Development of baroreflex influences on heart rate variability in preterm infants

Early Human Development 53 (1998) 37–52

Development of baroreflex influences on heart rate variability in preterm infants Jon E. Mazursky a , Clay L. Birkett b , Kurt A. Bedell a , Schlomo c a, A. Ben-Haim , Jeffrey L. Segar * a

Department of Pediatrics, University of Iowa, Iowa City, IA 52242, USA Department of Internal Medicine, University of Iowa, Iowa City, IA 52242, USA c Department of Physiology and Biophysics, Rappapport Family Institute for Research in the Medical Sciences, Haifa, Israel b

Accepted 14 April 1998

Abstract To investigate developmental changes in autonomic cardiovascular reflexes in preterm infants, we used autoregressive power spectral analysis to analyze the effect of upright tilting on heart rate variability in preterm infants. Twenty-eight infants were studied in a longitudinal fashion beginning at 28–32 weeks postconceptional age (postnatal age 1–5 weeks). Each week, heart rate variability in the supine position and after 458 head-up tilt was analyzed by spectral analysis. With the initial study of each infant, there was no significant change in heart rate following head-up tilt compared with baseline (20.560.9 bpm). However, linear regression analysis revealed that with increasing postnatal age, the change in heart rate in response to tilting became more positive (mean slope of regressions 0.4560.12 bpm / week, P , 0.005). The power spectral density of R–R interval variability in the low—(LF; 0.02–0.15 Hz) and high—(HF; 0.15–1.5 Hz) frequency ranges were obtained and the values normalized by dividing each component by the total power. For measurements obtained in the supine position, the LF / HF ratio progressively decreased with increasing postnatal age, indicating a maturational change in sympathovagal balance. We used the difference in the LF / HF ratio between tilt and the recumbent position as a measure of the change in autonomic input to the heart in response to unloading of the arterial baroreceptors. No significant change in these ratios were observed when infants were first studied between 28 and 32 weeks postconceptional age, suggesting that the cardiac baroreflex is poorly developed at this stage of development. However, with postnatal maturation, the LF component of the power spectrum became progressively larger with tilt relative to the basal state, such that the difference between *Corresponding author. Tel.: 11 319 3567244; fax: 11 319 3568669. 0378-3782 / 98 / $ – see front matter PII: S0378-3782( 98 )00038-3

 1998 Elsevier Science Ireland Ltd. All rights reserved.

38

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

LF / HF tilt and LF / HF base became progressively more positive (P , 0.006). These findings suggest that in premature infants, cardiac baroreceptor reflexes become more functional with postnatal development.  1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Spectral analysis; Arterial baroreflex; Autonomic nervous system; Neonates

1. Introduction It is well recognized that short-term alterations in arterial pressure modify the discharge of afferent baroreceptor fibers located in the regions of the carotid sinus and aortic arch [23]. This results in alterations in efferent parasympathetic and sympathetic nerve activities which influence heart rate (HR) and peripheral vascular resistance, both effects acting to buffer changes in arterial pressure [1,23] Studies in adults have shown that passive head-up tilting results in a small, transient decrease in arterial blood pressure, which in turn evokes neurally mediated peripheral vasoconstriction and cardiac acceleration [8,9]. Postural changes have similarly been performed in infants to assess the integrity of the cardiovascular response to baroreceptor activation. Several investigators have demonstrated in healthy term and preterm infants that head-up tilting produces a significant HR response [15,33,44] and that the magnitude of the response is proportional to the degree of tilting [44]. These results suggest that a functional baroreflex is present in the neonate. In contrast, other investigators have been unable to demonstrate consistent responses of HR to tilting, and concluded that the HR component of the baroreflex is poorly developed during the neonatal period [21,50]. Fluctuations or variability in HR reflect rhythmic changes in efferent excitatory and inhibitory autonomic activity generated via neural regulatory mechanisms. Computerized spectral analysis of HR variability has generally shown these rhythmical oscillations to be concentrated in two frequency ranges [4,25,30]. The low frequency (LF) component (0.02–0.15 Hz) reflects vasomotor activity associated with baroreceptor activity and thermoregulatory functions, and is influenced by cardiac sympathetic and parasympathetic afferent activity [24,30,32]. The high frequency (HF) peak, occurring above 0.15–0.2 Hz, corresponds to the respiratory frequency and is mediated by parasympathetic activity [30,32]. Studies in adults have demonstrated the capability of spectral analysis to assess changes in the sympathovagal balance induced by an orthostatic stimulus [28,30,35]. Tilt to the upright position is felt to be accompanied by cardiac sympathetic excitation and vagal withdrawal to the sinus node pacemaker [24,30,32]. In normotensive adults, tilting results in an increase in the LF component and a decrease in the HF component, the size of the changes being correlated with the degree of tilt [28]. Studies in infants utilizing spectral analysis have demonstrated maturational changes in efferent parasympathetic and sympathetic activity to the heart [12,13,45]. However, the role of arterial baroreceptors in modifying cardiac sympathetic and parasympathetic activity early during development has not been well studied. Therefore, the present study was designed to characterize the developmental changes

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

39

in sympathovagal balance in response to arterial baroreceptor unloading in preterm infants. More specifically, we longitudinally examined the changes in the HR power spectrum in response to head-up postural change in maturing preterm infants.

2. Materials and methods

2.1. Subjects The study protocol was approved by the University of Iowa Human Subjects Review Committee. Informed consent was obtained from the parents of all patients before enrollment. Infants with congenital anomalies, intraventricular hemorrhage, cardiac disease (including presence of patent ductus arteriosus), perinatal asphyxia, or evidence of maternal illicit drug use were excluded from study. Infants studied while mechanically ventilated had capillary pH values greater than 7.32 and pCO 2 values between 43–60 mmHg. No infants were receiving adrenergic, anticholinergic or paralytic agents at the time of study, although the majority of infants were receiving a methylxanthine (primarily caffeine citrate) for treatment of apnea of prematurity. Clinical care was not affected by participation in this study. Estimating a 30% difference in the HR variability response, as measured by the difference in the normalized power of the LF component of the spectra between rest aad tilt [28,30,35], a test power of 0.80, and an alpha of 0.05, we calculated that 30 patients were required for study [51]. Therefore, a total of 32 infants, born at 24 to 31 weeks gestation were enrolled during the first 1 to 4 weeks of life and studied serially on a weekly basis until discharge to home or transfer to another hospital (range 3 to 11 weeks). The majority of patients were initially studied beginning at 29 or 30 weeks postconceptional age. Two infants were started at 28 weeks, and no infant was enrolled beyond 32 weeks postconceptional age. To minimize possible influences of circadian rhythms on HR variability and power spectral densities (PSDs), all studies were performed between 10:00 a.m. and 3:00 p.m. Infants were studied during periods of sleep with few small movements and regular respirations, usually 1–2 h after being enterally fed. No further attempt was made to distinguish the behavioral (sleep) state of the infant.

2.2. Study protocol For each study, the infant was secured using a Velcro-strap harness in a supine position to a customized plexiglass board which could be smoothly tilted to a 458 head-up position over 2–3 s and locked into place. The electrocardiograms (ECG) and respiratory cycles were monitored using a modified Hewlett Packard 78304A cardiorespiratory monitor (Palo Alto, CA, USA) with standard ECG electrodes attached to the infants chest and abdomen. Prior to tilting, a 1 min baseline ECG was recorded with the infant in the recumbent position. The infant was then elevated to the 458 head-up position, and once having achieved the full tilt position, the ECG was obtained for an additional minute. The use of spectral analysis requires that the data

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

40

be stationary. For these reasons, periods of data were selected that were absent of sighs, startles, gross body movements or acute changes in breathing patterns, which may impose marked instability on the power spectrum. To further minimize skewing of the data by unrecognized irregular breathing or nonstationary data, the tilting maneuver was performed in triplicate, allowing a 5 min stabilization period in the flat supine position between each trial and the corresponding PSDs averaged for each week. The analog ECG signal was digitized using a Vetter Model 3000A PCM recording adapter (AR Vetter, Rebersburg, PA, USA) and stored on videotape using a Sony VHS Hi-Fi SLV-585HF / 686HF VCR. Off-line analysis of the ECG was performed on a 486 personal computer. ECG data were played back from the VHS videotape and digitized at 1 kHz using an analog to digital converter. The computer program automatically derived the time appearance of the R-waves from the ECG signal. The consecutive R–R intervals, as a function of the beat number, formed the tachogram signal. The PSD was estimated for this tachogram using an autoregressive (AR) model [2,30]. Specifically, the AR spectra is computed using the following equation:

O a(k)S(n 2 k) 1 GU(n) P

S(n) 5

k 51

where the signal of the n th index is the linear combination of the product of P model parameters; a(k) are the P unknown parameters, multiplied by the respective values of the signal S(n–k) together with a noise element GU(n). Only stationary tachograms devoid of dysrhythmias were analyzed. In contrast to the fast Fourier transform, AR modeling uses the raw data to identify a best-fitting model, from which the final spectrum, with a variable number of peaks is derived [32]. The model order was assessed using a maximum likelihood estimation, which reduces maximally the error between the predicated signal and the actual signal. In most cases, a model order of 10 was found to be adequate. By using AR modeling, many of the limitations of using the fast Fourier transform, including the need for frequency domain windowing or filtering and application only to periodic phenomena are avoided [30]. After excluding the nonoscillatory (DC) component (the component below 0.02 Hz, and felt to contain the contribution of nonrhythmic variations [24,30]), the PSDs were normalized and expressed as a percentage of total energy value (the total area under the curve). The PSD was divided into LF (0.02–0.15 Hz) and HF (0.15–1.5 Hz) bands. Studies in animals and humans suggest that HR changes at frequencies above 0.15 Hz are primarily caused by modulation of cardiac vagal efferent activity, while HR fluctuations at frequencies below 0.15 Hz are influenced by sympathetic and parasympathetic activity [24,30,32]. LF / HF ratios were computed each week for each infant during baseline and tilt. Values from each baseline and tilt series (performed in triplicate) were averaged together, such that one LF / HF ratio for baseline and tilt was obtained each week. These data were not included in the final analysis unless adequate baseline and tilt PSDs could be obtained at least twice for a given week.

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

41

2.3. Data analysis Only infants with 3 or more weeks of data were included in the final analysis. The difference in the LF / HF ratio between tilt and baseline (LF / HF tilt 2LF / HF base ) was used as an index of the change in the sympathovagal balance in response to tilt. Because each infant provided repeated data each week, a regression line for each subject was determined as a function of postconceptional age. A one sample t-test was used to test for a significant difference between the mean of the slopes of the regression lines and 0. Similar analyses were performed to determine if maturational changes in baseline LF and HR were present. Some infants were receiving mechanical assisted ventilation during part of their enrollment period. Because mechanical ventilation invokes reflex mechanisms which may in turn alter the relative tone of the autonomic nervous system [3,41,46], analysis was also performed with the data obtained while infants were mechanically ventilated excluded. Differences were considered significant when P,0.05. Stepwise regression analysis with slopes of the regression lines for LF / HF tilt 2LF / HF base or the change in HR as dependent variables and infant characteristics identified in Table 1 as independent variables was performed to identify predictors of a functional autonomic reflex response. Spearman rank correlation was used to identify the strength of association between the slopes of the regression lines for LF / HF tilt 2LF / HF base or the change in HR and infant characteristics.

3. Results A total of 28 infants completed the study, with 198 recording sessions analyzed. Twenty sessions were excluded for poor quality signals. Two infants (of the original 32 infants) were excluded due to worsening clinical status, while two others had fewer than 3 weeks of data. The mean gestational age at birth was 27.260.3 weeks, while the mean postnatal age upon study enrollment was 1862 days (range 5 to 28 days). The duration of study for each infant ranged between 3 to 11 weeks (median 7 weeks). Additional patient characteristics are shown in Table 1. When infants were initially studied, at postconceptional ages ranging between 28 to 32 weeks, the mean resting HR was 15562 bpm. With postural change to the 458 head-up position, no significant change in HR occurred (20.560.9 bpm, range 213 to 18 bpm). The mean resting HR at the time of the last study (postconceptional age range 35–39 weeks) was 15264 bpm, not significantly different from the HR value obtained at the initial study. However, at this stage of development, 458 head-up tilting produced a small but statistically significant increase in HR compared with the resting value (1.860.7 bpm, P,0.05). Linear regression analysis (Fig. 1) which takes into account the data obtained each week for each individual infant, demonstrated that the change in HR in response to tilting is significantly altered with increasing postnatal age, becoming progressively more positive (mean slope of regressions 0.4560.12 bpm / week, mean r-value of regressions 0.4760.05, P, 0.005). Regression analysis on data obtained from infants who show an increase in

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

42

Table 1 Characteristics of infants studied Infant name

Sex

GA (weeks)

BW (g)

Apgars (1 min/ 5 min)

Delivery

Mat Tob

PA (days)

MV (weeks)

Weeks in study

Methylxanthine

PCA at discharge (weeks)

M.C. H.F. C.J. N.W. A.F. A.F. A.W. K.B. K.N. B.S. D.F. D.F. D.G. T.D. B.B. J.O. S.R. J.R. C.C. M.S. A.M. J.W. L.S. M.B. O.B. R.L. A.L. E.T.

F F M f F F M F M F M M M M M F F M M F F M M F M F M F

30 29 26 30 27 27 28 25 28 27 28 28 29 29 27 28 29 29 24 28 27 25 27 27 27 26 25 25

1420 1110 700 1510 1230 1240 960 830 960 920 1150 920 980 1190 1110 815 1490 1400 760 1080 880 740 1140 1200 1210 820 790 640

9/9 4/8 5/7 9/9 4/8 4/7 3/9 6/7 3/7 7/8 5/6 3/5 5/6 5/7 4/6 3/7 7/8 7/9 5/6 6/7 1/6 5/7 3/7 6/8 7/7 2/5 4/6 1/4

V C C C V V V C V C V V V V C C V V V V V V V V V V C C

2 2 1 2 1 1 1 1 1 1 2 2 1 1 2 2 2 2 2 1 2 2 2 1 1 2 2 1

5 8 25 12 16 18 13 16 9 22 8 12 10 11 8 13 5 7 42 6 11 29 28 20 22 33 31 40

0 0 3 0 0 0 1 1 0 1 0 2 0 0 1 1 1 0 4 3 0 8 0 0 0 0 4 2

6 8 11 7 6 7 5 8 9 8 7 7 5 3 8 10 7 7 10 7 11 11 6 9 9 6 10 6

– Ca Ca Ca Ca Ca Ca Ca Ca Ca Ca Ca – Ca Ca Ca Ca Ca Ca Ca Ca Ca Ca Ca Ca Ca Th Ca

36 38 39 35 36 36 Tr 36 38 38 36 36 Tr Tr 36 38 36 36 42 35 40 43 36 38 38 37 40 Tr

GA, gestational age; BW, birth weight; V, vaginal delivery; C, cesarean section; Mat Tob, maternal tobacco use; PA, postnatal age at initial study; MV, mechanical ventilation; Ca, caffeine citrate; Th, theophylline; PCA, postconceptional age; Tr, transfered to local hospital.

HR with head-up tilt at either the first or last study (thus omitting infants having only less of a decrease in HR between the first and last studies) also demonstrates a mean slope significantly different from zero (0.3060.11 bpm / week, mean r-value 0.5360.06, P,0.01). Similar results were found when the data obtained while infants were mechanically ventilated are excluded (mean slope 0.5760.14 bpm / week, mean r-value 0.5860.05, P,0.005). The tachograms and corresponding power spectrum of HR variability at 08 and 458 tilt in a neonate at 29 weeks postconceptional age (postnatal age of 3 weeks) and again at 35 weeks are shown in Figs. 2 and 3. At rest, there is a major spectral component within the LF range (,0.15 Hz). However, rather than displaying a narrow respiratory peak within the HF range as seen in adults, the HF power spectrum is dispersed, as previously shown in premature infants [12]. The normalized

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

43

Fig. 1. Regression analysis of the change in heart rate in response to head-up tilt as a function of postconceptional age. Solid lines are regressions derived from individual infants. The (x,y) coordinates for the start of each line represent the postconceptional age and the change in heart rate, respectively, at the initial study. The length of each line signifies the length of time (in weeks) each patient was followed and studied. The hatched line represents the mean of the slopes for all regressions.

area of the LF component is slightly less in the premature neonate early in postnatal life compared with 6 weeks later. With tilt, the power spectrum shows no significant change at 29 weeks, in that the LF / HF ratio does not change between the two positions (Fig. 2). However, by the time the neonate reaches 35 weeks postconceptional age, the LF component of the power spectrum increases in response to head-up tilt in a fashion similar to that expected in older children and adults (Fig. 3). In this infant, the LF / HF ratio increased following tilt from 2.78 to 7.67. To determine if the basal sympathovagal balance controlling HR variability changes with maturation, we examined the postnatal changes in the LF band of the spectrum, in which most of the power is concentrated for infants. Upon initial study, the mean percentage power in the LF band was 84.261.3%, while that at the end of the study was 80.661.9%. By paired t-test alone, this represents a significant decrease in LF component with maturation (P,0.05) This maturational effect is also demonstrated when the data are additionally analyzed by linear regression, which takes all the weeks of study into account. By this method, the mean slope of the regressions was 20.4460.24% / week (mean r-value 0.4860.07, P,0.05). Similar results were obtained when the data from ventilated infants were excluded (20.6760.31% / week, mean r-value 0.6260.05, P,0.05). We used the difference in the LF / HF ratio between tilt and the recumbent position as a measure of the change in autonomic input to the heart in response to unloading of the arterial baroreceptors. When first studied at 28–31 weeks postconceptional age,

44

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

Fig. 2. Tachograms (top panels) and corresponding heart rate power spectra (bottom panels) from an infant at 29 weeks postconceptional age (3 week postnatal age) in the recumbent position (08) and during passive 458 head-up tilt. Note that there is no change in the frequency components during tilt (458) compared with rest (08).

no significant difference in the LF / HF ratios was found between infants in the recumbent position or after tilting (mean LF / HF tilt 2LF / HF base 5 20.5160.27 units). However, with postnatal maturation, the LF component of the power spectrum becomes larger during tilt compared to baseline, such that the difference between LF / HF tilt and LF / HF base became significantly more positive (Fig. 4). The mean slope of the linear regressions for LF / HF tilt 2LF / HF base vs. postnatal age was 0.3260.11 units / week (mean r-value 0.6560.06, P,0.006) for all data and 0.3860.14 (mean r-value 0.5360.04, P,0.01) when data from ventilated infants are excluded. To determine if any significant relationships between the infant characteristics identified in Table 1 and the slopes of the linear regressions for the change in HR and LF / HF tilt 2LF / HF base existed, Spearman correlation coefficients were calculated (Table 2). No significant positive or negative correlations were identified. Multiple

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

45

Fig. 3. Plots show the effect of tilting on the tachograms and corresponding heart rate power spectra obtained from the same infant a depicted in Fig. 1, now at 35 weeks postconceptional age (9 week postnatal age). The low frequency component of the power spectra at rest (08) is less than that obtained at 29 weeks. During passive tilt (458), the low frequency component increases, indicating a change in the sympathovagal balance.

regression analysis also failed to identify any variables (infant characteristics) as predictors of slopes of the calculated regressions.

4. Discussion The primary aim of our investigation was to study in prematurely born infants the effects of postnatal maturation on changes in HR and in the autonomic input to the heart in response to baroreceptor unloading. The magnitude and direction of the HR response to head-up tilt changed with increasing postconceptional age, becoming increasingly stimulus appropriate. Using power spectral analysis of HR variability, we found that in infants at 28–31 weeks postconceptional age, the LF / HF ratio did

46

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

Fig. 4. Regression analysis of the difference in the LF / HF ratio between tilt the recumbent position (LF / HF tilt 2LF / HF base ) as a function of postconceptional age. Solid lines are regressions derived from individual infants. The (x,y) coordinates for the start of each line represent the postconceptional age and the (LF / HF tilt 2LF / HF base ), respectively, at the initial study. The length of each line signifies the length of time (in weeks) each patient was followed and studied. The hatched line represents the mean of the slopes for all regressions.

not change with tilt, suggesting that early in development, unloading of baroreceptors does not alter the dynamic balance of sympathetic and vagal activities to the heart. However, linear regression analysis demonstrated that with increasing postnatal age, the LF component of the power spectrum increased with head-up tilt (and conversely the HF component decreased). These findings demonstrate that baroreceptor reflexes undergo change with maturation, becoming more functional with postnatal development.

Table 2 Correlation coefficients of slope of LF / HF and HR linear regressions with characteristics of infants

Sex GA BW Apgar (1 min) Apgar (5 min) Delivery Mat Tob MV

LF / HF slope

P-value

HR slope

P-value

0.21 0.30 0.05 20.04 0.12 0.29 20.16 0.04

0.28 0.12 0.80 0.83 0.55 0.14 0.43 0.84

0.14 0.20 0.07 20.32 20.13 0.13 20.01 0.04

0.49 0.31 0.71 0.11 0.51 0.53 0.96 0.83

LF, low frequency; HF, high frequency; HR, heart rate; GA, gestational age; BW, birth weight; Mat Tob, maternal tobacco use; MV, mechanical ventilation.

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

47

Numerous studies have investigated and characterized arterial baroreflex function in developing animals of various species, although the findings are inconsistent [17,31,39]. Studies in near-term fetal and newborn sheep have shown that baroreflex control of HR and renal sympathetic nerve activity is functional early during development [39]. In fact, in these studies, the sensitivities of the responses were greatest in near-term fetal sheep, and decreased after birth. On the other hand, studies in other species have found that baroreflex sensitivity increases with postnatal age [10,17]. In the human neonate, neural control of cardiovascular function has been assessed most often by measuring alteration in HR in response to postural changes [15,21,33,44,50]. Several studies have demonstrated in healthy term and preterm infants that head-up tilting (to unload arterial baroreceptors) produces a significant tachycardia [15,33,44], and that the magnitude of the response is proportional to the degree of tilting [44]. In contrast, other investigators have been unable to demonstrate consistent responses of HR to tilting [21,50]. In the present study, we found a small but statistically significant increase in HR in head-up tilt with the last study in each infant, but not the first. Interestingly, a number of infants had a relative bradycardia with head-up tilt which became less marked with increasing age. Reasons for cardiac slowing in response to head-up tilt in this population are unclear. However, the finding that HR decreased less in these infants with advancing age may suggest a developing influence of baroreceptor activity on HR, thereby attenuating the bradycardic response with head-up tilt. Using venous occlusion plethmography, Waldman et al. [50] found in nondistressed term and preterm infants that head-up tilt produced on average a 25% decrease in limb blood flow, suggestive of an increase in peripheral vascular resistance, although no significant tachycardia was observed. Together, these findings suggest that vascular sympathetic responses to baroreceptor stimulation are present early in life, while autonomic control of HR may be functionally delayed. To our knowledge, this is the first published report using power spectrum analysis of HR variability during orthostatic tilt in developing preterm infants. The finding that the LF / HF ratio is not altered by tilt in extremely preterm infants during the first few weeks of life is consistent with the suggestion that baroreflex control of cardiac function is relatively immature at this stage of human development. By studying these infants in a longitudinal fashion, we found that with increasing postnatal age, reflex mediated changes in HR and autonomic input to the heart become prominent, demonstrating that the neural reflexes involved in control of cardiovascular function are developmentally regulated. These findings are supported by data from Merrill et al. [27] describing the presence of a relationship between changes in HR and umbilical venous pressure with intravascular transfusion in anemic human fetuses greater than 32 weeks gestation, but not in those between 19 and 32 weeks gestation. The mechanisms responsible for the developmental changes observed in the present study are unclear, but may involve maturational changes in peripheral arterial baroreceptor responses, central baroreflex integration, or efferent innervation of the heart. Structural changes in the carotid sinus wall, differences in membrane characteristics of baroreceptor endings [11] and changes in ion pumps operating

48

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

within the baroreceptor membrane [20] can influence baroreceptor responses during development. Maturational changes in circulating levels of, or receptor affinity for, various hormonal factors such as angiotensin II, vasopressin and prostaglandins may also affect the baroreflex response at peripheral and central sites [6,7,11,40]. Cardiac sympathetic and parasympathetic innervation of the heart, which mature at different rates, may further influence the end organ response to changes in afferent baroreceptor nerve activity [5,49]. Finally, cardiopulmonary (low pressure) receptors may have a potential role in regulating the HR response. In the adult, unloading of these receptors, as occurs with head-up tilt, results in reflex tachycardia and sympathoexcitation. However, studies in fetal and newborn animals demonstrating that cardiopulmonary receptors participate little in regulating cardiovascular and autonomic responses to changes in intravascular volume and blood pressure argue against a significant role for these mechanoreceptors [26,29,42]. Analysis of HR variability has previously been used by a number of investigators to study maturational changes in autonomic activity [12,13,22,45]. Studies of fetal ECG tracings obtained on the maternal abdominal wall have shown that younger fetuses have a greater total energy of the power spectrum compared with more mature fetuses, consistent with the evolution of a stable and mature autonomic nervous system [22]. Maturational changes in the power spectra of HR variability have also been shown by comparing preterm to term infants [12,13]. There is a progressive decline in the LF / HF power ratio associated with both increasing postnatal and gestational age, indicating an increase in parasympathetic contribution to control of resting HR with maturation [12,13]. Our results are consistent with those previously reported, despite the fact that few of our infants were studied beyond 36–37 weeks postconceptional age. Clairambault et al. [13] found that changes in the HF component of the spectrum were greatest at 37–38 weeks, suggesting a steep increase in vagal tone at this age. Thus, we may have seen a greater maturational change in the LF / HF ratio during baseline conditions and in response to tilt had we been able to study our group of infants out to this age. While HR variability may be influenced by a number of factors, including behavioral or sleep state [36,47], maturational changes are clearly seen in either of the two sleep states [13,47]. In the present study, recordings were made only when it was determined that the infant was sleeping; no attempt was made to differentiate between active and quiet sleep. While it is possible that differences in sleep state could contribute to the developmental changes in HR variability, we feel it is unlikely that sleep state alone accounts for the maturational changes in the power spectrum in response to head-up tilt. In addition, the effect of sleep state on HR variability appears decreased in premature infants [13]. No attempt was made to analyze the respiratory signals or discern the effect of respiratory frequency on the HR power spectrum. Respiratory sinus arrhythmia (RSA), which is dependent in part on the respiratory rate, is mediated by parasympathetic innervation to the heart and is present in the HF region of the HR power spectrum. Previous studies investigating the contribution of RSA to HR variability in healthy term newborns have been inconsistent, finding both minimal [14,16,38] or large [13,19,43] influences of RSA on total HR variability. In near-term human

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

49

fetuses, the presence of breathing movements has been shown to have no effect on the distribution of power within various frequency regions [18]. Premature human infants also do not show a well defined respiratory peak in the HR spectrum, but rather a dispersed peak likely resulting from immature neural pathways [12]. With the relatively fast and irregular respiratory frequency of infants at the ages we studied (0.5–1.0 Hz), the contribution of RSA to HR variability, including cardiac aliasing [37] would be expected to be less than that seen later in development. The RSA contribution to HR variability is significantly related to the respiratory rate and the power of the RSA peak, lower respiratory rates being associated with higher RSA contribution [19,43]. The effect of respiration of HR variability in preterm infants, if present, would primarily influence HF bands of the HR power spectra, while having little effect on HR variability at baroreflex related frequencies, shown to be around 0.07 Hz in neonates [16]. The change in the normal respiratory pattern with maturation, and its relative contribution to HR variability may in part contribute to the decrease in the LF / HF ratio of the HR power spectrum with advancing postnatal age [12]. On the other hand, given that respiration remained stable for the short time period infants were tilted, it is unlikely that the maturation changes in the HR power spectra occurring with head-up tilt were related to acute changes in respiratory influences. Because the presence of respiratory distress syndrome has been shown to be associated with a marked decrease in HR variability [34,48], primarily in the LF range, no infant was studied during the acute or resolving phase of this type of pulmonary disease. In addition, it should be noted that the majority of the data were collected while infants were breathing spontaneously. Nonetheless, we cannot exclude the possibility that maturational changes in lung compliance and resistance contribute to developmental changes in HR variability and the HR power spectrum. Mechanical ventilation, which has been shown to influence HF HR variability in a manner similar to spontaneous RSA [46], was required for some of the infants studied during early postnatal life. Even with these data excluded, however, the results from this study are unaltered. Because so few infants were studied while on mechanical ventilation, no conclusions can be made about the effect of positive pressure ventilation on the relative tone of the autonomic nervous system early in life. Implicit in this study design are assumptions that thermoregulatory, hormonal and respiratory influences on HR variability did not change during the transition from supine to the tilt position. On the other hand, interaction among these neuroendocrine reflexes may contribute to the developmentally regulated autonomic response seen with postural change. The results of these studies demonstrate that autonomic input to the HR in response to baroreceptor unloading is minimally altered in preterm infants early in life, and suggest that baroreceptor reflexes undergo functional changes with maturation. The physiologic relevance of the statistical differences we found in the spectral measurements of autonomic input to the heart in response to positional change cannot be ascertained. However, we speculate that impaired autonomic regulation of HR and arterial pressure may contribute to the development of ischemic or hemorrhagic complications related to alterations in blood pressure and organ blood flow known to occur in extremely premature infants.

50

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

Acknowledgements The authors gratefully acknowledge the assistance of T. James R.N., K. Johnson R.N., J. Miller R.N., and V. Seward in obtaining the ECG recordings. This study was supported by National Institutes of Health Grant RR000359 from the General Clinical Research Center Program and a grant from the Children’s Miracle Network. J.L. Segar is supported by a Clinical Investigator Development Award HL-02865. J.E. Mazursky was supported by a National Institutes of Health Training Grant HL-07121.

References [1] Abboud F, Thames M. Interaction of cardiovascular reflexes in circulatory control. In: Shepherd JT, Abboud FM, editors. Handbook of physiology. Section 2, Vol III, Part 2. Bethesda, MD: American Physiological Society, 1983:675–753. [2] Aharon-Peretz J, Harel T, Revach M, Ben-Haim SA. Increased sympathetic and decreased parasympathetic cardiac innervation in patients with Alzheimer’s Disease. Arch Neurol 1992;49:919– 22. [3] Aibiki M, Koyama S, Ogli K, Shirakawa Y. Reflex mechanisms for changes in renal nerve activity during positive end-expiratory pressure. J Appl Physiol 1988;65:109–15. [4] Akselrod S, Gordon D, Madwed JB, Snidman NC, Shannon DC, Cohen RJ. Hemodynamic regulation: investigated by spectral analysis. Am J Physiol 1985;249:H867–75. [5] Assali NS, Brinkman CR, Wood Jr R, Danavino A, Nuwayhid B. Ontogenesis of the autonomic control of cardiovascular function in the sheep. In: Longo LD, Reneau DD, editors. Fetal and newborn cardiovascular physiology. New York: Garland STPM Press, 1978:47–91. [6] Bishop VS, Hasser EM, Nair .C. UC. Baroreflex control of renal nerve activity in conscious animals. Circ Res 1987;61:I76–81. [7] Bishop VS, Haywood JR. Hormonal control of cardiovascular reflexes. In: Zucker IH, Gilmore JP, editors. Reflex control of the circulation. Boca Raton, FL: CRC Press, 1991:253–71. [8] Borst C, Brederode JFMV, Wieling W. Mechanisms of initial blood pressure response to postural change. Clin Sci 1984;67:321–7. [9] Borst C, Wieling W, Brederode JFMV, Hond A, Rijk LGD, Dunning AJ. Mechanisms of initial heart rate response to postural change. Am J Physiol 1982;243:H676–81. [10] Buckley NM, Gootman PM, Gootman GD, Reddy LC, Weaver LC, Crane LA. Age-dependent cardiovascular effects of afferent stimulation in neonatal pigs. Biol Neonate 1976;30:268–79. [11] Chapleau MW, Hajduczok G, Abboud FM. Resetting of the arterial baroreflex: peripheral and central mechanisms. In: Zucker IH, Gilmore JP, editors. Reflex control of the circulation. Boca Raton, FL: CRC Press, 1991:165–94. [12] Chatow U, Davidson S, Reichman BL, Akselrod S. Development and maturation of the autonomic nervous system in premature and full-term infants using spectral analysis of heart rate fluctuations. Pediatr Res 1995;37:294–302. [13] Clairambault J, Curzi-Dascalova L, Kauffmann F, Medigue C, Leffler C. Heart rate variability in normal sleeping full-term and preterm neonates. Early Hum Dev 1992;28:169–83. [14] Dykes FD, Ahmann PA, Baldzer K, Carrigan TA, Kitney R, Giddens DP. Breath amplitude modulation of heart rate variability in normal full term neonates. Pediatr Res 1988;20:301–8. [15] Finley JP, Hamilton R, MacKenzie MG. Heart rate response to tilting in newborns in quiet and active sleep. Biol Neonate 1984;45:1–10. [16] Giddens DP, Kitney RI. Neonatal heart rate variability and its relation to respiration. J Theor Biol 1985;113:759–80. [17] Gootman PM. Developmental aspects of reflex control of the circulation. In: Zucker IH, Gilmore JP, editors. Reflex control of the circulation. Boca Raton, FL: CRC Press, 1991:965–1027.

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

51

[18] Groome LJ, Mooney DM, Bentz LS, Singh KP. Spectral analysis of heart rate variability during quiet sleep in normal human fetuses between 36 and 40 weeks of gestation. Early Hum Dev 1994;38:1–10. [19] Hathorn MKS. Respiratory modulation of heart rate in newborn infants. Early Hum Dev 1989;20:81– 99. [20] Heesch CM, Abboud FM, Thames MD. Acute resetting of carotid sinus baroreceptors. II. Possible involvement of electrogenic Na 1 pump. Am J Physiol 1984;247:H833–9. [21] Holden K, Morgan JS, Krauss AN, Auld PAM. Incomplete baroreceptor responses in newborn infants. Am J Perinatol 1985;2:31–4. [22] Karin J, Hirsch M, Akselrod S. An estimate of fetal autonomic state by spectral analysis of fetal heart rate fluctuations. Pediatr Res 1993;34:134–8. [23] Kirscheim HR. Systemic arterial baroreceptor reflexes. Physiol Rev 1976;56:100–77. [24] Malliani A, Pagani M, Lombardi F. Clinical and experimental evaluation of sympatho–vagal interaction: power spectral analysis of heart rate and arterial pressure variabilities. In: Zucker IH, Gilmore JP, editors. Reflex control of the circulation. Boca Raton, FL: CRC Press, 1991:937–64. [25] Malliani A, Pagani M, Lombardi F. Importance of appropriate spectral methodology to assess heart rate variability in the frequency domain. Hypertension 1994;24:140–2. [26] Merrill DC, McWeeny OJ, Segar JL, Robillard JE. Impairment of cardiopulmonary baroreflexes during the newborn period. Am J Physiol 1995;268:H134–H1351. [27] Merrill DC, Robillard JE, Weiner CP, Segar JL. Gestational-age dependent maturation of autonomic reflexes in the human fetus (Abstract). Pediatr Res 1993;33:249A. [28] Montano N, Ruscone TG, Porta A, Lombardi F, Pagani M, Malliani A. Power spectrum analysis of heart rate variability to assess the changes in sympathovagal balance during graded orthostatic tilt. Circulation 1994;90:1826–32. [29] O’Mara MS, Merrill DC, McWeeny OJ, Robillard JE. Ontogeny and regulation of arterial and cardiopulmonary baroreflex control of renal sympathetic nerve activity (RSNA) in response to hypotensive (NH) and hypotensive hemorrhage (HH) postnatally (Abstract). Pediatr Res 1995;37:31A. [30] Pagani M, Lombardi F, Guzzetti S, et al. Power spectral analysis of heart rate and arterial pressure variabilities as a marker of sympatho–vagal interaction in man and conscious dog. Circ Res 1986;59:178–93. [31] Palmisano BW, Clifford PS, Coon RL, Seagard JL, Hoffmann RG, Kampine JP. Development of baroreflex control of heart rate in swine. Pediatr Res 1989;27:148–52. [32] Parati G, Saul P, Di Rienzo M, Mancia G. Spectral analysis of blood pressure and heart rate variability in evaluating cardiovascular regulation. Hypertension 1995;25:1276–86. [33] Picton-Warlow CG, Mayer FE. Cardiovascular responses to postural changes in the neonate. Arch Dis Child 1970;45:354–9. [34] Prietsch V, Knoepke U, Obladen M. Continuous monitoring of heart rate variability in preterm infants. Early Hum Dev 1994;37:117–31. [35] Radaelli A, Bernardi L, Valle F, et al. Cardiovascular autonomic modulation in essential hypertension. Hypertension 1994;24:556–63. [36] Radvanyi MF, Morel-Kahn F. Sleep and heart rate variations in premature and full term babies. Neuropadiatrie 1976;7:302–12. [37] Rother M, Witte H, Zwiener U, Eiselt M, Fischer P. Cardiac aliasing—a possible cause for the misinterpration of cardiorespirographic data in neonates. Early Hum Dev 1989;20:1–12. [38] Schechtman VL, Harper RM, Kluge KA. Development of heart rate variation over the first 6 months of life in normal infants. Pediatr Res 1989;26:343–6. [39] Segar JL, Hajduczok G, Smith BA, Merrill DC, Robillard JE. Ontogeny of baroreflex control of renal sympathetic nerve activity and heart rate. Am J Physiol 1992;263:H1819–26. [40] Segar JL, Merrill DC, Smith BA, Robillard JE. Role of endogenous angiotensin II on resetting of the arterial baroreflex during development. Am J Physiol 1994;266:H52–9. ´ H, Sjovall ¨ [41] Sellden H, Ricksten S-E. Sympathetic nerve activity and central haemodynamics during mechanical ventilation with positive end-expiratory pressure in rats. Acta Physiol Scand 1986;127:51–60.

52

J.E. Mazursky et al. / Early Human Development 53 (1998) 37 – 52

[42] Smith FG, Klunkefus JM, Robillard JE. Effects on volume expansion on renal sympathetic nerve activity and cardiovascular and renal function in lambs. Am J Physiol 1992;262:R651–8. [43] Thompson CR, Brown JS, Gee H, Taylor EW. Heart rate variability in healthy term newborns: the contribution of respiratory sinus arrhythmia. Early Hum Dev 1993;31:217–28. [44] Thoresen M, Cowan F, Walløe L. Cardiovascular responses to tilting in healthy newborn babies. Early Hum Dev 1991;26:213–22. ´ L, Stoelinga G, Van Geign H. Spectral analysis of [45] van Ravenswaaij-Arts CMA, Hopman J, Kollee heart rate variability in spontaneously breathing very preterm infants. Acta Paediatr 1994;83:473–80. ´ LAA, Stoelinga GBA, Van Geign HP. The [46] van Ravenswaaij-Arts CMA, Hopman JCW, Kollee influence of artificial ventilation on heart rate variability in very preterm infants. Pediatr Res 1995;37:124–30. ´ LAA, van Amen L JP, Stoelinga GBA, van Geijn [47] van Ravenswaaij-Arts CMA, Hopman JCW, Kollee HP. Influences on heart rate variability in spontaneously breathing preterm infants. Early Hum Dev 1991;27:187–205. ´ LAA, Van Amen JPL, Stoelinga GBA, van Geijn [48] van Ravenswaaij-Arts CMA, Hopman JCW, Kollee HP. The influence of respiratory distress syndrome on heart rate variability in very preterm infants. Early Hum Dev 1991;27:207–21. [49] Vapaavouri EK, Shinebourne EA, Williams RL, Heymann MA, Rudolph AM. Development of cardiovascular responses to autonomic blockade in intact fetal and neonatal lambs. Biol Neonate 1973;22:177–88. [50] Waldman S, Krauss AN, Auld PAM. Baroreceptors in preterm infants: their relationship to maturity and disease. Dev Med Child Neurol 1979;21:714–22. [51] Zar JH. Biostatistical analysis. Engelwood Cliffs, NJ: Prentice Hall, 1984:1–718.