Music therapy and weight gain in preterm infants: Secondary analysis of the randomized controlled LongSTEP trial

1 INTRODUCTION

Appropriate growth is an important aspect of health and development during pregnancy, infancy, and childhood^{1, 2} and is associated with better clinical and neurodevelopmental outcomes for preterm infants.^{3, 4} The growth of preterm infants is influenced by many elements, including genetic factors, prenatal conditions, degree of prematurity, morbidities, nutrition, and developmental supportive care.^5-8 Monitoring the growth of this sensitive population is complex and requires specific tools and expertise.^9-13 Developmentally supportive care (DC) interventions were developed to reduce stress and discomfort and therefore may reduce energy expenditure and allow caloric intake to be used for growth.^{14, 15} Enhancing parent–infant bonding may increase breastfeeding duration and rate, reduce infants' stress, and stabilize vital signs that may lead to better growth.^{14, 15}

Music therapy (MT), one of the DC modalities, is intended to support infant development and parent–infant relationships by empowering parents in their parental roles and in understanding their infant's behavior.^{16, 17} MT may enhance the growth of preterm infants through its favorable effect on physiological stability, including heart rate and breathing,¹⁸ may promote maturation of sucking behavior,^19-21 and enhance parent–infant bonding.^18-22

Previous studies evaluating the effect of MT on weight gain among preterm infants produced inconsistent results.^{7, 18} Lack of study power and retrospective design,²² uncontrolled studies,²³ short assessment periods,^19-21 different types of MT interventions,^{19, 21, 24} and evaluation of different growth parameters⁷ have contributed to the challenges in evaluating the effect of MT on weight gain among preterm infants.

The present study assessed the effect of MT on weight gain of preterm infants during hospitalization in the neonatal intensive care unit (NICU) by analyzing data accumulated during the Longitudinal Study of Music Therapy's Effectiveness for Premature Infants and their Caregivers (LongSTEP) study, a large, multinational, randomized clinical trial.¹⁷

2 METHODS

3 RESULTS

A total of 213 participants were enrolled and randomly assigned to the LongSTEP study. Participants assigned to MT received a mean (SD) of 10.0 (6.0) sessions. Eighty-seven of 105 (83%) received per-protocol MT (≥6 sessions). Mothers were present at 84 sessions (SD 6.0), fathers at 2.2 (SD 3.4), and both parents at 2.7 (SD 3.7). Mean (SD) session length was 30.8 (11.2) min. Data from 201 infants were analyzed in the current study. We excluded infants who had missing weight measurements (11 in the SC group and one in the MT group; Figure 1). The baseline characteristics of the infants in the MT and SC groups are described in Table 1. They were not statistically different.

Weight parameters (weight, weight percentile, weight Z-scores) at birth, study enrollment, and discharge are presented in Table 2. No differences in weight parameters were recorded between the groups at any time point. Weight gain velocity was also similar between study groups (Table 2). No statistical difference in EUGR, represented by change in Z-scores from birth to discharge, was recorded between MT and SC (Table 2). Table 3 presents the linear mixed model for weight parameters. During the study period, weight parameters were not correlated with the study group (SC vs. MT). However, weight parameters were correlated with birth weight, male sex, and length of stay.

EUGR was significantly influenced by gestational age at birth and male sex; both were associated with an increased risk of EUGR at discharge (Table 4). Singleton birth, presence of neurological morbidities, antenatal steroid treatment, other neonatal morbidities (systemic infection, BPD, ROP, or NEC), parental factors (duration of skin-to-skin care, bonding, anxiety, and depression questionnaire scores), and type of enteral nutrition did not significantly influence the weight parameters in the study cohort.

4 DISCUSSION

In this subanalysis of data from the LongSTEP study,¹⁷ we found that MT during NICU hospitalization was not associated with better weight parameters compared to preterm infants who received SC. This finding does not support previous studies that showed that MT was associated with better weight gain.^19-23

Adequate growth of preterm infants is correlated with health, well-being, and favorable neurodevelopmental outcomes. Growth is affected by many factors, including environmental, genetic, demographic, comorbidities, and standard of care.^{5, 32} Therefore, it is not surprising that MT alone did not significantly influence weight parameters in the cohort. Of note, is that the use of pacifier-activated music in previous studies^19-21 could lead to reinforcement of sucking behavior, whereas MT targeted in between feedings as in the LongSTEP study did not do so. The mother listening to music while expressing milk may increase quantity and macronutrients, whereas music that is targeted solely at the infant might not have this effect. However, the effect of those music interventions on growth was not well-demonstrated.⁷

Extreme prematurity increased the risk for EUGR in our cohort. However, in contrast to other studies,⁸ systemic infection during the NICU stay, gastrointestinal, respiratory, and neurological comorbidities were not significantly associated with EUGR. This could be because most of our cohort had mild comorbidities that did not substantially affect weight gain.

We found that male preterm infants were at higher risk for EUGR. Our findings are in agreement with those of other studies that demonstrated more morbidities and higher mortality among male preterm infants, which may be associated with EUGR.^33-35

The LongSTEP study investigators choose to diagnose BPD according to the definition of most study centers.²⁵ We are aware that new definitions are gaining wider use but did not implement them in the current study to avoid different results than those in the LongSTEP study.^{16, 17}

Preterm birth imposes an unnatural separation between infants and their families, which contributes to increased stress and anxiety among babies and parents.^{36, 37} Lack of sufficient, consistent presence of a primary caregiver (usually the parents) is correlated with failure to thrive.³⁸ The NICUs in our cohort were located in countries where public support and welfare initiatives ensured consistent parental presence during hospitalization.¹⁶ Furthermore, all NICUs had implemented DC modalities to support infants and families.^{15, 39, 40}

Duration of skin-to-skin care as well as parental bonding, anxiety, and depression scores did not affect weight gain patterns in our cohort. This could be explained by the strong parental presence and high compliance with skin-to-skin care interventions (almost daily among most participants) and by the support system DC NICUs offer. It is also possible that parents who consented to participate in an MT trial have special characteristics.⁴¹

Of note, despite DC interventions, the EUGR rate of babies born <1500 g (very low birth weight) in our cohort (n = 109) was approximately 40%, which is consistent with the rates reported in large cohort studies that included NICUs; many without a special emphasis on DC.^{8, 42} This further demonstrates the complex, multifactorial influences on growth patterns.

Growth remains the most important key to assessing nutrition adequacy of preterm infants.⁴ Our data included information on human milk feeding, which is an important yet partial aspect of neonatal nutritional care. Both nutritional care and growth assessment require expertise to accurately interpret the data, adjust treatment, and allow optimal conditions for sufficient growth.⁴³

Some parameters of nutrition recommendations for preterm infants are not strongly validated.⁴⁴ Growth of preterm infants during their NICU stay is divided into three periods: adaptation to extrauterine life, transition from parenteral to enteral feeding, and full enteral feeds and stable growth.⁴⁵ The vast differences in postnatal age of study enrollment (Table 2) and partial data regarding nutritional intake (Table 1) made it difficult to interpret weight parameters. Furthermore, definition and classification of EUGR are not yet validated as having a prognostic yield and are highly variable depending on the growth reference used. Both the American Academy of Pediatrics⁴⁶ and the European Society of Pediatric Gastroenterology, Hepatology and Nutrition⁴⁴ recommend using a growth chart based on a large, robust database. Therefore, we chose the Fenton fetal–infant growth reference.³⁰ EUGR is useful for monitoring growth both in clinical and research settings, but it should be interpreted cautiously and in conjugation with other growth measures, such as velocity.⁴

Site-specific differences in weight gain patterns and in aspects of care known to affect growth, such as gestational age at discharge¹³ and human milk feeding,^{47, 48} highlight the influence of local practices and populations on growth outcomes.⁴⁴ Unfortunately, we could not draw any firm conclusions because the sample size per site was small. Nevertheless, local practices should be assessed regularly as part of NICU policies.

The strength of this study is the prospective nature of the cohort. In addition, the data concerning DC modalities offer a unique, holistic insight into aspects not usually addressed in literature related to growth or nutritional assessments.^{49, 50}

The main limitation is that weight gain does not equal growth. Length and head circumference measurements are fundamental for appropriate growth assessment; however, these were not included in the LongSTEP study. Furthermore, although some general nutritional guidelines were adopted by all sites, data on nutritional intake were limited. Of note, the length and timing of interventions varied significantly (Table 1), which may influence the effect of MT on weight gain parameters. Furthermore, the study was not powered to detect a difference in weight gain between the MT and SC groups; therefore, we might not have detected the effect of MT on weight gain, which larger studies might find.

In conclusion, MT for preterm infants and families was not associated with better weight gain parameters compared to the control group. Whether MT and/or weight gain patterns will correlate with future neurodevelopment and attachment is yet to be determined by the LongSTEP study results. Degree of prematurity and male sex remains the most significant risk factors for unfavorable weight outcomes, advocating focus on the very preterm infants and offering individually adjusted care to improve outcomes.

Future studies assessing the effect of MT on development and growth should be designed with early and long durations of MT intervention and include extremely, very, and late preterm infants, detailed data on nutritional intake, and assessment of complete anthropometric data, before, during, and after the study period. The study protocol should include guidelines for nutritional support to avoid differences in nutritional support among sites.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

The main trial received ethics approval from the Regional Committees for Medical and Health Research Ethics in Norway and was supplemented by ethics approvals in each participating country. The current study was included in the approvals granted for the LongSTEP study (ClinicalTrials.gov Identifier: NCT03564184).