Effect of nasal noninvasive respiratory support methods on pharyngeal provocation-induced aerodigestive reflexes in infants

Abstract

The pharynx is a locus of provocation among infants with aerodigestive morbidities manifesting as dysphagia, life-threatening events, aspiration-pneumonia, atelectasis, and reflux, and such infants often receive nasal respiratory support. We determined the impact of different oxygen delivery methods on pharyngeal stimulation-induced aerodigestive reflexes [room air (RA), nasal cannula (NC), and nasal continuous positive airway pressure (nCPAP)] while hypothesizing that the sensory motor characteristics of putative reflexes are distinct. Thirty eight infants (28.0 ± 0.7 wk gestation) underwent pharyngoesophageal manometry and respiratory inductance plethysmography to determine the effects of graded pharyngeal stimuli (n = 271) on upper and lower esophageal sphincters (UES, LES), swallowing, and deglutition-apnea. Comparisons were made between NC (n = 19), nCPAP (n = 9), and RA (n = 10) groups. Importantly, NC or nCPAP (vs. RA) had: 1) delayed feeding milestones (P < 0.05), 2) increased pharyngeal waveform recruitment and duration, greater UES nadir pressure, decreased esophageal contraction duration, decreased distal esophageal contraction amplitude, and decreased completely propagated esophageal peristalsis (all P < 0.05), and 3) similarly developed UES contractile and LES relaxation reflexes (P > 0.05). We conclude that aerodigestive reflexes were similarly developed in infants using noninvasive respiratory support with adequate upper and lower aerodigestive protection. Increased concern for GERD is unfounded in this population. These infants may benefit from targeted oromotor feeding therapies and safe pharyngeal bolus transit to accelerate feeding milestones.

improvement of respiratory outcomes in premature infants through utilization of noninvasive respiratory support [nasal continuous positive airway pressure (nCPAP) or variable-flow nasal cannula (NC)] has been reported (24, 30, 30a) and is considered as a first line of respiratory therapy, or as a step-down approach upon extubation (20, 22, 24, 27). Due to relatively low expense, ease of administration, and ability to provide developmental support, these noninvasive respiratory support measures are promising and increasingly popular among low-resource nations, where the prevalence of premature birth or volume of high-risk infants is increasing. Comparative clinical trials have been equivocal in their findings with regard to efficacy, effectiveness, or aerodigestive consequences (2, 4, 31). However, no studies have systematically evaluated the feeding or swallowing reflexes or airway-protective reflexes such as may be activated to protect against airway aspiration in this type of cohort. Therefore, development of oral feeding plans while on noninvasive respiratory support can be highly variable. Furthermore, worsening respiratory distress while on these types of respiratory support is commonly thought to be due to aspiration or gastroesophageal reflux disease (GERD). Swallowing difficulties are frequently associated with the above situations, and the pharynx is the locus of provocation in most of these scenarios. Anticipation and prevention of aspiration is a skill that is acquired, developed, and further modulated by multiple sensory motor neural pathways involving multiple cranial nerves (V, VII, IX, X, XI, XII) and the brain stem (13, 15). Supranuclear pathways relay to the thalamus, basal ganglia, and cerebral cortex (pre- and postcentral gyrus), which participate in further modulation of perception, volitional, and central regulatory swallowing and feeding behaviors. We have previously shown in healthy premature infants that pharyngeal stimulation activates several airway-protective reflexes that modulate safe swallowing sequences along with respiratory adaptation (11–14, 17). These adaptive responses include pharyngeal reflexive swallowing (PRS), pharyngo-upper esophageal sphincter contractile reflex (PUCR), pharyngo-lower esophageal sphincter relaxation reflex (PLESRR), deglutition apnea, and adaptive respiratory changes (11–13, 17).

This study was undertaken to determine the impact of different oxygen delivery methods (noninvasive respiratory support as a means to provide adequate oxygenation) on swallowing function and recruitment of aerodigestive mechanisms. Our aims were to characterize and compare the aerodigestive adaptive responses evoked upon pharyngeal stimulation in a cohort of infants receiving different types of oxygen delivery methods, i.e., nCPAP, NC, or none. We hypothesized that the sensory motor characteristics of proximal aerodigestive reflexes (PRS, PUCR, PLESRR, deglutition apnea, and adaptive respiratory changes) evoked upon pharyngeal stimulation in infants with bronchopulmonary dysplasia (BPD) receiving airway support via NC or nCPAP are distinct when compared with healthy infants on room air (RA, serving as the comparison group).

MATERIALS AND METHODS

Participants.

Subjects were recruited under our neonatal esophagus and airway interaction study protocol. Procedures were performed in the neonatal intensive care unit setting at Nationwide Children's Hospital. Informed parental consent and Institutional Review Board approval were obtained with compliance to the Health Insurance Portability and Accountability Act. Subject safety was monitored by a registered nurse and physician at the crib side during evaluation.

Study design and protocol.

Inclusion criteria for all subjects included: 1) enterally fed at the time of evaluation and 2) physiologically stable for at least 48–72 h before evaluation and without any new illness onset at the time of evaluation. Exclusion criteria included genetic and congenital birth defects. Infants were included in the study arm (BPD) if they met the following criteria: 1) born preterm and requiring mechanical ventilation, 2) diagnosis of BPD defined as supplemental oxygen requirement at 36 wk postmenstrual age as per National Institutes of Health consensus criteria, and 3) continued dependence on noninvasive respiratory support (NC or nCPAP) at the time of evaluation. Inclusion criteria for the control arm were: 1) absence of the diagnosis of BPD and 2) absence of any respiratory support at the time of evaluation; thus, subjects were in RA.

Pharyngoesophageal manometry, respiratory inductance plethysmography, and nasal airflow thermistor methods were used to test aerodigestive reflexes upon pharyngeal stimulation (Fig. 1) as previously described (12, 13, 17). Per protocol, if the subject had a nasogastric or orogastric tube for feeding at the time of evaluation, it was removed before manometry catheter placement. A silicone catheter with pressure ports and sleeves spanning from the pharynx to stomach (Dentsleeve International, Mui Scientific, Ontario, Canada) was attached to a pneumohydraulic micromanometric water perfusion system (Stationary Solar Gastro, v.8.21 Medical Measurement Systems, Dover, NH), and pharyngoesophageal pressure waveforms were recorded. Concurrent respiratory inductance plethysmography (Respitrace, Viasys, Conshocken, PA) and nasal airflow thermistor (Integra Life Sciences, Plainsboro, NJ) were used to detect ventilatory changes with inspiratory and expiratory phases of breathing (8). The infant was allowed to adapt to catheter placement before initiation of the experimental protocol, which included abrupt graded sterile water infusions by syringe (volumes of 0.1, 0.3, and 0.5 ml in triplicate) to the pharynx via the pharyngeal infusion port.

Fig. 1.

Aerodigestive reflexes evoked upon pharyngeal stimulation. Broken lines represent onset of pharyngeal stimulus. At lower volumes typical primary responses include the following solitary responses pharyngo-upper esophageal sphincter contractile reflex (PUCR) is characterized by upper esophageal sphincter (UES) contraction (A) or pharyngeal reflexive swallowing (PRS) is characterized by pharyngeal contraction, UES relaxation, esophageal body propagation, lower esophageal sphincter (LES) relaxation, and associated deglutition apnea (B). C: as the volume increases multiple PRS reflexes are evident. It is here that we can characterize initial and terminal responses. Note the initial response (light gray box) was defined as the first response to pharyngeal stimulation until a pause in pharyngeal signaling and/or esophageal body propagation. The terminal response (dark gray) is any pharyngeal contractile activity until a terminal swallow (esophageal body propagation) restores respiratory and digestive normalcy.

Analytical definitions apriori.

Aerodigestive reflexes were evaluated as described previously (12, 13, 17). Initial response was defined as the first aerodigestive response to pharyngeal provocation until a pause in pharyngeal signaling and/or esophageal propagation. Terminal response was defined as the final aerodigestive response that restored quiescence in the respiratory and esophageal baseline waveform rhythms (Fig. 1). Sensory threshold volume is defined as the least stimulus volume that resulted in a pharyngo-esophageal motor response (12).

Details of the initial pharyngo-esophageal and respiratory response characteristics included: 1) prevalence of pharyngeal or upper esophageal sphincter (UES) reflexes (%) and proportion (%) of primary response (PRS, PUCR, or none). Response was counted as occurring within 5 s of pharyngeal infusion, as defined previously (12). PRS was defined as pharyngeal contraction associated with UES relaxation, and esophageal peristalsis occurring upon pharyngeal provocation, whereas PUCR was defined as UES contractile response (12); 2) response latency (s) was measured as the duration from pharyngeal infusion onset to manometric response onset (12, 17); 3) occurrence of deglutition apnea (%), i.e., a pause in respiratory rhythm captured during respiratory inductance plethysmography and nasal airflow thermistor signal recordings when concurrently associated with PRS (13, 28); 4) respiratory phase at deglutition apnea onset (%), i.e., occurring during inspiration (upstroke), exhalation (downstroke), or interphase (in between respiratory cycles) (13, 28); and 5) pharyngeal contractile waveform recruitment (n), as the number of pharyngeal peaks until esophageal propagation commenced.

Details of the terminal pharyngoesophageal and respiratory response characteristics analyzed were: 1) total response duration (s), measured from stimulus onset to restoration of aerodigestive quiescence, 2) esophageal response duration (s) measured from the onset of esophageal body contractile waveform activity to offset of esophageal body activity, 3) pharyngeal contractile waveform recruitment (sensory motor) as the number of pharyngeal peaks until a PRS ensues restoring respiratory and digestive quiescence, 4) distal esophageal area under the contractile waveform curve was defined as duration of contraction (s) times peak distal esophagus amplitude (mmHg), 5) esophageal response duration (s) defined as onset of esophageal body propagation to offset of terminal esophageal body propagation, 6) UES nadir pressure (mmHg) of the terminal swallow defined as the lowest pressure point of UES relaxation, 7) esophageal propagation prevalence (%), i.e., complete, incomplete, or failed, 8) esophageal contractile pressure (mmHg) of the proximal, middle, and distal esophagus defined as the peak contractile pressure at each esophageal body location, and 9) esophageal contractile pressure duration (s) of the proximal, middle, and distal esophagus, defined as the onset of contraction to offset of contraction at each esophageal body loci.

Overall response characteristics, for both initial and terminal pharyngoesophageal responses, included: 1) pharyngeal response duration (s) as onset to offset of pharyngeal recruitment activity, 2) mean pharyngeal frequency (Hz) calculated as pharyngeal contractile peak recruitment divided by the duration of pharyngeal contractile peak recruitment, and 3) respiratory rhythm disturbance duration as period from onset of first abnormal breath to restoration of normal breath (Fig. 1C).

UES characteristics included: 1) basal tone (mmHg), 2) response type prevalence (%) i.e., relaxation, contraction, or no response, 3) contractile magnitude (mmHg) defined as the peak contractile pressure, 4) contractile duration (s) or onset to offset of UES contraction. If there was a terminal PRS, then the following UES characteristics measured were: 5) postpharyngeal contraction prevalence (%), i.e., present or absent. Postpharyngeal contraction was defined as the occurrence of UES contraction after the pharyngeal reflexive swallow contractile waveform pattern in the pharynx, 6) postpharyngeal contraction duration (s) was defined as UES nadir onset to peak UES contractile pressure, and 7) postpharyngeal contraction magnitude (mmHg) defined as the peak UES contractile pressure after pharyngeal contractile waveform.

Lower esophageal sphincter (LES) characteristics included basal tone (mmHg) and response type prevalence (%), i.e., relaxation, contraction, or no response. If LES relaxation occurred then the magnitude of relaxation was measured, i.e., nadir relaxation pressure (mmHg) and duration of relaxation (s) (17, 26).

Statistical analysis.

Demographic data were compared using Chi square tests, two-sample t-tests, and Wilcoxon rank-sum tests. To account for repeated measures that may be correlated, linear mixed models were used to compare continuous outcome variables between the groups of respiratory support, i.e., NC, nCPAP, and none. Generalized estimating equation models were used to compare categorical outcomes between groups. Because of the potential impact of prior respiratory or feeding interventions to time of evaluation on aerodigestive reflexes, linear mixed models were used to determine if these variables needed to be controlled for (Table 1). If found to be significant, they were included in the linear mixed and generalized estimating equation models comparing respiratory groups.

Table 1.

Effect of respiratory and feeding interventions before time of evaluation on aerodigestive reflexes (UES, LES, esophageal body, and overall response characteristics)

Characteristic	Ventilation Duration, days	P Value	nCPAP Duration, days	P Value	Cumulative (ventilation + nCPAP) Duration, days	P Value	Gavage Feeding Duration, days	P Value
UES
Basal pressure, mmHg	0.01 ± 0.05	0.9	0.09 ± 0.06	0.1	0.02 ± 0.03	0.5	0.08 ± 0.05	0.1
Contractile reflex occurrence	1.02 (1.00–1.03)	0.02	1.00 (1.00–1.02)	0.6	1.01 (1.00–1.01)	0.06	1.00 (1.00–1.02)	0.4
Contraction response latency, s	0.01 ± 0.02	0.5	−0.02 ± 0.01	0.2	−0.01 ± 0.01	0.2	−0.03 ± 0.02	0.07
Contraction response time at peak, s	0.02 ± 0.02	0.3	−0.00 ± 0.01	0.9	−0.00 ± 0.01	0.6	−0.02 ± 0.01	0.1
Contraction peak amplitude, mmHg	0.03 ± 0.09	0.7	0.20 ± 0.09	0.1	0.03 ± 0.06	0.6	0.04 ± 0.10	0.8
Contraction response duration, s	0.02 ± 0.04	0.6	0.09 ± 0.02	<0.01	0.01 ± 0.02	0.6	0.03 ± 0.02	0.1
LES
Basal pressure, mmHg	−0.02 ± 0.04	0.7	0.01 ± 0.05	0.9	−0.10 ± 0.03	0.07	−0.02 ± 0.04	0.7
Relaxation reflex prevalence	1.00 (1.00–1.00)	0.4	1.00 (1.00–1.02)	0.9	1.00 (1.00–1.00)	0.8	1.00 (1.00–1.01)	0.8
Relaxation response latency, s	−0.01 ± 0.01	0.4	−0.00 ± 0.01	1	−0.00 ± 0.01	0.4	0.00 ± 0.01	1
Relaxation time, s	−0.00 ± 0.01	0.9	−0.00 ± 0.01	0.9	−0.00 ± 0.01	0.9	0.00 ± 0.01	0.9
Relaxation nadir, mmHg	0.02 ± 0.02	0.4	0.03 ± 0.02	0.3	0.00 ± 0.01	0.5	0.02 ± 0.01	0.2
Relaxation nadir duration, s	0.01 ± 0.04	0.8	−0.09 ± 0.06	0.1	−0.02 ± 0.04	0.6	−0.06 ± 0.06	0.3
Esophageal body
Propagation occurrence	1.00 (1.00–1.01)	0.1	1.00 (1.00–1.03)	0.4	1.00 (1.00–1.01)	0.09	1.00 (1.00–1.01)	0.2
Overall Response
Response occurrence	1.00 (1.00–1.02)	1	1.00 (1.01–1.02)	0.6	1.00 (1.00–1.00)	0.6	1.00 (1.00–1.00)	0.3
No. of pharyngeal peaks	0.02 ± 0.01	0.09	−0.01 ± 0.02	0.7	0.01 ± 0.01	0.2	0.00 ± 0.01	0.7
Total response time, s	−0.05 ± 0.05	0.4	−0.05 ± 0.06	0.5	−0.03 ± 0.03	0.3	0.03 ± 0.05	0.5

Data presented as means ± SE or odds ratio(95% confidence interval). UES, upper esophageal sphincter; LES, lower esophageal sphincter RA, room air; NC, nasal cannula; nCPAP, nasal continuous positive airway pressure. If P value was found to be significant (P < 0.05), the variable was controlled for in the statistical model comparing respiratory groups (NC, nCPAP, or RA). Only two respiratory interventions found to be significant: 1) The odds of a UES contractile reflex occurrence increased by 2% for each additional day of ventilation, and 2) nCPAP duration, days had a significant effect on UES contraction response duration in that the response duration increased by 0.09 s for each additional day of nCPAP.

Data are presented as medians(interquartile range), means ± SE, percentage (%), or as otherwise indicated. P values <0.05 were considered statistically significant. All statistical analysis was performed using SAS version 9.3 (SAS Institute, Cary, NC).

RESULTS

Subject characteristics.

Participants were 38 infants born at 28.0 ± 0.7 wk gestation (birth wt of 1.2 ± 0.01 kg, 23 male) and were assigned a median APGAR score of 6 (1–9) at 1 min and 8 (1–9) at 5 min. Demographic and aerodigestive milestone characteristics are presented and categorized into three groups based on respiratory support status at evaluation [NC (n = 19), nCPAP (n = 9), and RA (n = 10)] (Table 2). The prevalence of intraventricular hemorrhage was two infants (11%) in NC (both grade I), four infants (44%) in nCPAP (2 in grade I, 1 in grade II, and 1 in grade III) and no infants (0%) in RA (P < 0.05 vs. nCPAP) groups. Of the total 38 infants, two (5%) received acid-suppressive medication and none (0%) received caffeine at the time of evaluation. The median oxygen flow rate of the NC group was 0.4(0.1–2.0) l/min, and median positive end-expiratory pressure of the nCPAP group was 8(6–8) cmH₂O. Of the NC group, the duration between the last day of nCPAP and the time of evaluation averaged 27 ± 17 days with median(interquartile range) 28(4–45) days.

Table 2.

Subject demographics

Characteristics	RA (n = 10)	NC (n = 19)	nCPAP (n = 9)
At birth
Gestational age, wk	33.9 ± 1.2*†	26.1 ± 0.5	25.4 ± 0.7
Weight, kg	2.3 ± 0.3*†	0.7 ± 0.1	0.8 ± 0.1
Length, cm	44.6 ± 2.2*†	33.4 ± 1.1	30.8 ± 1.0
Head circumference, cm	28.9 ± 2.4	23.5 ± 0.6	22.0 ± 0.9
At evaluation
Postmenstrual age, wk	40.6 ± 1.0	42.4 ± 1.0	40.2 ± 0.8
Weight, kg	3.3 ± 0.2†	3.2 ± 0.2	2.8 ± 0.2
Length, cm	48.2 ± 0.8†	47.3 ± 0.6	45.6 ± 0.7
Head circumference, cm	35.2 ± 0.8†	34.2 ± 0.5	32.7 ± 0.6
Duration of ventilation, days	0 (0–0)*†	54 (22–62)	60 (47–83)
Duration of nCPAP, days	0 (0–1)*†	32.5 (18.5–49.5)	52 (29–63)
Duration of ventilation + nCPAP, days	0 (0–3)*†	69 (54–105)	100 (77–135)
Feeding method, %
Gavage only:gavage + oral:oral only	0:30:70*†	16:68:16	89:11:0
Duration of tube feeding, days	0 (0–6)*†	101 (77–114)	105 (84–111)
During hospital stay
Duration of ventilation, days	0.4 ± 0.4*†	45.1 ± 6.6	60.2 ± 11.7
Duration of nCPAP, days	2.0 ± 1.8*†	32.1 ± 5.5	51.1 ± 11.5
Supplemental oxygen, days	1.6 ± 1.1*†	62.8 ± 7.2	48.8 ± 10.0
Gavage feeds only, days	1.0 ± 1.0*†	64.6 ± 8.6	68.7 ± 13.2
Gavage and oral feeds, days	6.6 ± 5.6*†	49.9 ± 5.7	54.1 ± 13.6
At discharge
Feeding method, % (oral:gastrostomy)	80:20	68:32	56:44
Length of hospital stay, days	17.1 ± 5.6*†	126.7 ± 11.3	136.9 ± 12.8

Data presented means ± SE; n, no. of infants. Units are median(interquartile range) or %. P = nonsignificant (NS) for NC vs. nCPAP.

^*P < 0.05 vs NC.

^†P < 0.05 vs. nCPAP.

Effects of pharyngeal stimulation on the initial and terminal pharyngoesophageal and respiratory responses.

Stimulus threshold volume (ml) to initiate response among infants on RA was 0.2(0.15–0.2), among NC was 0.1(0.1–0.2), and for nCPAP was 0.3(0.2–0.3) (P < 0.05 for NC vs. nCPAP). Representative recordings characterizing the stimulus effects on respiratory waveforms, pharyngeal contractile rhythms, UES relaxation, esophageal body, and LES relaxation are shown (Fig. 2). Data from a total of 271 pharyngeal infusions (66 in RA, 139 in nasal cannula, and 66 in nCPAP groups) were analyzed for detailed characteristics of pharyngo-UES, esophageal peristalsis, LES, and respiratory rhythm. Respiratory rhythm disturbance durations were 25.0 ± 3.8 s for infants on RA, 25.6 ± 2.6 s for NC, and 21.1 ± 3.4 s for nCPAP (P > 0.05 among all groups).

Fig. 2.

Effect of respiratory support on aerodigestive reflexes in response to 0.3 ml sterile water pharyngeal stimulation. Notice the adequate LES relaxation and duration when relaxation occurs in the nasal cannula (NC) and nCPAP groups. Also, note the respiration scales (abdominal respiration, thoracic respiration, tidal ventilation) are different from Fig. 1 to better accentuate the respiratory rhythm disturbances (shaded gray boxes).

Upon pharyngeal stimulation, the initial pharyngoesophageal response frequencies in RA, NC, nCPAP groups were 71.2, 79.1, and 66.7%, respectively (P > 0.05 for all groups). To test the effect of abrupt pharyngeal stimulus on initial pharyngoesophageal and respiratory response mechanisms under various nasal respiratory support conditions, we analyzed the frequency recruitment of pharyngeal reflexive swallow, pharyngo-UES contractile reflex or none (Fig. 3A), response latency to the PRS or PUCR (Fig. 3B), number of pharyngeal waveforms recruited (Fig. 3C), overall pharyngeal response duration (Fig. 3D), prevalence of pause in breathing during swallowing or deglutition apnea (Fig. 3E), and the respiratory phase in which deglutition apnea occurred (Fig. 3F).

Fig. 3.

Effect of respiratory support on initial pharyngeal provocation-induced responses. A: primary response type: although P is not significant (NS), PUCR is more frequent with increased degree of respiratory support. B: primary reflexive swallow latency: latencies were similar. C: pharyngeal recruitment: recruitment increased with respiratory support. D: pharyngeal response duration: response duration increased during NC but was similar during nCPAP. E: deglutition apnea prevalence: prevalence was not affected with any modes of respiration. F: phase of deglutition apnea: nCPAP has more deglutition apnea during inspiration. *P < 0.05 vs. NC. †P < 0.05 vs. nCPAP.

The frequency of subsequent pharyngoesophageal response that restored respiratory and esophageal normalcy among RA, NC, nCPAP groups was 88.5, 75.5, and 78.7%, respectively (P > 0.05 for all groups). To understand the sequential participatory mechanisms that may have the ability to restore aerodigestive quiescence, we analyzed the frequency recruitment of pharyngeal waveforms (Fig. 4A), duration of cumulative pharyngoesophageal waveform activity (Fig. 4B), distal esophageal contractile amplitude-duration product (area under curve, mmHg·s, Fig. 4C), duration of overall esophageal body waveform activity (Fig. 4D), UES nadir pressure (Fig. 4E), and frequency patterns of esophageal waveform propagation (Fig. 4F). Basal midesophageal pressure was −0.09 ± 0.9 mmHg in RA, 1.3 ± 0.7 mmHg in NC, and 2.6 ± 0.9 mmHg in nCPAP (P > 0.05 for all groups).

Fig. 4.

Effect of respiratory support on terminal pharyngeal-induced provocation responses. A: pharyngeal recruitment is not affected. B: total response duration is not affected. C: distal esophagus area under the curve is lower in NC. D: esophageal response duration is lower in NC. E: UES nadir pressure: respiratory support may interfere with complete UES relaxation. F: esophageal propagation: infants with respiratory support may have less complete propagation sequences. *P < 0.05 vs. NC. †P < 0.05 vs. nCPAP.

Effect of pharyngeal stimulus on the UES contractile and LES relaxation characteristics.

The effects of pharyngeal stimulus on UES contractile and LES relaxation characteristics are displayed in Table 3.

Table 3.

Sphincter characteristics

Characteristic	RA (n = 66)	NC (n = 139)	nCPAP (n = 66)
Upper esophageal sphincter
Basal pressure, mmHg	11.9 ± 2.2	15.6 ± 1.5	10.4 ± 2.2
Contractile reflex prevalence, %	6.1	10.8	12.5
Contraction response latency, s	3.2 ± 0.9	2.5 ± 0.3	3.2 ± 1.1
Contraction response time at peak, s	3.8 ± 0.8	4.5 ± 0.8	4.8 ± 1.0
Contraction peak amplitude, mmHg	27.8 ± 8.0	40.7 ± 0.1	28.8 ± 6.4
Contraction response duration, s	5.2 ± 1.9	3.9 ± 1.1	2.1 ± 1.3
Lower esophageal sphincter
Basal pressure, mmHg	22.6 ± 2.2	19.9 ± 1.7	20.9 ± 1.9
Relaxation reflex prevalence, %	79.7	73.4	59.4*
Relaxation response latency, s	5.2 ± 0.4	5.5 ± 0.3	5.1 ± 0.5
Relaxation time, s	1.8 ± 0.4	1.9 ± 0.2	2.3 ± 0.4
Relaxation nadir, mmHg	−1.0 ± 1.0	0.1 ± 0.8	−0.1 ± 1.2
Relaxation nadir duration, s	15.5 ± 2.2	15.3 ± 1.6	17.8 ± 2.4

Data are means ± SE; n, no. of stimuli. Units are %.

^*P < 0.05 compared with NC, all others P = NS.

DISCUSSION

Prevalence of BPD and chronic lung disease among infants is increasing, as well as aerodigestive morbidities resulting in nutritional and growth failure, neurodevelopmental delays, and economic burden. Improving survival by providing respiratory support using noninvasive methods is increasingly prevalent; however, concurrent digestive support strategies to optimize enteral nutrition, growth, and feeding regulation are lacking in this prototype of patients either during the time in the intensive care unit or beyond discharge. Broadly, in this population, aerodigestive morbidities manifest as dysphagia, life-threatening events, aspiration, pneumonia, atelectasis, GERD, as well as morbidities resulting from the consequences of escalating intensive care and medical and surgical therapies (9, 15, 18). Understanding the regulation of reciprocal aerodigestive interactions is important in modifying each of these morbidities, as well as with restoring better feeding adaptation strategies. This is the first study undertaken to clarify aerodigestive protective mechanisms in convalescing human infants with severe BPD receiving three different types of respiratory support.

Our study model is unique in that purported stimuli are focal, well targeted to the pharynx, and measurable, and the response characteristics are evaluable. Using objective measurements as previously validated and published in infants (12, 13, 17), we tested the effect of noninvasive respiratory support on aerodigestive reflexes in infants with BPD in a prospective manner. We found several similarities and differences with the sensory motor characteristics of airway protection mechanisms. Cardinal findings among infants receiving noninvasive respiratory support (compared with the control group) were as follows: 1) feeding milestones were delayed (as evidenced by gavage feeding durations in Table 2), 2) UES nadir pressure was greater, esophageal response duration was lesser, distal esophageal contractile amplitude was lesser, and complete esophageal propagation rates were lesser, and 3) UES contractile reflex characteristics and LES relaxation reflex characteristics were similarly developed. While UES and LES were adequate for most of the characteristics (compared with controls), it is likely that these infants have proximal aerodigestive protection capability in the event of proximally migrating GER. Multiple pharyngoesophageal and respiratory adaptation similarities were noted between NC and nCPAP.

The use of CPAP has been shown to inhibit gastroesophageal reflux in adults and newborn lambs (5, 19). The adult study proposed that the mechanism behind this could be due to passive elevation of basal midesophageal and LES pressures (19). Another study in adults showed that those on CPAP had decreased LES relaxation magnitude and duration (29). Yet another study in adults found an increase in midesophageal resting pressure while on CPAP (6). It was proposed that the reduction in GER is related to direct mechanical compression of the esophagus. In the animal study of newborn lambs, basal midesophageal and LES pressures were similar, but again LES relaxation magnitude and duration were decreased with CPAP (5). However, in our study, basal midesophageal pressure, basal LES pressure, LES relaxation magnitude, and duration were similar, but LES relaxation was less frequent in the nCPAP group. Given the similarities in midesophageal and LES characteristics, it is unclear and controversial if the use of CPAP to treat reflux in the infant population would be beneficial; therefore, further studies are warranted. Because LES pressures are comparable in all three groups, there is little concern for GERD under the given conditions and age at evaluation. However, if GERD is suspected, testing requires objective evaluation using pH impedance methods so as to clarify spatiotemporal characteristics of the refluxate with symptom correlation.

Implications of this study are several. 1) Among those receiving noninvasive respiratory support, although UES nadir pressures and esophageal body propagation characteristics were statistically significant, they are likely not clinically significant, since differences were small and the nCPAP group closely resembles the RA group. The addition of impedance concurrent with manometry may offer benefit to determine bolus transit in instances of incomplete or failed esophageal body propagation. Achievement of an independent feeding- and airway-protective function requires integrated physiological interactions between multiple systems and protective reflexes. In this study we only recognized some of the putative protective reflexes involved with airway safety. Further studies are desperately needed to address the multiorgan interactions between the brain, airway, and foregut. 2) Frequency of pharyngeal stimulation-induced responses was similar between all groups, including NC vs. nCPAP, suggesting appropriate abrupt pharyngeal adaptation to provocation such as may happen with GER, cough, expectoration, or pharyngeal bolus. This suggests that infants are capable of initiating PRS as a response to facilitate clearance of material from the pharyngeal introitus. Variability with pharyngeal waveforms is modified or decreased during NC and nCPAP, thus supporting greater inhibition of swallow-CPG interactions. 3) Controlled and repetitive oromotor feeding therapies (nonnutritive and nutritive oral stimulation) during noninvasive respiratory support may be helpful, since this may promote sensory motor integration, skill development, and aerodigestive neuroplasticity. Such mechanisms possibly reduce development of oral aversion and oropharyngeal dysphagia while improving pulmonary outcomes. We hypothesize that such repetitive therapeutic exercises may result in the development of muscle memory, neuromotor coordination, and appropriate aerodigestive protective reflexes. Further studies are therefore warranted to clarify these observations. Indeed, although this is a new approach and represents a change in practice, since oral feeding is controversial in this group, we have recently shown in a pilot study the safety and feasibility of feeding infants cautiously while on nCPAP (7). Given the response to targeted pharyngeal provocation, we believe that infants have the ability to protect and clear the bolus with appropriate proximal peristaltic mechanisms, even under the studied conditions of nasal CPAP. Intragastric gavage tube bolus feeding protocols may be optimal in providing more physiological nutrient delivery to the stomach, and help facilitate gastroduodenal interdigestive motility cycles and migrating motor complexes (1, 10, 16, 23). Furthermore, intragastric bolus feeding permits gastric venting in between feeding delivery periods, thus protecting infants from excessive gastric distention from possible swallowing of air. Transpyloric or continuous feeding methods have been shown to modify gastric capacity in adults with critical illness (3); however, further studies are needed to evaluate how these feeding strategies modify gastric accommodation in critically ill infants.

Limitations of our study and potential for future research can also be recognized. 1) Subject selection and participation was random and not based on exact definition of respiratory support that can be similarly included for all three groups at evaluation. In prospective studies, such an allocation can be harder given the heterogeneity of the evolving disease(s) or of their confounding factors (like other comorbidities such as intraventricular hemorrhage). Such studies will require a larger patient sample with well-structured longitudinal studies to evaluate the effects of novel therapies. 2) Individual functions (feeding methods, growth, length of hospital stay, respiratory milestones) were similar among the NC and nCPAP groups; we speculate that this observation may be related to the management approaches. In the current study, our intent was to define the frequency recruitment and characteristics of aerodigestive reflexes in the severe BPD prototype, and longitudinal studies are needed using this approach. 3) Discharging infants with chronic lung disease on NC oxygen with variable flows or concentrations is an emerging trend, as well as a method of weaning from positive pressure or noninvasive ventilation. However, in studies comparing high-flow NC vs. nCPAP, no differences in outcomes related to length of stay or BPD are noted (2, 21, 27, 31), yet feeding strategies or functional clinical outcomes have not been examined. Furthermore, safety to feed while on NC/nCPAP with variable flows remains a concern among providers.

In conclusion, aerodigestive pulmonary disease manifests as dysphagia, life-threatening events, aspiration, pneumonia, atelectasis, GERD, and morbidities resulting from the consequences of escalating intensive care or medical or surgical therapies. Regulation of airway-digestive interactions is important not only with modifying each of these morbidities but also with the development of better therapeutic strategies. This is the first study undertaken to clarify aerodigestive protective mechanisms in convalescing human infants with severe BPD receiving three different types of respiratory support. We have developed validated methods to examine neurosensory and neuromotor vagal reflex activity in infants. The current study lends support to provide mechanistic basis and rationale for supporting “controlled and regulated” oral feeding during nCPAP or high-flow nasal cannula. Further studies are needed to assess the economic burden, aerodigestive, and pulmonary outcomes as well as functional developmental outcomes.

GRANTS

This study was supported in part by funding from National Institute of Diabetes and Digestive and Kidney Diseases Grant P01-DK-068051 (S. R. Jadcherla, I. M. Lang, and R. Shaker).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.R.J. conception and design of research; S.R.J. and K.A.H. performed experiments; S.R.J., K.A.H., S.S., B.J.C., and J.L.S. analyzed data; S.R.J., K.A.H., S.S., B.J.C., J.L.S., and R.S. interpreted results of experiments; S.R.J., K.A.H., S.S., B.J.C., J.L.S., and R.S. drafted manuscript; S.R.J., K.A.H., S.S., B.J.C., J.L.S., and R.S. edited and revised manuscript; S.R.J., K.A.H., S.S., B.J.C., J.L.S., and R.S. approved final version of manuscript; K.A.H. and S.S. prepared figures.