Gastrointestinal Factors Associated with Risk of Bronchiectasis in Children

Summary/Abstract

Objective:

To evaluate gastrointestinal (GI) risk factors for bronchiectasis in children. We hypothesized that upper GI tract dysmotility would be associated with increased risk of bronchiectasis.

Study Design:

Subjects in this retrospective cohort study included those evaluated for persistent pulmonary symptoms in the Aerodigestive Center at Boston Children’s Hospital who underwent chest computed tomography (CT) between 2002 and 2019. To determine gastrointestinal predictors of bronchiectasis, baseline characteristics, comorbidities, enteral tube status, medications received, gastroesophageal reflux burden, adequacy of swallow function, esophageal dysmotility, gastric dysmotility, and neutrophil count on bronchoalveolar lavage (BAL) were compared between patients with and without bronchiectasis. Proportions were compared with Fisher’s exact test and binary logistic regression with stepwise selection was used for multivariate analysis. ROC analyses were utilized to compare BAL neutrophils and bronchiectasis.

Results:

Of 192 subjects, 24% were found to have evidence of bronchiectasis on chest CT at age 7.9±0.5 years. Enteral tubes (OR 5.77, 95% CI 2.25–14.83, p<0.001) and increased BAL neutrophil count (OR 5.79, 95% CI 1.87–17.94, p=0.002) were associated with increased risk while neurologic comorbidities were associated with decreased risk (OR 0.24, 95% CI 0.09–0.66, p=0.006). Gastroesophageal reflux was not found to be a significant risk factor. Neutrophil counts >10% had 72% sensitivity and 60% specificity for identifying bronchiectasis.

Conclusions:

Enteral tubes were associated with significantly increased risk of bronchiectasis but gastroesophageal reflux was not. Providers should consider obtaining chest CT to evaluate for bronchiectasis in children found to have unexplained elevated BAL neutrophil count.

Introduction

Bronchiectasis is a concerning finding on high resolution chest computed tomography (HRCT) and indicates severe lung damage that can be difficult to treat effectively^{1, 2}. This condition not only contributes to a significant decrease in quality of life for patients and their families but it also requires considerable healthcare resource use^3-5. Apart from bronchiectasis related to cystic fibrosis and recurrent pneumonia, there is little understanding about its etiology in pediatrics but prevalence appears to be increasing and more than 20% have no clear cause identified^{3, 6, 7}. Diagnostic work-up and management for these patients varies widely and it is likely these factors impact clinical outcomes as some studies have suggested reversal of disease with treatment^{1, 8, 9}. However, little is known about risk factors for disease and better understanding of these risks might allow for targeted approaches to prevent ongoing pulmonary damage in pediatric patients who are at risk^{10, 11}.

In particular, no prior studies have examined the gastrointestinal risk factors contributing to bronchiectasis; however, with the development of aerodigestive programs, the role of the gastroenterologist in multidisciplinary care has become more central to the diagnosis and management of these patients^12-14. One of the primary diagnoses considered is swallowing dysfunction with resultant aspiration which has been shown to play a role in pediatric and adult bronchiectasis^{8, 15}. Other diagnoses are also critical to consider including upper gastrointestinal dysmotility and gastroesophageal reflux disease (GERD) since these conditions can have significant symptom overlap and are commonly considered in children with bronchiectasis^16-19. The aim of this study was to determine gastrointestinal risk factors for bronchiectasis including oropharyngeal dysphagia, upper gastrointestinal dysmotility and gastroesophageal reflux disease.

Materials and Methods

We identified patients who were evaluated for persistent pulmonary symptoms in the Aerodigestive Center at Boston Children’s Hospital and underwent HRCT between 2002 and 2019. Persistent pulmonary symptoms were defined recurrent or persistent wet and/or productive cough, recurrent respiratory infections, or persistent aspiration on videofluoroscopic swallow studies, and recommendations for HRCT were made by the patient’s pulmonologist. Patients with cystic fibrosis were excluded. All radiology reports of the pediatric radiologists’ interpretation of the HRCT were reviewed in detail by the same research team to determine if they reported abnormal bronchial dilation as previously reported^{20, 21}. Since there might have been changes in reporting over time, however, subjects were only categorized as having bronchiectasis if this diagnosis was explicitly included in the radiologist’s report. Other HRCT findings including air trapping were also noted.

Charts were reviewed to determine clinical characteristics prior to performance of HRCT including demographics, comorbidities and medication use. Comorbidities of interest included those that might impact pulmonary health and/or upper GI symptoms and also included presence of surgically placed enteral tubes (gastrostomy or gastrojejunostomy) and fundoplication in the time period prior to HRCT. Airway malacia was defined as any bronchomalacia or tracheomalacia diagnosed on flexible laryngoscopy and/or direct laryngoscopy and rigid bronchoscopy (DLB). Medication use was defined as treatment with acid suppression medications including H2 receptor antagonists (H2RA) and proton pump inhibitors (PPI), macrolide prokinetics including erythromycin and azithromycin, and inhaled corticosteroids (ICS).

Diagnostic testing results were reviewed in detail to determine the presence of upper GI disorders (dysmotility, oropharyngeal dysphagia, GERD). These included results of multichannel intraluminal impedance-pH, videofluoroscopic swallow studies (VFSS), high resolution esophageal manometry, nuclear medicine esophagrams and gastric emptying scans, upper GI fluoroscopy imaging, and pathology from upper GI endoscopy. We also looked at cell counts from flexible bronchoscopy with bronchoalveolar lavage (BAL) with a focus on those performed prior to the diagnosis of bronchiectasis. Impedance studies were considered abnormal if they showed abnormal reflux burden, defined as ≥73 reflux episodes during a standard 24-hour study, or abnormal acid exposure, defined as pH <4 for ≥6% of the study²². VFSS were considered abnormal if they showed aspiration or laryngeal penetration for any texture studied. High resolution esophageal manometry results were considered abnormal if they showed ineffective esophageal peristalsis, esophagogastric junction obstruction or absent peristalsis. Nuclear medicine and fluoroscopic esophagrams were considered abnormal if they showed any esophageal transit delay without anatomic abnormality. Gastric emptying scans were considered abnormal if they showed ≥60% remaining on a 1-hour scan or ≥10% remaining on a 4-hour scan²³. Esophagitis was defined by >0 eosinophils per high powered field seen in any esophageal biopsy. Abnormal cell count on BAL was defined as any neutrophil count ≥10%^{24, 25}. Lipid laden macrophages were not included in our analysis because there was variability in whether this test was performed at our institution during the study time period and results do not correlate with meaningful outcomes^{26, 27}. BAL culture results were considered abnormal if they showed growth of any bacteria apart from normal flora. All patients with esophageal atresia were included in the esophageal dysmotility group due to near universal esophageal dysfunction in this condition^{28, 29}.

The primary outcome of this study was the presence of bronchiectasis on HRCT. Subjects were compared based on baseline characteristics, comorbidities, gastroesophageal reflux burden, adequacy of swallow function, esophageal dysmotility, and gastric dysmotility. Descriptive statistics were used to summarize patient characteristics and clinical factors, presented as mean ± standard error for continuous variables and percentage (frequency) for categorical variables. Proportions were compared with Fisher’s exact test and continuous variables compared with the Student’s t-test. ROC analyses were utilized to compare BAL neutrophil percent and bronchiectasis along with reflux episode number and bronchiectasis. Logistic regression was used to assess the association of demographics, comorbidities, procedures, medications and gastrointestinal motility factors with the presence or absence of bronchiectasis. Univariable assessments were made by conducting logistic regression with one variable at a time with case-wise deletion when missing data were present. Multivariable analysis with stepwise selection was used to determine independent correlates of bronchiectasis. All variables with P<0.10 on univariable assessment were included in the initial stepwise model and P<0.05 was required for retention in the final model. The data were assumed to be missing at random. Full information maximum likelihood estimation, available in Mplus version 8.2 (Los Angeles, CA), was used to impute the missing data leading to retention of all the subjects for analysis^{30, 31}. Results are reported as odds ratio (OR) with 95% confidence interval (CI). All statistical tests were 2-sided with P<0.05 considered statistically significant. Data were analyzed using SPSS version 27 and SAS version 9.4.

The present study was approved by our Institutional Review Board.

Results

192 subjects were included and underwent HRCT at mean age 7.9 ± 0.5 years. On HRCT 24% (46) were found to have evidence of bronchiectasis. Twenty-three percent (33) of those without documented bronchiectasis were noted to have air trapping on HRCT. Subject characteristics and comorbidities are shown in Table 1. Those with bronchiectasis were more likely to have an enteral tube and less likely to have neurologic comorbidities, as shown in Table 1. Enteral tubes were placed at mean 5.1 ± 0.7 years before HRCT with 80% (51/64) of enteral tubes having been placed greater than 1 year before HRCT. Of the 23 patients with bronchiectasis and enteral tube, 8 had a neurologic diagnosis, 8 had congenital heart disease, 3 had laryngeal cleft, 2 had tracheomalacia or bronchomalacia, 1 had ciliary dyskinesia, and 1 had interstitial lung disease.

Table 1:

Subject Characteristics

	Bronchiectasis (n=46)	No Bronchiectasis (n=146)	p-value
Subject Characteristics and Comorbidities
Age (years)	9.26 ± 0.99	7.51 ± 0.53	0.12
Male Sex	50% (23)	58% (85)	0.4
Female Sex	50% (23)	42% (61)	0.4
History of Prematurity	30% (14)	27% (40)	0.71
Tracheomalacia, bronchomalacia*	13% (6)	17% (25)	0.65
Laryngomalacia	0% (0)	3% (4)	0.57
Interstitial Lung Disease	2% (1)	1% (1)	0.42
Reactive Airway Disease	9% (4)	8% (11)	0.76
Laryngeal Cleft	24% (11)	20% (29)	0.54
Esophageal Atresia	11% (5)	12% (17)	1.0
Ciliary Dyskinesia	2% (1)	4% (6)	1.0
Congenital Heart Disease	26% (12)	18% (26)	0.29
Neurologic Disease	20% (9)	37% (54)	0.03
Enteral Tube	50% (23)	29% (41)	0.01
Fundoplication	13% (6)	10% (14)	0.58
Treatment with PPI	67% (31)	75% (108)	0.35
Treatment with H2RA	27% (12)	39% (58)	0.16
Treatment with Macrolide Prokinetic	20% (9)	21% (31)	1.0
Treatment with Inhaled Corticosteroid	64% (29)	49% (72)	0.09

^*Among the n=121 with bronchoscopy.

Diagnostic testing rates and results are shown in Table 2. Subjects with bronchiectasis underwent more testing overall compared to those without bronchiectasis and were more likely to have impedance studies, upper GI fluoroscopy and upper GI endoscopy. However, comparison of individual diagnostic testing results showed no significant differences between those with and without bronchiectasis apart from an increased proportion of subjects with significantly elevated BAL neutrophil count in those with bronchiectasis.

Table 2:

Diagnostic Testing and Gastrointestinal Motility

	All Subjects (n=192)	Bronchiectasis (n=46)	No Bronchiectasis (n=146)	P value
Diagnostic Testing Performed
Impedance	22% (43)	35% (16)	19% (27)	0.03
VFSS	76% (145)	80% (37)	74% (108)	0.44
Upper GI Fluoroscopy	37% (70)	50% (23)	32% (47)	0.04
4-hr Gastric Emptying Scan	6% (12)	9% (4)	6% (8)	0.49
1-hr Gastric Emptying Scan	27% (52)	28% (13)	27% (39)	0.85
Nuclear Medicine Esophagram	13% (24)	13% (6)	12% (18)	1.0
Esophageal Manometry	12% (22)	7% (3)	13% (19)	0.29
Upper GI Endoscopy	75% (144)	87% (40)	71% (104)	0.03
Bronchoscopy with BAL	50% (96)	63% (29)	46% (67)	0.06
Diagnostic Testing Results Among Subjects Tested
Impedance Abnormal	47% (20/43)	50% (8/16)	44% (12/27)	0.76
Impedance Abnormal by Event #	22% (9/43)	38% (6/16)	11% (3/27)	0.06
Impedance Abnormal by pH	37% (15/43)	44% (7/16)	30% (8/27)	0.51
VFSS with Aspiration	32% (46/145)	32% (12/37)	32% (34/108)	1.0
Upper GI Fluoroscopy Abnormal	24% (17/70)	22% (5/23)	26% (12/47)	1.0
4-hr Gastric Emptying Scan Abnormal	67% (8/12)	50% (2/4)	75% (6/8)	0.55
1-hr Gastric Emptying Scan Abnormal	40% (21/52)	46% (6/13)	39% (15/39)	0.75
Esophageal Dysmotility on Nuclear Medicine Esophagram	63% (15/24)	50% (3/6)	67% (12/18)	0.64
Esophageal Dysmotility on Manometry	68% (15/22)	100% (3/3)	63% (12/19)	0.52
Endoscopy with Esophagitis	26% (38/144)	30% (12/40)	25% (26/104)	0.53
BAL with Elevated Neutrophil Count	40% (48/121)	72% (21/29)	40% (27/67)	0.007
Gastrointestinal Motility Among Subjects Tested
Gastroesophageal Reflux	47% (20/43)	50% (8/16)	44% (12/27)	0.76
Oropharyngeal Dysphagia	32% (46/145)	32% (12/37)	32% (34/108)	1.0
Esophageal Dysmotility	77% (37/48)	80% (8/10)	76% (29/38)	1.0
Gastric Dysmotility	45% (27/60)	44% (7/16)	45% (20/44)	1.0
Any Upper GI Tract Dysmotility	76% (121/160)	80% (32/40)	74% (89/120)	0.53

Subjects had varied forms of upper GI dysmotility, as shown in Table 1. Those with upper GI dysmotility were more likely to have enteral tubes (OR 3.27, 95% CI 1.59-6.70, p=0.001). Figure 1 shows Venn diagrams of the types of upper GI tract dysmotility in the cohort, separated by the presence or absence of bronchiectasis. Motility disorders frequently occurred together; 17 patients had oropharyngeal dysphagia and esophageal dysmotility, 10 patients had esophageal dysmotility and gastric emptying delay, and 10 had oropharyngeal dysphagia and gastric emptying delay.

Figure 1: Venn Diagram of Overlapping Upper GI Tract Dysmotility.

There was overlap between each of the types of upper GI tract dysmotility. Those without bronchiectasis are shown in panel A and those with bronchiectasis are in panel B.

ROC analysis did not show a significant relationship between reflux episode number and bronchiectasis (AUC 0.54, 95% CI 0.32-0.76, p=0.68). None of the other reflux parameters were associated with the finding of bronchiectasis or abnormal neutrophil count (all p>0.22).

Risk factors for bronchiectasis from univariable and multivariable analyses are shown in Table 3. Enteral tube status and increased BAL neutrophil count were independently associated with increased risk while neurologic comorbidities were associated with decreased risk. A significant association between enteral tube status and bronchiectasis was also seen specifically in patients with neurologic comorbidities independent of aspiration status (OR 3.73, 95% CI 1.66-8.4, p=0.002). Because GERD, oropharyngeal dysphagia and dysmotility were not significant in the univariate analyses, they were not included in the final model.

Table 3A:

Univariable assessment of risk factors for bronchiectasis by logistic regression

	Odds Ratio (95% CI)	P value
Demographics
Age	1.04 (0.99 – 1.09)	0.11
Female sex	1.39 (0.72 – 2.71)	0.33
History of Prematurity	1.16 (0.56 – 2.40)	0.69
Comorbidities
Tracheomalacia, bronchomalacia	0.73 (0.28 – 1.9)	0.65
Interstitial Lung Disease	3.22 (0.20 – 52.6)	0.41
Reactive Airway Disease	1.17 (0.35 – 3.86)	0.80
Laryngeal Cleft	1.27 (0.58 – 2.79)	0.56
Ciliary Dyskinesia	0.52 (0.06 – 4.42)	0.55
Congenital Heart Disease	1.63 (0.75 – 3.56)	0.22
Neurologic Disease	0.41 (0.19 – 0.92)	0.03
Procedures
Enteral Tube	2.48 (1.25 – 4.89)	0.009
Fundoplication	1.41 (0.51 – 3.92)	0.51
Medications
Treatment with PPI	0.73 (0.35 – 1.49)	0.38
Treatment with H2RA	0.54 (0.26 – 1.12)	0.10
Treatment with Macrolide Prokinetic	0.90 (0.39 – 2.07)	0.81
Treatment with Inhaled Corticosteroid	1.75 (0.89 – 3.46)	0.11
Gastrointestinal Motility
Endoscopy with Esophagitis	1.56 (0.71 – 3.39)	0.27
BAL with Elevated Neutrophil Count*	3.89 (1.51 – 10.05)	0.005
Esophageal Dysmotility	0.94 (0.41 – 2.16)	0.89
Gastric Dysmotility	1.13 (0.45 – 2.87)	0.80
Gastroesophageal Reflux	2.35 (0.90 – 6.17)	0.08
Oropharyngeal Dysphagia	1.43 (0.74 – 2.79)	0.29
Any Upper GI Tract Dysmotility	1.92 (0.90 – 4.09)	0.09

^*N=71 did not have bronchoscopy and 25 of those with bronchoscopy did not have neutrophil count reported.

Analysis of bronchoalveolar lavage results revealed a higher neutrophil percent in subjects with bronchiectasis (33.7 ± 6 vs 14.8 ± 2.6, p=0.006). ROC analysis (Figure 2) showed a significant relationship between neutrophil percent and bronchiectasis (AUC 0.71, 95% CI 0.59-0.82, p=0.001). When using the abnormal cutoff of 10% neutrophils, the sensitivity for identifying bronchiectasis was 72.4% (95% CI 52.8-87.3) and specificity was 59.7% (95% CI 47-71.5). A significant association between bronchiectasis and elevated neutrophil count remained even when looking at the 70 subjects that had their bronchoscopy performed before their HRCT (OR 2.9, 95% CI 1.1-7.4, p=0.049).

Figure 2: ROC Curve Comparing Bronchoalveolar Lavage Neutrophil Percent with Bronchiectasis.

ROC analysis demonstrates a significant relationship between neutrophil percent and bronchiectasis (AUC 0.71, 95%CI 0.59-0.82, p=0.001).

Discussion

In the present study, we present the largest known cohort to date of children with non-cystic fibrosis bronchiectasis and well-characterized upper GI motility function in order to evaluate gastrointestinal factors associated with risk of bronchiectasis in children. Our findings are timely in the context of recent renewed interest in advancing our understanding of the pathophysiology, natural history and potential therapies for children with bronchiectasis^{11, 32, 33}. We utilized multivariable analysis with adjustment for comorbidities along with full information maximum likelihood estimation to reconcile potential predictors of bronchiectasis with markers of disease severity in an attempt to reduce confounding in this relatively sick patient population. While we did not find an association between upper GI dysfunction and bronchiectasis in the present study, we can make several key conclusions: 1) enteral tubes are associated with increased risk of bronchiectasis, 2) risk appears to be lower in those with neurologic comorbidities, 3) many patients are treated with acid suppression but this does not appear to decrease risk of bronchiectasis, and 4) providers should consider obtaining chest imaging to evaluate for bronchiectasis in children with elevated BAL neutrophil count.

Overall, the patients in this cohort underwent a significant number of diagnostic tests, an expected finding in a referral population of symptomatic patients. Many were found to have motility disorders, as would be expected in our Aerodigestive Center population. Among these, gastroesophageal reflux and oropharyngeal dysphagia with aspiration are commonly considered in the workup of bronchiectasis and these were more common in our cohort, but other disorders including esophageal dysmotility and gastric emptying delay were seen as well. Many of the subjects in the cohort received medications including acid suppression, prokinetics and inhaled corticosteroids. None of these approaches appeared protective in our multivariable model.

Prior work has suggested that upper GI tract dysmotility may pose a greater risk of pulmonary injury when there is impaired airway protection, as in the case of patients with oropharyngeal dysphagia³⁴. Piccione et al previously showed an association between aspiration and bronchiectasis but provided conflicting results on whether gastroesophageal reflux might also play a role⁸. We did not find a significant relationship for aspiration in the present cohort, but it is possible that this is related to our older pediatric population, where swallowing dysfunction associated with developmental immaturity may have been outgrown³⁵. Other groups have reported increased incidence of bronchiectasis in aerodigestive center populations and therefore the underlying conditions commonly seen in this population might put these patients at increased risk¹².

Children with pulmonary disease and gastroesophageal reflux disease are commonly treated with acid suppression medications and in some cases undergo fundoplication due to concern about reflux-related lung injury. Even recent reports of bronchiectasis in children recommend consideration of reflux therapies despite unclear benefit in this population¹⁹. However, our results showed the lack of association between bronchiectasis and abnormal impedance as a measure of reflux burden. We also found lack of a protective effect from acid suppression and from fundoplication, which might suggest uncertain benefit of these approaches and that any theoretical benefit to these interventions will not outweigh the risks^36-39. These results support the finding that the presence of pathological gastroesophageal reflux alone is likely not a common cause of bronchiectasis and are in agreement with prior data about the lower yield of impedance testing in children with respiratory symptoms^{18, 34}. We also note that inhaled corticosteroid was associated with increased risk of bronchiectasis in this cohort; however, this might be due to confounding by reverse causation since those at higher risk for the outcome were likely placed on this therapy. These medications are not currently recommended by practice guidelines, except for asthma-phenotype^{32, 40}.

Another common reason for GI consultation in this population of children with respiratory symptoms is for placement of enteral tubes^{19, 41, 42}. Enteral tube status showed the highest increase in bronchiectasis risk in our cohort. It is possible that enteral tube status is a marker of overall illness severity, but many of the patients had their tube in place for 5 years or more before undergoing chest CT, suggesting the tubes were not likely placed because of the severity of pulmonary disease seen at the time of CT. Prior studies have also shown increased pulmonary risk associated with enteral tubes, including hospitalization risk and increase in gastroesophageal reflux, which might suggest mechanisms by which bronchiectasis might occur^{43, 44}. Subjects in this study had enteral tubes placed for a variety of clinical indications, but given these ongoing risks, it is important to take advantage of more recent developments in blenderized tube feeding, the use of motility medications and, when necessary, transpyloric feeding status^{45, 46}. We cannot determine any causal relationship between enteral tubes and bronchiectasis with our retrospective study design and all decisions about safe feeding approaches must be made in the appropriate clinical context, but our findings do highlight the importance of GI involvement in the diagnosis and management of bronchiectasis in children.

It is notable that neurologic comorbidity was associated with lower risk in our cohort but the potential mechanism for this finding is unknown. One possibility is that patients with neurologic impairment may not undergo as many pulmonary CTs if clinicians feel the trajectory of their illness may not change⁴⁷. However, our multivariable analysis suggests that enteral tube status confers additional risk in neurologic patients even when controlling for aspiration. Some clinicians advocate for enteral tube placement and avoidance of oral feeding in neurologic patients with concern for poor airway protection; however, our findings suggest universal implementation of this approach may not be beneficial⁴⁸. The mechanism for this outcome is not completely clear, however, and further studies will be needed to clarify the reasons behind this finding. Our findings are also limited by only having a single HRCT for each patient, which might not completely capture additional changes that might be seen over time with progressive neurologic comorbidities.

With the growth of multidisciplinary aerodigestive centers, many patients with persistent pulmonary symptoms undergo triple airway endoscopy as an initial part of their workup before consideration of chest CT⁴⁹. This approach commonly includes flexible bronchoscopy with bronchoalveolar lavage and the results are typically interpreted in the multidisciplinary aerodigestive context^{8, 12-14}. Prior studies have shown an association between bronchiectasis and BAL neutrophil count and the results our ROC analysis are similar^{24, 50, 51}. While this might not be a surprising result, our results might suggest that elevated BAL neutrophil count with no other potential explanations (e.g. immunodeficiency, intercurrent illness) should raise concern for bronchiectasis. In this context, providers might consider obtaining chest CT, as earlier diagnosis might positively impact treatment strategy and outcomes.

Limitations of the present study include the retrospective nature of the study design. Since we included all patients from our Aerodigestive Center that underwent HRCT, it is likely that this referral population was sicker than a general pediatric population but since HRCT is not yet standard for all children with aerodigestive disorders our results are likely relevant to clinical practice. We also used HRCT of bronchiectasis as our primary outcome but we acknowledge that other forms of pulmonary injury might also be associated with upper GI dysmotility and should be addressed in future studies. An additional limitation is that not all subjects included in the cohort underwent all diagnostic tests and therefore diagnoses that require diagnostic testing such as airway malacia and true rates of test positivity in all bronchiectasis patients is not known. It is possible that there was increased selection bias by including more subjects with a higher pre-test probability of having abnormal motility testing results given the tertiary referral population and that different results might have been found with a more varied cohort. Another limitation is the small sample size with each test. The approach of full information maximum likelihood estimation was used to handle the missing data and retain all the subjects for analysis. Although full information maximum likelihood is known to produce asymptotically unbiased parameter estimates, the fact remains that our sample size is small and further multicenter studies to look at risk factors may be needed. Lastly, we were unable to assess longitudinal effects since we only had one chest CT available for each subject but it will be important to address this in future prospective studies to better understand approaches that might improve outcomes.

Table 3B:

Multivariable assessment of risk factors for bronchiectasis by logistic regression

	Odds Ratio (95% CI)	P value
Neurologic Disease	0.24 (0.09 – 0.66)	0.006
Enteral Tube	5.77 (2.25 – 14.83)	<0.001
Treatment with Inhaled Corticosteroid	2.81 (1.16 – 6.77)	0.02
BAL with Elevated Neutrophil Count	5.79 (1.87 – 17.94)	0.002

Abbreviations

BAL

Bronchoalveolar lavage

CT

Computed tomography

GERD

Gastroesophageal reflux disease

GI

Gastrointestinal

H2RA

H2 receptor antagonist

HRCT

High resolution chest CT

ICS

Inhaled corticosteroids

PPI

Proton pump inhibitor

VFSS

Videofluoroscopic swallow study