Effect of aspiration on the lungs in children: a comparison using chest computed tomography findings

The participants were 126 disabled children with severe motor disturbances and developmental retardation who received outpatient care for dysphagia at the Osaka Developmental Rehabilitation Center (Osaka, Japan) from 2012 to 2014 and who underwent FEES to evaluate swallowing function. Eighty-five children underwent FEES evaluations, after having undergone chest CT, as part of the standard medical practice, in the preceding week. They were included if their CT results could be analyzed. The exclusion criteria were as follows: (1) chest CT was administered more than 1 week before the FEES (36 participants) and (2) poor image quality due to body movements or other factors that made analysis difficult (5 participants).

All participants received nutrition orally: by total oral intake in 43 (50.6%) children and by the combined use of tube feeding in 42 (49.4%) children. Pneumonia history was confirmed from medical records and other clinical notes.

History of pneumonia was confirmed, based on the presence or absence of pneumonia in the patient’s medical records in the year before the FEES. Patients with this information in their records were considered as having a history of pneumonia.

Chest CT images were obtained using the Asteion TSX-021B CT scanner (Toshiba Medical Systems, Tokyo, Japan). The imaging parameters were 120 kVp, 1.375:1 pitch, and an automatically adjusted current of 100–150 mA. Axial images were acquired at 2-mm slice thickness.

The CT images were assessed by a respiratory organ specialist (A.U.) and pediatrician (T.K.). They had more than 20 years of clinical experience and were blind to the FEES results. When their assessments did not match, they both re-evaluated the images and discussed them until a consensus was reached. When assessment of an image was difficult because of artifacts in the images resulting from a patient’s inability.

to control body movement during imaging, the participants were excluded (n = 5). Based on the method described by Butler et al. [14], the images were assessed for parenchymal bands, bronchiolectasis, bronchial wall thickening, bronchiectasis, atelectasis, tree-in-bud pattern, intraluminal airway debris, fibrosis, and air trapping. Participants with any of these CT findings were considered as having pneumonia.

The aspiration and nonaspiration groups were compared for history of pneumonia and chest CT findings of pneumonia.

Among 85 participants, the FEES findings revealed food aspiration in 48 (56.5%) participants, saliva aspiration in 26 (30.6%) participants, and both food and saliva aspiration in 20 (23.5%) participants. Among 54 participants with aspiration, 40 (74.1%) participants had silent aspiration. History of pneumonia within 1 year was observed in 33 (38.8%) participants.

The CT image analysis revealed signs of pneumonia in 54 (63.5%) of 85 participants. The CT image findings were parenchymal bands (six [7.1%] participants), bronchiolectasis (27 [31.8%] participants), bronchial wall thickening (46 [54.1%] participants), bronchiectasis (two [2.4%] participants), atelectasis (17 [20.0%] participants), tree-in-bud pattern (eight [9.4%] participants), intraluminal airway debris (four [4.7%] participants), and other findings (four [4.7%] participants).

In the food aspiration/nonfood aspiration groups, 5/1 (10.4%/2.7%) participants had parenchymal bands; 13/14 (27.1%/37.8%) participants, bronchiolectasis; 27/19 (56.3%/51.4%) participants, bronchial wall thickening; 1/1 (2.1%/2.7%) participant, bronchiectasis; 11/6 (22.9%/16.2%) participants, atelectasis; 4/4 (8.3%/10.8%) participants, tree-in-bud pattern; 4/0 (8.0%/0%) participants, intraluminal airway debris; and 2/2 (4.0%/5.9%) participants, other findings.

In the saliva aspiration/non-saliva aspiration groups, 2/4 (7.7%/6.8%) participants had parenchymal bands; 11/16 (42.3%/27.1%) participants, bronchiolectasis; 20/26 (76.9%/44.1%) participants, bronchial wall thickening; 0/2 (0/3.4%) participants, bronchiectasis; 9/8 (34.6%/13.6%) participants, atelectasis; 1/7 (3.9%/11.9%) participants, tree-in-bud pattern; 3/1 (11.5%/1.7%) participants, intraluminal airway debris; and 3/1 (11.5%/1.7%) participants, other findings (Table 1).

A history of pneumonia, based on medical records from the previous year, was not correlated with the presence or absence of food or saliva aspiration (Table 2; p = 0.545 and p = 0.964, respectively). Food aspiration and chest CT findings of pneumonia were not correlated (p = 0.120). However, a significant correlation was observed with saliva aspiration (Table 3, p < 0.01). Logistic regression analysis revealed that saliva aspiration had a significant effect on bronchial wall thickening (odds ratio [confidence interval]: 4.231 [1.485–12.055]) and atelectasis (odds ratio [confidence interval]: 3.375 [1.124–10.131]). Multivariate logistic regression analysis with age and sex included as the moderator variables revealed similar effects. By contrast, food aspiration did not have a significant effect on any finding (Table 4).

We used FEES to assess aspiration in this study. The gold standard methods for evaluating swallowing functions are FEES and VFSS. Their detection rates of aspiration do not have a significant error [19, 20], and they are useful in children and adults [21‐23]. In our findings, 56% of the tested participants had aspiration, which did not substantially differ from the results in previous research [1]. Furthermore, 74% of the participants with aspiration exhibited silent aspiration. This finding is consistent with that of another report [5] demonstrating that silent aspiration was common among patients with aspiration. This finding in the present study indicated that using an FEES study is valid for assessing aspiration.

A contrast agent is unnecessary in FEES; therefore, food that is eaten normally can be used as the test food. Another major advantage of FEES is that aspiration can be confirmed, despite the absence of contrasted saliva, secretions, or other matter. Our results showed that, compared with food aspiration, saliva aspiration detected by FEES was correlated with inflammatory findings in the lungs. This finding indicated that FEES, which is able to assess saliva aspiration, is a useful swallowing function test.

In the present study, CT was used to evaluate inflammatory findings in the lungs; CT involves more radiation exposure and takes longer to conduct than does chest radiography and requires that body movements be controlled for a certain amount of time. Despite these issues, CT can detect dorsal and asymptomatic inflammatory findings. On account of its superior ability in detecting early stage lung changes and inflammatory findings, we considered it as the optimal method to carefully examine the relationship between aspiration and chest findings. Furthermore, for the current study, chest CT was conducted in a medical center that specializes in the care of disabled children. This center has specialized staff who are accustomed to interacting with disabled children and technicians who routinely image such patients. Therefore, except for the few patients who were excluded, we obtained images with sufficient precision for the assessments.

A history of pneumonia in the previous year was not correlated with the presence or absence of food or saliva aspiration. Weir et al. [24] examined the relationship between pneumonia history and the presence or absence of food aspiration (detected by VFSS) in children. The authors found a low correlation between food aspiration detected by VFSS and history of pneumonia, based on the World Health Organization definition, in the year before VFSS [24]. We used FEES but not VFSS to evaluate aspiration, but our results were similar to those of Weir et al. [24]. However, pneumonia was determined based on clinical symptoms such as fever or confirmed in medical records; therefore, a variation in the diagnoses could have affected the accuracy of the assessments.

Furthermore, because the time difference between the aspiration assessment and the onset of pneumonia was as much as 1 year, the assessments could have been affected by other factors that may have appeared between the onset of pneumonia and the assessment (e.g., a food-related intervention or onset of a different disease). Therefore, to examine the relationship between aspiration detected during testing and pneumonia, we used chest CT findings, which are associated with little diagnostic variation, and FEES performed within 1 week after CT.

In the present study, we examined the relationship between the presence or absence of aspiration detected by FEES, and the presence or absence of pulmonary inflammation in the chest CT image obtained within 1 week before FEES (inclusive of the day of the test). Compared with findings in previous studies [6, 7, 9, 12‐14, 24] that examined aspiration (detected by testing) and history of pneumonia, the present study had little time difference between the swallowing function test and the chest evaluation, which indicated the swallowing function at the time of the test was compared with the state of the lungs. Thus, food aspiration detected by testing was not correlated with inflammatory findings on chest CT. The study by Butler et al. [14] involved an older population and CT timing, and revealed no correlation between food aspiration detected by FEES and inflammatory findings in chest CT conducted within 1 year of the test. Our findings were derived from children; however, they support the findings of Butler et al. [14]. Thus, our findings suggested that food aspiration detected during testing does not necessarily represent a risk of pneumonia in disabled children. In addition, regarding food aspiration, our findings suggested that it is necessary to consider not only the presence or absence of aspiration, which is the conventional evaluation standard, but also the amount and frequency of aspiration.

We observed a significant correlation between saliva aspiration and chest CT findings of pneumonia. Furthermore, in a multivariate logistic regression analysis using age and sex as the moderator variables, saliva aspiration detected during testing was significantly correlated with the chest CT findings of bronchial wall thickening and atelectasis. Thus, CT may detect inflammatory changes of the lungs at an early stage. We observed atelectasis and bronchial wall thickening, which presumably occur because of repeated aspiration and inflammation of the bronchi caused by aspiration [25, 26]. In the present study, we used the advantages of FEES to examine, without contrast, the aspiration of food or other secretions such as saliva in the pharynx. Furthermore, to minimize stimulation of the oral cavity induced by test food ingestion or by inserting the endoscope, saliva aspiration was immediately assessed after inserting the endoscope and before the ingestion of the test food. Therefore, saliva aspiration detected by the test may reflect a chronic decrease in swallowing function or may reflect regularly occurring aspiration. In fact, 77% of participants who exhibited saliva aspiration also had food aspiration. Link et al. [27] reported that larger secretion pools were associated with an increased incidence of aspiration pneumonia in children. In another study [28], the researchers demonstrated that the extent of the pooling of secretions in the laryngopharynx was predictive of aspiration. These findings indicated that the presence or absence of saliva aspiration detected immediately after endoscope insertion could be used to predict pneumonia risk.

In the present study, the PAS was used as the assessment criterion for swallowing function, along with FEES and VFSS. However, this scale evaluates the presence or absence of aspiration from a few mouthfuls of test food ingested at the test venue; thus, it does not reflect the amount of aspiration normally present, the content of aspirated matter (i.e., apparent or silent), or the possibility of expectorating the aspirated matter. The PAS is an excellent indicator of swallowing function; however, it may be insufficient when considering pneumonia as an outcome. It may also be necessary to evaluate saliva aspiration, as well as the test food. The findings of the present study indicated that the management of swallowing dysfunction that is based only on the presence of food aspiration may result in excessive restriction or stopping of oral feeding, which may not allow an individual to maintain or develop swallowing function or may result in overmanagement that harms an individual’s food-related quality of life. Further investigations focusing on how to interpret swallowing function tests are needed.

The participants of this study were children with developmental disabilities who were unable to understand or follow instructions during CT. Therefore, it was impossible for them to control their body movements or maintain maximum inspiration during imaging. The imaging conditions were less than ideal; therefore, unrevealed findings may exist. To address this possibility, we excluded five participants whose images contained artifacts because of body movement or were otherwise difficult to interpret. By only evaluating participants whose images the assessors could interpret, we avoided lowering the precision of the image evaluations as much as possible.

In this study, it could not be determined whether the pneumonia was caused by aspiration or community acquired-infection. It was difficult to accurately diagnose whether the pneumonia was due to aspiration. However, 1) all subjects had swallowing impairment; and 2) all CT findings examined in this study were inflammatory findings due to aspiration. Based on these findings, the probability of pneumonia having been caused by aspiration is quite high in the subjects of this study.

In order to avoid missing saliva aspiration as much as possible, measures were taken, such as securing an appropriate field of view at the time of evaluation and confirmation of findings by multiple evaluators. However, the possibility of missing saliva aspiration is a limitation of this study.

We examined whether the presence or absence of aspiration during testing was associated with chest CT findings of inflammation and history of pneumonia in disabled children. Pneumonia history was not correlated with the presence or absence of food aspiration or with pulmonary CT findings. By contrast, the presence of saliva aspiration was significantly correlated with the pulmonary CT findings of bronchial wall thickening and atelectasis. When evaluating the swallowing function of disabled children, it may be difficult to assess risk of pneumonia by only examining food aspiration, whereas the presence or absence of saliva aspiration could indicate such a risk. Furthermore, our results suggest the need for new benchmarks to evaluate pneumonia risk, such as the assessment of the type or amount of aspirated matter.