Dysphagia refers to difficulties in swallowing solids or fluids and is estimated to affect 1% of the pediatric population, although its incidence may increase to up to 80% among children with cerebral palsy and other neurodevelopmental disorders.1

In children, dysphagia is usually a functional disorder, most commonly involving the mouth and pharynx. It is associated with predisposing conditions, such as neurologic abnormalities, musculoskeletal disorders, malformations of the oral cavity, pharynx, or esophagus, prematurity, prolonged non-oral feeding and chronic lung disease.2

The diagnosis is made through a clinical evaluation and instrumental methods. A bedside swallow assessment should be conducted by a speech language pathologist.3,4 Diagnostic tests are indicated if any abnormalities are detected in the clinical assessment, and the videofluoroscopic swallow study (VFSS) is considered the imaging gold standard for evaluation of dysphagia.

The videofluoroscopic swallow study is a dynamic imaging evaluation of swallowing function in which a series of lateral projection images is acquired while the patient swallows contrast medium in various consistencies and volumes. It allows assessment of the safety and efficiency of oral intake in the patient in real time.5 It can also be used to assess the effectiveness of the therapeutic approach implemented to correct the observed impairment and to adjust treatment to the recommended postural adjustments and/or adaptive feeding devices.6 Based on the severity observed in the test, oral intake recommendations are provided to the patient, alone or in addition to swallowing therapy/rehabilitation.

Although the radiation dose in VFSSs is relatively low, since the test involves exposure to ionizing radiation and pediatric patients are particularly sensitive to its effects, reducing the dose according to the “as low as reasonably achievable” (ALARA) principle should be considered a priority.7

To guide dose optimization according to the ALARA principle, the International Commission on Radiological Protection(ICRP) has promoted the use of diagnostic reference levels (DRLs) to prevent the delivery of radiation amounts in excess of the effective dose that do not contribute any additional useful information.7 In general, DRL values are set at the third quartile (75th percentile) of the median values used in a sample of procedures, although use of the median has also been recommended. Use of local DRLs, when available, is recommended, using national DRLs in their absence8 or, if neither are available, European DRLs.

Air kerma-area product (PKA) values are not always available to set DRLs, in which case it is possible to the dose-area product (DAP) as a value equivalent to the PKA. Since radiology terms are complex, in this article we will use the terms DAP, radiation dose or received dose interchangeably to refer to the DAP, which is the value that is actually obtained in VFSSs.

In Europe, the first pediatric imaging DRLs (PiDRLs) were published in 2018, but only a couple of isolated studies were available to develop the reference levels for fluoroscopy.9 In addition, there are no national DRLs for pediatric patients in Spain, which precludes adequate optimization of the technique and comparisons across centers in Spain.

For this reason, we conducted a study with the aim of describing the radiation dose and the duration of exposure in patients that underwent VFSSs at a tertiary care children’s hospital and comparing them to the European DRLs, in addition to assessing the association between VFSS results and different clinical variables.

We conducted a retrospective and descriptive study through the review of the data included in the documentation and reports of VFSSs conducted in a tertiary care hospital over a seven-year period (2015–2022).

The clinical variables under consideration were: age (and the age groups recommended for setting PiDRLs), underlying disease, degree of neurologic impairment through the Gross Motor Function Classification System (GMFCS).10

Modified swallow studies were performed with a Siemens Iconos R200 series 4501 system, which allows lateral plane imaging. Fluoroscopy was performed at a voltage of 50–56 kV, with a current of 0.2 to 0.4 mA and at a rate of 12.5 (FPS) frames per second. The object-image distance was 1 m. Collimation was performed by a radiology department nurse under the supervision of a radiologist. The different consistencies were prepared with a water-soluble contrast medium (Visipaque) that is flavorless to improve palatability. A thickener (Resource Clear) was used to prepare nectar-thick and spoon-thick consistencies, which, along with thin liquid, were the tested consistencies. The test protocol included these three consistencies, all administered at three different volumes in increasing order, measured with a syringe, with volumes based on age (2, 4 and 6 mL in children < 3 years; 3, 5 and10 mL in older children). If a volume proved not safe on account of aspiration, the test did not progress to greater volumes. The same applied to consistencies: if aspiration occurred with nectar-thick consistency, the following step with thin liquid was not performed, as it would be less safe. In infants aged less than 4 months, the test started with the liquid consistency and, if there was evidence of dysphagia, the nectar consistency was tested.

During the test, one or both caregivers were allowed into the room for the administration of the different consistencies, in addition to toys or electronic devices to produce a more relaxed atmosphere so the test could be completed quicker.

Once the VFSS concluded, the following was determined: presence or absence of dysphagia, severity of dysphagia (mild: small-volume [<10% ingested volume] penetrations or aspirations, moderate: penetrations or aspirations of 10% to 50% of ingested volume, severe: penetrations or aspirations of more than half of the ingested volume), type of dysphagia (oral, pharyngeal, esophageal or a combination thereof), type of abnormality (efficiency, safety or both), detection of aspiration/penetration (aspiration, penetration, or both), type of aspiration (silent, with secondary cough, both silent and with cough), score in the Bolus Residue Scale (BRS),11 radiation dose (in cGy/cm2), and duration of radiation exposure (in seconds).

The statistical analysis was conducted with the software package SPSS version 26 (SPSS Inc; Chicago, IL, USA.). We used the Shapiro-Wilk test and the Kolmogórov-Smirnov test to assess normality. In the descriptive analysis, we summarized quantitative data with the median and interquartile range, as the variables did not follow a normal distribution, producing charts and tables to represent the results. We used the Pearson correlation coefficient to compare quantitative variables, the χ2 test to compare qualitative variables, and the Kruskal-Wallis test with the Bonferroni correction to assess for differences between groups in comparisons of multiple quantitative and qualitative variables. We considered P values of less than 0.05 statistically significant.

We obtained the informed consent of the parents for participation in the study. The study protocol was approved by Clinical Research Ethics Committee of Aragón (January 12, 2022; minute no. 01/2022; code PI12/512).

The study included data for 322 patients, whose clinical characteristics are detailed in Table 1, where it can be seen that more than two thirds had underlying neurologic disease.

A total of 221 patients (68.6%) received a diagnosis of dysphagia, classified as mild in 95 cases (43%), moderate in 86 (38.9%), and severe in 40 (18.1%). Classifying cases by the affected phase of the swallow, 143 patients had oropharyngeal dysphagia (65%), 36 pharyngeal dysphagia (16%), 13 oral dysphagia (6%), 9 esophageal dysphagia (4%), 3 pharyngoesophageal dysphagia (1%), and in 17 patients (8%), all phases were affected. The test detected aspirations/penetrations in 179 patients (55.6%), with aspiration only in 45 patients (14%), penetration only in 73 patients (22.7%) and both in 61 patients (18.9%). It is worth noting that most aspirations were silent, that is, occurred in absence of apparent secondary cough or choking and could only be detected by videofluoroscopy (76 patients, 71.7% of aspirations), with patients with neurologic disease accounting for 90.3% (n = 65) of these cases.

Neurologic impairment was associated with a higher frequency of dysphagia (81.5%) and severe dysphagia (90% of patients with severe dysphagia were in the neurologic disease group) compared to gastrointestinal or respiratory disease, and these differences were statistically significant (P <  .05). We also found an association between more severe neurologic impairment assessed with the GMFCS and an increase in the proportion of patients with aspiration, which reached 75.5% in patients classified as grade V (Fig. 1). These differences were statistically significant (P < .05).

Figure 1.

Percentage distribution of patients with aspirations according to the Gross Motor Function Classification System category.

As regards the primary outcomes, the median DAP was 9.2 cGy/cm2 (interquartile range [IQR], 5.1−15) and the median exposure time was 123 s (IQR, 85−167.3). We assessed the correlation between the quantitative variables (age, exposure time and DAP) with the Pearson correlation coefficient, r. We found statistically significant correlations between exposure time and dose (P < .05, two-tailed) with a correlation coefficient of 0.56 and a coefficient of determination, R2, of 0.31; that is, a moderate average correlation, with the DAP increasing by 0.09 cGy/cm2 with each additional second of exposure during the procedure and 31% of the DAP variance explained by the duration of the procedure.

We also found a statistically significant association between age and DAP (P <  .01 bilateral), with an r of 0.3 and an R2 of 0.09; that is, a weak positive correlation. The results showed that age explained 9% of the DAP variance (Fig. 2).

Figure 2.

Scatter plot with best-fit line for the association between dose-area product (DAP) and age.

We used the Kruskal-Wallis test with the Bonferroni correction to analyze the association between primary and secondary variables. In the comparison by age group, we found statistically significant differences in the received dose (P < .05; Fig. 3), but not in the duration of exposure, with the dose increasing with increasing age. Specifically, we found differences between patients aged less than 1 month and all other groups (P < .05), and between patients aged 1 month to 4 years and all other groups (P < .05). We found no significant differences between the rest of the groups (with a P = .051 in the group aged 4–10 years versus the group aged 14–18 years).

Figure 3.

Median dose-area product (DAP) by age group.

Table 2 summarizes the results of the analysis of the primary outcomes and the rest of the variables. Starting by the type of underlying disease, we found significant differences between the group of patients with neurologic disease compared to patients with gastrointestinal or respiratory disease (P < .05 for both DAP and time of exposure).

In relation to the degree of neurologic impairment (assessed by means of the GMFCS), we found differences in both exposure time and dose (P < .05), with both increasing with increasing neurologic impairment. Specifically, there were significant differences in exposure time between grades I and III as well as grades I and V (P < .05). When it came to the dose of radiation, there were significant differences between grades I and III, grades I and IV, grades I and V, and grades II and V (P <  .05).

We also found differences in the primary outcomes based on the severity of dysphagia. When it came to exposure time, there were significant differences between patients without dysphagia and patients with dysphagia of any severity (P <  .05), as well as between patients with mild versus moderate dysphagia (P <  .05). With respect to dose, there were significant differences between patients without dysphagia and patients with dysphagia of any severity (P <  .05), as well as between patients with mild versus moderate dysphagia (P <  .05).

The analysis by type of abnormality, considering efficiency and safety, also identified significant differences. Specifically, there were differences in both dose and time of exposure between patients without impairment and patients in which both safety and efficiency were affected (P <  .05), as well as between patients in whom only safety was affected versus patients in whom both efficiency and safety were affected (P <  .05). The dose was higher in patients with reduced efficiency and safety compared to patients in whom only efficiency was affected. Exposure was longer in patients with reduced swallow efficiency compared to patients without abnormalities and to patients in whom only safety was affected. All these differences were statistically significant (P <  .05). We did not find differences in the remaining intergroup comparisons.

When we compared the primary outcome results to the BRS scores, we found significant differences, with both the dose and the time of exposure increasing with increasing BRS score. Specifically, we found significant differences in the DAP and the time of exposure between patients with a BRS score of 0 compared to patients with scores of 3, 4, and 6 points (P <  .05). We found no other differences between groups.

Recent studies have reported that dysphagia is present in 75%–85% of neurologic patients, which is consistent with the findings of our study.12 In addition, the frequency of aspiration observed in our patients (55.6%) was similar to the frequency described in the previous literature (approximately 50%).2,13 The association between the degree of neurologic impairment (assessed by means of the GMFCS) and both the severity of dysphagia and the frequency of aspiration has been consistently reported in recent studies.13–16 Neurologic patients are more likely to exhibit dysphagia and have more severe dysphagia, and it is worth noting that increased impairment is associated with poorer patient cooperation during the procedure, which in turn increases exposure time and therefore the received dose, a finding that has been reported in several studies.14–18

Another salient finding that, while expected, may seem paradoxical was the longer duration of the procedure and higher received dose in patients with moderate dysphagia compared to those with severe dysphagia. This is explained that the test is cut short in patients with severe dysphagia upon detection of swallow abnormalities with a serious impact on safety, such as clinically significant aspiration, which precludes assessment at riskier consistencies and therefore results in shorter exposures and lower doses compared to patients with moderate dysphagia.18

The findings of our study are consistent with those of similar studies that have found a direct correlation between exposure time and the DAP, confirming that longer screening times result in an increased radiation dose (r = 0.56; R2 = 0.31),17–22 which underscores the importance of minimizing the duration of the exam. On the other hand, the weak but significant association between age and the DAP, also reported by Ko et al.,22 could be related to the increase in the dimensions of the area to be screened with increasing patient age, in addition to the increase in the volumes and consistencies administered for examination of older patients. Other studies that included data on weight or body surface area also found a positive correlation between them and DAP.18–22 In fact, Weir et al.18 found that body length, and, specifically, the proportion of the head and neck relative to the total body length, were associated with the received dose.

Another relevant aspect was the comparison of the observed DAP with the recommended European PiDRLs9 and to similar studies,18,20,21,23–27 as we obtained significantly lower values (Table 3). However, the absence of national DRLs and of specific protocols for the performance of VFSS in children complicates comparisons with other centers.

Using a lower frame rate (12.5 FPS) was one of the key methodological adjustments in our study. This modification results in a substantial reduction of the received dose, which is particularly relevant in pediatric patients, compared to the 30 FPS applied in most centers.11,19,23 This results in a reduction of practically a two-thirds in the radiation delivered over a given time interval, but, on the other hand, it yields only a third of the images that would be acquired in other centers. Bonilha et al.17 demonstrated that reducing the frame rate allowed a relevant decrease in dose without compromising the diagnosis, despite the limited number of images available for detailed analysis. In our hospital, to prevent loss of images and facilitate their review, the images are recorded. The debate whether the frame rate of 30 FPS applied in adults should be maintained in pediatric VFSSs is still heated and ongoing, with some centers using 15 FPS13,22,28 while others conduct studies seeking to demonstrate that a rate of 30 FPS offers greater benefits to the patients, even with the resulting radiation dose.23

The optimization of the procedure through strategies like adequate collimation is another crucial factor in minimizing the radiation dose. Collimation consists in the reduction of the area of the x-ray beam to the patient’s anatomy that requires visualization, thus reducing the volume of tissue that is irradiated.29 An adequate balance must be maintained between excessive collimation, which results in a field of view that is too small, and insufficient collimation, in which the expanded field does not offer any additional benefits yet increases the radiation dose unnecessarily. Other strategies used in our hospital (help from family members, distraction, pleasant environment) are key elements in optimizing the process.

The association of the level of experience of radiology technicians and the supervising radiologist with exposure time and DAP has also been studied, with evidence of lower doses and shorter exposure in centers with more experience.30,31 In our study, a radiologist was present during the performance of VFSSs, so there was greater control of radiation doses and fluoroscopy times compared to centers where the procedure is performed by otolaryngologists or rehabilitation physicians alone.32

In conclusion, the VFSS continues to be the gold standard for swallowing assessment, as it allows for objective evaluation. Optimizing the process through strategies like collimation or a lower frame rate and leveraging the experience of the center are crucial factors in reducing the radiation dose received by patients. Given the heterogeneity in the radiation doses reported in the literature, it would be advisable to implement standardized protocols for pediatric VFSS in order to reduce the amount of radiation to which patients are exposed, facilitate the comparison of data, and undertake the establishment of national diagnostic reference levels.

This research did not receive any external funding.