Hormonal and Neuromuscular Responses to Breastfeeding: A Pilot Study

Article Hormonal and Neuromuscular Responses to Breastfeeding: A Pilot Study Biological Research for Nursing 2017, Vol. 19(4) 399-408 ª The Author(s) 2017 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/1099800417697380 journals.sagepub.com/home/brn Madalynn Neu, PhD, RN1, Zhaoxing Pan, PhD2, Ashley Haight, BS3, Karen Fehringer, PhD, OTR/L4, and Katrina Maluf, PT, PhD5 Abstract Objectives: Difficult breastfeeding in the first weeks after birth may result in muscle tension in infants and activation of the maternal hypothalamic–pituitary–adrenal (HPA) axis and sympathetic nervous system (SNS). Our primary objective was to examine the feasibility of collecting neuroendocrine markers of maternal HPA axis and SNS activation (salivary cortisol and a-amylase [sAA]) and electromyographic (EMG) markers of infant distress during feeding in the first 2 weeks after birth. We also examined the relationships of these indices to each other and to mother–infant interactive behaviors during feeding. Methods: We recruited mothers in the postpartum unit of a teaching hospital and observed a feeding in the dyad’s home. Cortisol and sAA were sampled before feeding, 10 min into feeding, at feeding end, and 20 min after feeding. Infant muscle activity was recorded continuously with an EMG data logger. We used the Nursing Child Assessment Feeding Scale to measure mother– infant interaction. Results: The 20 mothers reported no disruption to breastfeeding and no change in infant behavior due to collection measures. Mean cortisol levels decreased significantly; there was no significant change in sAA levels. Relationships were found between interactive behavior and trends in neuroendocrine biomarkers. Longer bursts of infant muscle activity were associated with higher levels of maternal cortisol during feeding but not mother–infant interactive behaviors. Conclusions: Maternal salivary biomarkers and their association with feeding behaviors can be a useful tool for clinical longitudinal research beginning soon after birth. Infant EMG data may be useful for assessing maternal arousal. Keywords feeding, cortisol, a-amylase, electromyography, mother–infant interaction Because feeding is one of the most frequent nurturing mother– infant activities in the first months of life, early feeding is an important way to establish mother–infant relationship patterns. Studies have confirmed the importance of responsive mother– infant interactions in development of emotion regulation, trust, cognition, social competence, and secure attachment (Crockenberg, Leerkes, & Barrig, 2008; Kochanska et al., 2010). Mother and infant cocreate their relationship through patterns of exchanges. Repeated interaction instructs each dyad member to expect particular behaviors and adapt to the other during subsequent interactions (Black & Aboud, 2011; Fogel, 2000). Sensitive maternal behavior during feeding includes recognition of, and prompt response to, infant cues. Other maternal behaviors such as securing position; smiling and talking to the infant without overstimulating; and providing opportunities for physical, emotional, and social learning are considered adaptive and nurturing. Infants signal needs and respond to mothers’ interactive behavior (Oxford & Findlay, 2015). Although breastfeeding is reported to be superior to formula feeding in promoting positive and responsive mother–infant interaction (Bystrova et al., 2009), difficulties can occur that inhibit achievement of an optimal experience for mother and infant (Stuebe et al., 2014; Yalcin & Kuskonmaz, 2011). Maternal and/ or infant inability to participate in responsive feeding interaction may result in problems such as increased duration of infant crying, reduced maternal and infant sleep, persistent feeding difficulty, and poor mother–infant interaction (Neu et al., 2014; Yalcin & Kuskonmaz, 2011). A variety of physiologic, demographic, and maternal emotional factors can influence early breastfeeding and mother– infant interactional behavior. Research has recently identified 1 College of Nursing, University of Colorado Anschutz Medical Campus, CO, USA 2 Department of Pediatrics, University of Colorado Anschutz Medical Campus, CO, USA 3 School of Physical Therapy, University of Colorado, CO, USA 4 Colorado School of Public Health, University of Colorado Anschutz Medical Campus, CO, USA 5 School of Exercise and Nutritional Sciences, San Diego State University, CA, USA Corresponding Author: Madalynn Neu, PhD, RN, College of Nursing, University of Colorado, 13120 E. 19th Ave. Box C288-19, Aurora, CO 80045, USA. Email: madalynn.neu@ucdenver.edu 400 700 genes that affect maternal bonding, postpartum depression, social behaviors, nervous system developmental events, and signaling pathways for prolactin, oxytocin (OT), endogenous opioids, and steroid receptors (Gammie, Driessen, Zhao, Saul, & Eisinger, 2016). b-endorphins (BEs) activate reward centers in the brain, reducing stress and providing euphoric feelings in the mother that initiate positive maternal feelings and promote a desire for physical contact with the infant soon after birth. Breastfeeding and the associated skin-to-skin contact continue to initiate release of BEs and the associated rewarding maternal feelings (Bystrova et al., 2009; Sakala, Romano, & Buckley, 2016). Successful breastfeeding depends on the interaction of BEs, prolactin, epinephrine, norepinephrine, OT, and cortisol. Although moderate increases in epinephrine and norepinephrine assist the birth process, levels of these hormones decrease after birth to allow maternal relaxation and promote calm interaction with the infant (Hillman, Kallapur, & Jobe, 2012). Adequate prolactin levels are necessary for milk production and continuation of breastfeeding (Sakala et al., 2016). In addition, the neuropeptide OT is associated with maternal sensitivity, responsiveness, recognition of emotions and mood of the infant, and breastfeeding (Feldman, 2012; MacKinnon et al., 2014). While the interaction of optimal levels of these factors promotes successful breastfeeding, when they are out of balance, breastfeeding can be negatively impacted. For instance, adequate cortisol levels are necessary to stimulate prolactin, but excess cortisol inhibits prolactin production and thus lowers milk supply (Sakala et al., 2016). The demographic factors of cultural customs and socioeconomic status often influence the choice to breastfeed (Jones, Power, Queenan, & Schulkin, 2015). Maternal emotional status, such as symptoms of depression or anxiety, also affects breastfeeding success (Stuebe et al., 2014). Abnormal breast shape or mammary tissues, polycystic ovary disease, and obesity are other obstacles to breastfeeding (Marasco, 2014; Stuebe et al., 2014). Investigating the myriad factors that negatively affect breastfeeding was well beyond the scope of the present study, however. Instead, we chose to use measures that can be obtained in saliva (cortisol and a-amylase) and thus would be less invasive than blood samples in order to collect information on maternal arousal during breastfeeding, which is associated with mother–infant interactional feeding behavior. Although the literature shows that OT interacts with cortisol to lower stress in women without depression, and the hormone is important in facilitating maternal nurturing behaviors (Cox et al., 2015; Feldman, 2012), we did not include measures of salivary OT in the present study because of concerns about the reliability and validity of the measure (Horvat-Gordon, Granger, Schwartz, Nelson, & Kivlighan, 2005). Cortisol is the hormonal end product of the hypothalamic– pituitary–adrenal (HPA) axis, and salivary cortisol is a valid biomarker of HPA-axis function (Aardal & Holm, 1995). HPAaxis activation triggers a cascade that results in cortisol secretion, which mobilizes energy in order to respond to real or perceived Biological Research for Nursing 19(4) threats. Salivary a-amylase (sAA) is an enzyme produced in the acinar cells, which is an indirect biomarker of sympathetic nervous system (SNS) function (Nater & Rohleder, 2009). Reactivity of cortisol and/or sAA, or an increase or decrease in the levels of the biomarker following exposure to a stimulus, indicates sensitivity to that stimulus (Ji, Negriff, Kim, & Susman, 2015). Including both cortisol and sAA in studies is desirable because they respond differently to various challenges (Laurent, Ablow, & Measelle, 2012). The stable-state diurnal sAA pattern opposes that of cortisol, with sAA levels increasing (Giesbrecht, Granger, Campbell, Kaplan, & the APrON Study Team, 2012) and cortisol levels decreasing (Lupien & McEwen, 2009). sAA is rapidly activated, and increases may indicate either psychosocial stress or feelings of control, approach, and positive mood. Conversely, increases in cortisol levels are associated with negative affectivity, withdrawal, attachment stress, and lack of control (Laurent et al., 2012). Mothers who have difficulty breastfeeding their infants may perceive rejection, inadequacy, and disapproval and may cease to breastfeed (Zanardo et al., 2012). These feelings are associated with HPA-axis and SNS activations (Bosch et al., 2009; Dickerson & Kemeny, 2004). Activation of HPA-axis and SNS may lead to maternal withdrawal or hostility in interactions with the infant (Laurent et al., 2012; Out, BakermansKranenburg, van Petit, & van Ijzendoorn, 2012). Infants frequently exposed to maternal withdrawal or hostility are at risk for negative emotional arousal and self-regulatory disruption (Crockenberg et al., 2008; Gerhold, Laucht, Texdorf, Schmidt, & Esser, 2002; Mäntymaa et al., 2015). Activation of the HPA axis and SNS, as evidenced by increasing levels of cortisol and sAA, may be associated with observed mother–infant behavior during feeding. Most research on mother–infant interaction involving term infants involve infants older than 3 months of age (Golik et al., 2013; Karacetin, Demir, Erkan, Cokugras, & Sonmez, 2011; Richter & Reck, 2013), which misses the critical period immediately following birth. Rapid changes in the dyadic relationship occur during this time, likely establishing interaction patterns. Recognition of very early interactional difficulties might allow for targeted interventions that can prevent subsequent serious problems. Thus, in the present study, we examined the feasibility of collecting and assessing maternal salivary cortisol and sAA levels and their associations with dyadic interactional feeding behavior in the first 2 weeks of breastfeeding. Infant physiologic arousal, whether low or high, likely also plays an important role in the behavior of both the infant and mother during breastfeeding. However, sAA is at low levels at birth and is thus not a good indicator of SNS activity in the early weeks of an infant’s life (Davis & Granger, 2009). In addition, the collection of saliva for cortisol or sAA is not possible while the infant is feeding. However, evaluation of muscle tension using electromyography (EMG) may provide an indirect measure of infant distress. Surface EMG has been used previously in infants to detect pain, muscle movement, 401 Neu et al. and muscle tension (Teulier, Ulrich, & Martin, 2012; Worley, Fabrizi, Boyd, & Slater, 2012), yet it has not been evaluated as a biomarker during infant feeding. Therefore, we assessed muscle tension during feedings as an indicator of infant psychophysiological distress. Our primary objective in the present study was thus to examine the feasibility of collecting neuroendocrine markers of maternal HPA and SNS activation (salivary cortisol and sAA) and EMG markers of infant distress during feeding in the first 2 weeks after birth. We also examined the relationships of these indices to each other and to mother–infant interactive behaviors during feeding. Results will be helpful in planning future longitudinal research. Method Setting and Sample We recruited mother–infant dyads from the postpartum unit at a hospital in the western United States for this descriptive comparative study. Inclusion criteria for infants were term birth (i.e., 38–42 weeks’ gestation) with Apgar scores of 7–10 at 1 min and 8–10 at 5 min and less than 4 days of hospitalization after delivery (discharge policy of hospital). Inclusion criteria for mothers were English-speaking and at least 18 years of age. In this report, we include data only from mothers who were breastfeeding. Exclusion criteria for infants were congenital anomaly or small for gestational age (i.e., <10th percentile for infants born at term). Mothers were excluded if they were taking steroid medications or had a diagnosis of mental or chronic physical illness. For correlation analysis, sample size of 25 achieves 80% power to detect a difference of .52610 between the null hypothesis correlation of 0 and the alternative hypothesis correlation of.52610 using a two-sided hypothesis test with a significance level of .05. Measures Demographic Data. Recorded demographic data were age, ethnicity, gender of the infant, occupation and education of the mother, vaginal birth or C-section, length of stay of the infant in the hospital, medications taken in the last 24 hr, and type of infant feeding (breast or formula). Assessment of Mother–Child Interactive Behaviors. We used the Nursing Child Assessment Feeding Scale (NCAFS) to evaluate mother–child interaction during breastfeeding. The NCAFS is an observational tool with four maternal subscales (Sensitivity to Cues, Response to Distress, Social–Emotional Growth Fostering, and Cognitive Growth Fostering) and two infant subscales (Clarity of Cues and Responsiveness to Caregiver). Maternal subscales combined comprise 50 items, covering observable behaviors that describe maternal caregiving during a feeding. Scores range from 0 to 50 for the maternal subscales and from 0 to 26 for the infant subscales. Higher scores indicate more optimal maternal or infant behaviors. The NCAFS is used widely with infants from birth through 12 months of age. The scale has adequate internal and test–retest reliability. Concurrent validity with Bradley and Caldwell’s Home Observation for Measurement of the Environment Scale (1978) was r ¼ .48, .36, and .54 with caregiver, child, and dyadic scores, respectively (Oxford & Findley, 2015). Oxford and Findley (2015) published normative data for 853 dyads with infants 0–5 months of age. Physiological Measures. We assayed saliva samples within 3 months of collection, processing all samples from a single subject in the same assay. We used a commercial, expandedrange, high-sensitivity enzyme immunoassay kit (Salimetrics, State College, PA) to determine salivary cortisol levels. The kit detects cortisol levels in the range of 0.083–82.77 nmol/L (0.003–3.0 mg/dl). We ran samples in duplicate and reported the means. Inter- and intraassay coefficients of variability were 3.9% and 9.5%, respectively. To determine sAA levels, we used a commercial kinetic assay kit (Salimetrics), running the samples in duplicate and diluting them to a 1:200 ratio. Interand intraassay coefficients of variability for sAA were 4.0 and 11.1%, respectively. We quantified muscle activity as the rate and duration of bursts in the EMG signal throughout the feeding period. Our primary EMG outcomes focused on the temporal characteristics of muscle activity because it is not possible to obtain a reference contraction to compare the normalized amplitude of muscle activity across infants. Less frequent and longer bursts of muscle activity are typical of postural contractions and sustained muscle tension, whereas more frequent bursts of brief muscle activity are typically observed during voluntary or spontaneous movements. We quantified EMG bursts as Hunter and Enoka (2003) previously described. Briefly, the EMG signal was rectified and low-pass filtered at 2 Hz. The processed signal was then differentiated to identify rapid changes in signal amplitude and divided by the average of the rectified EMG to account for differences in absolute amplitude among infants. The start of a burst was identified when the differentiated EMG signal exceeded 5 standard deviations (SD) of the mean EMG during a 30-s period of muscle rest; the end of a burst was identified as the time at which the signal decreased most rapidly (i.e., most negative value) before the start of the next burst. Individual bursts were constrained to occur >2s apart, with a minimum duration of 0.5s. Burst duration was calculated as the time between the start and end of each identified burst, and the burst rate was calculated as the number of bursts per minute. Procedures The institutional review board at a 600-bed university teaching hospital in the western United States approved the study. Nurses from the Clinical and Translational Research Center (CTRC) at the recruitment hospital screened and approached mothers in the postpartum unit within 24 hr after delivery. The CTRC nurses explained the study to mothers who met the eligibility criteria and obtained informed consent for mother and infant. Enrollment lasted 3 months. A research assistant 402 (RA) telephoned each participating mother, explained the study protocol in detail, and scheduled a feeding observation appointment. Because both salivary cortisol and sAA show measurable levels between 11 a.m. and 1 p.m. (Nater & Rohleder, 2009), we scheduled observations to fall within this window. Feeding observations took place in the homes of the dyads and most lasted approximately 20 min. The RA videotaped the feeding observations for NCAFS scoring. Before videotaping, the RA asked the mother to feed the infant as usual, except to avoid watching television, texting, and/or having other children in the room when possible. The RA also asked the mother to refrain from talking to her (the RA) during the observation except during saliva collection. We chose to videotape in the dyad’s home, reasoning that the mother would be more comfortable with her new baby in her own home rather than in a laboratory environment. However, we could not hide the camera from the mother. The RA tried to be as unobtrusive as possible while videotaping at the same time insuring that the faces and as much of the bodies as possible of the members of the dyad were included. A certified NCAFS examiner later reviewed and scored the videotape. The NCAFS often is scored using video recordings, so the coder can stop the recording or reverse to confirm behaviors. The examiner used the NCAFS scoring sheets to note dyad behaviors. The RA collected saliva from the mother (1) immediately before the feeding, (2) 10 min after the feeding began, (3) when the feeding ended, and (4) 20 min after the feeding ended. Peak responses occur 20 and 10 min after a potential stressor for cortisol and sAA, respectively (Engert et al., 2011). We chose these time points for saliva collection to capture changes in cortisol and sAA during the feeding. Mothers sat quietly holding the infant while waiting for the final saliva collection. Mothers denied drinking caffeinated or alcoholic beverages or exercising within 3 hr prior to the observation, smoking, or consuming anything other than water for the hour prior; and using medications except vitamins in the previous 24 hr (Granger, Kivlighan, el-Sheikh, Gordis, & Stroud, 2007). To obtain saliva, mothers placed and kept a SalivaBio oral swab (SOS; Salimetrics, Carlsbad, CA) under their tongues for 2 min. The SOS filters large macromolecules and other particulate matter from the sample. The RA placed each sample in the Salimetrics collection tube and put the tube in an iced cooler. Within 2 hr after the observation, samples were placed in a freezer at 20 C, where they stayed until assay. The RA recorded infant muscle activity continuously throughout each feeding session using surface EMG. Before the feeding, the RA placed a concentric Ag/AgCl electrode (1.6-cm diameter) on the skin overlying the anterior deltoid of the infant’s arm that was facing away from the mother’s body during feeding and a ground electrode on the infant’s elbow. The RA secured electrodes with a Coban wrap to minimize movement artifacts in the EMG signal. EMG data were sampled with a single-channel EMG data logger (micrEMG, OT Bioelectronics, Torino, Italy) from the start to the end of the feeding session, as defined by the mother, without interruption during nonfeeding infant behaviors such as crying or sleeping. Biological Research for Nursing 19(4) Signals were amplified (980 V/V) and band-pass filtered (10– 500 Hz) prior to sampling (1,024 Hz) and stored on an internal secure digital (SD) card for subsequent off-line processing. We selected the microEMG device for ecological EMG monitoring in the home because it is minimally obtrusive (22 g, 57 mm  37 mm  16 mm), inexpensive, and portable. Results from a recent study indicate that the microEMG and the laboratorybased Telemyo 900 EMG system perform similarly, supporting the use of the microEMG for recording muscle activity during free living situations (Walters, Kaschinske, Strath, Swartz, & Keenan, 2013). After the observations, mothers completed the demographic survey. The RA also asked the mothers if they enjoyed feeding, if they typically did other activities during the feeding, if the videotaping and RA presence caused distress or was disturbing, if saliva collection had caused her discomfort, or whether the EMG monitoring had changed infant-feeding behavior. We gave a US$25 gift card to the mother and offered a complimentary breastfeeding consultation from an occupational therapist (K.F.) who specialized in infant feeding. Statistical Methods We used SAS 9.4 for statistical analysis. Because length of feeding varied from dyad to dyad, we did not collect samples at fixed time intervals; therefore, we chose slope as the best way to analyze cortisol and sAA. Based on the estimated slope, we classified the trend of each individual response profile as increasing (positive slope) or decreasing (negative slope) by fitting a linear regression model to the hormone data of each individual mother. To assess the average trending of 20 mothers, we used a linear mixed-effect model with random intercept and linear terms, which basically fits an individual linear regression line to individual data for each subject while providing a regression line for estimating the average trend of all the subjects. The dependent variable in this model was cortisol or sAA level, and the predictor was minutes from the first saliva sample (Time 0). We then correlated the slope estimate from this model with other measures. Because parity (multiparas and primipara) may be associated with differences in cortisol or sAA slope during feeding, we introduced parity status and its interaction with measurement time into the linear mixed-model analysis of trending. For both cortisol and sAA, the interaction term was not statistically significant, so we did not conduct a separate analysis for parity. We used independent t tests to compare NCAFS measures and Fisher’s exact test to compare categorical measures between the two trending categories. We calculated descriptive statistics for EMG outcomes to characterize total feeding time, burst rate, and burst duration. To assess associations between EMG and neuroendocrine indices of infant and maternal stress, we used Pearson’s correlation coefficients. We used the same analysis to examine the relationship of EMG indices to mother–infant behavior as assessed by NCAFS scores. 403 Neu et al. Table 1. Baseline Characteristics of Participants. Characteristic Mean (SD) Maternal age (years) Infant age (days) on study day 30.5 (6.1) 11.9 (4.8) n (%) 7 (35) Primipara Type of delivery Vaginal C-section Maternal education Some high school High school graduate College degree (4 year) Graduate degree Father involved, yes Race African American or Black Hispanic White 16 (80) 4 (20) 3 (15) 4 (20) 2 (10) 11 (55) 17 (85) 3 (15) 4 (20) 13 (65) Note. N ¼ 20. SD ¼ standard deviation. Results We recruited 28 mother–infant dyads. Of these, three mothers could not be contacted after enrollment, three mothers were missing cortisol data points due to feedings of less than 10 min, and two mothers bottle-fed their infants. Because of the small sample of bottle-fed infants, we included only breastfeeding dyads in this analysis. Thus, data for 20 mothers were available for the cortisol and sAA analyses. For the correlation of NCAFS and biological data, the sample was reduced to 19 due to one feeding videotape that was defective. Table 1 shows demographic data for the 20 dyads. The mean total NCAFS score of 56.7 (+7.7), mean maternal score of 39.6 (+5.3), and mean infant score of 17.1 (+3.3) were within 1 SD of the published norms (Oxford & Findley, 2015). Mothers reported that they enjoyed feeding, but most typically combined other activities with the feeding, such as watching television, talking on the phone or texting, talking to their other children, listening to music, reading, or working. However, five mothers (25%) reported concentrating only on the baby or feeding. NCAFS total mean scores did not differ between dyads of mothers who concentrated only on the infant (54.8 +10.4) and mothers who occupied themselves with other activities (57.4 +6.8) in this small sample. Only three mothers requested a feeding consultation. Another five mothers had scores in the lowest 10th percentile, according to NCAFS norms (Oxford & Findlay, 2015) but did not request a feeding consultation. Feasibility Our primary objective was to examine the feasibility of collecting neuroendocrine markers of maternal HPA and SNS activation (salivary cortisol and sAA) and EMG markers of infant distress during feeding in the first 2 weeks after birth. Mothers stated that the saliva collection was not disruptive to the feeding and denied feeling distressed by collection. They also stated that the EMG electrode did not change their infant’s usual feeding behavior. Thus, collection of multiple saliva samples was feasible. Visual inspection of raw EMG records often revealed substantial movement artifacts (identified as transient spikes in the EMG record exceeding the maximum voltage range [+10 V]) and poor signal quality (i.e., low signal-to-noise ratio) that could not be reliably interpreted as bursts of muscle activity. We removed these portions of the EMG signal during this visual inspection using custom interactive software developed in Mat Lab (Math Works, Inc., Natick, MA), leaving an average of 64% of the total feeding time available for analysis. We excluded complete EMG records for five infants due to poor signal quality. Mean EMG burst rate calculated for the remaining EMG data set was 6.8 + 4.0 bursts/min, and mean burst duration was 3.4 + 0.8 s. Physiologic Biomarkers and Associations With Mother–Infant Interactional Feeding Behavior Our secondary objective was to examine the relationship of maternal salivary cortisol and sAA and infant EMG indices to each other and to mother–infant interactive behaviors during feeding. Because there were variabilities between participants in terms of baseline values, direction and rate of change, and timing of salivary sampling, we used slope as the summary outcome. Figure 1 shows the regression lines for the 20 subjects with salivary cortisol and sAA data. We noted two outliers in the sAA regression. The linear mixed-model regression analysis showed a nonsignificant upward trend in sAA with or without outliers, so we included the outliers in the figure to show sample variability. The downward trend in salivary cortisol was significant, F(1, 19) ¼ 9.89, p ¼ .005. The sAA levels of seven mothers decreased during the feeding, while levels of 13 mothers increased. Table 2 shows the differences in NCAFS subscale scores, when dyads were grouped according to whether sAA increased or decreased. Only the difference in maternal sensitivity to infant cues was significant, showing that mothers with higher scores in sensitivity had decreasing levels of sAA versus mothers with lower scores, whose sAA levels increased. Most effect sizes were small (d < .5). Although the difference was not statistically significant, mothers of infants who were responsive to their mothers’ facial expressions and actions had decreasing sAA levels while levels of mothers with infants who were less responsive increased. The effect size for this difference was moderate (0.69). The salivary cortisol levels decreased for 15 mothers and increased for five. Mothers with higher scores on NCAFS sensitivity to infant cues showed a trend for decreasing levels of salivary cortisol versus mothers with lower scores, whose cortisol levels increased, T(1.99, 17), p ¼ .06, as shown in Table 2. Mothers of infants with higher scores on the display of clear feeding cues had increasing cortisol levels versus mother of 404 Biological Research for Nursing 19(4) B 700 * 600 Salivary Corsol (mcg/dL) Salivary Amylase (U/mL) A 500 * 400 300 200 100 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0 10 20 30 40 50 0 60 Minutes from the first saliva sample Regression line: Amylase = 74.4775 +0.7298*me 10 20 30 40 50 60 Minutes from the first saliva sample Regression line: Corsol = 0.2457 – 0.00165*me Figure 1. Raw data of (A) salivary a-amylase (sAA; U/ml) and (B) cortisol (nmol/L). Dashed lines connect data points from the same mother. The solid lines are the predicted population mean curves (n ¼ 20) from linear mixed-model regression analysis. The time points at which measures were collected were (1) immediately before the feeding, (2) 10 min after the feeding began, (3) when the feeding ended, and (4) 20 min after the feeding ended. Figure 1A shows trends (slope ¼ .73, p ¼ .16) in sAA levels (not significant) during the data collection period. Figure 1B shows trends in cortisol levels (slope ¼ .002, p ¼ .005) during the data collection period. *Two outliers showed large increases in sAA levels from baseline. Sensitivity analysis after excluding outliers produced a consistent positive but nonsignificant trend. Table 2. Nursing Child Assessment Feeding Scale (NCAFS) Scores in Relation to Direction of Trend in Change in Salivary a-Amylase (sAA) and Cortisol Levels. Trend Decreased NCAFS Subscale and Measure n Subscale Score M (SD) Maternal sensitivity to cues sAA 6 15.50 (0.55) Cortisol 15 14.53 (1.36) Maternal response to child’s distress sAA 6 9.50 (1.64) Cortisol 15 9.93 (1.22) Maternal social–emotional growth fostering sAA 6 11.00 (1.90) Cortisol 15 10.93 (2.12) Maternal cognitive growth fostering sAA 6 5.17 (1.17) Cortisol 15 5.13 (1.36) Infant clarity of cues sAA 6 11.17 (2.14) Cortisol 15 11.33 (2.19) Infant responsiveness to caregiver sAA 6 6.17 (0.75) Cortisol 15 5.67 (1.88) Trend Increased n Subscale Score M (SD) p Value 13 4 13.62 (1.39) 13.00 (1.41) 0.005 0.06 13 4 10.00 (1.15) 9.50 (1.73) 13 4 Cohen’s Effect Sizea Normed NCAFS Score for 0–5 Monthsb 1.94 1.10 13.63 13.63 0.45 0.57 0.36 0.29 10.14 10.14 10.54 (2.60) 9.75 (3.30) 0.70 0.39 0.20 0.43 11.48 11.48 13 4 4.69 (2.14) 3.75 (3.20) 0.62 0.46 0.29 0.60 5.52 5.52 13 4 11.85 (2.03) 12.75 (0.50) 0.51 0.03 0.32 1.05 11.99 11.99 13 4 5.15 (2.19) 4.75 (2.06) 0.16 0.41 0.69 0.47 7.03 7.03 Note. N ¼ 19. SD ¼ standard deviation. a Effect size was included as presentation of standardized differences between the groups and is helpful for assessing the clinical meaning of the observed difference. b The table for 12 years of maternal education was used because 85% of the maternal sample had completed high school (Oxford & White, 2015, p. 112). infants with lower scores, whose levels decreased, T(2.29, 16.9), p ¼ .04. Although the differences were nonsignificant, the moderate effect sizes for maternal social–emotional and cognitive growth fostering and infant responsiveness showed that mothers with decreasing cortisol levels had higher dyad growth-fostering scores and more responsive infants than mothers with increasing levels. There were no significant relationships between infant EMG burst rate or duration and NCAFS scores (r ¼ .33 and .10, respectively; p > .05). Similarly, infant EMG burst rate was not significantly correlated with neuroendocrine indices of maternal stress during feeding (mean salivary cortisol, r ¼ .35, p > .05; or mean sAA, r ¼ .21, p > .05). Infant EMG burst duration was significantly correlated with mean levels of maternal 405 Neu et al. salivary cortisol (r ¼ .60; p < .03) but was not associated with sAA (r ¼ .00; p > .05). Discussion Aim 1 was to determine the feasibility of collecting maternal saliva samples and infant EMG data during a feeding session in the first 2 weeks after birth. Although we found that it is feasible to collect maternal saliva samples during an infant feeding session, we could not use samples from three mothers because feedings were too short. In future research, regular timing of salivary sample collection and covarying feeding time may prevent subject loss. We excluded a high percentage of EMG recordings due to insufficient signal quality, suggesting limited feasibility of obtaining reliable EMG recordings from the arm muscles of feeding infants in the first 2 weeks of feeding. Future studies should consider monitoring activity of a more task-relevant muscle such as the masseter or frontalis. We did not select these muscles in the present study due to concerns that placement of EMG electrodes on the face might interfere with feeding or be unacceptable to the mother. Monitoring maternal trapezius muscle activity during feeding may be a more feasible alternative for assessing distress in the mother–infant dyad, as this muscle has been shown to be sensitive to psychological stressors in adults (Shahidi, Haight, & Maluf, 2013). Salivary Cortisol and sAA Findings Aim 2 was to gather descriptive information on levels of cortisol and sAA in mothers of infants less than 3 weeks of age during feedings. Overall, salivary cortisol and sAA levels showed that the HPA axis and SNS responded to the feeding inversely, but the trend in sAA was nonsignificant. Although we conducted observations at a time when sAA levels were expected to be high enough to be reactive, the levels actually displayed little reactivity and were generally low. Collection, even during late morning or early afternoon, may not reveal changes in sAA levels in the first 2 weeks after birth. The significant decrease in cortisol that we observed during the feeding session suggests that feeding evoked positive feelings for most mothers. In a prior study of pregnant women, Giesbrecht, Granger, Campbell, Kaplan, and the APrON Study Team (2012) found that not only momentary feelings of depression but also feelings of energy or arousal were associated with increased sAA levels. In the present study, when we classified sAA reactivity according to NCAFS scores, a small group of mothers with high sensitivity scores displayed a decrease in sAA levels, suggesting that they were emotionally aroused at the start of feeding but relaxed as they provided sensitive care to their responsive infants. The upward trend in cortisol levels in the small group of mothers in the present study with low dyad NCAFS scores suggests that mothers who were less sensitive to their infants’ cues and scored lower on fostering social–emotional and cognitive growth during the feeding session may have found feeding unpleasant. Their infants were less responsive, which may have engendered withdrawal or somewhat negative feelings. They also may have felt that the feeding was somewhat uncontrollable. Laurent, Ablow, and Measelle (2012) reported an increase in cortisol levels during attachment stress in mothers of older infants. It is interesting that, in the present study, mothers with increasing cortisol levels and the lowest NCAFS scores had infants who provided the clearest feeding cues. Perhaps these infants needed to provide clearer feeding cues to get their needs met. As stated previously, measuring all the physiologic factors that influence successful breastfeeding was well beyond the scope of this study, but other research has suggested connections among genetics, prolactin, BEs, OT, and cortisol. A genetic contribution is evident for maternal bonding and social behavior as well as production of necessary prolactin, OT, BEs, and steroid receptors (Bystrova et al., 2009; Gammie et al., 2016; Sakala et al., 2016). Low OT levels may have contributed to less sensitivity and fewer growth-fostering behaviors in mothers during feeding. Research has also found that OT inhibits HPA-axis secretion of cortisol (Cox et al., 2015). With inadequate levels of OT, cortisol levels would be more likely to increase and could inhibit prolactin production, which would decrease milk supply, making breastfeeding more difficult and less rewarding (Sakala et al., 2016). Because maternal BE production is maintained after the birth by rewarding and pleasurable feelings and, in turn, promotes pleasurable and rewarding feelings, difficult breastfeeding could reduce BE production and further diminish maternal pleasure during feeding. Infant sucking also stimulates release of maternal OT and prolactin (Sakala et al., 2016). Infants exhibiting low arousal and weak suck, disinterest in feeding, and/or low responsivity to the mother could result in less reward to the mother, decreasing BEs, and negative arousal in the mother indicated by increasing cortisol levels (Bosch et al., 2009; Dickerson & Kemeny, 2004; Zanardo et al., 2012). EMG Findings Compared to findings in a previous study of fatiguing isometric contractions in adults (Hunter & Enoka, 2003), infants in the present study displayed a higher frequency of shorter duration bursts of arm muscle activity during feeding. This pattern of muscle activity is consistent with spontaneous arm movements commonly observed in young infants. Our findings indicate that the frequencies of bursts of muscle activity were unrelated to mother–infant interactive feeding behaviors and thus are unlikely to reflect infant distress during feeding. Interestingly, however, longer periods of sustained muscle activity in infants were associated with higher levels of maternal salivary cortisol during feeding. This finding may provide preliminary support for the use of EMG burst duration as a biomarker of HPA axis–mediated distress in the infant– mother dyad. 406 Other Findings Mothers in this study breastfed their infants, at least for the first 1–2 weeks postbirth. The low NCAFS scores of five (26%) of the mothers may have been due to physiologic factors or, perhaps, may have reflected a settling-in period. Short hospital stays after birth seem to leave mothers in need of assistance with breastfeeding after discharge (Yalcin & Kuskonmaz, 2011). Also of interest is that only about one third of mothers reported concentrating only on their infant for the entire feeding or part of the feeding. Mean NCAFS scores did not differ between dyads of mothers who concentrated only on the infant and mothers who did not. With the proliferation of technological innovations, there may be long-term consequences of use of these innovations on the mother–infant relationship during feedings. This finding may thus indicate an important area for future research. Limitations This study was a pilot with a small sample and involved observation of only one feeding in the early weeks of life. Future research with a larger sample and/or more frequent observations would yield more information on mother–infant interactive behaviors and activation of the HPA axis and SNS during feeding over time. Mothers in this study exclusively breastfed. A larger longitudinal study could include mothers who use formula or mixed breast and formula feeding. The mothers in this study were primarily highly educated and fathers were involved. Findings may be different with a sample of mothers with less education and/or single mothers in high-stress situations. We did not collect additional variables—such as the amount of holding the mother has done, especially skin-toskin, social support, maternal body mass index, and nipple pain—that might have influenced feeding behavior and hormonal activity in this sample, but they would be important to include in future research. Because the SNS is activated by both positive and negative emotional arousal, researchers might consider using short surveys or qualitative interviews to provide insight into the mother’s feelings during feedings. Conclusion Despite these limitations, our findings demonstrate the feasibility of collecting maternal saliva for biomarker assay and infant EMG during feedings in the first 2 weeks of life and to report meaningful trends in these indices that are associated with mother–infant interactive feeding behavior. These observations indicated several areas for future research. The relationship between EMG burst duration and increasing maternal cortisol level during feeding suggests there may be value in further research into using infant EMG as a biomarker of HPA axis–mediated distress in the infant–mother dyad. Salivary cortisol and sAA are quite easy to collect and, coupled with behavioral observation, can be used as biomarkers of arousal in Biological Research for Nursing 19(4) mothers for future longitudinal research that has the potential to identify unfavorable feeding interaction trajectories early in the relationship. Clinically, this is important because intervention in the early weeks of life can prevent long-term and established difficulties and preserve breastfeeding in the dyad. Authors’ Note The contents of this article are solely the responsibility of the authors and do not necessarily represent official NIH views. Acknowledgments The authors would like to acknowledge Peggy Emmett, MA, MT (ASCP), and Mary Harrington, MT (ASCP), of the Children’s Hospital Colorado CTRC Core Lab for their valuable assistance with the salivary cortisol and sAA assays. Authors Contribution Madalynn Neu contributed to conception, design, and acquisition; drafted the manuscript; critically revised the manuscript; gave final approval; and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Zhaoxing Pan contributed to design, analysis, and interpretation; drafted the manuscript; critically revised the manuscript; gave final approval; and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Ashley Haight contributed to design and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Karen Fehringer contributed to design and interpretation, drafted the manuscript, critically revised the manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Katrina Maluf contributed to design, analysis, and interpretation; drafted the manuscript; critically revised the manuscript; gave final approval; and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by NIH/NCATS Colorado CTSA # UL1 TR000154 and UL1 TR001082 and intramural funding from University of Colorado College of Nursing. References Aardal, E., & Holm, A. C. (1995). Cortisol in saliva—Reference ranges and relation to cortisol in serum. European Journal of Clinical Chemistry and Clinical Biochemistry, 33, 927–932. Black, M. M., & Aboud, F. E. (2011). Responsive feeding is embedded in a theoretical framework of responsive parenting. Journal of Nutrition, 141, 490–494. doi:10.3945/jn.110.129973 Bosch, J. A., de Geus, E. J., Carroll, D., Goedhart, A. D., Anane, L. A., van Zanten, J. J., . . . & Edwards, K. M. (2009). A general enhancement of autonomic and cortisol responses during social evaluative threat. Psychosomatic Medicine, 71, 877–885. doi:10.1097/PSY. 0b013e3181baef05 Neu et al. Bradley, R. H., & Caldwell, B. M. (1978). Screening the environment. American Journal of Orthopsychiatry, 48, 114–129. Bystrova, K., Ivanova, V., Edhborg, M., Mattiesen, A. S., RansjoArvidson, A. B., Mukhamedrakhimov, R., . . . Wodstrom, A. M. (2009). Early contact versus separation: Effects on mother-infant interaction one year later. Birth, 36, 97–109. doi:10,1111/j.1523536x.2009.00307.x Cox, E. Q., Stuebe, A., Pearson, K., Grewen, K., Rubinow, D., & Meltzer-Brody, S. (2015). Oxytocin and HPA stress axis reactivity in postpartum women. Psychoneuroendocrinology, 55, 164–172. doi:10.1016/j.psyneuen.2015.02.009 Crockenberg, S. E., Leerkes, E. M., & Barrig, J. P. S. (2008). Predicting aggressive behavior in the third year from infant reactivity and regulation as moderated by maternal behavior. Developmental Psychopathology, 20, 37–54. doi:10.1017/S0954579408000023 Davis, E. P., & Granger, D. A. (2009). Developmental differences in infant salivary alpha-amylase and cortisol responses to stress. Psychoneuroendocrinology, 34, 795–804. doi:10.1016/jpsyneuen. 2009.02.001 Dickerson, S. S., & Kemeny, M. E., (2004). Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory research. Psychological Bulletin, 130, 355–391. doi:10.1037/ 0033-2909.130.3.355 Engert, V., Vogel, S., Efanov, S. I., Duchesne, A., Corbo, V., Ali, N., & Pruessner, J. C. (2011). Investigation into the cross-correlation of salivary cortisol and alpha-amylase responses to psychological stress. Psychoneuroendocrinology, 36, 1294–1302. Feldman, R. (2012). Oxytocin and social affiliation in humans. Hormones and Behavior, 61, 380–391. doi:10.1016/j.yhbeh.2012.01. 008 Fogel, A. (2000). Beyond individuals: A relational-historical approach to theory and research on communication. In M. L. Genta (Ed.), Mother-infant communication (pp. 123–161). Rome, Italy: Carocci. Gammie, S. C., Driessen, T. M., Zhao, C., Saul, M. C., & Eisinger, B. E. (2016). Genetic and neuroendocrine regulation of the postpartum brain. Frontiers in Neuroendocrinology. 42, 1–17. doi:10.1016/j.yfrne.2016.05.002 Gerhold, M., Laucht, M., Texdorf, C., Schmidt, M. H., & Esser, G. (2002). Early mother-infant interaction as a precursor to childhood social withdrawal. Child Psychiatry and Human Development, 32, 277–293. Giesbrecht, G. F., Granger, D. A., Campbell, T., & Kaplan, B., & the APrON Study Team. (2012). Salivary alpha-amylase during pregnancy: Diurnal course and associations with obstetric history, maternal demographic, and mood. Developmental Psychobiology, 55, 156–167. doi:10.1002/dev.21008 Golik, T., Avni, H., Nehma, H., Greenfield, M., Silvan, Y., & Tauman, R. (2013). Maternal cognitions and depression in childhood behavioral insomnia and feeding disturbances. Sleep Medicine, 14, 261–265. Granger, D. A., Kivlighan, K. T., el-Sheikh, M., Gordis, E. B., & Stroud, L. R. (2007). Salivary alpha-amylase in biobehavioral research: Recent developments and applications. Annals of the New York Academy of Science, 1098, 122–144. doi:10.1196/ annals.1384.008. 407 Hillman, N. H., Kallapur, S., & Jobe, A. H. (2012). Physiology of transition from intrauterine to extrauterine life. Clinics in Perinatology, 39, 769–783. doi:10.1016/j.clp.2012.09.009 Horvat-Gordon, M., Granger, D. A., Schwartz, E. B., Nelson, V. J., & Kivlighan, K. T. (2005). Oxytocin is not a valid biomarker when measured in saliva by immunoassay. Physiology and Behavior, 84, 445–448. Hunter, S. L., & Enoka, R. M. (2003). Changes in muscle activation can prolong the endurance time of a submaximal isometric contraction in humans. Journal of Applied Physiology, 94, 108–118. doi:10.1152/japplphysiol.00635.2002 Ji, J., Negriff, S, Kim, H., & Susman, E. J. (2015). A study of cortisol reactivity and recovery among young adolescents: Heterogeneity and longitudinal stability and change. Developmental Psychobiology, 58, 283–302. doi:10.1002/dev.21369 Jones, K. M., Power, M. L, Queenan, J. T., & Schulkin, J. (2015). Racial and ethnic disparities in breastfeeding. Breastfeeding Medicine, 10, 186–196. doi:10.1089/bfm.2014.0152 Karacetin, G., Demir, T., Erkan, T., Cokugras, F. C., & Sonmez, B. A. (2011). Maternal psychopathology and psychomotor development of children with GERD. Journal of Pediatric Gastroenterology and Nutrition, 53, 380–385. Kochanska, G., Woodard, J., Kim, S., Koenig, J. L., Yoon, J. E., & Barry, R. A. (2010). Positive socialization mechanisms in secure and insecure parent-child dyads: Two longitudinal studies. Journal of Child Psychology and Psychiatry, 51, 998–1009. doi:10.1111/j. 1469-7610.02238.x Laurent, H. K., Ablow, J. C., & Measelle, J. (2012). Taking stress response out of the box: Stability, discontinuity, and temperament effects on HPA and SNS across social stressors in mother-infant dyads. Developmental Psychology, 48, 35–45. doi:10.1037/ a0025518 Lupien, S. J., & McEwen, B. S. (2009). Effects of stress throughout the lifespan on the brain, behavior, and cognition. Nature Reviews Neuroscience, 10, 434–445. doi:10.1038/nrn2639 MacKinnon, A. L., Gold, I., Feeley, N., Haydon, B., Carter, C. S., & Zelkowitz, P. (2014). The role of oxytocin in mothers’ theory of mind and interactive behavior during the perinatal period. Psychoneuroendocrinology, 48, 52–63. doi:10.1016/j.psyneuen,2014.06. 003 Mäntymaa, M., Puura, K., Luoma, L., Latva, R., Salmelin, R. K., & Tamminen. (2015). Shared pleasure in early mother-infant interaction: Predicting lower levels of emotional and behavioral problems in the child and protecting against the influence of parental psychopathology. Infant Mental Health Journal, 36, 223–237. doi:10.1002/imhj.21505 Marasco, L. A. (2014). Unsolved mysteries of the human mammary gland: Defining and redefining the critical questions from the lactation consultant’s perspective. Journal of Mammary Gland Biology and Neoplasia, 19, 271–288. doi:10.1007/s10911-015-9330-7 Nater, U. M., & Rohleder, N. (2009). Salivary amylase as a noninvasive biomarker for the sympathetic nervous system: Current state of research. Psychoneuroendocrinology, 34, 486–496. doi:10.1016/j.psyneuen.2009.01.014 Neu, M., Schmiege, S., Pan, Z., Fehringer, K., Workman, R., Marcheggianni-Howard, C., & Furuta, G. (2014). Interactions during 408 feeding with mothers and their infants with symptoms of gastroesophageal reflux. Journal of Alternative and Complementary Medicine, 20, 493–499. doi:10.1089/acm.2013.0233 Out, D., Bakermans-Kranenburg, M. J., van Petit, J., & van Ijzendoorn, M. H. (2012). Salivary a-amylase and intended harsh caregiving in response to infant crying: Evidence for physiological hyperreactivity. Child Maltreatment, 17, 295–305. doi:10.1177/ 1077559512464427 Oxford, M., & Findlay, D. (2015). Caregiver/parent-child interaction feeding manual. Seattle, WA: NCAST. Richter, N., & Reck, C. (2013). Positive maternal interaction behavior moderates the relation between maternal anxiety and infant regulatory problems. Infant Behavior and Development, 36, 498–506. Sakala, C., Romano, A. M., & Buckley, S. J. (2016). Hormonal physiology of childbearing, an essential framework for maternalnewborn nursing. Journal of Obstetric, Gynecologic and Neonatal Nursing, 45, 264–275. doi:10.1016/j.jogn.2015.12.006 Shahidi, B., Haight, A., & Maluf, K. (2013). Differential effects of mental concentration and acute psychosocial stress on cervical muscle activity and posture. Journal of Electromyography and Kinesiology, 23, 1082–1090. doi:10.1016/jelekin.2013.05.009 Stuebe, A. M., Horton, B. J., Chetwynd, E., Watkins, S., Grewen, K., & Meltzer-Bordy (2014). Prevalence and risk factors for early, Biological Research for Nursing 19(4) undesired weaning attributed to lactation dysfunction. Journal of Women’s Health, 23, 404–412. doi:10.1089/jwh.2013.4506 Teulier, C., Ulrich, B. D., & Martin, B. (2012). Functioning of peripheral Ia pathways in infants with typical development: Typical responses in antagonist muscle pairs. Experimental Brain Research, 208, 581–593. doi:10.1007/s00221-0102506-x Walters, T. J., Kaschinske, K. A., Strath, S. J., Swartz, A. M., & Keenan, K. G. (2013).Validation of a portable EMG device to assess muscle activity during free-living situations. Journal of Electromyography and Kinesiology, 23, 1012–1019. doi:10.1016/ j.jelekin.2013.06.004. Worley, A., Fabrizi, L., Boyd, S., & Slater, R. (2012). Multi-nodal pain measurements in infants. Journal of Neuroscience Methods, 205, 252–257. doi:10.1016/j.jneumeth.2012.01.009 Yalcin, S. S., & Kuskonmaz, B. B. (2011). Relationship of lower breastfeeding score and problems in infancy. Breastfeeding Medicine, 6, 205–208. doi:10.1089/bfm.2010.0092 Zanardo, V., Savona, V., Cavallin, F., D’Antona, Giustardi, A., & Trevisanuto, D. (2012). Impaired lactation performance following elective delivery at term: Role of maternal levels of oxytocin and prolactin. Journal of Maternal-Fetal & Neonatal Medicine, 25, 1595–1598. doi:10.3109/14767058.2011.648238