Early childhood is a critical period for physical growth and cognitive development¹. Children must be able to consume sufficient nutrients during this time to achieve optimal physical and neurological growth^2,3. Dysphagia during early childhood interrupts a child from learning normal swallowing skills and can lead to long-term medical complications, including malnu-trition, recurrent pneumonia, chronic lung disease, and developmental delay of the brain and all body systems^4-6. As such, there is a growing need for future research to better understand the nature of dysphagia among pediatric populations.

Aspiration, which is defined as the entry of bolus below the true vocal folds, is a major concern for children with dysphagia because aspiration is reported as the primary etiology leading to recurrent wheezing, recurrent pneumonia, and severe impairment of lung function in pediatric populations^7-9. Thus, the main purpose of swallowing evaluations is to identify the pathophysiology that contributes to or explains a risk of aspiration¹⁰.

Temporal measurements from Videofluoroscopic Swallowing Study (VFSS) have been used to examine pathophysiological factors that increase the risk of aspiration in adult populations^11-13. Temporal measure-ments provide quantitative values to examine physio-logical events during oropharyngeal swallowing. For example, the duration measures represent the amount of time each physiological swallowing event lasts¹⁴. On the other hand, the onset of movement measures represent the time at which a specific physiological event initiates in relation to bolus position^15,16. Quan-titative data of temporal measurements allow clinicians to identify pathophysiological factors associated with aspiration risk by comparing temporal characteristics to the temporal range of normal swallowing¹⁷. Namely, temporal measurements can serve as a tool for iden-tifying patterns of swallowing that may classify an individual as being at high risk for aspiration.

As young children have differences in anatomy and are less neurologically developed than adults, their swallowing mechanism and signs of dysphagia may be different from those of adults^18,19. However, in spite of the benefits of temporal measurements, it is still unclear whether temporal measurements are applicable to pediatric populations as a means of identifying pathophysiology associated with aspiration during oropharyngeal swallowing.

Some research has used temporal measurements to understand swallowing functions in pediatric popula-tions^19,20. However, it still remains unknown whether an investigation of temporal characteristics of the entire swallowing process can describe relations among physiological events in each swallowing stage in pedia-tric populations. Also, previous research focused par-ticularly on either infants during bottle feeding²¹ or children at a wide age range^20,22 to examine swallo-wing functions. Young children are good candidates for swallowing research because they have not yet acquired a full maturity of oropharyngeal structures, and they also have been introduced to various foods after a transitional period from bottle feeding to spoon feeding²³. Therefore, the present study examined tem-poral characteristics that represent pathophysiolo-gical factors of high risk of aspiration across the entire swallowing process, including both the oral and pharyngeal stages in young children ranging in age from 1 year to 3 years and 11 months.

The specific aims of this study were to examine the difference in each temporal measure between young children with aspiration and without aspiration and to investigate the effect of bolus consistency on the temporal measures. The objective numeric data with different bolus consistency of the current study will serve as a reference to distinguish swallows with aspi-ration from swallows without aspiration. In addition, the identification of pathophysiological factors among temporal characteristics will aid clinicians in develo-ping appropriate and targeted interventions in the hope of reducing the risk of aspiration and its medi-cal complications.

1. Participants

The study included 71 children who were referred to VFSS in the department of pediatric rehabilitation medicine from January 2011 to December 2018 at the tertiary hospital in South Korea. Among 71 children, 42 met the following inclusion criteria: a) range in age from 1 year to 3 years and 11 months, b) normal oral and pharyngeal structures, c) no history of VFSS, and d) fed by spoon during VFSS. Of the excluded 29 children, two were younger than 1 year in age, 15 had anomalies related to swallowing disorders (e.g., cleft palate/lip, laryngeal cleft, laryngomalacia, tracheal/ esophageal stenosis, and tracheoesophageal fistula), six were only bottle fed, and six failed to swallow during VFSS. The children included 23 females and 19 males. The mean age was 24.7 months (standard devi-ation, 8.8 months; range, 12 to 46 months).

The children were divided into two groups according to the presence of aspiration. Aspiration was defined as the entry of bolus below the true vocal folds. A child who had aspirated on at least one bolus (i.e., on at least one trial among four trials) was classified into group 1, or aspirators. Sixteen children were aspirators whose mean age was 21.2 months. A child who had not shown aspi-ration on any bolus (i.e., no aspiration on all four trials) was classified into group 2, or non-aspirators. Twenty- six children were non-aspirators. The mean age of the non-aspirators was 26.3 months.

Etiologies and other clinical information (i.e., enteral feeding, tracheostomy, and history of pneumonia) were summarized according to the group in Table 1. The most prevalent etiology in the patients was neurolo-gical disorders (86%, 36 of 42), including hypoxic isc-hemic encephalopathy (14), cerebral palsy (3), encep-halitis (5), brain tumor (2), neuromuscular disease (2), seizure (1), spinal cord disease (1), pachygyria (1) and unspecified (7). Three patients (7%) had genetic syn-drome, including Edward syndrome and Down syndrome. Two patients (5%) had cardiorespiratory disorders, including primary pulmonary hypertension and heart transplant status. An etiology of one child (2%) was not identified.

Table 1 . Demographics and clinical information.

Variables	Aspirators (n=16)	Non-aspirators (n=26)
Age, range (mo)	21.2±8.3, 12-42	26.3±8.4, 14-46
Sex
Female	9 (56)	14 (54)
Male	7 (44)	12 (46)
Etiology
Neurological disorders	15 (94)	21 (81)
Cardiorespiratory disorders	1 (6)	1 (4)
Genetic syndromes	0 (0)	3 (11)
Unspecified	0 (0)	1 (4)
Enteral feeding
No tube	2 (12)	9 (35)
NG tube	6 (38)	12 (46)
PEG	1 (6)	1 (4)
No report	7 (44)	4 (15)
Tracheostomy
No	12 (75)	19 (73)
Yes	1 (6)	4 (15)
No report	3 (19)	3 (12)
History of pneumonia
No	5 (31)	18 (70)
Yes	3 (19)	4 (15)
No report	8 (50)	4 (15)

Values are presented as number (%)..

NG tube: nasogastric tube, PEG: percutaneous endoscopy gas-trostomy..

2. VFSS procedures

VFSS was recorded at the swallowing clinic, depart-ment of pediatric rehabilitation medicine in the tertiary hospital in South Korea. Swallowing was measured using either digital videofluoroscopy or C-arm at 30 frames per second. Each pediatric participant was seated upright in a chair for the examination. The fluoroscopic tube was focused in the lateral plane on the oral cavity (the lips anterior to the pharyngeal wall posterior) and the nasopharynx (superior) to below the upper esophageal sphincter (UES) area (inferior). Each participant swallowed two boluses of different consistencies: thin liquid (i.e., zero-level of Interna-tional Dysphagia Diet Standardisation Initiative (IDDSI)) and puree (i.e., four-level of IDDSI). The thin liquid consisted of a diluted barium solution (35% w/v). The puree was a mixture of Yoplait^Ⓡ and barium sulfate. Each type of consistencies was fed twice (i.e., the total number of swallowing trials was four). All participants were fed with a spoon. The volume of bolus for each trial was 2 ml. With digital videofluoroscopy, frame- by-frame images were acquired to digital imaging files (.avi) using a computer-based image processing sys-tem equipped with a digital computer frame grabber board (Pinnacel Studio MovieBox DV, Pinnacle Sys-tem, Inc., Mountain View, CA, USA; Pegasus HD/SD Board, Grass Valley Inc., Honorine, France) and image processing software (Pinnacle Studio 9.0, Pinnacle Sys-tem, Inc.; EDIUS 4.5, Grass Valley Inc., Hillsboro, OR, USA). The X-ray voltage was set at a 40-kV peak, which allowed the soft tissue of the laryngeal and pharyngeal structures to be visible. With a C-arm (Mo-bile C-ARM, Ziehm Vision, Ziehm Imaging GmbH Inc., Nurnberg, Germany), swallowing video was recorded as digital imaging files (.avi) using software (SKY HD RED). The X-ray voltage was set at a 63-kV peak. In addition, the investigator collected demographic infor-mation (i.e., date of birth, age, sex, and birth weight), clinical information (i.e., height/weight on VFSS date, etiology, history of prematurity, history of pneumo-nia, onset of dysphagia, congenital anomaly, current diet, types of tube feeding, and presence of T-can-nula), and VFSS information (i.e., VFSS date, consis-tencies used during VFSS, presence of aspiration/ penetration, consistencies penetrated/aspirated, and esophageal stage report) by reviewing the medical re-cord. A retrospective review of VFSS data and medical chart was performed with the approval of institutional review board of Seoul National University Hospital (IRB: H-1906-144-1043) and Ohio University (IRB: 19-E-215).

3. Experimental procedures

Adobe Premiere Pro CS5.5 with a 100-millisecond video timer was used for slow-motion and frame-by- frame analyses of VFSS videos. First, the investigator verified the presence of aspiration for the liquid and puree swallows. Second, the investigator documented swallowing problems during oropharyngeal swallowing such as drooling, poor bolus formation, premature spil-lage, residue in the valleculae and/or pyriform sinus, and time to aspiration. Then, the investigator condu-cted temporal measurements.

The current study focused on oropharyngeal temporal measures to understand the swallowing physiology during oropharyngeal swallowing. The valleculae were considered to be a referent point triggering pharyngeal swallow in young children^19,21. Thus, this study used the valleculae to calculate the measures that represent initiation of pharyngeal swallowing and initiation of laryngeal closure. Operational definitions of each measure were listed according to the swallo-wing stage as follows.

Oral Transit Time (OTT): the time from tongue tip elevation to the arrival of the bolus head at the valle-culae¹².

Delayed Pharyngeal Swallow (DPS): the time of the initiation of laryngeal elevation in relation to the fra-me at which the bolus head reaches the valleculae²⁴.

Pharyngeal Transit Time (PTT): the time from the arrival of the bolus head at the valleculae to the bolus tail passing the upper esophageal sphincter (UES)¹². Initiation of Laryngeal Closure (ILC): the time from the bolus head reaching the valleculae to the first contact of the arytenoids and the epiglottis²⁵. Duration to achi-eve Laryngeal Closure (D to achieve LC): the time from the initiation of laryngeal elevation until the first contact of the arytenoids and the epiglottis²⁶. Laryngeal Clo-sure Duration (LCD): the time from the initiation of laryngeal closure until the final contact of the aryte-noids and the epiglottis^12,16. Duration of UES Opening (DUESO): the time from the initiation of UES opening to the closure of the UES¹⁴.

The following eight references were observed to cal-culate the measures: (1) initiation of tongue elevation, (2) bolus head reaching the valleculae, (3) bolus tail passing the UES, (4) initiation of laryngeal elevation, (5) initial contact of arytenoids and epiglottis, (6) final contact of arytenoids and epiglottis, (7) initiation of opening of the UES, and (8) closure of the UES.(Fig. 1.)

Figure 1. Reference points for the temporal measures.

4. Statistical analysis

A two-level hierarchical linear model (HLM) was used to determine 1) whether there is a difference between the aspirators and the non-aspirators in the oropha-ryngeal temporal measures and 2) the effect of bolus consistency on the oropharyngeal temporal measures. The within-participant variable was modeled at Level 1 and the between-participant variable was modeled at Level 2. Consistency variables (i.e., liquid and puree) were used as a Level 1 predictor, and group variables (i.e., aspirator and non-aspirator) were used as a Level 2 predictor. The dependent variables were the oro-pharyngeal temporal measures, and each was mode-led separately. The interaction between the group and consistency was also analyzed to investigate if there was group-by-consistency interaction for the oropha-ryngeal temporal measures. An F-test was used to deter-mine the significance of all analyses. A Holm-Bonfe-rroni correction was used to adjust P-value for the multiple comparisons with the oropharyngeal temporal measures. SPSS v25.0 (IBM Corp., Armonk, NY, USA) was used for all the statistical analyses.

1. Reliability

Intra-rater and inter-rater reliabilities were com-pleted for the temporal measures using intraclass correlation coefficient (ICC). For intra-rater reliability, the first rater randomly selected 10% of swallows with aspiration and 10% of swallows without aspiration and analyzed the swallows. A significant correlation bet-ween the first and second ratings was found (r=0.99 for OTT, r=0.99 for DPS, r=0.99 for PTT, r=0.99 for ILC, r=0.94 for D to LC, r=0.99 for LCD, r=0.97 for DUESO, P<0.01). For inter-rater reliability, a second rater analyzed the same swallows. Significant correla-tions between the raters were found (r=0.99 for OTT, r=0.99 for DPS, r=0.99 for PTT, r=0.99 for ILC, r=0.86 for D to LC, r=0.98 for LCD, r=0.98 for DUESO, P<0.01).

2. Temporal measures between aspirators and non-aspirators

Table 2 presents the mean and standard error of OTT and DPS. There was no significant difference in the OTT between the aspirators and the non-aspirators. In the oropharyngeal transit stage, the DPS showed a significant difference between the two groups. The aspirators showed 0.51 s longer DPS as compared to the non-aspirators (F[1,26]=24.81, P<0.01).

Table 2 . Comparison of the oropharyngeal temporal measures between the two groups.

	Aspirator	Non-aspirator	P-value	Holm-Bonferroni adjusted P-value
OTT	0.50±0.05	0.38±0.04	.121	.013
DPS	0.61±0.08	0.10±0.06	.000*	.007
PTT	1.05±0.09	0.64±0.06	.001*	.006
ILC	0.76±0.08	0.26±0.06	.000*	.008
D to LC	0.15±0.01	0.16±0.01	.382	.025
LCD	0.63±0.04	0.57±0.03	.329	.017
DUESO	0.28±0.03	0.34±0.02	.065	.050

Values are presented as mean±standard error..

OTT: oral transit time, DPS: delayed pharyngeal swallow, PTT: pharyngeal transit time, ILC: initiation of laryngeal closure, D to LC: duration to achieve laryngeal closure, LCD: laryngeal closure duration, DUESO: duration of upper esophageal sphincter opening..

*P.

Table 2 provides the mean and standard error of PTT, ILC, D to LC, LCD, and DUESO. There was a significant difference between the aspirators and the non-aspi-rators in the PTT. The aspirators had 0.41 s longer PTT than the non-aspirators (F[1,23]=14.98, P<0.01). In the ILC, the means between the two groups were signifi-cantly different. The aspirators had 0.50 s longer ILC than the non-aspirators (F[1,26]=23.04, P<0.01). In the D to LC, LCD, and DUESO, there were no signific-ant differences between the aspirators and the non- aspirators.

3. Temporal measures between liquid and puree swallows

The second aim of this study was to examine the effect of bolus consistency on the temporal measures. The means and standard errors of the oropharyngeal temporal measures according to bolus consistency are described in Table 3. Liquid swallows were prone to have longer DPS and ILC than puree swallows. However, there was no significant difference between the two consistencies.

Table 3 . Comparison of the oropharyngeal temporal measures by the bolus consistency.

	Liquid	Puree	P-value	Holm-Bonferroni adjusted P-value
OTT	0.42±0.05	0.46±0.05	.577	.025
DPS	0.44±0.07	0.28±0.07	.054	.008
PTT	0.93±0.07	0.76±0.07	.075	.010
ILC	0.60±0.07	0.43±0.07	.040	.007
D to LC	0.15±0.01	0.16±0.01	.801	.050
LCD	0.63±0.03	0.57±0.03	.079	.013
DUESO	0.33±0.02	0.29±0.10	.114	.017

Values are presented as mean±standard error..

OTT: oral transit time, DPS: delayed pharyngeal swallow, PTT: pharyngeal transit time, ILC: initiation of laryngeal closure, D to LC: duration to achieve laryngeal closure, LCD: laryngeal closure duration, DUESO: duration of upper esophageal sphincter opening..

After the analyses to answer the first and second research questions, the interaction between the group and consistency was analyzed. No significant intera-ctions were observed between the group and consistency in the oropharyngeal temporal measures.

1. Temporal measures between aspirators and non-aspirators

The first aim of this study was to investigate the difference in the temporal characteristics using the oro-pharyngeal temporal measures between young children with aspiration and those without aspiration. The result indicated that delayed initiation of pharyngeal swallowing was associated with a high risk of aspira-tion. Prolonged DPS in the aspirators seemed to be associated with oral dysfunction. Previous studies repor-ted that dysfunctions in the oral stage had a negative effect on the timely initiation of pharyngeal swallo-wing^10,13,27. In the current study, the aspirators showed poor lip closure, reduced tongue elevation, and poor bolus controls. These oral dysfunctions may have contri-buted to the delayed initiation of pharyngeal swallowing in the aspirators. It was interesting that some of the children in the non-aspirators did not aspirate des-pite showing oral dysfunction. DPS was a response measure that examined the onset of swallowing phy-siology in relation to bolus movement. It is believed that the bolus transition during the oral stage was pro-longed in the non-aspirator group, but it did not lead to aspiration. It may be related to compensatory behavior such as the movement for airway protection (i.e., initiation of laryngeal elevation) occurring faster and more effectively. The current study only focused on the time measures (i.e., duration and onset of phy-siological events). Further investigation needs to exa-mine the mechanisms of how this compensatory beha-vior occurs to protect the airway to understand the swallowing function of children with dysphagia but no aspiration.

Previous research in adult and pediatric popula-tions with swallowing disorders reported that increased PTT was related to pharyngeal muscular weakness^20,28,29. The result of this study may suggest that weak pha-ryngeal propulsion in the aspirators was associated with a risk of aspiration. Reduced peristaltic constri-ction of pharyngeal muscles resulted in residue in the pharynx and increased aspiration risk after the swal-low^12,30. Four children in the current study showed aspirations after the swallow. They also showed diffuse pharyngeal residue after the swallow. Two of the four children tried to do repeated swallows to clear the pharyngeal residue. However, it was not effective. The other two children did not even try to do repeated swallows. The present findings reflected that the pro-longed PTT may be related to both pharyngeal mus-cular dysfunction and sensory deficits that increased the risk for aspiration in children with dysphagia, as seen in adults with dysphagia^31,32.

ILC was the other measure that had a significant difference between the aspirators and the non-as-pirators. Delayed initiation of laryngeal closure has been reported as a main pathophysiological factor contributing to aspiration in the pediatric swallowing literature^20,22,33. Longer ILC in the aspirators of the current study demonstrated that the children who failed to achieve appropriate coordination of lary-ngeal closure as the bolus entered the pharynx were at high risk of aspiration.

The aspirators and non-aspirators did not differ in D to LC, LCD, and DUESO. The D to LC did not seem to be applied to the pediatric population. The D to LC in the current study ranged from 0.11 to 0.18 s regardless of the group. The D to LC was distributed across a relatively small range of mean values, as com-pared with the adult swallowing study (from 0.18 s to 0.36 s with liquid bolus)²⁶. The more superior and anterior location of the pediatric larynx may influence the reduced length of maximum hyolaryngeal excur-sion as compared to the adult larynx³⁴. Therefore, D to LC may be too short to identify ineffective com-pletion of laryngeal closure in the pediatric population.

Steele and Cichero¹⁰ stated that reduced LCD can be a risk factor for aspiration in combination with delayed initiation of pharyngeal swallowing and pro-longed pharyngeal transition in adults with dysphagia. This pattern among LCD, DPS, and PTT was incon-sistent in the aspirators of the current study. The inconsistent patterns can be related to vocal fold clo-sure. The vocal folds seemed to stay closed in several aspirators even though the entrance of the airway was opened. As it is difficult to visualize the vocal fold closure during VFSS, further research with other ins-truments, such as fiber-optic endoscopic examination, is needed to understand other physiologies related to the airway protection mechanism in pediatric popu-lations.

2. Temporal measures between liquid and puree swallows

The second research aim of the study was to exa-mine the effect of bolus consistency (i.e., liquid and puree) on temporal characteristics of swallowing phy-siology in the pediatric population with dysphagia. Although the difference between the consistencies was not statistically significant, liquid swallows appeared to have longer DPS and ILC than puree swallows, which was an opposite trend from the previous fin-dings in adults with normal swallowing³⁵ and adults with dysphagia³⁶.

One possible explanation for the present finding was the property of cohesiveness in the puree. Given that a small bolus volume was used (2 ml in the cur-rent study), the puree bolus may be more cohesive and move faster than liquid. As the liquid bolus spreads in the oral cavity, it may be hard for children with dysphagia to control the liquid bolus. A part of the liquid bolus reached the valleculae, whereas the entire puree bolus contained the valleculae during swallo-wing in the current study. It may be assumed that mechanoreceptors in the epiglottis and valleculae may receive relatively more sensory information with puree bolus than with liquid bolus, leading to a faster res-ponse of laryngeal closure^37,38.

Appropriate sensory input sent to the cortex and brainstem and motor function of oropharyngeal stru-ctures are required for safe and effective swallowing^38-40. The children in the current study had complex me-dical diagnoses. Thus, it is assumed that the disru-ptions of neurological development impeded appro-priate development in swallowing function. Children in the study might not be able to respond effectively to bolus with different viscosity. Future investigation is needed to examine the effect of bolus consistency on temporal characteristics in children without dys-phagia. Based on findings of the future research, it would be understood how children without dysphagia at the midpoint of anatomical and neurological deve-lopment could adjust swallowing physiology accor-ding to different bolus consistency by using objective temporal data.

3. Limitations

The findings of this study need to be interpreted with caution due to the following limitations. Given the small number of participants, it was not possible to have the same number of children in each group. In addition, these children suffered from different etio-logies. This study used pre-recorded VFSS data for analysis which can lead to missing data. Referral bias is possible because all data were derived from one tertiary hospital. Lastly, the comparison with norma-tive data was limited due to a dearth of literature on temporal measurements in normal children. Despite these limitations, the objective temporal data in this study provide a basis for understanding swallowing physiology in pediatric dysphagia. Furthermore, qua-ntitative measures can contribute to future comparisons with other studies that examine swallowing physio-logy in children without dysphagia or in children with various medical conditions.

This retrospective study examined a number of temporal measures of the entire swallowing process in young children with dysphagia using frame-by- frame analysis. Also, comparisons of the temporal measures between the aspirators and the non-as-pirators were performed. This study showed that the prolonged time values in DPS, PTT, and ILC were associated with the risk of aspiration. As clinical aspects, findings of the current study indicated that the risk of aspiration increased when a child showed delayed initiation of pharyngeal swallowing, delayed initiation of laryngeal closure, and poor bolus tran-sition in the pharynx during oropharyngeal swallowing. These findings will help clinicians suggest appropriate management strategies when they observe the patho-physiological factors associated with aspiration risk. In terms of the effect of bolus consistency on the temporal characteristics, there was no significant dif-ference between the liquid bolus and the puree bolus in the oropharyngeal temporal measures. The effect of bolus consistency on the temporal characteristics should be investigated further with various consis-tencies as well as bolus volumes.

The authors have no relevant financial or non- financial interests to disclose. 

This study was approved by the institutional review board of Seoul National University Hospital (IRB: H- 1906-144-1043) and Ohio University (IRB: 19-E-215). Informed consent from the patients was waived by both IRBs.

No funding was received for conducting this study.

In accordance with the principles of responsible authorship, the roles and contributions of each author are detailed as follows: study conceptuali-zation and design: Yunju Han, Youngsun Kim; resou-rces: Byung-Mo Oh; IRB administration at Seoul Na-tional University Hospital: Byung-Mo Oh, You Gyoung Yi; IRB administration at Ohio University: Yunju Han, Youngsun Kim; data collection: Yunju Han, You Gyoung Yi; data curation: Yunju Han; data analysis and interpretation: Yunju Han; data review and super-vision: Byung-Mo Oh, Youngsun Kim; draft manu-script preparation: Yunju Han. All authors reviewed the results and approved the final version of the manuscript.