Association of uncoordinated sucking pattern with developmental outcome in premature infants: a retrospective analysis

Background Stress signals during sucking activity such as nasal flaring, head turning, and extraneous movements of the body have been attributed to incoordination of sucking, swallowing, and respiration (SSR) in premature infants. However, the association of uncoordinated sucking pattern with developmental outcomes has not yet been investigated. The aim of this study was to investigate whether uncoordinated sucking pattern during bottle-feeding in premature infants is associated with the developmental outcomes at 8–12 and 18–24 months of age (corrected for prematurity). Methods We retrospectively reviewed the medical records and video recordings for the Neonatal Oral-Motor Assessment Scale (NOMAS) of premature infants and divided them into two groups based on the presence or absence of incoordination. The Bayley-III cognition composite scores of the incoordination-positive and incoordination-negative group were compared at 8–12 and 18–24 months of age. Results Seventy premature infants exhibited a disorganized sucking pattern according to the NOMAS. The average Bayley-III cognition composite scores at 8–12 months of age were 92.5 ± 15.6 and 103.0 ± 11.3 for the incoordination-positive (n = 22) and incoordination-negative groups (n = 48), respectively (p = 0.002). The average Bayley-III cognition composite scores at 18–24 months were 90.0 ± 17.9 and 100.7 ± 11.5 for the incoordination-positive (n = 21) and incoordination-negative groups (n = 46), respectively (p = 0.005). A multiple linear regression analysis indicated that the presence of uncoordinated sucking pattern, grade 3 or 4 germinal matrix hemorrhage–intraventricular hemorrhage, and moderate to severe bronchopulmonary dysplasia were independently associated with cognitive development at 18–24 months of age. Conclusions Uncoordinated sucking pattern in premature infants was independently associated with a higher risk of abnormal developmental outcome in the cognitive domain of the Bayley-III at both 8–12 and 18–24 months. There may be a need for periodic follow-up and early intervention for developmental delay when incoordination of SSR that results in stress signals on the NOMAS is observed before 40 weeks postmenstrual age.


Background
Feeding skills in a newborn requires complex functions that are directed by the central nervous system (CNS), including the coordination of sucking, swallowing, and respiration (SSR) [1][2][3][4]. Preterm infants often have difficulty in coordinating SSR [5][6][7][8]. It is reported that incoordination of SSR is associated with a longer transition time to full oral feeding in premature infants with tube feeding [9]. Additionally, lack of muscle strength and/or endurance could precipitate the inability to sustain sucking in premature infants [10].
Two hypotheses exist regarding the difficulty in sucking encountered by neonates: (i) CNS immaturity that is selflimiting as the child ages and (ii) neurologic dysfunction due to disruption of the CNS [1,3,[11][12][13][14]. Distinguishing between these two potential causes of sucking difficulty is important to adequately counsel parents and commence early intervention to address developmental issues.
The Neonatal Oral-Motor Assessment Scale (NOMAS) is a noninvasive assessment tool to distinguish a dysfunctional or disorganized sucking pattern from a normal sucking in clinical practice [15,16]. A dysfunctional sucking pattern in NOMAS is defined as abnormal jaw and tongue movements that result in the interruption of feeding, such as a flaccid tongue, excessive excursion of the jaw, or no movement. It is rarely observed [17] but is frequently accompanied by congenital anomalies, severe brain lesions, and neuromuscular disease [14,18]. Palmer, the developer of the NOMAS system, suggested that a dysfunctional sucking pattern is strongly associated with neurological dysfunction, rather than CNS immaturity [18]. However, despite the high prevalence of a disorganized sucking pattern on NOMAS among premature infants [19], its association with developmental outcome is still unclear.
Previous studies have suggested that some items belonging to the disorganized sucking patterns could be more relevant to neurologic dysfunction [19,20]. Recently, the Dutch NOMAS working group categorized disorganized sucking pattern in the original NOMAS into three clusters according to the presence of an arrhythmical sucking pattern, inability to sustain sucking, and incoordination of SSR, which results in stress signals [21]. However, each cluster does not consider the severity of sucking difficulty.
Palmer et al. [15] considered "incoordination of SSR that results in stress signals" in NOMAS to be reflected by nasal flaring, head turning, and extraneous movements of the body or limbs during sucking activity. The Dutch NOMAS working group added several symptoms, such as choking, coughing, gagging, yelping, and grunting, to the findings of stress signals. However, whether these symptoms are more relevant than other symptoms of sucking difficulty to developmental outcome is not yet known.
The aim of this study was to investigate whether uncoordinated sucking pattern during bottle feeding in premature infants is associated with the developmental outcomes at 8-12 and 18-24 months of age (corrected for prematurity).

Subjects
We performed a retrospective review of the medical records of premature infants born between January 2014 and December 2016 at Seoul National University Hospital (SNUH) and referred for consultation to the Division of Pediatric Rehabilitation for sucking difficulty. The results of the NOMAS evaluation were incorporated into the electronic medical records, which were analyzed retrospectively by the authors. We included those infants who were born very premature (< 32 weeks) or very low birth weight (BW) (< 1500 g), assessed as having a disorganized sucking pattern in the NOMAS before 40 weeks postmenstrual age (PMA), and evaluated using the cognitive domain of the Bayley Scale of Infant Development, third edition (Bayley-III) [22], at 8-12 months of age (corrected for prematurity). Premature infants with the absence of Bayley-III at 8-12 months of age or the evaluated NOMAS after 40 weeks PMA were excluded. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Ethical approval was obtained from the SNUH Institutional Review Board (no. 1711-128-901).

NOMAS evaluation
All infants born earlier than 32 weeks gestational age (GA) or less than 1500 g with feeding difficulty had been referred to the Division of Pediatric Rehabilitation, and an occupational therapist evaluated their NOMAS score within 72 h from commencement of feeding in the hospital where this study was conducted. The evaluation was performed by analyzing the video recording of sucking activity [4]. Video recording included a close-up lateral view of the mouth, jaw, and neck. Only bottle-feeding was analyzed. The video recording started before the lip reached the bottle nipple and was stopped after recording for more than 2 min since the initiation of the oral feeding. Bottlefeeding was performed by a neonatal intensive care unit (NICU) nurse during the video recordings. Sucking difficulties were categorized into normal, disorganized, and dysfunctional sucking pattern according to the classification of the original NOMAS version [15].

Conversion to NOMAS cluster system
For this study, the disorganized sucking pattern was divided into three clusters (cluster 2, 3, or 4) according to the cluster system presented by the Dutch group, as shown in Table 1 [21]. The existing NOMAS video recordings were re-evaluated and reclassified according to the cluster system by a rehabilitation doctor (one of the coauthors of this paper) who has been certified by Marjorie Meyer Palmer, the original developer of the NOMAS. The occupational therapist, who had more than 5 years of experience in rehabilitative interventions for swallowing function in infants at the NICU, also blindly evaluated the NOMAS cluster using the same video recordings. The first evaluator (rehabilitation physician) was blinded to clinical factors, but not to diagnosis, GA, and PMA. The second evaluator (occupational therapist) was blinded to all clinical factors.

Medical records acquisition
Premature infants included in this study were evaluated using the Bayley Scales of Infant Development, third edition (Bayley-III), cognition composite score at 8-12 and 18-24 months of age (corrected for prematurity). In addition to 18-24 months of corrected age, 8 to 12 months corrected age was chosen because it is known to be a good time to identify the suspicion or presence of cerebral palsy or other neurologic abnormality [23]. It is also known to be an appropriate time for the first developmental assessment to be performed, usually the Bayley Scales of Infant Development, because the children show little stranger anxiety at this age and most are cooperative [23].
Through a retrospective electrical medical record review, GA, sex, brain sonographic finding (such as germinal matrix hemorrhage-intraventricular hemorrhage [GMH-IVH], PMA at initiation of oral feeding (defined as oral consumption of at least 5 mL of milk), PMA at the time of NOMAS evaluation, small for gestational age (SGA) or appropriate for gestational age (AGA), severity of bronchopulmonary dysplasia (BPD), birth weight (BW), development of sepsis, and 5 min Apgar score were obtained.

Statistical analysis
Cohen's kappa was obtained for the interrater reliability of the two evaluators. The cognitive domain of Bayley-III was compared between the incoordination-positive (cluster 4) and incoordination-negative groups (clusters 2 and 3) at 8-12 and 18-24 months of age using the independent t-test. Pearson's correlation coefficient r was calculated between continuous variables (GA and BW) and cognition composite score at 18-24 months. Multiple linear regression was performed with the Bayley-III cognition composite score at 18-24 months as the dependent variable. Analyses were performed using IBM SPSS version 20 software (IBM Corp., Armonk, NY, USA). A probability value < 0.05 was considered significant in all analyses.

Subjects
A total of 71 premature infants had both a Bayley-III cognitive domain and NOMAS score. One premature infant belonging to cluster 5, who had a grade 3 GMH-IVH and was a cardiopulmonary resuscitation survivor, showed delayed cognitive development at 8-12 months of age and was excluded from the analysis since the infant did not belong to the disorganized sucking pattern in the NOMAS cluster system. Clinical characteristics of the premature infants included in the analysis (n = 70) are shown in Table 2. Among the premature infants with a disorganized sucking pattern in NOMAS, there were four premature infants in cluster 2, 44 premature infants in cluster 3, and 22 premature infants in cluster 4.

Interrater reliability for the NOMAS cluster system
The two evaluators agreed on the cluster levels for 65 of 70 recordings (Cohen's kappa = 0.853). Disagreement occurred between clusters 2 and 3 (1 infants) and between clusters 3 and 4 (4 infants). The reliability of the two evaluators on the presence of incoordination items (cluster 4 vs. clusters 2 or 3) was higher with a Cohen's kappa value of 0.864, with four disagreements being documented.
The "unable to sustain" suckle pattern includes the following:  (Table 4).

Discussion
Premature infants frequently experience difficulty in sucking, thereby delaying the transition from tube to full oral feeding [24][25][26]. According to the NOMAS cluster system, sucking difficulty can be classified as arrhythmical sucking pattern, inability to sustain sucking pattern, incoordination of SSR sucking pattern, and dysfunctional sucking pattern (Table 1). Although the dysfunctional sucking pattern is known to be associated with poor developmental outcome, the relationship with developmental outcome is not well known in other patterns. The results of this study suggest that incoordination of SSR sucking pattern could be more associated with the development in the cognitive domain both at 8-12 and 18-24 months than arrhythmical sucking or inability to sustain pattern. Most of the preterm infants with an immature sucking pattern can successfully bottle-feed as the SSR matures, and the sucking pattern of term infants is characterized by the rhythmic alternation of suction and expression/ compression [11]. In the arrhythmical or inability to sustain pattern group, it could be assumed that SSR coordination has already been formed in preterm period, but sucking difficulty has occurred due to lack of oromotor strength and endurance [11], and this group showed a faster development than the group with less developed SSR coordination.
Previous studies have attempted to predict developmental delay through neonatal oromotor function. In a study of 27 premature infants without brain lesions, the risk of developmental delay increased when the premature infants exhibited a disorganized sucking pattern at 37 weeks PMA [20]. However, this study only examined the presence of disorganized sucking pattern and did not distinguish between specific patterns among disorganized sucking pattern.
The results of this study are compatible with those of the study by Nieuwenhuis et al. [19], who classified the disorganized sucking patterns in the NOMAS into two categories: disorganized due to arrhythmic sucking and disorganized due to lack of coordination of SSR. They concluded that uncoordinated sucking patterns, but not arrhythmic sucking patterns, were associated with abnormal fidgety movement at 14 weeks post-term [19]. However, in the previous study, they did not discern dysfunctional sucking pattern from incoordination sucking pattern in the NOMAS. Additionally, the study did not distinguish the inability to sustain sucking pattern separately.
Regarding brain ultrasound finding, GMH-IVH grades 3 and 4 are widely considered predictors of developmental delay, whereas the implications of GMH-IVH grades 1 and 2 remains controversial [27,28]. In concert with previous studies, GMH-IVH grades 3 and 4 were identified as predictors of the cognition composite score in a multiple linear analysis in the present study. In the present study, BPD was also analyzed as a statistically significant predictor of developmental delay. Mizuno and colleagues reported that infants with BPD demonstrated not only poorer feeding coordination, but also poorer feeding endurance and performance [29], which might have affected development [30].
In the NOMAS, stress signals, including nasal flaring, head turning, and extraneous movement, are regarded as the symptoms of incoordination of SSR (Table 1), as described by Palmer et al. [16,18]. However, the relationship between incoordination of SSR and stress signals was suggested through clinical observation and not by direct measurement of SSR. To demonstrate the incoordination of SSR, recordings of intraoral pressure (rhythmic alternation of suction and expression/compression) [11,24], pharyngeal pressure [24,25], nasal thermistor flow [26,30], and thoracoabdominal plethysmography [24] have been used, but these methods are not widely used, particularly in the clinical settings, due to their complexity and invasiveness. However, in terms of research, the relationship between SSR incoordination and clinically observed stress symptoms could be investigated by directly measuring SSR. There might be certain stress signals that are more relevant to SSR incoordination than other signals.
There are a few limitations to this study. First, we did not include more objective signs that could be assessed  in premature infants with sucking difficulty, including episodes of desaturation, apnea, and bradycardia. Those measurable signs might complement NOMAS, which is composed of observational findings. Second, because this study was conducted retrospectively, we only evaluated the sucking pattern before 40 weeks PMA since most premature infants are discharged from the NICU before term age. In our study, clusters 2 and 3 (i.e., arrhythmia and inability to sustain sucking without stress signals) were less relevant to neurodevelopmental outcomes. However, if those symptoms persist during the postterm period, the clinical relevance could be changed. For example, Wolthuis-Stigter et al. [10] reported that the inability to sustain sucking at 46 weeks significantly increased the odds of abnormal neurodevelopmental outcomes at 2 years of age. Therefore, for future studies, a prospective study design would be necessary for improving our understanding of the clinical significance of sucking difficulties that are observed in premature infants.

Conclusion
Uncoordinated sucking pattern that results in stress signals (e.g., head bobbing, extraneous movements of the body or limbs during sucking, choking, gagging, coughing, yelping, and grunting) in premature infants was independently associated with a higher risk of abnormal developmental outcome in the cognitive domain both at 8-12 and 18-24 months. There might be a need for periodic follow-up and early intervention for developmental delay when incoordination of SSR that results in stress signals in NOMAS is observed before 40 weeks PMA.