Abstract
BACKGROUND:
Feeding tolerance among premature infants is unpredictable using clinical parameters. Ghrelin, a peptide hormone, acts on the hypothalamus to increase hunger and gut motility. It is present in fetal tissues, promotes intestinal maturation, and is secreted in milk. We hypothesized that higher serum ghrelin levels on days 0–7 are associated with improved feeding tolerance and growth in premature infants.
METHODS:
Infants (< 1500 g birth weight, n = 36) were recruited on day (D) 0–7. Serum ghrelin was measured by ELISA on D 0–7, D 10–14, and D 24–32, and milk ghrelin in a feeding concurrent with each serum sample. Feeding tolerance was assessed as days to first and full enteral feeds. Growth was quantified as both weight and adipose and muscle deposition by ultrasound.
RESULTS:
Mean serum ghrelin levels decreased from D 0–7 to D 24–32. Higher ghrelin levels on D 0–7 were correlated with shorter time to first enteral feeding, but not with time to full enteral feeds, rate of weight gain, or rate of accretion of muscle or adipose tissue. Milk ghrelin was not related to serum ghrelin or growth. Abdominal and suprascapular muscle and adipose increased during the first month, but weight gain correlated only with the rate of accretion of abdominal adipose.
CONCLUSIONS:
Elevated serum ghrelin in the first days of life may contribute to gut motility and readiness to feed. Weight gain in premature infants may primarily indicate abdominal fat accumulation, suggesting that ultrasound measurement of muscle accretion is a better marker for lean body growth.
Introduction
Ghrelin is a 28-amino acid peptide that is secreted predominantly by the parietal mucosa of the gastric fundus [1]. In adults, secretion occurs in a pulsatile manner throughout the day with pre-prandial surges [2]. Ghrelin stimulates growth hormone release from the pituitary gland and modulates gastric motility and acid secretion. In the central nervous system, it acts on the hypothalamic arcuate nucleus and dopaminergic regions of the mesolimbic system to increase appetite and regulate feeding behavior.
Ghrelin may also play a role in gut maturation during fetal and neonatal life. Ghrelin-secreting cells have recently been detected in fetal stomach, duodenum, pancreas, and lung from the 10th week of gestation [3]. It has been detected at comparable levels in the fetal arterial and venous circulations even during the third trimester, when placental expression is minimal [2]. Postnatally, infants are exposed to exogenous ghrelin in human milk, suggesting that ingestion of human milk may modulate neuroendoc-rine pathways involved in the regulation of appetite and feeding behavior [3]. In animal models, exogenous enteral ghrelin administration affects intestinal mucosal mitosis, apoptosis, crypt fission, and histometry. Animals receiving formula supplemented with ghrelin demonstrate improved intestinal growth and maturation of mucosal architecture, when compared to controls. These findings suggest that milk-borne ghrelin passes through the stomach intact and may play a role in development and function of the gastrointestinal tract [4].
It is likely that circulating serum ghrelin, derived from both endogenous expression and ingestion of human milk, acts on both the central nervous system and gastrointestinal tract to promote maturation of feeding function. Previous reports indicate that serum ghrelin levels increase during the first four weeks after term birth [5, 6]. However, the effects of ghrelin levels on growth and development of premature infants are not known [7]. We hypothesized that higher ghrelin levels in premature infants may be associated with earlier feeding tolerance and improved growth.
The goals of this study are to define the normative pattern of serum ghrelin levels in preterm infants during the first month, and determine if serum ghrelin is: 1) associated with enteral feeding tolerance in preterm infants as assessed by their time needed to reach first or full enteral feeds, 2) affected by dietary exposure, and 3) correlated with rate of growth. We also evaluated the utility of point-of-care ultrasound measurements of adipose and muscle tissue for assessing growth in preterm infants.
Methods
Subjects
Infants (n = 36) were recruited into this prospective cohort study from October 2017 to February 2019 at the neonatal intensive care units (NICU) at Cohen Children’s Medical Center (CCMC) and North Shore University Hospital (NSUH). This study was approved by the Northwell Health Institutional Review Board and informed consent was obtained from the parents of all participating infants prior to enrollment.
Premature newborns (≤30 weeks gestation) were eligible for recruitment within the first 7 days after birth. Exclusion criteria were major congenital gastrointestinal or genetic anomalies. Gestational age, maternal age, mode of delivery, days of life (DOL) upon enrollment, maternal diabetes, day of first ent-eral feeding, day of reaching full enteral feeds, wei-ghts, and types of enteral feeds (mothers milk vs donor milk, vs formula) were recorded from the electronic medical record. Composition of parenteral nutrition, initiation of feeds, and advancement of feeds were determined for all infants using a standard NICU protocol.
Specimen collection and processing
Samples were collected during three time windows, synchronized to clinically indicated blood sampling: D 0–7, D 10–14, and D 24–32. Specimens were collected in pretreated microtainers containing 50μl pefabloc to prevent ghrelin degradation. All samples were centrifuged within 60 minutes of collection (5000 RPM at –4°C for 15 minutes). Supernatants were removed, acidified with 5μl hyd-rochloric acid, and frozen at –80°C until analysis.
Samples of the milk (formula, mother’s milk, or donor milk) being fed to infants concurrently to the serum samples, and also on D 21, 45, and 60, were collected. Milks were not taken if infants were nil per os on the day of collection. Milk specimens were frozen at –80°C within 1 hour of collection.
All samples were analyzed at the Lilling Family Neonatal Research Lab, Feinstein Institute for Medical Research. Ghrelin was quantified using a human Ghrelin ELISA kit (ELH-GHRL-1, RayBiotech, Pea-chtree Corners, GA) according to the manufacturer’s instructions.
Ultrasound measurements of body composition
Point-of-care ultrasound studies of suprascapular and abdominal muscle and adipose tissue were performed according to the method described by McLeod et al. [8]. Studies were done once a week during the first month of life and on the D 45 and 60 using high-resolution (14 MHz) linear array transducer with skin preset (Zonare Z-One PRO, Mindray, Bio-Medical Electronics Co Ltd, Shenzhen, China), by an investigator (DK) blinded to feeding regimens. A high frequency transducer was selected to allow for optimal spatial and contrast resolution, using average gain of ∼70dB. The left side of the body was used in all subjects. The transducer was held orthogonal to the skin with examiner’s hand lightly resting on the patient which allowed for fine, controlled scanning. Warmed gel was put between the transducer and the skin, with an ample ultrasound gel ‘stand-off’ pad to minimize tissue compression. The transducer was positioned to image the Y-configuration of the left scapula along its longitudinal axis. The skin and subcutaneous fat appear as an echogenic layer in the image, and the muscle appears hypoechogenic with echogenic horizontal muscle fibers overlying highly echogenic bone (Fig. 1A). Measurements of the skin with subcutaneous fat and muscle were taken at the distal scapula tip [8]. Images were ‘frozen’ before measurements were taken. Each measurement was repeated three times and the average value recorded. The same procedure was used for the abdominal adipose and rectus muscle tissue at the level of the umbilicus (Fig. 1B).
Ultrasound visualization of suprascapular (A) and abdominal (B) adipose and muscle tissues. SSAT, suprascapular adipose tissue; SSMT, suprascapular muscle tissue; AAT, abdominal adipose tissue; AMT, abdominal muscle tissue.
To evaluate if ghrelin level is associated with feeding tolerance, correlation between ghrelin measurements at day 0–7 with time to first enteral feed and full enteral feed were each assessed using Spearman’s rank-order correlation, a non-parametric alternative to the Pearson correlation. Spearman correlations were used to determine the percent change in serum ghrelin from D 0–7 to D 24–32.
The crude rates of change from D 0–7 to D 24–32 for each ultrasound measure (suprascapular adipose, suprascapular muscle, abdomen adipose and abdomen muscle) were correlated with each other, and with percent weight change using Spearman rank correlation. Rate of change (per day) was measured as the absolute difference in each parameter from D 0–7 to D 24–32, divided by the number of days from D 0–7 to D 24–32 samples.
To assess the relationship between weight and change in ultrasound measures over time, a two-way factorial repeated measures analysis of variance (RMANOVA) was also performed, where DOL was a within-subjects effect, weight at each time point was another; presence of an interaction between weight and time on each ultrasound measurement was also evaluated.
To assess the association between serum ghrelin measurements and ghrelin in milk across time, RMANOVA was utilized. Reported negative values for milk ghrelin levels were converted to zero for analyses.
The significance of change in serum ghrelin over time and each ultrasound measure over time was each assessed using a random intercept and slope model. Serum ghrelin was log-transformed to better meet normality assumptions.
For all analyses, a result yielding p-value < 0.05 was considered statistically significant. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC).
Results
Demographic characteristics of the cohort are shown in Table 1. Mean serum ghrelin levels decreased from D 0–7 to D 24–32 (p = 0.0004) (Fig. 2A).
Demographic characteristics of the cohort
Demographic characteristics of the cohort
A. Plot of loge-transformed serum ghrelin during the first 30 days, with fitted trend line across patients. Mean serum ghrelin levels decreased from D 0–7 to D 24–32 (p = 0.0004). B. Scatter plot of loge-transformed serum ghrelin (D 0–7) with time to first enteral feed. Serum ghrelin (D 0–7) was negatively correlated with time to first enteral feed (r=–0.51, p = 0.0027).
Serum ghrelin on D 0–7 was negatively correlated with time to first enteral feed (r = –0.51, p = 0.0027) (Fig. 2B). In contrast, serum ghrelin was not correlated with time to full enteral feeds, rate of weight gain, or rate of accretion of muscle or adipose tissue. Milk ghrelin was not related to serum ghrelin or growth.
Suprascapular muscle, suprascapular adipose, abdominal muscle, and abdominal adipose tissue each increased significantly during the first 4 weeks. Muscle and adipose tissue accumulated at distinct rates (p = 0.0029 and p = 0.0132 for the suprascapular and abdominal sites, respectively). Weight gain was moderately positively correlated with the rate of accretion of abdominal adipose (rho = 0.65, p = 0.017), but not with abdominal muscle, suprascapular muscle, or suprascapular adipose (rho = 0.21, 0.48 and 0.46, respectively, and p≥0.1 for all) (Fig. 3).
Relationship between rate of change in weight and ultrasound measurements. Rate of weight gain was strongly correlated with rate of accretion of abdominal adipose (rho = 0.65, p = 0.017), but not with abdominal muscle, suprascapular muscle, or suprascapular adipose (rho = 0.21, 0.48, and 0.46, respectively).
We found that infants with higher early serum ghrelin levels more rapidly demonstrated signs of feeding readiness, which are dependent on intestinal motility. Serum ghrelin on D 0–7 is likely derived primarily from endogenous production in response to intrauterine conditions. Previous studies have reported that cord blood ghrelin levels are higher in premature infants with growth restriction, when compared to infants with intrauterine growth appropriate for gestational age [3]. Thus, ghrelin may be a compensatory mechanism for infants in an unfavorable intrauterine environment that primes receptiveness to postnatal enteral nutrition.
Ghrelin circulates in both acylated and unacylated forms. The presence of the acyl chain (octanoate residue) is required for binding to the growth hormone secretagogue receptor (GHSR-1a), which mediates many of the actions of ghrelin [9–11]. Thus, the acetylated form has been termed “active” by some authors. However, it is now known that unacylated ghrelin also plays a role in energy balance [12, 13], including adipogenesis via GHSR-1a independent pathways. Moreover, animal data suggest that degradation of acylated to unacylated ghrelin occurs more rapidly in fetuses and neonates than in adults [14]. The unacylated form accounts for more than 90%of ghrelin immunoreactivity in human plasma [15]. Therefore, total serum ghrelin levels rather than acylated ghrelin levels were measured for the current studies.
Contrary to expectation, higher early total serum ghrelin levels were not associated with improved feeding tolerance or weight gain in premature infants. This finding is consistent with the possibility that high early serum ghrelin is a marker for infants vulnerable to growth failure. Taken together, our findings suggest that infants with high total serum ghrelin demonstrate accelerated readiness for feeding, but that growth responsiveness may be impaired by other metabolic consequences of intrauterine stress.
Previous studies have shown that ghrelin is excreted in term human milk at higher [16, 17] or similar [18–20] concentrations to maternal serum, and that milk ghrelin is decreased in the presence of high maternal BMI or gestational diabetes [21]. It is possible that maternal factors affecting ghrelin expression in milk after preterm delivery are distinct from those after term delivery [22]. Our finding that levels of ghrelin in milk were not correlated with levels in infant serum supports the idea that endogenous expression in response to pre- and perinatal stressors is the predominant source of ghrelin in premature infants. Potential limitations to this study are the limited size of the cohort, wide variability of ghrelin measurements, and the possibility that milks sampled at the time of serum specimens may not be representative of the overall diet.
The optimal metrics for body growth in premature infants are not known, and our findings suggest that ultrasound measurements may be a more informative tool than weight alone. Accurate assessment of hospitalized infants’ growth and development is essential. Different methods have been described, including skin fold measurements, dual energy x-ray absorption, and magnetic resonance imaging. However, these methods are either painful, imprecise, costly, expose infants to radiation, or lack portability [23, 24]. In contrast, ultrasound is inexpensive, portable, and radiation-free [8, 25–27]. In this study, we have shown that ultrasound can be used as a bedside, clinical tool to quantify and characterize neonatal growth. Further, we found that muscle and adipose accumulate at different rates [8, 29]. Weight gain was most strongly correlated with the rate of accretion of abdominal adipose, suggesting that ultrasound measurements of muscle accretion may be better marker for meaningful growth in lean body mass [30].
In summary, serum ghrelin levels in the first week may be a marker for early feeding readiness in premature infants but are not related to longer term feeding tolerance or growth. Weight gain, which is currently used as the gold standard for body growth in premature infants, likely reflects primarily the gain in abdominal adipose tissue. Ultrasound measurements of muscle accretion are feasible and may be a better marker for growth of lean body mass.
Disclosures
The authors have no real or apparent conflict of interest to disclose. No external financial support was received for this study.