Breastfeeding in Infants who Aspirate May Increase Risk of Pulmonary Inflammation

Abstract

Objective:

To evaluate management strategies and pulmonary outcomes for breastfed infants with oropharyngeal dysphagia.

Study Design:

We performed a retrospective cohort study of breastfed infants diagnosed with oropharyngeal dysphagia with documented aspiration or laryngeal penetration on videofluoroscopic swallow study (VFSS). Medical records were reviewed for VFSS results and speech-language pathologist (SLP) recommendations following VFSS, results of chest x-ray (CXR), results of bronchoalveolar lavage (BAL) within 1 year of VFSS, and aspiration-related hospitalizations occurring before or within 1 year of VFSS. Subjects were categorized as cleared or not cleared to breastfeed based on the VFSS. Proportions were compared with Chi-square and Fisher’s exact tests and means with Student’s t-tests.

Results:

Seventy-six infants (4.7 ± 0.4 months old) were included; 50% (38) had aspiration and 50% (38) had laryngeal penetration. After VFSS, 70% (53) were cleared to breastfeed while 30% (23) were not cleared to breastfeed. Patients with aspiration were less likely to be cleared to breastfeed (p=0.006); however, 55% (21/38) of those with aspiration were still cleared to breastfeed. Infants cleared to breastfeed had significantly more pulmonary hospitalizations (p=0.04) and were also at increased risk of elevated neutrophil count (p=0.02) and culture growth on BAL (p=0.01). Significantly increased abnormal neutrophil count was also found in those cleared to breastfeed with laryngeal penetration (p=0.01).

Conclusions:

Infants with oropharyngeal dysphagia counseled to continue breastfeeding had increased risk of BAL inflammation and more pulmonary hospitalizations compared to those that were told to stop breastfeeding.

Introduction

Oropharyngeal dysphagia with aspiration is common in infants with feeding difficulties and can cause symptoms of coughing, choking, respiratory distress, noisy breathing, feeding difficulties, brief resolved unexplained events (BRUE), and poor growth^1–4. Multiple studies suggest that interventions such as thickened liquids, position changes and adjusting flow rate to prevent aspiration in bottle-fed infants can improve symptoms of concern, decrease pulmonary hospitalizations, and reduce risk of bronchiectasis but little is known about management of aspiration in breastfed infants^{1, 2, 5–9}.

Management of oropharyngeal dysphagia can be a challenge in infants as most aspiration in this age group is silent; therefore, instrumental assessment is required to make the diagnosis and confirm appropriate treatment^{3, 10, 11}. Aspiration can be diagnosed using videofluoroscopic swallow study (VFSS) or fiberoptic endoscopic evaluation of swallowing (FEES). Breastfeeding FEES can be used to specifically evaluate breastfeeding safety and studies have found that adjustments in positioning can improve swallow function in breastfeeding infants^{12, 13}. Unfortunately, this test is not readily available at all institutions, requires specialized expertise and patient cooperation, and has its own limitations related to visualization of aspiration as it is happening^14–16. Because of these limitations, many providers rely on VFSS to assess swallow function in breastfed infants. However, because VFSS is typically performed using a bottle, how these results extrapolate to feeding at the breast is not known^{17, 18}.

Apart from diagnostic and management strategies, little is known about potential risks of breastfeeding for infants with oropharyngeal dysphagia⁸. While breastfeeding is recommended for infants with normal swallow function, the preferred feeding method may differ for infants with aspiration or laryngeal penetration as this practice might worsen pulmonary outcomes^{8, 19}. The American Academy of Pediatrics (AAP) recommends exclusive breastfeeding for at least the first 6 months of life since breastfeeding has many short and long-term benefits for infants including immune system support, prevention of infections and improved neurodevelopmental outcomes for infants^{19, 20}. There is also benefit to the mother including reduced risk of breast and ovarian cancer, type 2 diabetes and high blood pressure^{21, 22}. However, AAP breastfeeding guidelines do not address management of children with oropharyngeal dysphagia¹⁹. For these infants, a careful balance of the potential benefits of breastfeeding with potential risks of aspiration may need to be considered^{1, 6, 23, 24}.

While breastfeeding is often recommended despite aspiration on VFSS, there are some studies to suggest that aspiration of breast milk is not benign. Prior studies have evaluated pulmonary inflammation and lung compliance in rabbit lungs instilled with human breastmilk and reported increases in inflammation and decreases in lung compliance on par with the damaging effects of formula aspiration^{25, 26}. In a single retrospective study of breastfed aspirating infants, infants who were breast fed continued to have coughing in 32% of patients and 10% had a diagnosis of a confirmed respiratory infection over the three months of follow up. The authors concluded based on this short follow up, that despite symptoms, breastfeeding was safe to continue. This study was limited by the lack of recognition of high rates of silent aspiration, the short follow up period, and the lack of objective biomarkers of pulmonary health⁸.

The lack of data in this area may complicate the decision-making process about how best to counsel parents of breastfeeding infants who are diagnosed with oropharyngeal dysphagia. Multiple factors may influence the feeding plan for any infant and family, including symptom severity and medical comorbidities^{10, 27}. Patients often present to multiple providers for these symptoms including International Board Certified Lactation Counselors and/or Certified Lactation Counselors, speech language pathologists (SLP), pulmonologists, gastroenterologists, otolaryngologists and primary care providers^28–30. To overcome the limitations of the literature, including a lack of studies that look at persistent breastfeeding and the relationship to biomarkers of lung inflammation, we performed this study to evaluate management strategies after VFSS and resultant, objective pulmonary outcomes for breastfed infants with oropharyngeal dysphagia. We hypothesized that infants with oropharyngeal dysphagia that continue to breastfeed may have increased risk of pulmonary inflammation and hospitalization.

Methods

We performed a retrospective cohort study of breastfed infants diagnosed with oropharyngeal dysphagia with either aspiration or laryngeal penetration on videofluoroscopic swallow study (VFSS) performed at age <12 months in 2015, 2018 and 2021; these years were chosen to incorporate pre- and post-COVID practice patterns and to incorporate changes in thickening patterns at our institution over time. A control group of bottle-fed infants with abnormal VFSS results from 2015 was also included for comparison of outcomes between bottle-fed and breastfed infants.

Medical records were reviewed for VFSS results and speech-language pathologist (SLP) recommendations following VFSS exams. VFSS were considered abnormal if they showed aspiration or laryngeal penetration for any consistency studied; at our institution, penetration-aspiration scales (PAS) were only added to the VFSS reports in 2020 so limited data using standardized metrics is available³¹. The frequency of breast-feeding along with supplemental bottle feedings with expressed breastmilk were also obtained via review of the SLP’s VFSS report. Thickened liquid consistencies were reported according to IDDSI terminology³². Breastfeeding fiberoptic endoscopic swallow studies (FEES) were not included as FEES studies were not readily available at our institution during the study period.

Diagnostic records were reviewed to evaluate for evidence of pulmonary inflammation that may be related to aspiration. These included results of chest x-ray (CXR) and results of flexible bronchoscopy with bronchoalveolar lavage (BAL) within 1 year of VFSS for the sub-groups of subjects that had these tests performed. These timeframes were chosen based on the typical time during which infants with persistent dysphagia undergo these additional diagnostic tests and over the age which the AAP encourages continued breastfeeding^{10, 19, 27}. Abnormal cell count on BAL was defined as any neutrophil count ≥10%^{33, 34}. BAL culture results were considered abnormal if they showed growth of bacteria apart from normal upper respiratory flora and were reported by our institution’s microbiology lab as rare (1–10 colonies), few (>10 colonies in first quadrant), moderate (growth including the second quadrant), or abundant (growth including the third or fourth quadrant). Lipid laden macrophages were not included as performance of this test was phased out during the study period because results do not clearly correlate with meaningful outcomes^{35, 36}.

To determine pulmonary hospitalization risk, records were reviewed for aspiration-related hospitalizations occurring before or within 1 year following VFSS. These included hospitalizations for respiratory symptoms based on diagnoses included in hospitalization discharge summaries.

Charts were reviewed to determine relevant clinical characteristics including demographics, comorbidities and medications. Comorbidities of interest included those that might impact swallow function and pulmonary health. Medications of interest included inhaled corticosteroids (ICS) prescribed before or within 1 year of VFSS and antibiotics prescribed within 1 year after VFSS so as to reflect potential ongoing pulmonary risk after management changes were recommended/not recommended following VFSS.

The primary aim of this study was to compare adverse pulmonary outcomes, including evidence of pulmonary inflammation and risk of pulmonary hospitalization, in infants with oropharyngeal dysphagia who were either cleared or not cleared to breastfeed. Subjects were categorized as cleared to breastfeed if no recommendation to not breastfeed was made by SLP after VFSS or they were categorized as not cleared to breastfeed if breastfeeding was discouraged by SLP after VFSS. 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 Chi-square or Fisher’s exact test and means with the Student’s t-test. All statistical tests were 2-sided with P<0.05 considered statistically significant. Data were analyzed using SPSS version 27.

The present study was approved by our Institutional Review Board.

Results

The cohort included 76 breastfed subjects with VFSS performed at mean of 4.7±0.4 months of age. Baseline characteristics are in Table 1. Indications for VFSS included coughing and choking for 17 subjects, BRUE spells for 15 subjects, cough and respiratory distress for 11 subjects, cough for 9 subjects, cough and spit-up for 7 subjects, respiratory distress for 6 subjects, choking for 5 subjects, feeding difficulties for 3 subjects and congestion for 3 subjects. Overall, 70% (53) were cleared to breastfeed by the SLP while 30% (23) were not cleared to breastfeed.

Table 1:

Cohort Characteristics

	Breastfed Infants with Abnormal VFSS (n=76)
Baseline Characteristics
Age in Months at VFSS	4.65 ± 0.36
Weight (kg)	4.5 ± 0.74
Weight z-score	−1.45 ± 0.31
Male Sex	59% (45)
History of Prematurity	30% (23)
Gestational Age in Months	32.89 ± 0.79
Congenital Heart Disease	13% (10)
Neurologic Comorbidities	8% (6)
Laryngeal Cleft	5% (4)
Breastfeeding Frequency before VFSS
Once daily	21% (16)
Multiple Times per Day	64% (49)
Exclusively Breastfed	15% (11)
Pulmonary Hospitalization
Any Pulmonary Hospitalization before VFSS	13% (10)
VFSS Findings
Laryngeal Penetration a	50% (38)
Deep	83% (29/35)
Shallow	17% (6/35)
Consistent	50% (18/36)
Inconsistent	50% (18/36)
One Episode	6% (2/36)
Aspiration	50% (38)
Silent Aspiration	87% (33/38)
Recommendation after VFSS
Continue Breast/Bottle feeding without Management Change	16% (12)
Nipple Flow Rate Change Alone	25% (19)
Thickened Bottles b	59% (45)

^aNote: Denominators represent subjects for whom frequency/depth of laryngeal penetration was reported.

^bNote: There was no safe consistency identified for 7 subjects.

In the cohort overall, 50% (38) had CXR 2.3±0.9 months from VFSS. Indications for CXR included tube placement for 12 subjects, aspiration for 9 subjects, cough and wheezing for 6 subjects, pneumonia for 5 subjects, respiratory distress for 3 subjects, laryngomalacia for 2 subjects, and stridor for 1 subject. Of these, 50% (19) had abnormal findings: 32% (12) atelectasis, 8% (3) bronchial wall thickening, 5% (2) increased interstitial markings, and 5% (2) bronchovascular crowding.

Twenty-five percent (19) underwent BAL 7.2±0.8 months from VFSS. Indications for BAL included aspiration for 7 subjects; aspiration and chronic cough for 4 subjects; recurrent respiratory infections for 4; aspiration, wheeze and cough for 2; and cough/congestion for 2 subjects. Between subjects that did or did not undergo BAL, there were no differences in pulmonary hospitalizations, comorbidities, proportion with aspiration or consistencies recommended after VFSS, but those that underwent BAL were older at VFSS (5.9±0.7 vs 4.2±0.4 months, p=0.04). Of those that underwent BAL, 32% (6) had a neutrophil count ≥10%, suggesting pulmonary inflammation. The mean neutrophil count was 12.2±4.6%. Thirty-seven percent (7) had positive BAL cultures. During the time before or within 1 year after their VFSS, 25% (19) of the cohort required hospitalization attributable to aspiration. Of these, 13% (10) required pulmonary hospitalization before their VFSS and 14% (11) required pulmonary hospitalization after their VFSS.

A comparison of baseline characteristics and breastfeeding frequency prior to VFSS between those cleared or not cleared to breastfeed is in Table 2. Comorbidities did not differ between groups but a lower proportion of patients were cleared to breastfeed during recent years. Breastfeeding frequency prior to VFSS was similar between the groups.

Table 2:

Comparison of Baseline Characteristics between Subjects Cleared or Not Cleared to Breastfeed after VFSS

	Cleared to Breastfeed after VFSS (n=53)	Not Cleared to Breastfeed after VFSS (n=23)	P-value
Baseline Characteristics
Age in Months at VFSS	5.0 ± 0.4	3.9 ± 0.7	0.17
Weight (kg)	4.02 ± 1.02	5.62 ± 0.71	0.2
Weight z-score	−1.7 ± 0.39	−0.87 ± 0.45	0.17
2015 Group (n=34)	82% (28)	18% (6)	0.049
2018 Group (n=31)	65% (20)	35% (11)
2021 Group (n=11)	45% (5)	55% (6)
Male Sex	68% (36)	39% (9)	0.02
History of Prematurity	34% (18)	22% (5)	0.29
Gestational Age in Months	32.7 ± 0.9	33.4 ± 1.9	0.77
Congenital Heart Disease	10% (5)	22% (5)	0.15
Neurologic Comorbidities	10% (5)	4% (1)	0.44
Laryngeal Cleft	4% (2)	13% (3)	0.14
Breastfeeding Frequency before VFSS
Once daily	23% (12)	17% (4)	0.82
Multiple Times per Day	64% (34)	65% (15)
Exclusively Breastfed	13% (7)	17% (4)

A comparison of VFSS findings and management changes recommended after VFSS is in Table 3. Recommendations on breastfeeding safety did not significantly differ depending on depth or frequency for subjects with laryngeal penetration. Subjects with aspiration were less likely to be cleared to breastfeed; however, 55% (21) of those with aspiration were still cleared to breastfeed. There was no safe consistency identified for 2 subjects in the cleared to breastfeed group and 5 subjects in the not cleared to breastfeed group (4% vs 22%, p=0.02); these patients required enteral tube placement. Six subjects had elevated side-lying position tested at VFSS. Of these, 2 with laryngeal penetration were counseled to discontinue breastfeeding, 2 with laryngeal penetration were cleared to continue breastfeeding, and 2 with aspiration were cleared to continue breastfeeding. Of the subjects with records of their feeding following VFSS, 39% (9/23) in the not cleared to breastfeed group were reported to be receiving primarily expressed breastmilk by bottle compared to 54% (23/43) in the cleared to breastfeed group (p=0.17).

Table 3:

Comparison of VFSS Findings and Recommendations between Subjects Cleared or Not Cleared to Breastfeed after VFSS

	Cleared to Breastfeed after VFSS (n=53)	Not Cleared to Breastfeed after VFSS (n=23)	P-value
VFSS Findings
Laryngeal Penetration	60% (32)	26% (6)	0.006
Deep	83% (24/29)	17% (5/29)	0.56
Shallow	100% (6/6)	0% (0/6)
Consistent	89% (16/18)	11% (2/18)	0.69
Inconsistent	81% (13/16)	19% (3/16)
One Episode	100% (2/2)	0% (0/2)
Aspiration	40% (21)	74% (17)	0.006
Silent Aspiration	81% (17/21)	94% (16/17)	0.36
Recommendation after VFSS
Continue Breast/Bottle feeding without Management Change	19% (10)	9% (2)	0.33
Nipple Flow Rate Change Alone	24% (13)	26% (6)	0.89
Thickened Bottles a	57% (30)	65% (15)	0.61
Slightly Thick	30% (9/30)	0% (0/15)	0.05
Mildly Thick	53% (16/30)	60% (9/15)	0.6
Moderately Thick	13% (4/30)	40% (6/15)	0.06
Puree	3% (1/30)	0% (0/15)	1.0

^aNote: There was no safe consistency identified for 2 subjects in the cleared to breastfeed group and 5 subjects in the not cleared to breastfeed group (4% vs 22%, p=0.02).

Table 4 shows pulmonary outcomes. There was no statistically significant difference in CXR abnormalities. However, subjects cleared to breastfeed were at increased risk of BAL inflammation compared to those not cleared to breastfeed, including increased proportion with abnormal neutrophil count and higher mean neutrophil count, even within the laryngeal penetration subgroup. Subjects cleared to breastfeed were also more likely to have positive BAL cultures, including 2 Haemophilus influenzae, 2 Moraxella catarrhalis, 1 Streptococcus pneumoniae, 1 Streptococcus pyogenes, 1 Moraxella catarrhalis and Haemophilus influenzae, and 1 Streptococcus pneumoniae and Haemophilus influenzae. The single positive culture in the not cleared to breastfeed group grew Haemophilus influenzae. There were no significant differences in rates of BAL inflammation or culture positivity by breastfeeding frequency (p>0.12). There were no significant differences in rates of BAL inflammation or culture positivity by frequency or depth in those with laryngeal penetration (p>0.4). One subject had 1% eosinophils seen on BAL and was in the group that was not cleared to breastfeed.

Table 4:

Comparison of Pulmonary Outcomes between Subjects Cleared or Not Cleared to Breastfeed after VFSS

	Cleared to Breastfeed after VFSS (n=53)	Not Cleared to Breastfeed after VFSS (n=23)	P-value
Chest X-Ray (CXR)
CXR Obtained	55% (29)	39% (9)	0.32
Months from VFSS to CXR	2.5 ± 0.9	1.8 ± 1.2	0.65
CXR Abnormal	55% (16)	33% (3)	0.25
Bronchoalveolar Lavage (BAL)
BAL Obtained	23% (12)	30% (7)	0.57
Months from VFSS to BAL	7.6 ± 0.9	6.6 ± 1.5	0.6
Mean Neutrophil %	18.9 ± 6.5	0.71 ± 0.71	0.009
Laryngeal Penetration Subgroup (n=12)	22.4 ± 8.1	0 ± 0	0.01
BAL with ≥10% Neutrophils	50% (6/12)	0% (0/7)	0.02
Laryngeal Penetration Subgroup (n=12)	56% (5/9)	0% (0/3)	0.09
BAL Culture Positive	58% (7/12)	11% (0/7)	0.01
Rare Growth	43% (3/7)	0 (0)
Moderate Growth	43% (3/7)	0 (0)
Abundant Growth	14% (1/7)	0 (0)
Laryngeal Penetration Subgroup (n=12)	56% (5/9)	0% (0/3)	0.09
Inhaled Corticosteroid (ICS)
ICS Prescribed	21% (11/53)	4% (1/23)	0.07
Antibiotics
Antibiotics Prescribed	17% (9)	13% (3)	0.67
Any Pulmonary Hospitalization before VFSS	13% (7)	13% (3)	0.98
Any Pulmonary Hospitalization after VFSS	19% (10)	4% (1)	0.16
Number of Pulmonary Hospitalizations after VFSS	0.26 ± 0.09	0.04 ± 0.04	0.04
Number of Pulmonary Hospitalization Nights after VFSS	0.62 ± 0.33	0 ± 0	0.06

Medications compared between the groups are shown in Table 4. Overall, 16% (12) of the subjects in the cohort were treated with ICS and 16% (12) prescribed antibiotics. Forty-two percent (5) of subjects were prescribed both ICS and antibiotics. Subjects treated with ICS were more likely to have been older at the time of their VFSS (7.6 ± 0.8 vs 4.1 ± 0.4 months, p<0.001) and to have elevated BAL neutrophil count (73% vs 6%, p<0.001). There were no significant differences between those treated or not treated with antibiotics.

Pulmonary hospitalizations before and after VFSS are in Table 4. Those cleared to breastfeed had more post-VFSS pulmonary hospitalizations compared to those not cleared to breastfeed. There was no difference seen in pre- or post-VFSS pulmonary hospitalizations by breastfeeding frequency (p>0.29).

There were 171 subjects in the bottle-fed group. Neurologic comorbidities were more common in this group (p=0.008) but baseline characteristics did not otherwise differ from the breastfed group, as in Supplemental Table 1. In comparison to the bottle-fed group, those that were cleared to continue breastfeeding had increased BAL culture positivity and similar if not increased risk of pulmonary inflammation. These results are in Supplemental Table 2.

Discussion

Despite many known benefits of breastfeeding, it is important to understand the risks of continued breastfeeding for infants with oropharyngeal dysphagia and aspiration, in order to best counsel families. We sought to address gaps in the current literature by comparing objective pulmonary outcomes for infants who were or were not cleared to breastfeed after being diagnosed with oropharyngeal dysphagia. Our results suggest several important findings. First, recommendations after VFSS are variable in breastfed infants despite a history of pulmonary inflammation and hospitalizations. Second, in patients with oropharyngeal dysphagia, persistent breastfeeding is associated with lung inflammation and increased pulmonary hospitalization.

The most concerning finding of our study is the increased risk of pulmonary inflammation in infants cleared for continued breastfeeding. This included both increased neutrophil number and proportion with abnormal neutrophil count on BAL in addition to increased risk of positive cultures, all of which suggest active pulmonary injury which may have long-term consequences for these infants’ respiratory health. In fact, the only infants found to have elevated neutrophil counts in this cohort were the ones cleared for continued breastfeeding despite aspiration or laryngeal penetration on VFSS. It is notable that even in the sub-group with laryngeal penetration in the absence of aspiration, often considered a milder form of swallowing dysfunction, patients still had increased pulmonary inflammation. This did not differ significantly by frequency/depth of the laryngeal penetration seen on VFSS, which is consistent with other studies on the clinical impact of laryngeal penetration^{4, 5}. For infants with elevated neutrophil count, this finding has been associated with bronchiectasis even in infants and suggests that significant pulmonary injury may have already occurred^{6, 33}. While it might be expected that those that underwent bronchoscopy had increased symptom severity, there were no differences seen for any of the variables between groups apart from age at VFSS, suggesting that those that did not undergo BAL might also have evidence of pulmonary inflammation if tested.

Chest x-ray abnormalities were not significantly different between the groups, likely due to the lower sensitivity of this diagnostic study. As other studies have shown, this suggests that clinicians cannot rely on reassuring chest x-ray results when evaluating the pulmonary impact of ongoing aspiration^{27, 37}. Findings of bronchiectasis and other signs of aspiration-related lung injury can be radiographically diagnosed by chest CT but none of the subjects in this relatively-young cohort underwent chest CT^{6, 33}.

In regard to ICS use, these medications are commonly prescribed in even small children with chronic respiratory symptoms, those who have asthma risk factors, and those who are hospitalized with pulmonary symptoms regardless of etiology. The asthma literature indicates that patients with eosinophilia are most likely to respond to inhaled corticosteroids, but the anti-inflammatory mechanism of corticosteroids, extends beyond eosinophils and they are also commonly prescribed in patients with oropharyngeal dysphagia^{10, 38, 39}. There was no clear pattern for when these medications were started in the current study, so it is unclear if they were prescribed in response to inflammation or to provide additional protection in infants that continued to breastfeed. The higher proportion of ICS prescribed to breastfed infants suggests that perhaps those that continued to breastfeed were more symptomatic. It is unknown if ICS might protect against aspiration-related lung injury and their use is variable in clinical practice, but future studies may be required to determine if they might have a role in breastfed infants with oropharyngeal dysphagia and if this outweighs potential risks of prolonged ICS use^{10, 40}. Subjects treated with ICS in this cohort were more likely to have elevated neutrophil count even when treated before BAL but it is unclear if this may be due to lack of effect or insufficient protection against inflammation from ongoing aspiration.

We also sought to compare pulmonary hospitalization risk between those cleared and not cleared to breastfeed since 13% of the cohort required hospitalization prior to their VFSS and this pulmonary hospitalization rate was similar among those both cleared and not cleared to breastfeed. Notably, during the year after VFSS, patients cleared to breastfeed had significantly more acute pulmonary hospitalizations compared to those not cleared to breastfeed. This might suggest that cessation of breastfeeding in infants with oropharyngeal dysphagia may decrease their chances of hospitalization but larger studies with longer follow-up periods may be required to determine this more definitively.

Our results may contrast with those of Hersh et al, which suggested breastfed infants with aspiration in their cohort were not likely to have pulmonary exacerbations⁸. However, pulmonary health was defined by clinical symptoms and only monitored for 3 months in that study. Given the high proportion of infants with silent aspiration, we focused on a longer time course and more objective markers of adverse pulmonary effects in the present study^{3, 11}. Based on our results, continued breastfeeding in infants with oropharyngeal dysphagia does not appear to be without risk.

It is notable that a sizeable proportion of the infants in this study were cleared to breastfeed despite being found to have aspiration or laryngeal penetration on their swallow study. This group even included a number of infants diagnosed with silent aspiration for whom symptom monitoring cannot provide a reliable assessment of pulmonary risk^{3, 11}. Our record review could not uncover any patterns as to why some infants were cleared to breastfeed even with significant oropharyngeal dysphagia. It may be that some of these infants were thought to have little experience with bottle feeding but the majority were not exclusively breastfed before the VFSS. It may also be that assumptions were made about change in position or flow at the breast possibly being more favorable. Some SLPs might have deferred to the medical team to make definitive recommendations in regards to continued breastfeeding but we had to assume that safety recommendations explicitly stated in the VFSS report were consistent with those shared with families and referring providers. It is also possible that those making recommendations about continued breastfeeding made assumptions about breastmilk having anti-inflammatory properties that may make it less harmful if aspirated given the paucity of research in this area^{41, 42}. However, our finding of elevated neutrophils on bronchoalveolar lavage in those cleared to breastfeed suggests this is not the case. The comparison of outcomes with our group of bottle-fed controls also suggests feeding at the breast is at least not less inflammatory for the lungs compared to bottle feeding. Our results are in agreement with prior reports of increased pulmonary inflammation and decreased lung compliance in rabbit lungs instilled with human milk^{25, 26}. It was notable that a lower proportion of patients were cleared to breastfeed during more recent years, however, perhaps suggesting a change in practice over time.

No subject in the present study underwent breastfeeding FEES and therefore the SLPs and referring providers likely had to extrapolate findings from VFSS to the swallowing physiology taking place during breastfeeding. To our knowledge, there is only one recent study that utilized VFSS to compare swallow function between bottle feeding and breastfeeding¹⁷. This study noted differences in multiple swallow components between these feeding methods but included only 25 infants with a relatively low proportion found to have aspiration or laryngeal penetration so accurate sensitivity analyses could not be performed. In contrast, older studies actually suggested more concordance between swallow function by bottle and breast using VFSS¹⁸. Given this, our data can be generalized to other institutions like ours that do not have ready access to breastfeeding FEES. For institutions that rely on FEES to assess breastfeeding infants, it may be that this approach can be used to identify modifications that may make breastfeeding safer, but it should be noted that this exam may also be limited by lower sensitivity compared with VFSS^{12, 14, 16}.

Pediatricians and feeding specialists along with other clinicians play a critical role in counseling families regarding breastfeeding. There has been a laudable push to recognize the many benefits of breastfeeding and support families to enable successful breastfeeding for the benefit of both mothers and children; however, it is important to recognize that feeding at the breast may not be the safest feeding method for infants with oropharyngeal dysphagia. Luckily, feeding thickened breast milk by bottle offers an alternative for families such that infants can receive nutritional and immunologic benefits of breastmilk while reducing risk of aspirating breast milk. This can be accomplished though thoughtful collaboration with teams of experts who care for children at risk for aspiration including speech language pathologists, pulmonologists and gastroenterologists^{7, 9}. In fact, many children who were not cleared to breastfeed in the present study were still receiving expressed breastmilk in follow-up. Given the potential for pulmonary inflammation in those that continue to breastfeed, clinical follow-up and monitoring is essential as reliance on symptoms alone is not adequate to assess risk.

There are a number of limitations to consider when interpreting our results. This was a retrospective study limited to data within our institution’s medical record system. Another limitation is that no subject underwent breastfeeding FEES so we recognize that what was seen during VFSS may not match what would have been seen by FEES. Future studies will be required to assess long term outcomes in aspirating breastfed patients with FEES assessment. We also had to make the assumption that breastfeeding continued in those cleared to breastfeed and not in those not cleared to breastfeed although it is possible that those recommendations were not consistently followed at home after the swallow study. Additionally, not all subjects underwent CXR or bronchoscopy. We acknowledge that neutrophils seen on BAL may be due to other causes, including reactive airway disease and viral infections, and not just aspiration. In our medication review, we only had access to our institution’s electronic medical records and therefore it is possible we were not able to capture all medications; however, both ICS and antibiotics for respiratory infections would typically be managed the pulmonary team for a child with dysphagia and would have been included in our system. Finally, validated penetration-aspiration scales were not available because of the timespan over which we collected data. Without these, there may have been inconsistencies in reporting the findings seen during VFSS that could have influenced the results.

Conclusions

Continued breastfeeding in infants with oropharyngeal dysphagia is associated with increased risk of pulmonary inflammation and pulmonary hospitalizations. In infants with known oropharyngeal dysphagia, the benefit to continued feeding at the breast needs to be weighed against the risk of pulmonary inflammation and, when possible, administration of thickened breast milk by bottle should be considered as an important alternative for preserving lung health.

Abbreviations

AAP

American Academy of Pediatrics

BAL

Bronchoalveolar lavage

CXR

Chest x-ray

FEES

Fiberoptic endoscopic evaluation of swallowing

ICS

Inhaled corticosteroid

SLP

Speech-language pathologist

VFSS

Videofluoroscopic swallow study

Data Availability Statement:

The data that support the findings of this study are available from the corresponding author upon reasonable request.