm/s.

Study protocol: In 21 of 23 patients studied during cardiac catheterization, routine right- and left-sided cardiac cathe- terization and oxygen consumption by the Fick method were performed. In all patients the largest BP cuff that could be accommodated on the child’s upper arm was positioned with the cuff bladder centered over the brachial artery. Three consecutive Dinamap readings at l-minute intervals were obtained with simultaneous recordings of central pressure and axillary pulse tracings. Measurements disrupted by patient movement or crying were discarded, and recordings were continued until 3 successive determinations were obtained with the patient in a quiet, resting state.

Data analysis: Noninvasive BP values were determined by averaging the 3 successive Dinamap values. The axillary pulse tracings were then calibrated using the method of Ste- fadouros et alI4 by assignment of systolic pressure to the peak and diastolic pressure to the nadir of the tracing. End-systolic pressure was estimated by linear interpolation to the level of the incisura (Fig. 1). Central aortic values were taken as the average of 20 consecutive beats.

Statistical analysis: For each patient, the difference be- tween aortic and Dinamap pressure values were calculated, as well as the respective percent errors (aortic minus Dinamap pressure divided by aortic pressure). Central and Dinamap pressures were compared by a simple linear regression (least-squares method). The random error of the 2 methods

100 -

B

E 80-

2

60 -

40 -

20 -

1 I I 1 I

20 40 60 00 100 120 140

PolN (mmHg)

FIGURE 2. The relation between central aortic (PAO) and Dinamap (PoIN) measurements for systolic pressure. The linear regression equation and correlation coefficient (r) are given. The line of identity (solid line) and least-squares regression line (dotted line) are shown.

was compared by examining the correlated variances and calculation of the sample correlation coefficient.15 To deter- mine if age, weight, cardiac index, systemic vascular resistance or heart rate influenced the accuracy of the Dinamap mea- surements, coefficients of correlation for each of these vari- ables versus the percent error in measurement were calcu- lated.

Results

Pressure data: Pressure measurements for the in- dividual patients, as well as the difference and percent error are presented in Table I. Seventy-five percent of systolic Dinamap values were within 5 mm Hg of the aortic value, and 94% were within 10 mm Hg. Dinamap

140 7

120'

100 u

z

E 80.

2

60'

40'

I 1 I 1 4

20 40 60 80 100 120 140

PO,,, (mmHg)

FIGURE 3. The relation between central aortic (PAo) and Dinamap (PoIN) measurements for diastolic pressure. The linear regression equation, correlation coefficient, line of identity, and least-squares regression line are given as in Figure 2.

October 1,1983 THE AMERICAN JDURNAL DF CARDIOLOGY VolWne 52 889

TABLE I Comparison of Aortlc and Dinamap Pressure Data

Systolic Blood Pressure (mm Hg) Diastolic Blood Pressure (mm Hg) End-Systolic Blood Pressure (mm Hg)

Pt Ao Din Diff % Error Ao Din Diff % Error Ao Din Diff % Error

TY +7 -1

+-Zl +2

+16 0

-20 -8 +6

-x -4 -6

-15 -14

-8

::

+G

:: +11

-s -2

61 58

-1

zz T:

+5 :s4 +5 +5 -2 -3

z; :: +5 +6

-Y -z

$7 :: +1 +1 -2 -2 +1.4 +2.1

2.3 3.4 Mean for group +1.8 +2.5 +0.8 +0.8 SD for group 4.7 5.3 4.2 9.2

Ao = central aortic pressure; Din = Dinamap pressure; Diff = difference between central aortic and Dinamap pressure measurement; % error = percent error (difference between central awtic and Dinamap pressure measurements divided by central aortlc pressure reading); SD = standard deviation.

and central systolic values correlated highly (r = 0.979) with a regression equation of aortic pressure = 0.99 (Dinamap pressure) +3.0 (Fig. 2). Dinamap diastolic pressure values were within 5 mm Hg of central values in 84% of the cases and within 10 mm Hg in 97%. A high degree of correlation was found (r = 0.941), with a cal- culated regression line of aortic pressure = 1.09 (Dina- map pressure) -3.3 (Fig. 3). Ninety-seven percent of noninvasively determined end-systolic pressure values were within 5 mm Hg of the central value and 100% were within 10 mm Hg. The correlation between the 2 methods was very high (r = 0.993), with the central end-systolic pressure best predicted by the equation aortic pressure = 1.01 (Dinamap pressure) + 0.6 (Fig. 4). There was no significant difference in the correlated variances for any of the 3 pressure determinations.

The patients’ age, weight, heart rate, cardiac index and systemic vascular resistance correlated poorly with the percent error for each of the 3 pressures (Table II).

Discussion

Correlation of Dinamap and central aortic pressures: Accurate systolic and diastolic BP values were measured with the Dinamap device for the neo- nates, infants and young children in our study. The re-

sults correspond closely to the previous findings with this device in adults.6 In addition, the use of externally recorded axillary pulse tracings permitted accurate estimation of end-systolic pressure. No systematic error due to age, weight, heart rate, cardiac index or systemic vascular resistance was detected.

Use of axillary pulse: The axillary pulse tracing was used to calculate end-systolic pressure because of technical difficulties in recording adequate carotid pulse tracings in small patients. With the arm held fully ex- tended at the shoulder, pulsations of the axillary artery may be easily recorded as it passes over the head of the humerus. In our experience, axillary pulse tracings were easier to record than carotid pulse tracings and were much less distressing to the infant. The tracings re- corded are less influenced by respiratory activity and are free of swallowing artifact, permitting feeding to be used to pacify the infant during the study. The axillary pulse tracing closely resembled the central aortic trac- ings, with the incisura occurring at a comparable level.

Advantages of the Dinamap device: The inade- quacy of traditional mercury manometer-stethoscope BP measurement techniques in infants has prompted a search for other methods.16 The accuracy of the dia- stolic pressure measurements recorded with Doppler

070 NONINVASIVE BLOOD PRESSURE MEASUREMENTS IN INFANTS AND CHILDREN

80 . I/‘ P.~-l.ol P,,,*o.6

40 r -0.993

20 :

I I 1 J

20 40 80 80 100 120 140

PA p, (mmHg)

FIGURE 4. The relation between central aortic (PA,,) end-systolic pressure and the pressure calculated by linear interpolation of the ax- illaty pulse tracing (PApT) from Dinamap systolic and diastolic pressures. The linear regression equation, correlation coefficient, line of identity, and least-squares regression line are given as in Figure 2.

devices has been poor, with correlation coefficients of 0.5 to 0.7 compared with intraarterial measurements of diastolic pressure.l”J7 In addition to delivering accurate measurements of diastolic pressure, the automatic de- vice decreases the number of personnel needed for BP monitoring. Finally, the Dinamap does not use a transducer or microphone (as do other commercially available devices), eliminating problems of cuff dis- placement by patient movement.

Limitations of Dinamap: Vigorous respiratory ac- tivity, crying or struggling causes wide swings in arterial pressure and augment the beat-to-beat variability. The design of the Dinamap device necessitates a moderately stable physiologic state for accurate measurements. In this study, non-neonates were, by necessity, evaluated during light sedation in the catheterization laboratory. Although this degree of cooperation may not exist during routine BP measurements, we do not hestitate to administer a light sedative in our noninvasive labo- ratory to ensure high-quality echo- and phonocardio- grams. Finally, marked irregularity of cardiac rhythm may interfere with measurements. This is an uncommon problem in this age group and was not addressed in the present study.

Clinical implications: The availability of accurate, noninvasive BP determinations in neonates, infants and young children creates numerous clinical and research applications. It may eliminate the need for direct in- traarterial pressure measurements in some patients who require careful vital signs monitoring, such as those undergoing pharmacologic or diagnostic interventions. In addition, it should allow more sophisticated exami- nation of the physiologic changes that occur in the normal neonate and infant. Finally, the ability to obtain reliable end-systolic pressure values permits noninva- sive analysis of sensitive indexes of LV contractility,

TABLE II Correlation of Measurement Error With Clinical and Hemodynamic Factors

Systolic Blood Diastolic Blood End-Systolic Blood Pressure Pressure Pressure r Value r Value r Value

Age 0.216 0.271 -0.137 Weight 0.259 0.379 0.032 !P -0.064 0.201 0.020 0.342 -0.286 0.124

SVR 0.030 -0.080 0.376

Cl = cardiac index; HR = heart rate; r value = correlation coefficient; SW = systemic vascular resistance.

such as the end-systolic pressure-dimension and wall stress-shortening relations.1-4 This is particularly im- portant in small children, in whom noninvasive studies of myocardial function have been unreliable and diffi- cult to perform.

Acknowledgment: We thank Mary Gillan, Joanne Miccile, and Barbara Cesaro for assistance in preparation of this manuscript, Emily Flynn for the graphics, and the technical and nursing personnel of the catheterization laboratory and neonatal intensive care units for their cheerful assistance in data collection.

References

1. Borow KM, Neumann A, Wynne J. Sensitivity of end-systolic pressure- dimension and pressure-volume relations to inotropic state in humans. Circulation 1982;85:988-997.

2. Borow KM, Propper R, Blerman FZ, Grady S, lnatf A. Left ventricular end-systolic pfessuredimension relation in patients with thakssemia major: a new noninvasive method for assessing contractile state. Circulation 1982;88:980-985.

3. Borow KM, Green LH, Grossman W, Braunwald E. Left ventricular end- systolic stress-shortening and stress-length relations in humans. Normal :;I,I+;s and sensitivity to inotropic state. Am J Cardiol 1982;50:1301-

.I”_.

4. Borow K, Grady S, Henderson C, Papish S, Goorln A. Early detection of left ventricular contractile abnormalities in humans (abstr). Circulation 1982;88:suppl ll:ll-250.

5. Silas JH, Barker AT, Ramsay LE. Clinical evaluation of Dinamap’” 845 automated blood pressure recorder. Br Heart J 1980;43:202-205.

8. Borow KM. Newburner JW. Noninvasive estimation of central aortic pressure ubing the o>cillometric method for analyzing systemic artery pulsatile blood flow: comparative study of indirect systolic, diastolic, and mean brachial artery pressure with simultaneous direct ascending aortic pressure measurements. Am Heart J 1982;103:879-888.

7. Cappa BE, Pallln IM. Systolic blood pressure determination in the newborn and infant. Anesthesiolo

8. Gddrlng D, Wholknann B y 1952;13:848-849. . Flush method for blood pressure determinations

in newborn infants. J Pediatr 1952;40:285-289. 9. Moaa AJ, Llebllng W, Austin WO, Adams FH. An evaluation of the flush

method of determinino blood pressure in infants. Pediatrics 1957;20:

10.

11.

12.

13.

14.

15.

18.

17.

53-82. McLaughlin GW, Kirby RR, Kemmerer WT, deLemos RA. Indirect mea- surement of blood pressure in infants utilizing Doppler ultrasound. J Pediatr 1971~7!+300-303 _ , _ _ _ _ _ _ _ Kirkland RT, Kirkland JL. Systolic blood pressure measurement in the newborn infant with the transcutaneous Doppler method. J Pediatr 1972; 80:52-58. Hartmann AF, Kllnt R, Hernandez A, Goldring D. Measurement of blood pressure in the brachial and posterior tibia1 arteries using the Doppler method. J Pediatr 1973;82:498-500, Reder RF, Mmkh I, Cohen ML, Stelnfeld L. Evaluating indirect blood

measuement techniques: a comparison of three systems in infants %%&en. Pediatrics 1978;82:328-330. Stefadouroa MA, Dough&y MJ, Groaaman W, Craige E. Determination of systemic vascular resistance by a noninvasive technique. Circulation 1973;47:101-107. Bnadacor GW, Co&ran WG. Statistical Methods. Ames, Iowa: Iowa State University Press, 1980: 190. McLaln LG. Hvoertension in childhood: a review. Am Heart J 1976:92: 834-847. .’ Whyte RK, Elseed AM, Fraser CB, Shlnebourne EA, deSwlet M. Assess- ment of Doppler ultrasound to measure systolic blood pressures in infants and young children. Arch Dis Child 1975;50:542-544.