Modulation of Pharyngoesophageal Motility Characteristics During Concurrent Oral Feeding in Infants: Novel Pilot Observations

Abstract

Effect of interventions or diseases on the mechanisms of swallowing difficulties, and indications to change nipples, flows, or formula thickness in high-risk infants remains unclear. We evaluated and compared the effects of nipple flows and/or formula thickness on pharyngoesophageal timing and amplitude characteristics during oral milk challenge (OMC). A total of 48 OMC trials were undertaken in 12 infants [born at 36.4 ± 3.8 weeks gestation, studied at 48.0 ± 5.6 weeks postmenstrual age (PMA)] concurrent with high resolution pharyngoesophageal manometry (HRPM). We evaluated timing and amplitude of swallowing characteristics (pharyngeal and esophageal contractile strength, duration, activity, pharyngeal peak interval variability, bursts, and rhythm) and volume intakes. We compared within and between the entities: un-thickened formula vs. thickened formula and slower- vs. faster flow nipple. OMC durations averaged 71 ± 31 s and did not differ between interventions (p > 0.05). Pharyngeal swallow rhythms (timing) and contractile characteristics were dependent on the interaction between nipple flow rates and formula thickness (p < 0.05). Thickness decreased overall contractility and modified distal pharyngeal contractile activity (p < 0.05), which is critical to opening the upper esophageal sphincter. A structural equation model was developed that revealed the influence of PMA, respiratory status, volume extracted, nipple flow rate, and thickener presence on pharyngeal swallowing frequency with a comparative fit index of 0.83. Oropharyngeal stimulus-types trigger, integrate, and modify brainstem mediated oral and pharyngeal rhythms and modulate contractility and airway protection mechanisms. Effects of therapies, maturation and diagnoses are measurable using these novel approaches at crib-side.

Introduction

Independent oral eating is a complex process that matures during development. These skills involve the coordination of neuro-regulatory centers, pharyngoesophageal musculature and cardiorespiratory stability, which can be challenging among high-risk infants admitted to the neonatal ICU (NICU) [1–4]. Swallowing difficulties can result from prematurity, comorbidities, or inadequate oral experiences, thus disrupting the development of suck-swallow-peristalsis-breathe coordination [5, 6]. As new technological advances are being implemented in the NICU, the survival rates of these infants are increasing. As a result, the swallowing difficulties related to morbidities is increasing and is estimated to be between 25 and 80% [7–11]; therefore, the cost and burden of chronic feeding difficulties is continually increasing [2].

Precise diagnosis and characterization of mechanisms of swallowing difficulties remain unclear due to the lack of objective standards and consensus among evaluation methods. Video-fluoroscopy swallow studies (VFSS) and high-resolution pharyngoesophageal manometry (HRPM) are the testing modalities utilized in the evaluation of aerodigestive and swallowing difficulties. VFSS can assess the transit of the liquid bolus extracted from the bottle as it propels through the anatomic regions involved in swallowing until the upper esophageal sphincter (UES). VFSS assessments are limited by shorter duration of evaluation [12], the risks associated with radiation exposure [13–15], and the lack of application of consistent standardized practices for VFSS methods [16–18]. Furthermore, skilled personnel are necessary for VFSS as monitoring high-risk infants transported to the radiology suites is required [15, 19]. Moreover, high-risk infants have airway supports and assessment of swallowing skills can be challenging, as aerodigestive and cardiorespiratory comorbidity factors cannot be assessed with limited trials [12].

In contrast, HRPM measures the neuromuscular function, rhythms and relationships between the structures involved with oropharyngeal and pharyngeal-UES-esophageal phases of swallowing at the point-of-care [20–24]. Unlike VFSS, HRPM allows for sustained and prolonged evaluation of swallowing without any of the risks imposed by VFSS, while integration of cardiorespiratory measures can improve the evaluation of safety metrics and quality of data. However, structural anatomy cannot be visualized with HRPM, and skilled personnel are required to perform these studies as this is an emerging technological translation that can be cumbersome [6]. Given the lack of appropriate precision methods to assess swallowing in convalescing high-risk infants, commonly practiced non-evidence-based therapies include modifications to nipple flow rates or addition of thickeners, both of which may have unintended consequences (i.e., regression of oral feeding skill [25, 26], worsening of aerodigestive symptoms [25–27], sluggish gastrointestinal motility [28, 29], constipation from thickener use [30], growth and weight gain disturbances [31, 32], and long-term consequences [5, 28, 33]).

To expand upon our current understanding and diagnostic evaluation of infant swallowing difficulties in hopes of developing improved management strategies, HRPM was used to examine the effects of nipple flow rate and formula thickness on swallowing mechanisms during oral milk challenges (OMC). We hypothesized that changing the nipple flow rates or viscosity of the formula modifies the timing and amplitude characteristics of swallowing. More specifically, we hypothesized that the increased viscosity due to adding formula thickeners, will modify the oral bolus extraction and therefore the lingual peristalsis, making appropriate pharyngeal stimulation unlikely due to smaller bolus sizes or a delay in the bolus reaching the pharynx [2, 34]. Thus, the downstream pharyngeal motility would not be optimal or operational, leaving any thickened milk extracted pooling in the mouth, which would cause concern for safe airway management. In addition, deficiencies in oral lingual anatomical structures or their sensory-motor functions can minimize both extraction and bolus transport to the pharynx. Delayed or interrupted pharyngeal stimulation can lead to pharyngeal-UES-esophageal dysmotility mechanisms that may affect airway and cardiorespiratory functions acutely [35, 36].

Materials and Methods

Study Design, Participants and Setting

The study participants were a subset of infants that participated in a larger cohort study (Clinicaltrials.gov: NCT02583360) evaluating infants with swallowing difficulties [37]. Subjects were prospectively recruited, and studies were conducted at Nationwide Children’s Hospital, Columbus, OH with appropriate Institutional Review Board approval, compliance to Health Insurance Portability and Accountability Act guidelines, and written informed parental consent obtained. Data quality was secured with research electronic data capture tools (REDCap). Participants underwent pharyngoesophageal motility evaluation via HRPM to evaluate the effect of nipple flow rates (slower versus faster) and formula thickeners (un-thickened versus thickened) on oral feeding function. Inclusion criteria were: (a) VFSS referral for feeding-related aerodigestive symptoms with infant age < 60 weeks postmenstrual age (PMA), (b) on full enteral feeds with at least partial oral feeds, and (c) ≤ 1 L per minute of supplemental oxygen. Exclusion criteria were: (a) infants with known genetic, metabolic, or syndromic diagnoses, (b) severe neuropathology (≥ grade III intraventricular hemorrhage or moderate to severe perinatal hypoxic ischemic encephalopathy), (c) gastrointestinal malformations, (d) craniofacial malformations, (e) history of any surgical procedures, and (f) exclusively breastfeeding infants.

VFSS and HRPM Protocols

The methodological and analytical aspects of VFSS and HRPM protocols have been published by us [20–22, 37–39]. Briefly, a 6 Fr HRPM probe with 25 unidirectional pressure sensors (UniTip High-Resolution Catheter, Unisensor, USA) was attached to a portable HRPM system (Solar GI, Laborie Medical Technologies, Mississauga, ON, Canada). The HRPM probe was calibrated first, placed via nasogastric route and secured at the nares. After adaptation to probe placement, the monitored infant was then safely transported to the VFSS suite. At the VFSS suite, concurrent VFSS and HRPM recording began once the infant was settled on the VFSS table in the side-lying position or semi-reclined position via the Tumble Forms Feeder Seat Positioner (Patterson Medical, Illinois, USA) for bottle feeding. Infants were evaluated for the presence of penetration (2–5) or aspiration (6–8) using the Penetration-Aspiration Scale [37, 40] with different thicknesses of premixed liquid barium sulfate (Varibar^®, Bracco Diagnostics Inc, New Jersey, USA) and nipple flow rates based the patient’s clinical needs. VFSS was then performed in accordance with the established evaluation protocol and institutional interdisciplinary guidelines for quality and safety [17, 41]. VFSS imaging was conducted in lateral view during bottle feeding by an interdisciplinary team that consisted of a radiologist, an occupational therapist, and a speech language pathologist. To ensure accurate VFSS imaging, each test was recorded using three standardized metrics: field of view, magnification, and pulse repetition rate. Standardized collimation borders (lips anteriorly, inferior bony orbits superiorly, spinous processes posteriorly, cervical 5 and 6 vertebrae inferiorly) were used to establish a focused view of the oropharyngeal tract and to limit unnecessary radiation exposure [37, 42]. Magnification was performed with the As Low as Reasonably Achievable principle and did not exceed twice magnifications on a standard-three scale, if required during the study [17, 37, 42]. Imaging was captured at a rate of 30 frames per second [37, 43].

Following the VFSS observations, the infant and the study team returned to the procedure room to perform HRPM with the OMC trials. For the current report, our focus was to determine the changes in timing and amplitude characteristics of pharyngeal-and esophageal motility characteristics during OMC trials that were recorded during the HRPM. Thus, the goal for each infant was to offer four 1-minute OMCs that included two different bottle nipples (modified flow rate) and formula thickness (no thickener added, or 1.5 teaspoons/ounce of infant cereal added). The thickness of the formula used was nectar consistency [28] and was selected to simulate similar barium consistencies used during the VFSS. Infants were fed in a slightly upright or side-lying position with their current bottle (bottle and nipple) and with a bottle chosen by our study team. The study team increased or decreased the nipple flow rate based on the subject’s oral feeding performance and subject safety (respiratory status and if symptoms were noted). Once the bottle nipple and formula were chosen for each trial, the timer began when the infant latched on to the nipple. In sequential testing, each infant underwent the diagnostic testing protocol outlined above separately.

Data Analysis

Nipple flow rates were determined from prior published work from others by measuring the weight of the formula expressed in 1 min by each nipple attached to a breast pump and multiplying by the density of the formula (mL/g) [44–46]. We assigned each bottle used in the OMC trials with the nipple flow rate as previously calculated by others and categorized the bottle nipples into slower and faster categories. For the two bottle nipple systems that were provided to each infant, the bottle nipple that had the lower flow rate was classified as being in the slower group, and the other bottle nipple (which had the higher flow rate) was classified as being in the faster group. For example, if the flow rates for the two bottle nipple systems were 2 mL/min and 8 mL/min, then the bottle nipple with a flow rate of 2 mL/min was placed in the slower group and the bottle with the 8 mL/min nipple flow rate was placed in the faster group. Nipple flow rates that were unknown or indistinguishable were excluded from the data analysis. To permit rigorous comparison, infants with less than 4 OMC trials or OMC with poor documentation were excluded.

Automated algorithms using Python (v. 3.7.5) were used as previously published to detect pharyngoesophageal contraction peaks and durations [22, 23, 47] as shown in Fig. 1. Pharyngeal swallows during the OMC were classified as a solitary swallow indicating that the pharyngeal peak-to-peak interval duration was > 2 s or a swallow burst with the pharyngeal peak-to-peak interval duration ≤ 2 s [47]. Swallow function during OMC was measured by the following pharyngoesophageal timing and amplitude characteristics. Timing characteristics were: (a) pharyngeal contractile activity, percent (%), defined as the sum of all pharyngeal contraction durations divided by the total duration of the OMC and multiplied by 100 [47], (b) pharyngeal contractile frequency, Hz, defined as the number of pharyngeal contractions per second [47], (c) pharyngeal peak interval variability, seconds, as the standard deviation of pharyngeal peak-to-peak interval durations [39, 47], and (d) pharyngeal peak interval stability, number (#), calculated as the standard deviation of pharyngeal peak-to-peak intervals divided by the average pharyngeal peak-to-peak interval duration [39, 47].

Fig. 1.

Data analysis and comparison of HRPM characteristics during OMC with un-thickened and thickened formula PX-pharyngeal, UES-upper esophageal sphincter, ESO-esophageal, LES-lower esophageal sphincter, STO-stomach. P1-P6 represent 6 pharyngeal contractions. Orange boxes represent a swallow burst (Ex. P1-P4). A burst was defined as consecutive pharyngeal peaks within 2 s (D1) of each other. The number of pharyngeal swallows per burst was then defined as the average number of swallows (pharyngeal peaks) between all bursts. White boxes represent solitary swallows (Ex. P5 and P6). A solitary swallow was defined as pharyngeal peaks > 2 s (D2) apart. Note, that the number of pharyngeal swallows is decreased, and the pharyngeal contractile integral is increased with thickened formula.

Pharyngoesophageal amplitude characteristics were measured as (a) contractile integral, mmHg*cm*seconds (pressure amplitude*length*contraction duration), which is a measure of contractile strength, and was calculated by integrating the pressure amplitude over the length and duration of the contraction of the anatomical region [22, 37] and (b) contractile duration, seconds, as the onset of the contraction to the offset of the contraction [22, 37].

Statistical Analysis

Descriptive analysis for the demographic and clinical characteristics were reported as number (N) and percent (%) for categorical variables and Mean ± SD for continuous variables. We categorized the nipple flow rates into faster and slower groups based on previously calculated values in in vitro experiments [44–46]. Effect sizes were calculated using Cohen’s d measurement across thickened and un-thickened formulas and within each level of nipple flow categorization (all nipple flow rates combined, slower, and faster). Effect sizes were considered small (d=0.2), medium (d = 0.5), and large (d ≥ 0.8) [48, 49].

To test how the infants will respond to the change in nipple flow rate, regardless of the absolute value and magnitude differences between the two nipples, and formula thickness on different pharyngoesophageal timing and amplitude characteristics, we classified the nipples as slower or faster flow. Linear mixed effect models or generalized estimating equations were used when the distribution of the outcome was severely skewed from normality (skewness > 5). Linear mixed-effect models were used with a random intercept to account for dependency among observations since we had four repeated trials: two trials using the faster flow nipples with/without adding thickener, and two trials using the slower flow nipples with/without adding thickener. In addition, we used compound symmetry to account for the correlation between the observations from the same trial for each infant. All analyses were adjusted (included as a fixed effect) for formula thickness, nipple flow rates, the interaction between formula thickness and nipple flow rate, and the volume of formula extracted per minute determined by dividing the total volume extracted during the OMC by the total duration of the OMC. Bonferroni adjustment was used for post-hoc comparison. The Bonferroni adjustment for p-values was calculated by multiplying each p-value by the number of tests. Adjusted p-values that exceed 1 are reduced to 1, which allows us to compare the adjusted p-value to α = 0.05.

The previous analysis did not account for the differences in the bottle nipples (i.e. different nipple flow rates) as we assumed that all nipple flow rates within their slower or faster category were similar, or the overlap of the same nipple flow rate being placed into both groups. In addition, the magnitude change between the slower and faster nipples across the infants were different which could have also influenced the response. Therefore, the previous analysis was repeated using the reported flow rate of the nipple as a continuous variable to enable us to test whether the relationship between the reported flow rate and the pharyngoesophageal timing and amplitude characteristics are dependent on the formula thickness.

We also applied structural equation modeling [50] to examine the hypothesized sequential relationships among the infant’s characteristics or environmental factors such as infant’s PMA, nipple flow rate, or fluid thickness on the frequency of pharyngeal peaks (number of pharyngeal peaks per second). The frequency of pharyngeal peaks was chosen as the outcome metric because we previously found that infants with a robust pharyngoesophageal pattern (determined as an average solitary swallow per minute of 2.1 ± 1.1 and average number of pharyngeal swallows in burst per minute as 28.0 ± 6.8) during oral feeding had a higher rate of oral intake and increased prevalence of independent oral feeding at discharge [47]. The characteristics that we chose are clinically relevant, influencing factors, and included infant’s PMA at evaluation (maturation), the presence of respiratory support (airway status), volume extracted from the bottle nipple (oral efforts), nipple flow rate (bottle nipple systems), and the amount of thickener added to the formula (consistency/viscosity). Structural equation modeling is a statistical methodology that examines the potential causal relationships between one or more independent or dependent variables simultaneously. Independent and dependent variables can be classified as measured (observed, denoted by rectangular boxes) or latent (not directly observed or measured, denoted by circles). The relationships (estimated effect) between variables are represented by single or double-headed arrows in which the direction of the arrows determines which variable is the influencer (hypothesized direct relationship). If a double-curved headed arrow is present, it shows a covariance between the variables [50]. We reported the standardized coefficients for the model with the largest absolute value representing the strongest effect. To evaluate the fit of the model, the root means square error of approximation and comparative fit index (CFI) were calculated. CFI was used to examine the discrepancy between our model and baseline model in which all observed variables are uncorrelated [51]. All analyses were conducted using SAS, version 9.4, and statistical significance was determined at α = 0.05.

Results

Participant and OMC Characteristics

A total of 48 OMC trials were performed in 12 infants (10 male; born at 36.4 ± 3.8 weeks gestation) that underwent pharyngoesophageal motility testing at 48.0 ± 5.6 weeks PMA. OMCs averaged 71 ± 31 s per trial and did not differ between interventions (p > 0.05). The average flow rate for: all nipples was 13.7 ± 7.4 mL/min, the nipples categorized as slower flow was 8.7 ± 4.0 mL/min, and the nipples categorized as faster flow was 18.6 ± 6.8 mL/min. The difference between the slower and faster nipple flow rates was 9.9 ± 5.6 mL/min. The average pharyngeal length measured by HRPM was 3.4 ± 0.5 cm.

Clinical characteristics of participants were diverse (Table 1), in that: (a) average weight was 2.7 ± 0.8 kg at birth and 4.7 ± 1.0 kg at evaluation, (b) 33% were born premature, (c) 33% had chronic lung disease, (d) 75% had some oral feeding at evaluation, (e) 75% were breathing room air at evaluation, (f) 92% infants were exclusively oral fed at discharge, and (g) 83% were breathing room air at discharge. In addition, each infant underwent a VFSS study in which 3 (25%) had no evidence of airway entry, 5 (42%) had the presence of penetration, and 4 (33%) had the presence of aspiration. The overall length of hospital stays averaged 85 ± 70 days and ranged from 41 to 188 days.

Table 1.

VFSS results

Subject Number	Infant Characteristics	Reason for VFSS	Trial Number	Nipple Used	Viscosity	Laryngeal Penetration	Aspiration	†Pooling	*Feeding Method
									Time of Study	After 4 weeks
1	Preterm, BPD	choking/gagging w/feeds	1	Dr. Brown Premie	Thin	Y	Y	N	*Tube	PO
			2	Dr. Brown L3	Nectar	Y	Y	N
			3	Dr. Brown L4	Honey	N	N	N
2	Preterm, BPD, IVH	choking/gagging w/feeds	1	Dr. Brown UP	Thin	Y	Y	N	Tube	Tube
			2	Dr. Brown L3	Nectar	Y	N	N
3	Term, O2 need	persistent O2 need	1	Dr. Brown L1	Thin	Y	N	N	PO	PO
4	Term	minimal oral feeds, weight loss	1	Medela L1	Thin	Y	Y	N	PO	PO
			2	Dr. Brown Premie	Thin	Y	Y	N
			3	Medela Calma	Nectar	N	N	N
5	Term, SGA	stridor/choking/ gagging w/feeds	1	Tommee Tippee L1	Thin	N	Y	N	PO	PO
			2	Tommee Tippee L0	Thin	N	Y	N
			3	Tommee Tippee L2	Nectar	N	Y	N
			4	Tommee Tippee L3	Nectar	Y	N	N
6	Term	choking/gagging w/feeds	1	Parent Choice Premie	Thin	Y	N	N	PO	PO
			2	Parent Choice Slow Flow	Thin	Y	N	N
7	Term, Laryngomalacia	coughing/choking w/feeds	1	Avent L0	Thin	N	N	Y	PO	PO
			2	Avent L1	Nectar	N	N	Y
			3	Avent L0	Thin	Y	N	N
8	Preterm, IVH, OSA, O2 need	stridor w/feeds, persistent O2 need	1	Dr. Brown Premie	Thin	Y	N	N	Tube	PO
			2	Dr. Brown L1	Thin	N	N	N
9	Term	coughing w/feeds	1	Dr. Brown L1	Thin	Y	N	N	PO	PO
10	Term	coughing/choking w/feeds	1	Nuk Slow Flow	Thin	N	N	N	PO	PO
11	Preterm	coughing/choking w/feeds	1	Avent L1	Thin	N	N	N	PO	PO
			2	Syringe	Thin	N	N	N
12	Term	coughing/choking w/feeds	1	Dr. Brown L1	Thin	N	N	N	PO	PO

BPD- bronchopulmonary dysplasia, IVH- intraventricular hemorrhage, OSA- obstructive sleep apnea, PO- feeding by mouth only, SGA- small for gestational age,

^*Tube-partially oral fed.

^†No pooling was noted in any anatomical region in those subjects marked as, “N”, whereas pooling was noted for subject 7 in the valleculae and pyriform sinuses in both trials. None of the subjects had nasopharyngeal reflux or delayed swallow noted on their VFSS.

Timing and Amplitude of Pharyngoesophageal Motility Characteristics Compared among Slower and Faster Nipple Flow Rates

Comparisons were made between the nipples categorized as faster flow and slower flow for both formula thickness levels (un-thickened vs. thickened) for pharyngoesophageal timing and amplitude characteristics during OMC trials as shown in Table 2. Infants had decreased pharyngeal contractile activity and consumed less milk with the thickened formula regardless of the nipple flow rate group (p < 0.01). With the faster flow nipples, infants had increased pharyngeal contractile integrals and peak to peak durations while the pharyngeal contractile frequency decreased (p < 0.01). Similarities were noted within both, in un-thickened and in thickened formulas regardless of the type of nipple used (i.e., slower flow with un-thickened formula vs. slower flow with thickened formula, p > 0.05). In addition, we found that the effect of formula thickness on pharyngeal peak interval variability and number of pharyngeal swallows per burst depends on the nipple flow rate level (Fig. 2). Overall, pharyngeal peak interval variability increased significantly with thickened formula, compared to un-thickened formula; this increase is greater with the faster flow nipple (Fig. 2A). While the number of pharyngeal swallows per burst (Fig. 2B) decreased significantly with thickened formula, compared to un-thickened formula; this reduction is greater with the faster flow nipple. This indicates that both the nipple flow rate and formula thickness need to be considered when selecting an appropriate bottle.

Table 2.

Effect of nipple flow rate category (Slower or Faster) and formula thickness on pharyngeal swallowing function during oral milk challenge

Swallow Function	Slower Flow			Faster Flow
	Un-Thickened Formula	Thickened Formula	p-value	Un-Thickened Formula	Thickened Formula	p-value
Milk Intake Rate, mL/min	8 ± 2	4 ± 2	0.01	10 ± 1	1 ± 1	< 0.01
Volume Consumed, mL	9 ± 2	5 ± 2	< 0.01	11 ± 1	1 ± 1	< 0.01
Timing
Pharyngeal
Contractile Activity, %	37 ± 7	12 ± 6	< 0.01	49 ± 9	5 ± 4	< 0.01
Peak to Peak Duration, s	3.3 ± 0.9	7.2 ± 2.2	0.42	3.0 ± 0.7	12.4 ± 2.7	< 0.01
Contractile Frequency, Hz	0.7 ± 0.1	0.4 ± 0.1	0.17	0.7 ± 0.1	0.1 ± 0.1	< 0.01
Peak Interval Stability, number	0.7 ± 0.1	0.6 ± 0.1	1.00	0.6 ± 0.1	0.6 ± 0.1	1.00
Esophageal
Contractile Activity, %	27 ± 6	20 ± 7	1.00	28 ± 4	25 ± 8	1.00
Amplitude
Pharyngeal
Overall Amplitude, mmHg	47 ± 8	51 ± 8	1.00	47 ± 8	64 ± 11	0.66
Proximal Amplitude, mmHg	31 ± 6	32 ± 6	1.00	24 ± 6	44 ± 11	0.51
Distal Amplitude, mmHg	18 ± 2	18 ± 3	1.00	20 ± 2	20 ± 3	1.00
Contractile Duration, s	0.5 ± 0.1	0.7 ± 0.3	1.00	0.5 ± 0.1	0.9 ± 0.4	1.00
CI, mmHg.cm.s	60 ± 8	70 ± 11	0.21	58 ± 6	92 ± 10	< 0.01
Esophageal
Contractile Duration, s	3.3 ± 0.6	3.3 ± 1.5	1.00	2.7 ± 0.5	5.2 ± 1.4	0.46
Distal CI, mmHg.cm.s	295 ± 63	347 ± 101	1.00	296 ± 48	448 ± 103	1.00

CI- contractile integral (a measure of contractile strength). Data presented as LSMean ± SE. Note, pharyngo-esophageal amplitude characteristics were not impacted by nipple flow rate group.

Fig. 2.

Effect of formula thickness on pharyngeal timing characteristics with slower and faster bottle nipples categories Px-pharyngeal. Data plotted as mean and standard errors. *P < 0.05 Thickened vs. Un-thickened formula. The comparison between slower and faster nipple flow rate groups within each formula thickness category is represented by the vertical brackets and p-value in the figure. (a) Pharyngeal peak interval variability was defined as the standard deviation between pharyngeal swallows (in seconds). Note, the pharyngeal peak interval variability between pharyngeal contractions is increased with thickened formula regardless of the nipple used. (b) A burst was defined as consecutive pharyngeal peaks within 2 s of each other. The number of pharyngeal swallows per burst was then defined as the average number of swallows (pharyngeal peaks) between all bursts. Infants with a greater number of pharyngeal swallows per burst have a higher success of independent oral feeding [47]. In this study, the number of pharyngeal swallows per burst is greater with un-thickened formula. Also, within un-thickened formula the faster flow nipple has a significantly higher number of pharyngeal swallows per burst than the slower nipple.

Timing and Amplitude of Pharyngoesophageal Motility Characteristics Compared among Un-thickened and Thickened Formula

Comparisons between un-thickened and thickened formula were made for pharyngoesophageal timing and amplitude characteristics while the using the nipple flow rate as a continuous variable (Table 3). The effect sizes for each pharyngoesophageal swallowing characteristic have been reported for thickened formula vs. un-thickened formula under different nipple flow rates in Table 3. The number of pharyngeal swallows per burst, pharyngeal contractile activity, and pharyngeal contractile frequency was significantly decreased with the thickened formula (p < 0.01). The pharyngeal peak to peak duration, pharyngeal peak interval variability, pharyngeal proximal amplitude, pharyngeal contractile duration, and pharyngeal contractile integral (CI) increased with the thickened formula (all, p < 0.05). The relationship (slope) between the nipple flow rate and the different pharyngoesophageal swallowing characteristics with thickened and un-thickened formulas are shown in Table 4. The pharyngeal peak to peak duration and pharyngeal peak interval variability increased with thickened formula as the nipple flow rate increased (positive slope, p < 0.01). The pharyngeal contractile frequency decreased with thickened formula as the nipple flow rate increased (negative slope, p = 0.04). We plotted the relationship between the nipple flow rate (increasing the nipple flow rate by 1 mL/min) and the following variables: overall pharyngeal activity, distal pharyngeal activity, peak to peak duration, and pharyngeal peak interval variability for un-thickened and thickened formulas (Fig. 3). Note, that the pharyngoesophageal timing characteristics are less coordinated with thickened formula as the nipple flow rate increases.

Table 3.

Comparison of Un-Thickened vs. Thickened formula on (A) swallowing function characteristics and (B) effect size compared between nipple flow rates

Swallow Function	A. Comparisons of Means			B. Effect Size
	Un-Thickened Formula	Thickened Formula	p-value	All Flows	Faster Flow	Slower Flow
Timing
Pharyngeal Swallows per Bursts, number/min	26.0 ± 5.9	4.2 ± 3.0	< 0.01	−1.3	−2.1	−0.8
Solitary Swallows, number/min	2.7 ± 0.8	1.8 ± 0.3	0.36	−0.4	−0.5	−0.1
Pharyngeal
Contractile Activity, %	42.8 ± 7.5	8.6 ± 4.1	< 0.01	−2.1	−2.5	−1.3
Peak to Peak Duration, s	3.5 ± 1.1	9.1 ± 2.9	< 0.01	1.7	2.1	1.4
Contractile Frequency, Hz	0.7 ± 0.1	0.3 ± 0.1	< 0.01	−1.3	−2.5	−1.0
Peak Interval Variability, s	2.4 ± 0.3	6.3 ± 0.6	< 0.01	1.1	2.0	0.6
Peak Interval Stability, number	0.6 ± 0.1	0.6 ± 0.1	0.91	−0.1	0.2	−0.1
Esophageal
Contractile Activity, %	28.2 ± 4.8	22.7 ± 5.8	0.55	−0.5	−0.4	−0.4
Amplitude
Pharyngeal
Overall Amplitude, mmHg	47 ± 8	59 ± 8	0.05	0.1	−0.2	0.1
Proximal Amplitude, mmHg	27 ± 6	39 ± 8	0.05	0.3	0.3	0.2
Distal Amplitude, mmHg	19 ± 2	19 ± 3	0.68	−0.3	−0.4	−0.3
Contractile Duration, s	0.5 ± 0.1	1.0 ± 0.4	0.01	0.8	0.8	0.6
CI, mmHg.cm.s	58.8 ± 6.6	83.3 ± 10.1	< 0.01	0.6	0.8	0.4
Esophageal
Contractile Duration, s	3.0 ± 0.5	4.2 ± 1.1	0.34	0.4	0.6	0.4
Distal CI, mmHg.cm.s	293.6 ± 42.8	378.2 ± 83.8	0.53	0.3	0.4	0.4

CI- contractile integral (a measure of contractile strength). Data presented as Estimate ± SE or LSMean ± SE. Note, that the peak-to-peak duration is increased with thickened formula. The effect size was determined using Cohen’s d test for each variable for all nipple flow rates combined (thickened vs. un-thickened), within the faster flow nipple category (thickened vs. un-thickened), and within the slower flow nipple category (thickened vs. un-thickened). Cohen’s d test classifies effect sizes as small (d = 0.2), medium (d = 0.5), and large (d ≥ 0.8, in bold text) [48, 49]

Table 4.

Relationship (i.e. Slope) between the nipple flow rate and swallow function characteristics

Swallow Function	Slope				Difference between Slopes for UF vs. TF
	Un-Thickened Formula	p-value	Thickened Formula	p-value
					Difference of Differences	p-value
OMC Duration	−0.02 ± 0.88	0.82	−0.31 ± 0.32	0.66	0.28 ± 0.66	0.86
Timing
Swallows within Bursts, number/min	0.44 ± 0.42	0.29	−0.19 ± 0.40	0.64	0.63 ± 0.58	0.28
Solitary Swallows, number/min	0.02 ± 0.15	0.87	0.00 ± 0.04	0.97	0.02 ± 0.14	0.88
Pharyngeal
Contractile Activity, %	1.56 ± 0.68	0.04	−0.71 ± 0.76	0.69	2.28 ± 1.01	0.02
Peak to Peak Duration, s	−0.02 ± 0.02	0.21	0.03 ± 0.01	< 0.01	−0.05 ± 0.01	< 0.01
Contractile Frequency, Hz	0.00 ± 0.00	0.87	−0.03 ± 0.01	0.04	0.03 ± 0.01	0.05
Peak Interval Variability, s	−0.06 ± 0.06	0.32	0.26 ± 0.06	< 0.01	−0.32 ± 0.07	< 0.01
Peak Interval Stability, number	0.00 ± 0.01	0.89	−0.01 ± 0.00	0.18	0.00 ± 0.01	0.66
Esophageal
Contractile Activity, %	0.76 ± 0.55	0.11	−0.24 ± 0.63	0.18	1.00 ± 1.02	0.29
Amplitude
Pharyngeal
Overall Amplitude, mmHg	−0.29 ± 0.29	0.31	0.68 ± 0.32	0.04	−1.95 ± 0.79	0.01
Proximal Amplitude, mmHg	−0.86 ± 0.30	< 0.01	0.22 ± 0.31	0.48	−2.16 ± 1.05	0.04
Distal Amplitude, mmHg	0.20 ± 0.16	0.20	0.13 ± 0.14	0.34	0.14 ± 0.23	0.55
Contractile Duration, s	0.00 ± 0.01	0.79	0.01 ± 0.02	0.48	−0.01 ± 0.03	0.58
CI, mmHg.cm.s	0.00 ± 0.01	0.96	0.03 ± 0.01	< 0.01	−0.03 ± 0.01	< 0.01
Esophageal
Contractile Duration, s	0.04 ± 0.08	0.57	−0.02 ± 0.15	0.88	0.06 ± 0.21	0.76
Distal CI, mmHg.cm.s	0.1 ± 3.3	0.88	0.3 ± 9.4	0.46	−0.2 ± 11.4	0.71

UF-Un-Thickened Formula, TF- Thickened Formula, CI- contractile integral (a measure of contractile vigor), OMC-Oral Milk Challenge. Data presented as Estimate ± SE or LSMean ± SE. The slope in this table is defined as the relationship between the swallow function characteristic as the nipple flow rate increases by a unit of 1 mL/min. Interpretation: The pharyngeal contractile activity (percent of time the pharyngeal region was contracted) in the un-thickened group is increasing by 1.56% for every 1 mL increase in the nipple flow rate.

Fig. 3.

Simulated Model: effect of nipple flow rate on pharyngoesophageal timing characteristics with un-thickened and thickened formula Px-pharyngeal. Raw data for the nipple flow rates are shown for the un-thickened formula (gray circles) and for the thickened formula (black squares); the nipple flow rates were obtained from prior published work [44–46] and do not account for the reduction in the flow rate that may have resulted from the use of thickened formula. The equations for the linear regression lines are: e^([nipple flow rate estimate*x]+[thickness estimate*a]+[nipple flow rate*thickness*x]+[formula consumed per minute*0.277] + intercept), where x is the variable we are plotting, a is 1 for un-thickened and 0 for thickened formula, nipple flow rate*thickness is the interaction, and formula consumed per minute is the volume the infant took over the duration. a-b) Overall and Distal Pharyngeal Activity decrease as the flow rate increases by 1 mL/minute with thickened formula but increase with un-thickened formula. c-d) Pharyngeal Peak to Peak Duration and Pharyngeal peak interval variability increase as the flow rate increases by 1 mL/minute with thickened formula but decrease with un-thickened formula. Note, that the pharyngoesophageal timing characteristics are becoming less coordinated with thickened formula.

Structural Equation Modeling for Frequency of Pharyngeal Peaks

The structural equation model explained how the infant’s characteristics, nipple flow rate, and formula thickness influenced the frequency of pharyngeal peaks (Fig. 4). The nipple flow rate had a negative association (−0.18) on the volume extracted. Therefore, when the nipple flow rate, is increased by one standard deviation the volume extracted will decrease by −0.18 mL. The volume extracted had a positive association (0.53) on the frequency of pharyngeal peaks. The dashed line from the frequency of pharyngeal peaks to the volume extracted denoted an untested hypothesis that the frequency of pharyngeal peaks (infant’s skill) may also influence the volume. Around 36% of the variability in the frequency of pharyngeal peaks and 5% of the variability in volume taken by the infants was explained by our suggested model. The CFI for this model is 0.83 and the root mean square error of approximation is 0.26.

Fig. 4.

Structural equation modeling- characteristics influencing the frequency of pharyngeal peaks. All reported values represent the standardized effect and **p-value < 0.01. PMA-postmenstrual age. The purpose of this structural equation model is to identify which factors influence the frequency of pharyngeal peaks (number of pharyngeal peaks/second). The frequency of pharyngeal peaks is our outcome and is denoted by the gray box. Our influencing factors are the infant’s PMA at evaluation, the presence of respiratory support, volume extracted from the bottle nipple, nipple flow rate, and the amount of thickener added to the formula. The influencing factors are denoted by the white boxes. A relationship between variables (estimated effect) is denoted by the solid black arrows and the direction determines which variable is the influencer. For example, the nipple flow rate affects the volume extracted from the bottle nipple therefore the arrow is drawn from nipple flow rate to the volume extracted. Therefore, the interpretation is that when the nipple flow rate (mL/min) is increased by one standard deviation the volume extracted will decrease by −0.18 mL. The black dashed arrow denotes an untested hypothesis that the frequency of pharyngeal peaks can affect the volume extracted from bottle nipple. Note, that several relationships are at play during the act of orally feeding.

Discussion

Independent oral eating skills are critical milestones in high-risk infants that are convalescing in the neonatal ICUs; however, these skills are dependent on complex neurological-airway-digestive processes that mature or adapt during the time of development and recovery from primary illness. These skills involve communication and coordination between multiple organ systems, which can be challenging among this high-risk population regardless of their age or maturation. However, the mechanisms of swallowing difficulties in infants, and indications to change nipple flows or thickness of formula remain unclear. Such interventions may have unintended consequences. The effects of modified nipple flow or formula thickness on pharyngoesophageal timing and amplitude characteristics in relation to oral intake were compared. We tested the hypothesis that changing the nipple flow rate, or the viscosity of formula modifies the timing and amplitude of pharyngoesophageal motility characteristics during OMCs concurrent with the use of HRPM. Specifically, we determined how the characteristics of the formula bolus altered pharyngeal stimulation.

Salient Findings

The salient findings of the current study are as follows: (1) Swallow function as determined by pharyngeal timing and amplitude characteristics is dependent on both the nipple flow rate and fluid thickness; (2) The thickened formula decreases pharyngeal coordination compared to un-thickened formula when increasing the nipple flow rate; (3) The structural equation model hypothesizes and highlights several internal and external relationships involved in oropharyngeal swallowing.

Physiologic Basis for our Findings and Clinical Reasoning

In the current study, similar observations were noted in the comparison between the slower versus faster nipple flow rate groups and the continuous analyses. It was noted that the adaptation of oropharyngeal-esophageal motility appears to be more regulated with un-thickened formula regardless of the nipple used as there was a lack of statistical differences between the nipple flow rates. This fact is indicated by the presence of substrate reaching the pharynx to trigger timely (rapid) and repetitive pharyngeal contractility (rhythms, bursts, contractions) all of which contribute to the efficiency of pharyngeal motor function, which is measured by the pharyngeal contractile integral (PhCI). Conversely, increasing the thickness of formula slowed down the overall motility or influenced the infant’s adaptation to a different stimulus. Specifically, pharyngeal contractile frequency occurrence and activity decreased as the peak-to-peak duration, pharyngeal peak interval variability, and PhCI increased (p < 0.01). The lack of significance between the slower and faster nipple flow rate groups cannot be explained owing to the pilot nature of the study and multiple contributory variables (characteristics of nipple and their fabrication, orifice, variability of in vitro testing methods, timing of OMC and biological variables).

With the continuous nipple flow analysis, pharyngeal contraction duration was increased with thickened formula (Table 3, p = 0.01). The increased contractile duration demonstrates that, when exposed to boluses of thicker formula, receptors conveyed afferent signals to central pattern generating apparatus (Glossopharyngeal and Vagus Nuclei) for a longer time, thus resulting in stronger or more prolonged contractions needed to clear the stimulus. Thus, the triggering contractile mechanism was prolonged to promote clearance from the pharyngeal introitus. In addition, the increased contractile duration may contribute to the higher PhCI per individual contraction. The PhCI is a critical prerequisite for UES relaxation as it overcomes the resistance offered at the UES. Furthermore, these pharyngeal timing characteristics are dependent on UES and LES relaxation which are critical components for downstream activation of esophageal peristalsis and safe bolus transit [21, 52, 53]. Prolonged bolus dwell in the pharynx may inhibit or delay central pattern generator functions; thus, the regulation of upstream, downstream, and side-stream functions is impaired. Thus, in those infants with developmental neuropathology (where central regulation of swallowing and eating functions are affected) or cardiorespiratory disease (where side stream extra esophageal functions are affected), the presence of thickened formula dwelling in the pharynx for prolonged periods can potentially be unsafe. Additionally, extraction of thickened formula from the bottle can be difficult for these infants with developmental neuropathology or poor respiratory reserve.

The act of oropharyngeal swallowing in the study cohort is not only dependent on formula thickness and nipple flow rate but also several other factors as displayed in the structural equation model (Fig. 4). As previously mentioned, the stimulation of the pharynx by appropriate bolus formation and delivery are essential to the development of swallowing skills needed for independent oral feeding. However, these skills are closely tied to the infant’s postnatal maturation and are honed with consistent practice [5, 54]. Therefore, we included PMA as an influencing factor in our model. Our model suggests that the infants with a higher PMA will have increased volume extracted from the bottle. This could indicate that premature infants need more time to develop the swallowing skills necessary for adequate volume extraction. Furthermore, the presence of supplemental respiratory support is common in convalescing infants in the NICU. Hence, we simply included respiratory support in our structural equation model even though our inclusion criteria of ≤ 1 L per minute of supplemental oxygen could suggest reasonable lung function. It is also important to note that these infants were evaluated for swallowing difficulties which could be contributed to suboptimal respiratory management (i.e., the supplemental oxygen requirement may need increased). Further studies are needed to examine the effect of respiratory support on independent oral swallowing capabilities. Although, our structural equation model explores a few of the factors involved in oral feeding, future studies are needed using HRPM to expand upon and validate these concepts. Also, HRPM allows for multiple and varied OMC trials with minimal risk and is an objective tool to examine the swallowing, airway safety, and oral intakes to determine the most effective and safe feeding therapies. Such methodological advances offer new tools to multiple disciplines in implementing point-of-care precision strategies.

Comparisons to Other Studies

Previous investigations defined distinct pharyngoesophageal motility rhythms using cluster analysis and found that the subjects with a robust pharyngoesophageal pattern during oral feeding had a higher rate of oral intake and increased prevalence of independent oral feeding at discharge [47]. Also, the pharyngeal peak interval variability (SD of the pharyngeal peaks) was decreased while the pharyngoesophageal and deglutition apnea activity (percent of time that anatomical region was active over the total duration of the oral feed) was increased [47]. The current study found that pharyngoesophageal motility was affected by the nipple flow rate and formula thickness, but the overall pharyngeal coordination (pharyngeal peak interval variability and activity) was modified and potentially worse in the trials with thickened formula (Table 3; Fig. 3). Pharyngeal peak interval variability modifies sensory-motor responses and slows the functions of all the components of aerodigestive apparatus which can affect the quality of feeding. Equally important to note is the increase in PhCI associated with the thickened formula (current study, p < 0.01), and this observation was also noted when the infants were fed with the thickened barium sulfate during our prior study [37]. All the above findings indicate oropharyngeal stimulus-types trigger, integrate, and modify brainstem mediated oral and pharyngeal rhythms and modulate contractility and airway protection mechanisms. Basic science work [55–60] confirms our pathophysiologic reasoning in that pattern generation is distinct for different phases of swallowing, and these functions are developing/adapting in convalescing high-risk infants and are indeed the reasons for oral eating difficulties [39].

Limitations

The current study provides pilot data and preliminary insights into the effect of nipple flow rate and formula thickness on pharyngeal timing and amplitude characteristics during OMCs. Like any clinical study, there were limitations. The data presented is only from a subset of infants from our previous observational study cohort [37] and included only those who had complete recordings involving all 4 OMCs trials in the same infant. This rigor in the same infant’s comparative application was necessary as we needed high fidelity recordings and complete meticulous analysis for all 48 oral feeding challenge trials. To further clarify, these OMCs were not performed during concurrent VFSS; thus, measured characteristics from VFSS could not be obtained. Additionally, a true comparison could not be ascertained as it is not ethical to perform HRPM or VFSS studies in healthy infants (i.e., no control group). However, a control group can be construed as a group of infants that do not have any swallowing difficulties and future prospective studies can be powered using these controls to determine causal mechanisms.

Furthermore, the observed pharyngeal manometry findings may be related to neurological (sensory or motor or both) or muscular or central pattern coordination. To ascertain specific mechanisms requires provocation manometry. This study was not designed to determine such mechanisms rather we observed differences within and between the infants regarding presenting 4 different modes of stimulus during OMC. Therefore, this study leads to the development of newer hypotheses in that variations in presentation of sensory stimulation modifies pharyngeal muscular function, pharyngeal rhythmic (timing) patterns and oropharyngeal bolus transit or extraction.

Lastly, variable in vitro nipple flow rates within the same type of nipple existed due to the lack of reliable quality of nipple and bottle standards [44–46]. This contributory factor was mitigated by the rigor of the analysis in using the nipple flow rate as a continuous variable. In addition, the effect of formula, expressed human milk, the combination of formula and expressed human milk, or different types of thickeners on pharyngeal swallowing rhythms was not studied. Once more, these limitations highlight several influencing factors that contribute to an OMC in infants, especially in those who have swallowing difficulties.

Conclusions

We examined the effects of different nipple flows and formula thickness in individual infants undergoing novel HRPM during OMCs. Given the context of the current study, central and peripheral regulation of swallowing may play a part, which is a complex process governed by multiple factors (pattern generators, afferent modulation, efferent modulation, regional muscular tissue effectiveness, supranuclear factors, developmental maturation and growth trajectory). In this study, we only quantified the ultimate motor system properties controlling the swallow timing and inter-swallow intervals, amplitude and contractile integrals, and frequency of swallowing markers. We found that, the adaptation of oro-pharyngo-esophageal motility characteristics appear to be more regulated with un-thickened formula regardless of the nipple flow used. Thickened formula increased the pharyngeal peak interval variability of pharyngeal contractile rhythms (timing) and modified the pharyngeal contractile activity thus altering and prolonging the oropharyngeal motility sequences. Ultimately, these findings indicate that different types of oropharyngeal stimuli-trigger, integrate, and modify brainstem mediated oral and pharyngeal rhythms and modulate contractility and airway protection mechanisms differently. Careful consideration is needed when selecting nipple flow and formula thickness in the high-risk NICU populations. This work and the applied technologies are helpful for providing physiologic-evidence at crib-side for parents and multiple disciplines involved with infant feeding.

Abbreviations

HRPM

High resolution pharyngoesophageal manometry

OMC

Oral milk challenge

PhCI

Pharyngeal contractile integral

PMA

Postmenstrual age

UES

Upper esophageal sphincter

VFSS

Video-fluoroscopy swallow studies

Data Availability

The data, analytic methods, and study materials that support the findings of this study are available upon reasonable request from the corresponding author. The data, analytic methods, and study materials are not publicly available due to privacy or ethical restrictions.