Respiratory Outcomes for Ventilator-Dependent Children With Bronchopulmonary Dysplasia Free

Subjects were recruited from the BPD Collaborative Outpatient Registry as a convenience sample. Inclusion criteria included a diagnosis of BPD by the 2001 National Heart, Lung, and Blood Institute workshop definition, birth between 2016 and 2021, and discharge on home ventilator support.^13,15  Patients who died after tracheostomy placement, but before hospital discharge, were excluded. Twelve tertiary care centers with experience in home ventilator management contributed data, including Arkansas Children’s Hospital, Children’s Hospital of Philadelphia, Children’s Mercy Hospital Kansas City, Riley Children’s Hospital, Johns Hopkins Children’s Center, Boston Children’s Hospital, C.S. Mott Children’s Hospital, Nationwide Children’s Hospital, Lucille Packard Children’s Hospital Stanford, Children’s Hospital of Los Angeles, Monroe Carell Jr. Children’s Hospital at Vanderbilt, and Phoenix Children’s Hospital. Participating centers obtained local review board approval, and data use agreements were completed between institutions before compiling anonymized data housed at Nationwide Children’s Hospital. Caregivers gave informed consent for registry participation and chart review in accordance with local institutional review board policies.

Data were retrospectively extracted at a single timepoint for each subject between November 2021 and March 2022 from electronic medical records using a standardized collection instrument developed by members of the BPD Collaborative. Race/ethnicity were reported by caregivers. Birth weight percentiles by gestational age were derived from national norms published in 2003.¹⁶  Pulmonary hypertension refers to any pulmonary hypertension observed via echocardiogram or cardiac catheterization on or after 36 weeks PMA. Pulmonary antihypertensive medications include any outpatient use of phosphodiesterase type 5 inhibitors, endothelin receptor antagonists, or prostanoids. Airway lesions include any documentation of vocal cord lesions or paralysis, suprastomal collapse, subglottic stenosis, and/or tracheobronchomalacia. Airway surgeries in preparation for decannulation include cleft repair, airway reconstruction, tonsillectomy, and/or adenoidectomy. Feeding tubes include gastrostomy, gastrojejunostomy, or jejunostomy tubes. First respiratory clinic visit refers to the first outpatient visit after initial hospital discharge to a pediatric pulmonary clinic or NICU follow-up clinic that provides pulmonary care. All ages at key events (ie, tracheostomy placement, initial discharge to home, first respiratory clinic visit [as a potential marker of access to care], complete liberation from the home ventilator, and decannulation) were recorded in chronological age rounded to the nearest month.

Median age and associated percentiles for key events were generated using Kaplan-Meier analysis with right censored data to account for subjects who may not have experienced key events at the time of chart review (eg, decannulation) (Table 1 and Fig 1). Differences between centers for the timing of key events were assessed via log-rank tests (Table 2 and Fig 2). Stata IC 15.0 (StataCorp, College Station, TX) was used for all analyses. P values < .05 were considered significant.

Of the 155 recruited subjects, the majority were male (54.8%); almost one-half (47.1%) were Black/African American and 11.8% were Hispanic (Table 3). The mean gestational age was 25.9 ± 2.5 weeks and mean birth weight was 757 ± 371 g. The mean birth weight percentile corrected for gestational age was 32 ± 25%, with 26.5% meeting criteria for classification as small for gestational age based on a birth percentile less than or equal to the tenth percentile. The median age at the time of chart review was 32 months of age (mean ± SD: 35.0 ± 14.6 months). Of the 155 subjects, 99 (66.0%) had been weaned from the ventilator at the time of the study, including 49 (32.6%) who had been decannulated.

The study population had high rates of pulmonary vascular, airway, and gastrointestinal morbidities (Table 4). More than one-half of subjects (58.7%) had pulmonary hypertension after 36 weeks’ PMA with one-third (33.5%) requiring outpatient pulmonary antihypertensive medications. Additionally, almost one-half of subjects were discharged on diuretic therapy (48.0%). Airway lesions affected most subjects (73.4%), with large airway malacia and subglottic stenosis present in 48.4% and 31.0% of subjects, respectively. Most subjects (80.7%) had a cuffed tracheostomy at discharge. Almost every subject (96.8%) underwent surgical placement of a feeding tube with more than one-third (35.1%) requiring procedural management of gastroesophageal reflux (Nissen fundoplication, gastrojejunostomy tube, or surgical jejunostomy). There were 16.1% of subjects who were reported to have a syndrome or significant congenital anomaly, including omphalocele, congenital diaphragmatic hernia, or cyanotic heart disease, which could have contributed to their respiratory status. Readmissions were common in the study population, with 73.9% of subjects having at least 1 readmission for respiratory reasons within 12 months of initial hospital discharge. The median duration between initial hospital discharge and first readmission was 5 months using Kaplan-Meier analysis (n = 149). The mortality rate was 2.6% within the follow-up period. One death was secondary to a tracheostomy complication; the cause of the other 3 deaths was not noted but was not secondary to progression of BPD or pulmonary hypertension.

Median age of key events was determined through Kaplan-Meier analysis to account for subjects who may not have experienced specific events at the time of chart review (Table 1 and Fig 1). The median age of tracheostomy placement was 5 months of age; referring to mean gestational age of the study population as 25.9 weeks, the median corrected age of tracheostomy placement would be estimated as 48 weeks’ PMA for the study population. The median age of hospital discharge was 10 months of age, or 5 months after tracheostomy placement. However, 29% of subjects remained admitted for 6 months or longer after tracheostomy placement, and 7.1% remained admitted for a year or more. The median age of first respiratory clinic visit was 11 months of age, and 81.9% of subjects appear to have been seen within 1 month of initial discharge. The median age of liberation from mechanical ventilation was 27 months. To determine if center-level factors played a role in the timing of these key events, log-rank tests were performed. Of note, the median ages for all key events differed by center (P ≤ .001 for all events) (Table 2 and Fig 2); this was similarly true when considering only centers that contributed at least 10 subjects (P < .001 for all events). Additionally, subjects with a syndrome or significant congenital anomaly had earlier tracheostomy placement (P = .001) and initial discharge (P = .008) than those without such diagnoses, but there were no differences in ages of first visit (P = .08) or ventilator liberation (P = .21). Last, we also assessed age of key events relative to initial hospital discharge, and observed the median time to first visit after hospital discharge was 1 month and ventilator liberation was 16 months. These durations continued to differ by center (P = .006 for first visit; P = .029 for ventilator liberation).

The median age of decannulation was slightly over 4 years of age (49 months). Of the 49 subjects who were decannulated, almost one-half (49.0%) required airway surgery, most commonly airway reconstruction, followed by tonsillectomy and/or adenoidectomy. The mean age for decannulation for the 24 subjects requiring airway surgery was similar to the 25 subjects who did not require airway surgery (36.5 ± 2.4 vs 35.2 ± 2.3 months; P = .69). More tracheostomy decannulations occurred during the summer months (32.7% June-August), relative to the fall (26.5% September-November), spring (22.5% March-May), and winter (18.4% December-February). The distribution of season for decannulation did not differ from an equal probability by season model (likelihood-ratio χ²P value = .54). The median age for decannulation was found to differ significantly based on center (P < .001); this was also true when considering centers contributing more than 10 subjects (P < .001). Additionally, there were no differences in age of decannulation (P = .13) between subjects with a syndrome or significant congenital anomaly compared than those without such diagnoses. Last, we observed the median time to decannulation after hospital discharge was 35 months, and this duration continued to differ by center (P = .002).

In this multicenter registry-based study of infants with severe BPD on long-term mechanical ventilation in the home setting, we found that the median time for liberation from positive pressure ventilation was 27 months of age, which approximates an average duration of 17 months of home ventilation. We also found decannulation took place an additional 22 months after liberation from home ventilation. The duration of respiratory support is likely influenced in part by stochastic morbidity events and center-level factors and therefore may be modifiable.

Not unexpectedly, our study population was notable for high frequencies of comorbidities, including pulmonary hypertension (58.7%), feeding tubes (96.8%), and airway lesions (73.4%). Pulmonary hypertension is also more common among infants with severe BPD,¹⁷  and its presence has been independently associated with the need for tracheostomy.¹⁸  In our study, almost all infants required a feeding tube and some received fundoplication and jejunal feeding to manage their gastroesophageal reflux disease, which has been observed in other smaller studies.^19,20  Airway lesions are also more common among preterm infants with BPD,²¹  and fixed airway lesions may require surgical interventions before decannulation. Additionally, dynamic airway lesions are common in children with severe BPD and although this usually improves over time, dynamic obstruction from central airway collapsibility (tracheomalacia and bronchomalacia) may require tracheostomy and mechanical ventilation until improvement or resolution. Our results emphasize the need to discuss the additional possibility of feeding tube placement and need for airway reconstruction surgeries with families considering tracheostomy placement to manage severe BPD.

The high frequencies of comorbidities likely contributed to a high rate of rehospitalization within the first 12 months at home (73.9%) in our ventilated population, which is higher than what has been previously reported in infants with severe BPD alone, including nonventilated patients (39.4%).²²  Although we were not able to determine the specific indications for readmissions, single-center studies have identified respiratory infections and tracheostomy-related complications to be common among these infants.^19,23–25  These data highlight the importance of preventive counseling at initial discharge, particularly the recommendation to avoid exposures to respiratory viruses and receive appropriate antiviral prophylaxis, such as influenza and COVID-19 vaccinations and palivizumab.²⁶  However, despite the high morbidity rate, we found that the mortality rate among our cohort was low (2.6%) compared with other published single-center reports.^10,19,20,27  Of interest, none of the deaths was related to the primary respiratory disease, although 1 was related to a tracheostomy complication.

Of key importance in the management of these children was the variability among centers regarding tracheostomy placement, initial hospital discharge, first follow-up at clinic visit, timing of liberation from the ventilator, and decannulation. Although American Thoracic Society clinical practice guidelines exist,²⁸  our results confirm a previous study that highlighted diversity of management strategies among geographically distinct tertiary care BPD centers regarding management of severe BPD after hospital discharge,¹³  and extend our knowledge on the longer term respiratory outcomes of ventilator-dependent infants with severe BPD and tracheostomy.

There are several factors that we did not examine, which may explain the intercenter differences in the key events that we observed in this study. First, the decision to place a tracheostomy in infants with severe BPD is highly complex, including both center-level practices and caregiver decision-making. As seen in the multicenter Children’s Hospitals Neonatal Consortium database, the incidence of tracheostomy in infants with severe BPD correlates with patient volume at BPD centers, with larger volume centers performing more tracheostomies and discharging more patients from the hospital on chronic mechanical ventilation.⁵  Additionally, the timing of tracheostomy placement shows marked variability among different centers and has been reported to range from 41 to 51 weeks’ PMA in single-center studies.^{18–20,29–31}  The discharge process for infants with tracheostomy and chronic home ventilation may also vary between centers and depends on a host of factors including completion of required education and training of family caregivers, family psychosocial status, social determinants of health, and availability of skilled home nursing services locally.^32,33  In particular, social determinants of health may persist and change from birth to decannulation.^34,35  These barriers to discharge can cause significant delays in discharge despite infants being medically ready to go home. Standardized discharge processes have been shown to decrease hospital stay of tracheostomy and ventilator-dependent children.^36,37 

Outpatient care usually involves a multidisciplinary team led by a pulmonologist or neonatologist in a BPD center. This multidisciplinary approach can allow for a smoother transition of care from an intensive care setting to an outpatient environment.¹⁰  Each local center may have different levels of resources available to them to accomplish this level of care coordination, which may contribute to variability in the timing of first clinic visit. The wide variability in weaning from the ventilator can depend on clinical stability of the patient,^38,39  the frequency of follow-up in the outpatient setting, and perhaps the mode of ventilation or COVID-19 pandemic concerns as well. At some centers, telehealth visits have been used to successfully wean ventilator support in a timely fashion.^40,41  We also found that there was a wide variation in the timing of decannulation among our subjects in this study. This variability is likely due in part to local requirements, access to polysomnography, local otolaryngology practices, and preferences on the decision to decannulate.^10,19,20,31  The need for airway surgery was quite high prior decannulation. However, we found that the performance of such surgery did not necessarily delay decannulation. Last, there may be differences in disease severity among the patients with severe BPD in the outpatient cohort that account for some of the differences in their respiratory outcomes.

There are limitations to this study, including the retrospective nature of the study and a median follow-up age of 32 months, which may not have allowed capture of all events because not all subjects may have progressed to those events. With the median patient age of 32 months, we may not have captured all mortality events but would note that the risk for mortality for those with BPD is higher than the general population even into childhood.⁴²  Nevertheless, subjects in this study are from recent cohorts (2016–2021) and thus likely represent current practices. Also, our convenience sample has omitted patients who may have had outcomes could have affected or altered the reported results. Our sample also consists of a heterogenous group of infants with severe BPD encompassing phenotypes that manifest both from prematurity and in a subset, from other disease processes (eg, congenital diaphragmatic hernia, omphalocele) from which this report has only a few patients. Another limitation was that our cohort of subjects was drawn from major tertiary care children’s centers and may not capture infants and children who receive care at smaller centers in the United States. We would note that these reported results are from a heterogenous group of infants with severe BPD encompassing phenotypes that manifest both from prematurity and in a subset, from other disease processes (eg, congenital diaphragmatic hernia, omphalocele, congenital heart disease) from which this report has only a few patients.

Also, in the absence of a consensus definition for pulmonary hypertension, we used an unvalidated definition of pulmonary hypertension; two subjects without a diagnosis of pulmonary hypertension had received sildenafil as an outpatient. Finally, although we did note differences between centers in terms of timing of events, centers did not provide uniform numbers of subjects, and there was variation in the ages of the subjects at the time of chart review, which may lead to some bias.

In conclusion, we found a wide variability in respiratory outcomes of ventilator-dependent infants with severe BPD. This may be due to the number of patients treated at each center, the complexity of the cases, and the overall strategies used by clinicians when managing patients with severe BPD, including decision-making around tracheostomies. Further studies are needed to identify etiologies of intercenter practice variability and specific phenotypes in children with severe BPD who require tracheostomies and home mechanical ventilation. Finally, establishing a universal guideline for liberating children with BPD from home ventilation and tracheostomy may decrease morbidities and improve long-term health for this population, including growth and neurodevelopment, in addition to respiratory outcomes.

BPD

bronchopulmonary dysplasia

PMA

postmenstrual age