Early Human Development 90 (2014) 843–850

Contents lists available at ScienceDirect

Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev

Kangaroo care: cardio-respiratory relationships between the infant and caregiver Elisabeth Bloch-Salisbury a,b,⁎, Ian Zuzarte a, Premananda Indic a, Francis Bednarek b,1, David Paydarfar a,c a b c

Department of Neurology University of Massachusetts Medical School, Worcester, MA 01655, USA Department of Pediatrics, University of Massachusetts Medical School, Worcester, MA 01655, USA Wyss Institute for Biologically Inspired Engineering, Harvard University, Boston, MA, USA

a r t i c l e

i n f o

Article history: Received 23 April 2014 Received in revised form 15 August 2014 Accepted 25 August 2014 Available online xxxx Keywords: Preterm infants Skin-to-skin cohabitation Cardiorespiratory coupling Apnea Respiratory stability

a b s t r a c t Background: Kangaroo care, i.e., skin-to-skin cohabitation (SSC) between an infant and caregiver, is often used in neonatal intensive care units to promote bonding, breastfeeding and infant growth. The direct salutary effects of SSC on cardio-respiratory control in preterm infants remain equivocal; some reports suggest improved breathing stability, others indicate worsening of apnea, bradycardia and hypoxemia. Aim: The purpose of this study was to investigate physiological relationships between the infant and caregiver during SSC. We hypothesized that respiratory stability of the premature infant is influenced by the caregiver's heartbeat. Design: A prospective study was performed in eleven preterm infants (6 female; mean PCA 32 wks). SSC was compared to a preceding incubator-control period (CTL) matched for time from feed and condition duration. Abdominal respiratory movement, electrocardiogram, skin temperature and blood-oxygen levels were recorded from the infant and the caregiver. Results: During CTL, infant interbreath interval variance (IBIv; respiratory instability) was directly related to its own heart rate variance (HRv; rho = 0.770, p = 0.009). During SSC, infant IBIv and apnea incidence were each related to caregiver HRv (rho 0.764, p = 0.006; rho 0.677, p = 0.022, respectively). Infant cardiorespiratory coupling was also enhanced during SSC compared to CTL in the eupneic frequency range (0.7– 1.5 Hz, p = 0.018) and reduced for slower frequencies (0.15–0.45 Hz; p = 0.036). Conclusion: These findings suggest that during SSC, respiratory control of the premature infant is influenced by the caregiver's cardiac rhythm. We propose that the caregiver's heartbeat causes sensory perturbations of the infant via somatic or other afferents, revealing a novel cohabitation-induced feed-back mechanism of respiratory control in the neonate. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction “Kangaroo care” (KC) is a natural, cost-effective intervention used to reduce mortality and morbidity of premature and low birth-weight infants throughout the world [1–9]. In its simplest form the infant is clad only in a diaper and placed upright and prone on its caregiver's chest, skin-to-skin [2,3,6]. Research suggests that KC improves physiological function, promotes breastfeeding and mother-infant attachment, and reduces developmental risks of the infant [1,4,7–17]. However, the direct salutary effects of KC on infant cardio-respiratory control remain equivocal. Some reports suggest thermal regulation and cardiorespiratory stability are achieved during KC; i.e., a reduction in the frequency of pathophysiological pause in breathing (apnea) and decreased ⁎ Corresponding author at: Department of Neurology and Pediatrics, 55 Lake Avenue North S5-718, Worcester, MA 01655. Tel.: +1 508 856 6232. E-mail address: [email protected] (E. Bloch-Salisbury). 1 Deceased July 15, 2013.

http://dx.doi.org/10.1016/j.earlhumdev.2014.08.015 0378-3782/© 2014 Elsevier Ltd. All rights reserved.

heart rate (bradycardia) and improved oxygenation [4,12,16,18–22], whereas other studies report an increase in infant body temperature associated with an increase in the number of apnea, bradycardia and/or blood-oxygen desaturation events [23–25]. The purpose of this study was to investigate physiological relationships between the caregiver and the infant during KC to determine if cardio-respiratory stability of the premature infant was associated with that of its caregiver. Despite a wide range of outcome measures that have been reported, including those affecting the infant's cardio-respiratory responses, temperature, pain response and sleep [12,13,15–30], there has been no integrative study to determine the essential physiological interactions between the infant and caregiver during KC necessary for improved infant response. The mechanisms for the therapeutic efficacy of KC that stabilize respiration and improve heart rate remain elusive. We propose that during KC, wherein the infant's head and chest wall overlay the chest wall of the caregiver, mechanical perturbations of the caregiver's heartbeat affect underlying receptors of the cohabitating infant. These receptors might include cutaneous, musculoskeletal, visceral and

844

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850

vestibular-cochlear receptors that project to brain centers that are involved in respiratory control [31–35]. This study tested the hypothesis that respiratory stability of the premature infant is influenced by sensory perturbations of the caregiver's heart beat during KC. Though respiration and heart rate are under the control of two independent systems, coordination between the two has been shown to reflect health and maturation [31,32]. Cardio-respiratory coupling quantifies the interaction between respiratory and heart rhythms. The interdependence is related to reflex activity between pulmonary afferents and vagal outflow [33,34]. In infants, cardio-respiratory interaction is vital for providing optimal balance between gas exchange and metabolic demands [32,34]. A second purpose of this study was to determine whether cardio-respiratory coupling of the infant would be enhanced during SSC. 2. Methods 2.1. Human subjects A prospective study was performed on 11 preterm infants (GA b 35 wks) at the University of Massachusetts Memorial Healthcare Neonatal Intensive Care Unit (NICU). Exclusion criteria were congenital malformation, chromosomal disorders, congenital or perinatal infection of the central nervous system, intraventricular hemorrhage Ngrade II and hypoxic-ischemic encephalopathy. Infants treated with methylxanthines were included if on maintenance dosing and the drug had reached steady-state. One additional subject who developed necrotizing enterocolitis was excluded. Hospital records and chart review were used to obtain demographic and medical information. All mothers denied smoking, illicit drug and alcohol use throughout pregnancy. Informed consent was obtained from the mother of each infant enrolled; one father also provided informed consent to participate in the study. The study was approved by the University of Massachusetts's Medical School Institutional Review Board for Human Subjects. 2.2. Measures and procedures To test the hypothesis that respiratory stability of the premature infant is influenced by sensory perturbations of the caregiver's heart beat during Kangaroo Care infants participated in two conditions: 1) Skin-toskin cohabitation (SSC). Infants were clad only in a diaper and placed in a prone position on its parent-caregiver to provide skin-to-skin contact in accordance with Kangaroo-Care guidelines [3,6]. Blankets and a knitted cap (positioned above the infant's ear surfaced to the caregiver's body) were used to maintain warmth. Caregivers remained semi-supine (~ 15–30°) in a stationary, cushioned recliner chair. 2) Control Period (CTL). To provide a thermo-regulated environment, infants were studied in their assigned incubator; one infant with mature thermoregulation was studied in his open crib using routine coverings. All infants were placed prone with the mattress tilted head up at ~15–30° to best mimic the positioning during SSC stimulation. 2.2.1. Physiological measurements For each infant and caregiver, respiratory inductance plethysmography (RespIPleth) was used to record thoracic and abdominal respiratory movements (Somnostar PT; Viasys Healthcare, Yorbalinda, CA; Embla). Electrodes placed over the skin surface of the chest (Hewlett Packard), 3-lead configuration, were used to record electrocardiographic activity (ECG). Transcutaneous arterial-blood oxygen saturation (SaO2) was measured using a separate pulse oximeter attached to the infant's foot or wrist, and to the parent's index finger (Nellcor, Hayward, CA). Skin temperature was recorded continuously using disposable adhesive temperature probes attached to the infant's and parent's axilla via separate electronic monitoring thermometers (Physitemp TH-5, Clifton, NJ). Cardio-respiratory signals, SaO2, and skin temperature were

recorded continuously throughout both conditions for each infant and throughout SSC for the caregiver. 2.2.2. Environmental and behavioral measurements For each condition a sound meter (ExTech Instr) placed near the infant's head was used to measure changes in sound intensity (dBA). A light meter (AEMC, Industrial Process Measurements) placed by the infant's head was used to measure changes in light levels (lux). Overt behavioral data were recorded using a camera with a wide-angled lens (MicroCamera, Panasonic) within the infant's incubator or crib, or set to capture infant and parent during SSC stimulation, synchronized with the physiological, audiometry and light signals. 2.2.3. Data acquisition Infant and adult ECG signals were sampled at 2000 Hz, RespIPleth at 50 Hz, SaO2 at 10 Hz, transcutaneous pulse-oximeter plethysmographic activity at 100 Hz, and temperature, environmental light and sound signals at 20 Hz. The data were displayed during the experimental periods and stored on hard disk for offline analysis (Embla N7000, Broomfield, CO). Observations by the investigators were recorded as timestamped text comments along with the signals. Fig. 1 illustrates an example of recorded signals during SSC in one infant-caregiver dyad (Subject 9). 2.2.4. General procedures Studies were conducted during daytime hours; CTL periods were conducted between 7 a.m. and 1:00 p.m., SSC periods were conducted between 11:00 a.m. and 5:00 p.m. For each infant the CTL condition preceded the SSC condition to control for potential carryover effects of the SSC experimental condition. After initial set up of equipment and attachment of all sensors, infants were given their routine feed (gavage or bottle); feeds were conducted either in the isolette, infant crib or in the caregiver's arms. For the CTL condition, following feed the infant was placed in their isolette/crib for the inter-feed interval (3–4 h, depending on the infant's routine feeding schedule). For SSC, infants were held for the maximum holding period that the caregiver could manage not to exceed the inter-feed interval. Caregivers were asked to use the bathroom prior to holding their infant, and mothers who were pumping breast milk were asked to pump prior to SSC. All caregivers were instructed to hold the infant in accordance with Kangaroo-Care guidelines [3,6], for as long as they could sit in the semi-reclined position. We did not control for conversation at the bedside, but sought to allow conversation that parents typically provide during holding in order to study response in real, NICU-bedside setting. For each condition, following the feed there was an observation period of 30-min to assure integrity of the recordings, reduce potential confounds of digestion and to allow the infant to resume sleep. Analysis periods for each infant were determined by the maximum duration of the SSC condition (following the 30-min feed adjustment period) so that CTL and SSC periods were matched for time from feed and duration. 2.3. Data processing and analyses All data analyses involving manual measurements of physiological signals were completed offline. Investigators were masked to condition for analyses of infant data; caregivers were only studied during SSC. 2.3.1. Movement periods and condition time For each condition, movement periods were defined by movement that generated distortion in the transcutaneous pulse-oximeter plethysmographic signal that exceeded 5 s. Movement artifact was defined by gross body movements that obscured the cardio and/or respiratory signals. Video recordings and text comments written at the time of the study were used to further confirm periods of movement and to identify nursing interventions and/or technical contamination, including infant RespIPleth contaminated by caregiver respiratory signal (indexed by

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850

845

10 sec 100

Inf SpO2

95 100

CG SpO2

95

% %

Inf Rib Inf Abd CG Abd 8.5

Inf ECG

mV

-2.6 0.93

CG ECG

mV

-0.32

Inf Pleth CG Pleth 38

Inf Temp

36 38

CG Temp

36 70

Sound

ºC ºC dB

40 127 lux 124

Light

Fig. 1. Example of recorded signals during skin-to-skin cohabitation for one infant-caregiver dyad. INF = infant; CG = caregiver; SpO2 = blood-oxygen saturation; Rib = respiratory inductance plethysmography of chest wall; Abd = respiratory inductance plethysmography of abdominal muscle; ECG = electrocardiogram; temp = skin temperature; Pleth = plethysmographic signal of transcutaneous pulse-oximeter used to indicate movement periods; sound = ambient sound levels; light = ambient light levels. Arrows demarcate peak of inspiratory signal (Inf Abd) and R-R waves (ECG).

caregiver RespIPleth bleed on the infant signal) [36]. Periods with signal artifact were excluded from analyses. Valid condition time was defined as the period within each condition that was not distorted by movement artifact, infant hiccups that contaminated the respiratory signal, nursing interventions and technical contamination. Signals were analyzed with respect to valid recording time. Contiguous physiological measurements recorded throughout each condition were calculated for the non-contaminated portion of each signal (detailed below) for each subject. 2.3.2. Interbreath intervals (IBIs) For each infant, inductance plethysmography of abdominal movements was used to generate a time series of IBIs, determined from the peak of the inspiratory signal using automated peak-detection software (LabChart 7, ADI Instruments, Colorado Springs, CO). IBI peak detections are demarcated in Fig. 1. Statistical properties of the IBI histogram included the mean and variance of the IBI distribution (s), variance being a measure of breathing stability. Incidents of IBIs N10 s, calculated per unit of valid recording time, were defined as a pause in breathing rhythmicity. These pauses were always associated with a lack of effort in both the abdominal and rib plethysmographic activity. Therefore, pauses defined in this study were central (non-obstructive) apneas [37]. 2.3.3. Cardiac intervals Intervals between cardiac R waves were calculated using an automated peak detection program (MATLAB, MathWorks, Natick, MA). Mean and variance of heart rate (beats/min) were calculated for each infant during CTL and SSC, and for the caregiver during SSC. Fig. 1 illustrates peak detection of R waves for infant and adult ECG. 2.3.4. Oxygen desaturation For each infant SaO2 was calculated for CTL and SSC conditions. Oxygen desaturation time was calculated as the percentage of time in which O2 saturation fell below 85%.

2.3.5. Skin temperature, ambient sound and light levels Mean and variance of infant skin temperature, and ambient sound (dBA) and light (lux) levels were calculated during both conditions; caregiver skin temperature was also calculated during SSC. 2.3.6. Cardio-respiratory coupling To determine coupling between heartbeat and respiration, cardiorespiratory interaction was calculated in infants during CTL and SSC using spectral methods. The similarity between the infant respiration signal and the infant RR-interval (s) in the frequency domain was estimated by the magnitude square coherence function defined as

C xy ð f Þ ¼

2 P xy ð f Þ P xx ð f Þ P yy ð f Þ

where Pxx and Pyy are the power spectral densities of the infant respiration signal and infant RR-interval signal, respectively, and Pxy is the cross spectral density between the two signals. The cross-power spectrum was defined as the Fourier transform of the cross correlation function of the two signals. Both signals were resampled to 4 Hz to align them in time. MATLAB (MathWorks, Natick, MA) function ‘mscohere’ was used to estimate the magnitude squared coherence, using Welch's averaged periodogram method to find the power spectral densities [38]. The magnitude squared coherence between the resampled signals were estimated by segmenting each time series in overlapping windows (N = 1024 points; 50% overlap) and using a Hanning window length of 128 within each segment; each segment was approximate 4.26 min. The peak coherence value within each of the four frequency bands [FB; 0.01–0.15 Hz (LF); 0.15–0.45 Hz (HF1); 0.45–0.7 Hz (HF2); and 0.7–1.5 Hz (HF3, eupnea)] [39] were calculated for every segment and then averaged within each subject for both the CTL and SSC conditions. This provided a mean coherence value for each subject within each frequency band for each condition.

846

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850

2.3.7. Statistical analysis Statistical calculations were performed using commercially available software (SPSS, version 11.5, Chicago, IL). Pairwise t-tests were used to determine whether differences in physiological signals, and ambient sound and light levels existed between CTL and SSC conditions during valid recording times. Separate repeated-measures ANOVAs were used for multi-factorial analyses. Greenhouse-Geisser correction was used in instances where Mauchly's test indicated spherecity was violated; ε with unadjusted degrees of freedom is reported. Bonferroni adjustment was used for post hoc tests for factors with more than two levels. Nonparametric correlations (Spearman's rho) were used to establish association among infant and parent physiological signals and ambient sound and light for CTL and SSC conditions. All values are expressed as means and SD. P values less than 0.05 were considered statistically significant. 3. Results 3.1. Clinical characteristics Eleven premature infants (6 female) participated in the study at mean post-conceptional age (PCA) 31.9 wks (SD = 2.7 weeks; see Table 1 for demographics). Mean age of the caregiver providing SSC was 33.6 y (SD 6.8 y). All infants were studied in their incubator during CTL period except Subject 11 who was studied in his crib. All infants were held by their mother for SSC except Subject 8 who was held by her father. Six singletons, one twin pair (Subjects 5 and 6) and three unpaired twins (Subjects 2, 9 and 11) participated. Nine infants were on gavage feed, Subject 6 was on gavage and bottle feed, Subject 9 was on bottle feed. 3.2. Condition comparisons 3.2.1. Condition times and movement periods The total time caregivers held their infant skin-to-skin ranged between approximately 30 and 160 min (Table 1); the longer durations reflect those infants who were held for feeds during SSC (gavage or breastfed) vs infants who were fed in their isolette. The equivalent portion of each condition period for CTL and SSC (i.e., matched from end of feed plus 30 min post-feed) was included for analyses (Table 1). The mean condition time for CTL and SSC was 41.3 min (SD 21.2, Table 1). Signal artifact due to infant movement (mean 10.9% SD 6.2), hiccups, nursing interventions and technical contamination were similar between conditions; mean valid condition time used for CTL analyses was 35.9 min (SD 21.8) and for SSC analyses was 35.4 min (SD 19.2; paired t test, p = 0.74). There was a trend for less infant movement during SSC (mean 8.9% normalized/h; SD 4.8) than for CTL (mean 13.3% normalized/h, SD 7.4; p = 0.09).

3.2.2. Cardio-respiratory and temperature assessments Table 2 provides a summary of infant and adult physiological signals. Neither the infant IBI mean (index of respiratory rate) or IBI variance (index of breathing stability) were different between CTL and SSC conditions, nor were there any differences in the incidents of apneic pauses N10 s between these conditions. Due to technical problems, ECG was not recorded in Subject 2 during CTL. For the 10 infants, mean heart rate trended slightly higher during SSC (150.8 bpm, SD 14.5) than CTL (148.1 bpm, SD 14.4; p = 0.07); variance of the infant heart rate was not different between conditions. Among all infants, mean oxygenation was similar between conditions (mean 97.3%, SD 3.2). Time in which infant O2 saturation fell below 85% was also not different between conditions; only four infants presented with episodes of clinical desaturation, two in whom duration increased and two in whom duration decreased with SSC. Due to technical problems, infant skin temperature was not recorded in Subject 7 during SSC. Among the remaining 10 infants, mean skin temperature was similar between conditions (36.3 ° C, SD 0.5); skin temperature variance trended higher during SSC (0.008) than during CTL (0.003; p = 0.07). Caregivers mean heart rate during SSC was 72.0 bpm (SD 8.1); mean heart rate variance was 18.7 (SD 12.9). Caregivers mean temperature during SSC was 35.9 ° C (SD 0.6); temperature variance was 0.017 (SD 0.034). 3.2.3. Ambient sound and light levels Due to technical problems, ambient sound was not measured for Subjects 2, 4 and 5, and light was not measured for Subjects 2 and 4. Among the remaining subjects, ambient sound was significantly louder during SSC (mean 51.3 dBA, SD 2.6) than during CTL (mean 48.9 dBA, SD 1.3; p = 0.02). Sound variance was also significantly greater during SSC (mean 18.0, SD 7.9) than during CTL (mean 10.2, SD 8.4; p = 0.03). There was no difference in light levels between CTL (mean 72.9 lux, SD 59.6) and SSC (mean 85.8 lux, SD 72.0; p = 0.35); light variance was also not significantly different between conditions (p = 0.34). 3.3. Relationships among infants and caregivers 3.3.1. Cardio-respiratory associations Among infants, increased respiratory instability was directly related to increased cessations in breathing, reflected by a direct relationship between infant IBI variance and infant apneas N 10 s for both CTL (rho = 0.853, p = 0.001) and SSC (rho = 0.830, p = 0.002). Cardiorespiratory relationships within infant and between infant and caregiver for each condition are illustrated in Fig. 2. During CTL, infant respiratory instability (IBI variance) was directly related to infant heart rate variance (rho = 0.770, p b 0.01). In contrast, during SSC, increased infant IBI variance was not significantly associated with increased infant

Table 1 Subject characteristics. Subject

Symbol

GA (wks)

PCA (wks)

Birth weight (g)

Study weight (g)

Gender

Race/ethnicity

SSC caregiver age (yrs)

1 2 3 4 5 6 7 8 9 10 11 Mean (SD)

○ □ w ◆ ▲ Q ▼ △ ▽ ■ ●

29.57 32.86 25.86 29.29 29.43 29.43 32.43 27.43 34.86 26.57 30.86 29.87 (2.74)

30.86 33.71 27.43 30.14 32.71 33.57 33.43 28.43 35.71 30.29 34.86 31.92 (2.67)

1180 1870 430 1598 1010 1390 1625 1410 2210 795 1010 1321 (506)

1185 1635 500 1390 1260 1770 1545 1240 2120 1090 1590 1393 (127)

Male Female Female Male Female Female Male Female Female Male Male

Hispanic Caucasian African American Caucasian Caucasian Caucasian Hispanic Caucasian Caucasian Caucasian Caucasian

23 25 28 31 34 34 40 45 32 37 41 33.64 (6.82)

Condition time: matched vs [TCT] (min) 31.01 21.13 39.57 28.55 48.72 69.81 43.93 11.31 86.01 38.51 36.00 41.32 (21.18)

[33] [51] [113] [122] [83] [70] [90] [66] [89] [90] [158] [88] (35)

Caffeine treatment at time of study

Respiratory support at time of study

Yes No Yes Yes No No No Yes No Yes No

None None Nasal Canula O2 None None None None Vapotherm® O2 None Vapotherm® O2 None

Note: symbol = subject identification for Fig. 2; GA = gestation Age; PCA = post conceptional Age at time of study; SSC caregiver Age = age of caregiver for skin-to-skin cohabitation; matched = equivalent SSC and CTL analysis periods; TCT = total time infants were held SSC.

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850 Table 2 Infant and caregiver physiological responses during control (CTL) and skin-to-skin cohabitation (SSC). CTL Mean (SD) Infant Respiratory rate (per min) Interbreath interval (s) Interbreath interval variance Apnea N5 s (incidents/h) Apnea N10 s (incidents/h) Heart rate (per min) Heart rate variance Skin temperature (°C) Skin temperature variance SaO2 (%) SaO2 variance

55.14 1.21 0.94 39.37 4.09 148.14 47.29 36.34 0.0028 97.28 6.32

Caregiver Heart rate (per min) Heart rate variance Skin temperature (°C) Skin temperature variance

SSC Mean (SD)

(17.5) (0.45) (0.86) (36.59) (5.57) (14.36) (25.07) (0.46) (0.002) (2.32) (10.25)

n/a n/a n/a n/a

50.30 1.28 0.89 33.79 3.81 150.78 48.71 36.35 0.0079 97.30 3.75

(13.6) (0.36) (0.79) (28.58) (5.72) (14.50) (34.73) (0.49) (0.008) (4.13) (4.48)

72.01 18.71 35.88 0.027

(8.05) (12.91) (0.57) (0.07)

P value

0.106 0.410 0.857 0.551 0.890 0.072 0.906 0.933 0.067 0.986 0.454

heart rate variance (rho = 0.545, p = 0.08; note, this trend toward significance is due to an outlier, Subject 8, which when excluded resulted in rho = 0.394, p = 0.26; Fig. 2, panel C). Rather, during SSC, increased infant IBI variance was associated with increased caregiver heart rate variance (rho 0.764, p = 0.006). During SSC infant apnea incidence (N 10 s) was also directly related to caregiver heart rate variance (rho

0.677, p = 0.022), whereas during CTL infant apnea incidence was not related to caregiver heart rate (p = 0.221) nor to its own heart rate variance (p = 0.231). Although there was an increase in sound variance during SSC, infant IBI variance was not associated with sound variance for either condition. Increased infant IBI variance was associated with mean infant temperature during CTL (rho = 0.636, p = 0.048), but not during SSC (rho = 0.515, p = 0.128); infant IBI variance was not associated with infant temperature variance for either condition (p N 0.1). 3.3.2. Cardio-respiratory coupling Cardiorespiratory coupling was measured in 10 infants (Subject 2 was excluded due to technical problems with ECG recording). Repeated measures ANOVA was performed to test the effects of Condition (CTL and SSC) and Frequency Band [0.01–0.15 Hz (LF); 0.15–0.45 Hz (HF1); 0.45–0.7 Hz (HF2); and 0.7–1.5 Hz (HF3, eupnea)] on cardiorespiratory coupling within infants. There was no main effect of Condition (p = 0.345). There was a significant main effect of Frequency Band [F(1.458, 13.118) = 9.108, p = 0.006, Greenhouse-Geisser correction, ε = .486] due to greater coupling in HF3 (eupneic frequency range; mean 0.575, SD 0.076) compared to LF (apneic frequencies; mean 0.448, SD 0.064, p = 0.003). An interaction was also observed between Condition and Frequency Band [F(3, 27) = 3.895, p = 0.020]. Post hoc comparisons revealed that coherence between infant heart rate (RR interval) and infant respiration in the eupneic frequency range (HF3, 0.7–1.5 Hz) increased significantly during SSC (mean 0.59, SD 0.07) compared to CTL (mean 0.56, SD 0.08, p = 0.018) and that for slower frequencies coupling was reduced significantly for HF1 (0.15– 0.45 Hz) during SSC (mean 0.45, SD 0.08) compared to CTL (mean

Infant Interbreath Variance

A) CTL within infant

(rho 0.770, p=0.009)

3.0

B) CTL infant and caregiver (ns) 3.0

2.5

2.5

2.0

2.0

1.5

1.5

1.0

1.0

0.5

0.5

0.0

0

20

40

60

80

100 120 140

0.0

0

Infant Interbreath Variance

C) SSC within infant (ns) 3.0

2.5

2.5

2.0

2.0

1.5

1.5

1.0

1.0

0.5

0.5

20

40

60

80

100 120 140

Infant Heart Rate Variance

20

30

40

50

D) SSC infant and caregiver

3.0

0

10

Caregiver Heart Rate Variance

Infant Heart Rate Variance

0.0

847

0.0

(rho 0.764, p=0.006)

0

10

20

30

40

50

Caregiver Heart Rate Variance

Fig. 2. Cardio-respiratory associations within infant and between infant and caregiver dyad during SSC and CTL. Symbols reflect individual subjects (see Table 1 for symbol key).

848

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850

0.51, SD 0.10; p = 0.036) and approached significance for LF (0.01– 0.15, p = 0.087; Fig. 3). To determine whether cardio-respiratory coupling in the eupneic range (HF3) changes over time, separate one-way ANOVAs were used to test the effects of coherence segment order for SSC and CTL (see Section 2.3.6). Due to variable durations in condition time among subjects (e.g., range approx. 11–70 min), we separately tested order effects for the first 9 coherence segments (9 subjects) and for the first 14 coherence segments (8 subjects), as well as for the 4.26 min alternating segments to provide a non-overlap profile. For each of these analyses, there was no effect of order, suggesting cardio-respiratory coupling did not adapt over the course of each condition.

4. Discussion The present integrative study is the first to explore the physiological interactions between the infant and caregiver during KC that may be linked to infant cardio-respiratory response. The findings support our hypothesis that respiratory stability of the premature infant is influenced by sensory perturbations of the caregiver's heart beat during SSC. We found that during SSC, infant respiration (IBI variance) and apnea episodes (N10 s) were each directly related to the variability of the caregiver's heart rate (RR variance). In contrast, during CTL infant respiratory instability (IBI variance) was directly related to its own heart rate instability. In addition, we demonstrated that within infant cardio-respiratory coupling was enhanced in the eupneic breathing frequencies (0.7–1.5 Hz) and reduced in the slower frequencies (0.15– 0.45 Hz) during SSC compared to CTL. These data suggest there are important physiological interactions between the caregiver and infant that may optimize efficacy of cardio-respiratory control. A caveat to the study was the small sample size. Nonetheless we observed strong LF HF1 HF2 HF3

0.01-0.15Hz 0.15-0.45Hz 0.45-0.70Hz 0.70-1.50Hz

(eupnea)

Fig. 3. Infant cardio-respiratory coupling. Coherence was significantly greater during SSC than CTL for the eupneic frequency range (0.7–1.5 Hz), and significantly reduced for slower frequencies (0.15–0.45 Hz).

associations. The ability to detect a significant effect in the small sample suggests this is a robust phenomenon in the general population. The importance of vestibular stimulation on early human development has been well established [35,40–43]. In particular vestibular stimulation has vital influences on the respiratory oscillator that include reducing apneic events and enhancing eupneic breathing in premature infants [35,43,44]. It is unclear whether it is the rocking movement per se that exerts phasic input on the respiratory pattern generator, or mechanical perturbations that impinge upon underlying receptors of the vestibular system that help modulate respiratory control [35,43,44]. For example, stochastic vibrotactile stimulation without rocking the infant has also been shown to improve respiratory stability and bloodoxygenation in preterm infants, supporting the use of stochastic perturbation for therapeutic management of respiratory dysrythmias in neonates [45]. Stimulation by rocking [35,43], by whole body motions at frequencies close to breathing via an oscillating mattress [46–48], as well stochastic stimulation among a range of frequencies [45] likely stimulate both vestibular and somatic afferents, which may entrain or enhance respiratory rhythms by resetting and pacing the system's oscillation via phasic input to its respiratory pattern [35,43–45]. The present finding that infant respiratory stability was associated with caregiver's heart beat during SSC supports the notion that external, perturbations impinge upon the respiratory control centers to modulate breathing. Furthermore, that cardio-respiratory coupling in the frequency range of normal breathing was also improved suggests SSC may optimize cardio-respiratory synergies that promote healthy physiology in premature infants [32]. The exact phasic relationships and entrainment of infant breathing to external perturbations warrant further study. There may be optimal perturbations for enhancing respiratory stability and rhythms that provoke instability. For example, during SSC, there may be some threshold at which caregiver's heart beat improves or worsens infant respiratory stability, which may explain in part why some infants show improved cardio-respiratory response with SSC [4,12,16,18–22], whereas others do not [23–25]. In the present study we found that when caregiver's heart rate variance was N30 there was an increase in infant IBI variance (e.g., Fig. 2D) and apneic incidence, but it remains unclear whether caregiver heart beat drives infant breathing or vice versa. A constraint of the study was that unlike controlled stochastic oscillations of an experimental mattress that wax and wane within a welldefined range of frequencies [45], the perturbations of the caregiver's heartbeat is not well defined, varies among caregivers, and may include unpredictable bursts of bradycardia and tachycardia. Due to such variable rhythms of the caregiver's heart beat and to inherent irregularity of breathing and frequent episodes of signal artifact due to frequent movement in the premature infants, we were unable to determine the points in time when infant respiration shifted to stable or unstable rhythms in response to caregiver heartbeat, or when caregiver heartbeat may have in fact responded to changes in infant respiration (e.g., following an apneic pause). Accordingly, it is unclear during SSC whether infant breathing rhythms follow or precede certain frequencies of caregiver's heart rate, or if caregiver's heart rate is responding to change in the infant's breathing pattern, or both. We speculate there are times during SSC when the infant respiration follows the caregiver's heart rate, particularly during stable periods, as well as when the caregiver's heart rate may reflect periods of infant breathing, particularly during periods of apnea and instability when clinical monitors alarm and cause stress to the caregiver. Time-series algorithms that assess instantaneous changes, account for interruption of signal (e.g., due to movement artifact; nursing intervention), and examine bi-directional influence and not necessarily symmetric interrelationships may help discern these unique relationships [49]. It is not surprising that infant temperature and ambient noise were more variable during SSC than CTL. All but one infant required an incubator due to immature thermoregulation, so it was expected that infant temperature would be less variable in the thermo-

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850

regulated environment of an incubator. Analogously, noise level was dampened in the incubator compared to the open-NICU environment. Notably, neither variance of infant temperature nor variance of ambient noise was associated with infant IBI variance. Rather, during SSC infant respiratory instability was associated with the caregiver's heart beat and not with the variability of temperature or ambient noise, further supporting the role of perturbation of the caregiver's heart beat rather than other external influences. This study supports the hypothesis that respiratory stability of the premature infant is influenced by the caregiver's cardiac rhythm during SSC. We speculate that during SSC, perturbations of the caregiver's heartbeat affect underlying receptors of the cohabitating infant. However, it is unclear whether the infant is responding to heart beat somesthetic perturbations of the caregiver per se, or to auditory (e.g., cardio-respiratory sounds) and/or olfactory cues (e.g., secretions from areola glands [50]) as the infant's head overlays the chest wall of the caregiver. Future studies designed to separate these cues may help discern those sensory perturbations that optimize infant cardiorespiratory responses during skin-to-skin cohabitation. 5. Conclusion The present study used a highly controlled experimental design that matched analysis periods between CTL and SSC for duration and time from feed. All infants were studied prone, with head of mattress slightly elevated to help mimic the KC position. Respiratory muscle movement of the caregiver was recorded to ensure periods of cross-signal contamination were excluded from analysis of the infant's respiratory signal. CTL periods always preceded SSC to control for potential carryover effects of SSC on cardio-respiratory control. We did not find differences in group mean incidence of apnea, duration of blood-oxygen desaturation or levels of skin temperature between CTL and SSC conditions, suggesting the importance of examining individual variability rather than group mean values per se for understanding interactions between the infant and caregiver that may influence respiratory stability of the infant. The major findings that infant IBI variance and apneic episodes were directly related to caregiver heart rate variance during SSC supports our hypothesis that respiratory stability of the premature infant is influenced by sensory perturbations of the caregiver's heart beat when the infant is held prone on its caregiver's chest skin-toskin. Furthermore, that infant cardio-respiratory coupling was enhanced during SSC compared to CTL in the eupneic frequency range suggest there are important physiological interactions between the caregiver and infant during SSC that may optimize cardio-respiratory synergies that promote healthy physiology in premature infants. Funding support This study was supported by a Grant in Aid and a Scientist Development Grant from the American Heart Association (EBS) and the Wyss Institute for Biologically Inspired Engineering, Harvard University (DP). The study sponsors had no involvement in study design, collection, analysis and interpretation of data, manuscript writing, or decision to submit the manuscript for publication. Conflict of interest statement The authors declare no conflicts of interest. Acknowledgments The authors gratefully acknowledge the infants and parents for participating in this study and appreciate the cooperation of the UMass Memorial NICU medical team.

849

References [1] Conde-Agudelo A, Belizan JM, Diaz-Rossello J. Kangaroo mother care to reduce morbidity and mortality in low birthweight infants. Cochrane Database Syst Rev 2011(3). http://dx.doi.org/10.1002/14651858.CD002771.pub2. [2] Charpak N, Ruiz JG, Zupan J, Cattaneo A, Figueroa Z, Tessier R, et al. Kangaroo mother care: 25 years after. Acta Paediatr 2005;94(5):514–22. [3] DiMenna L. Considerations for implementation of a neonatal kangaroo care protocol. Neonatal Netw 2006;25(6):405–12. [4] Ibe OE, Austin T, Sullivan K, Fabanwo O, Disu E, Costello AMD. A comparison of kangaroo mother care and conventional incubator care for thermal regulation of infants b 2000 g in Nigeria using continuous ambulatory temperature monitoring. Ann Trop Paediatr 2004;24(3):245–51. [5] Ludington-Hoe SM. Thirty years of kangaroo care science and practice. Neonatal Netw 2011;30(5):357–62. [6] Nyqvist KH, Anderson GC, Bergman N, Cattaneo A, Charpak N, Davanzo R, et al. State of the art and recommendations. Kangaroo mother care: Application in a high-tech environment. Acta Paediatr 2010;99(6):812–9. [7] Pattinson RC, Bergh AM, Malan AF, Prinsloo R. Does kangaroo mother care save lives? J Trop Pediatr Dec 2006;52(6):438–41. [8] Vesel L, ten Asbroek AHA, Manu A, Soremekun S, Agyemang CT, Okyere E, et al. Promoting skin-to-skin care for low birthweight babies: Findings from the Ghana Newhints cluster-randomised trial. Trop Med Int Health 2013;18(8):952–61. [9] Worku B, Kassie A. Kangaroo mother care: A randomized controlled trial on effectiveness of early Kangaroo mother care for the low birthweight infants in Addis Ababa, Ethiopia. J Trop Pediatr 2005;51(2):93–7. [10] Rao S, Udani R, Nanvati R. Kangaroo mother care for low birth weight infants: A randomized controlled trial. Indian Pediatr 2008;45:17–23. [11] Bagby K, Bowen S. Kangaroo care increases breastfeeding rates. J Obstet Gynecol Neonatal Nurs 2012;41:S49. [12] Begum EA, Bonno M, Ohtani N, Yamashita S, Tanaka S, Yamamoto H, et al. Cerebral oxygenation responses during kangaroo care in low birth weight infants. BMC Pediatr 2008;8:51. [13] Feldman R, Eidelman AI. Skin-to-skin contact (kangaroo care) accelerates autonomic and neurobehavioural maturation in preterm infants. Dev Med Child Neurol 2003; 45(4):274–81. [14] Ghavane S, Murki S, Subramanian S, Gaddam P, Kandraju H, Thumalla S. Kangaroo mother care in kangaroo ward for improving the growth and breastfeeding outcomes when reaching term gestational age in very low birth weight infants. Acta Paediatr 2012;101(12):e545–9. [15] Kaffashi F, Scher MS, Ludington-Hoe SM, Loparo KA. An analysis of the kangaroo care intervention using neonatal EEG complexity: A preliminary study. Clin Neurophysiol 2013;124(2):238–46. [16] Ludington-Hoe SM, Johnson MW, Morgan K, Lewis T, Gutman J, Wilson PD, et al. Neurophysiologic assessment of neonatal sleep organization: Preliminary results of a randomized, controlled trial of skin contact with preterm infants. Pediatrics 2006;117:e909–23. [17] Schneider C, Charpak N, Ruiz-Pelaez JG, Tessier R. Cerebral motor function in very premature-at-birth adolescents: a brain stimulation exploration of kangaroo mother care effects. Acta Paediatr 2012;101(10):1045–53. [18] Bergman NJ, Linley LL, Fawcus SR. Randomized controlled trial of skin-to-skin contact from birth versus conventional incubator for physiological stabilization in 1200- to 2199-gram newborns. Acta Paediatr 2004;93(6):779–85. [19] Bier JAB, Ferguson AE, Morales Y, Liebling JA, Archer D, Oh W, et al. Comparison of skin-to-skin contact with standard contact in low-birth-weight infants who are breast-fed. Arch Pediatr Adolesc Med 1996;150(12):1265–9. Boju SL, Krishna MG, Uppala R, Chodavarapu P, Chodavarapu R. Short spell kangaroo [20] mother care and its differential physiological influence in subgroups of preterm babies. J Trop Pediatr 2012;58(3):189–93. [21] Heimann K, Vaessen P, Peschgens T, Stanzel S, Wenzl TG, Orlikowsky T. Impact of skin to skin care, prone and supine positioning on cardiorespiratory parameters and thermoregulation in premature infants. Neonatology 2010;97(4):311–7. [22] Mitchell AJ, Yates C, Williams K, Hall RW. Effects of daily kangaroo care on cardiorespiratory parameters in preterm infants. J Neonatal Perinatal Med 2013;6(3):243–9. [23] Fischer CB, Sontheimer D, Scheffer F, Bauer J, Linderkamp O. Cardiorespiratory stability of premature boys and girls during kangaroo care. Early Hum Dev 1998;52(2): 145–53. [24] Mori R, Khanna R, Pledge D, Nakayama T. Meta-analysis of physiological effects of skin-to-skin contact for newborns and mothers. Pediatr Int 2010;52(2):161–70. [25] Bohnborst B, Heyne T, Peter CS, Poets CF. Skin-to-skin (kangaroo) care, respiratory control, and thermoregulation. J Pediatr 2001;138(2):193–7. [26] Cong X, Ludington-Hoe SM, McCain G, Fu P. Kangaroo Care modifies preterm infant heart rate variability in response to heel stick pain: Pilot study. Early Hum Dev 2009; 85(9):561–7. [27] Feldman R, Rosenthal Z, Eidelman AI. Maternal-preterm skin-to-skin contact enhances child physiologic organization and cognitive control across the first 10 years of life. Biol Psychiatry 2014;75(1):56–64. [28] Ludington-Hoe SM, Nguyen N, Swinth JY, Satyshur RD. Kangaroo care compared to incubators in maintaining body warmth in preterm infants. Biol Res Nurs 2000; 2(1):60–73. [29] Ludington-Hoe SM, Anderson GC, Swinth JY, Thompson C, Hadeed AJ. Randomized controlled trial of kangaroo care: cardiorespiratory and thermal effects on healthy preterm infants. Neonatal Netw 2004;23(3):39–48. [30] Park H-K, Choi BS, Lee SJ, Son I-A, Seol I-J, Lee HJ. Practical application of kangaroo mother care in preterm infants: clinical characteristics and safety of kangaroo mother care. J Perinat Med 2014;42(2):239–45.

850

E. Bloch-Salisbury et al. / Early Human Development 90 (2014) 843–850

[31] Clark MT, Rusin CG, Hudson JL, Lee H, Delos JB, Guin LE, et al. Breath-by-breath analysis of cardiorespiratory interaction for quantifying developmental maturity in premature infants. J Appl Physiol 2012;112(5):859–67. [32] Garcia AJ, Koschnitzky JE, Dashevskiy T, Ramirez JM. Cardiorespiratory coupling in health and disease. Auton Neurosci Basic Clin 2013;175(1–2):26–37. [33] Harper R, Gozal D. Sleep and respiratory control. In: Haddad G, Abman S, Chernick V, editors. Chernick-Mellins basic mechanisms of pediatric respiratory disease. 2nd ed. Hamilton, Ontario: BC Decker; 2002. p. 412. [34] Yasuma F, Hayano J. Respiratory sinus arrhythmia – Why does the heartbeat synchronize with respiratory rhythm? Chest 2004;125(2):683–90. [35] Sammon MP, Darnall RA. Entrainment of respiration to rocking in premature infants – Coherence analysis. J Appl Physiol 1994;77(3):1548–54. [36] Sontheimer D, Fischer CB, Scheffer F, Kaempf D, Linderkamp O. Pitfalls in respiratory monitoring of premature infants during kangaroo care. Arch Dis Child 1995;72(2): F115–7. [37] Abu-Shaweesh JM, Martin RJ. State of the art series: Sleep disordered breathing – neonatal apnea: What's new? Pediatr Pulmonol 2008;43:937–44. [38] Welch PD. The use of fast Fourier transform for the estimation of power spectra: A method based on time averaging over short, modified periodograms. IEEE Trans Audio Electroacoust 1967;15:70–3. [39] Indic P, Bloch-Salisbury E, Bednarek F, Brown EN, Paydarfar D, Barbieri R. Assessment of cardio-respiratory interactions in preterm infants by bivariate autoregressive modeling and surrogate data analysis. Early Hum Dev 2011;87(7): 477–87. [40] Dieter JN, Emory EK. Supplemental stimulation of premature infants: A treatment model. J Pediatr Psychol 1997;22(3):281–95. [41] Nandi R, Luxon LM. Development and assessment of the vestibular system. Int J Audiol 2008;47(9):566–77.

[42] Schaefer M, Hatcher RP, Barglow PD. Prematurity and infant stimulation: A review of research. Child Psychiatry Hum Dev 1980;10(4):199–212. [43] Zimmerman E, Barlow SM. The effects of vestibular stimulation rate and magnitude of acceleration on central pattern generation for chest wall kinematics in preterm infants. J Perinatol 2012;32(8):614–20. [44] Korner AF, Kraemer HC, Haffner ME, Cosper LM. Effects of waterbed flotation on premature-infants – Pilot study. Pediatrics 1975;56(3):361–7. [45] Bloch-Salisbury E, Indic P, Bednarek F, Paydarfar D. Stabilizing immature breathing patterns of preterm infants using stochastic mechanosensory stimulation. J Appl Physiol 2009;107:1017–27. [46] Korner AF, Guilleminault C, Van den Hoed J, Baldwin RB. Reduction of sleep apnea and bradycardia in preterm infants on oscillating water beds: A controlled polygraphic study. Pediatrics 1978;61(4):528–33. [47] Jones RA. A controlled trial of a regularly cycled oscillating waterbed and a nonoscillating waterbed in the prevention of apnoea in the preterm infant. Arch Dis Child 1981;56(11):889–91. [48] Saigal S, Watts J, Campbell D. Randomized clinical trial of an oscillating air mattress in preterm infants: Effect on apnea, growth, and development. J Pediatr 1986; 5(109):857–64. [49] Mrowka R, Cimponeriu L, Patzak A, Rosenblum MG. Directionality of coupling of physiological subsystems: age-related changes of cardiorespiratory interaction during different sleep stages in babies. Am J Physiol Regul Integr Comp Physiol 2003; 285(6):R1395–401. [50] Doucet S, Soussignan R, Sagot P, Schall B. The secretion of areolar (Montgomery's) glands from lactating women elicits selective, unconditional responses in neonates. PLoS One 2009;4(10):e7579.