Long-term pulmonary function outcomes in children with pulmonary hypoplasia

Abstract

Objective:

To determine if the underlying cause of pulmonary hypoplasia results in different trajectories of lung growth by describing pulmonary function in 8- to 13-year-old children born with congenital diaphragmatic hernia (CDH), early intervention congenital lung malformation (EICLM), and giant omphalocele (GO).

Study Design:

We performed spirometry and plethysmography, and echocardiograms in 81 children aged 8- to 13-years with CDH, EICLM, and GO. Clinical and demographic data were collected at the study visit and from electronic medical records. Quantitative analyses of right ventricular (RV) function were retrospectively performed.

Results:

Fifty-two children with CDH, 17 with EICLM, and 12 with GO were included in the study. Most patients (51.9%) had abnormal lung function. Those with CDH and EICLM were more likely to have an obstructive process, while subjects with GO frequently had restrictive disease. Chest wall abnormalities, patch repair, larger diaphragm defect, and intrathoracic liver position were associated with abnormal lung function in patients with CDH. Abnormal lung function tended to be associated with lower right ventricular function, although this was not statistically significant.

Conclusions:

Abnormal lung function persists into late childhood in patients with CDH, EICLM, and GO. However, specific patterns of pulmonary function abnormalities occur in each disorder. This suggests that lung growth and remodeling likely differ depending on the underlying cause of pulmonary hypoplasia.

Introduction

Normal lung development begins as early as 22 days gestation when a lung bud projects from the ventral foregut and divides into right and left lung buds. ^1,2 A variety of aberrant processes occurring after this time can impact lung development and result in pulmonary hypoplasia.^3,4 Three highly morbid^5–7 causes of pulmonary hypoplasia are giant omphalocele (GO), congenital diaphragmatic hernia (CDH), and large congenital lung malformations (CLMs).

In CDH, incomplete diaphragm closure leads to the herniation of abdominal viscera into the thoracic cavity. Patients have reduced alveolar surface area and airway branching, as well as decreased pulmonary vascularization with increased pulmonary arterial smooth muscle.⁸ The term CLM encompasses several space-occupying thoracic lesions including congenital pulmonary airway malformation (CPAM), bronchopulmonary sequestration (BPS), and congenital lobar overinflation (CLO). To date, no study has exclusively assessed long-term pulmonary function outcomes in children with large, space occupying CLMs requiring either fetal intervention or resection within the first 48 hours of life (early intervention CLM, EICLM). Omphalocele is characterized by the extrusion of abdominal viscera, covered by a membrane, outside of the abdominal cavity. Giant omphalocele refers to a large evisceration that contains most of the liver.^9,10 In general, small omphaloceles are not associated with pulmonary hypoplasia. However, in some patients with GO, lung growth is impaired beginning in the fetal period.^9,10 Although infant pulmonary function has been reported in survivors of GO⁹, long-term pulmonary function has not been described to date. In each of these three disorders (CDH, EICLM, GO), the pathogenesis of pulmonary hypoplasia differs, and a comparison of long-term lung function outcomes has not been performed.

Pulmonary function, as measured by forced vital capacity (FVC) and forced expiratory volume in one second (FEV₁), normally follows three trajectories across the lifespan: (1) increase in pulmonary function throughout childhood due postnatal alveolarization (during the first few years of life) and increased alveolar and airway size (2) lung function plateau during early adulthood, and (3) steady lung function decline with aging.¹¹ Children with lung growth abnormalities that cause either a reduction in airway size or number, or a reduction in the number of alveoli, would be expected to achieve a lower peak lung function, placing them at risk for developing early-onset symptomatic lung disease, such as chronic obstructive pulmonary disease (COPD).^11–13 Therefore, identification of patients at risk for abnormal lung function during childhood is essential to allow for targeted interventions like tobacco smoke avoidance, treatment of concomitant diseases such as asthma, and reducing negative environmental and occupational exposures.^11–13

In July 2004, we established an interdisciplinary program at our institution to care for children with pulmonary hypoplasia. The current study aimed to describe pulmonary function in children aged 8 to 13 years with GO, CDH, and EICLM. We hypothesized that pulmonary function abnormalities would differ between the groups, based on the underlying cause of pulmonary hypoplasia. We also aimed to describe factors associated with abnormal lung function in this cohort.

Methods

The Institutional Review Board at the Children’s Hospital of Philadelphia approved the current study, and all parents or legal guardians gave written or verbal informed consent for their children to be included (IRB 18–015396).

Patient population

This is a single center cross-sectional study of pulmonary function among 8- to 13-year-old children diagnosed with pulmonary hypoplasia within the first 28 days of life and followed at the Children’s Hospital of Philadelphia Pulmonary Hypoplasia Program. For this study, research visits occurred from September 2016 to October 2018. All eligible patients were emailed a recruitment letter regarding the study. Parents of patients were then contacted by phone for enrollment. Patients were randomly selected to be contacted until the target enrollment of 100 participants was reached (convenience sample). The subset of patients who completed pulmonary function testing is included in this study.

Our program defined GO as a large abdominal evisceration covered by a membrane containing most of the liver.^9,10,14 For CDH patients, the defect size was graded from A to D, from smallest to largest, by direct visualization during operation, in accordance with published consensus criteria.¹⁵ Only CLMs that required prenatal intervention or resection within the first 48 hours of life due to respiratory distress were included (EICLM). We reviewed the medical records of children who enrolled in this study and included those who performed pulmonary function testing (spirometry and/or plethysmography). Demographic and clinical data were obtained during the initial study visit and retrospectively from the electronic medical record. For this study, we documented the following chest wall abnormalities: chest wall asymmetry, pectus carinatum/excavatum, and rib flaring. In our experience, nearly all patients with GO have a narrow chest, and therefore, this was not included as a specific chest wall abnormality to compare between the cohorts.

Echocardiogram

Prospectively acquired echocardiograms were performed by a single pediatric sonographer (AF) using standard pediatric views with 3–8 MHz transducers on a Phillips IE33 machine (Phillips, Andover, MA, USA) in accordance with our echocardiography laboratory’s standard imaging protocol at the time of the study visit and digitally stored in the Syngo Dynamics system (Siemens, USA). Additional quantitative analyses of right ventricular (RV) function were retrospectively performed at the time of this sub-analysis. Two pediatric cardiac sonographers (YW and DA), blinded to clinical characteristics, obtained offline measures of RV systolic function that predict adverse outcomes in CDH and other forms of developmental lung diseases.^16–20 RV global longitudinal strain (RVGLS) and RV free wall strain (RVFWS) measurements of systolic myocardial deformation were obtained using Tomtec software (Image Arena 4.6; Munich, Germany). Strain is reported as a negative number, with a greater absolute value indicating better systolic function.²¹ Absolute value of the strain measures was used in this study. RV systolic function was also assessed by tricuspid annular plane systolic excursion Z-scores (TAPSEZ) and RV fractional area change (RVFAC) calculated as the end-diastolic area minus the end-systolic area divided by the end-diastolic area.^22–24

Pulmonary function testing

Spirometry and plethysmography were performed using a Morgan Scientific Medisoft Plethysmograph and ComPAS software. All studies followed American Thoracic Society (ATS) and European Respiratory Society (ERS) technical standards. We assigned a grade to each spirometry study according to ATS/ERS guidelines, and only included studies that received grade A or B for forced expiratory volume in one second (FEV₁) and forced vital capacity (FVC).²⁵ Only acceptable plethysmography studies according to the ATS/ERS guidelines were included.²⁵ Race-neutral reference equations were used, consistent with an updated American Thoracic Society statement.²⁶ All pulmonary function testing was performed in the outpatient setting, and subjects were free from illness at the time of the study. Bronchodilator testing was not performed. In addition to reporting median PFT values, we identified abnormal pulmonary function patterns in subjects who completed valid spirometry and plethysmography, using a z-score less than or greater than 1.65 to define a particular value as abnormal for all measures, consistent with the updated ERS/ATS technical standard.²⁷ A subject with a decreased FEV₁/FVC was deemed to have an obstructive process. A decreased total lung capacity (TLC) and normal or increased FEV₁/FVC was defined as a restrictive process. A mixed restrictive/obstructive disorder was defined as decreased FEV₁/FVC and a low TLC. An increased residual volume (RV)/TLC accompanied by a normal FEV₁/FVC and normal or increased TLC was categorized as lung overdistension. A pattern of impaired chest wall mechanics was defined as a normal FEV₁/FVC, increased RV/TLC, and a low TLC. This category would include individuals with respiratory muscle weakness or an abnormally stiff chest wall. When assessing variables associated with abnormal lung function, we defined abnormal lung function as either: low FEV₁/FVC, low TLC, or increased RV/TLC.

Statistical analysis

Continuous variables were summarized using the median (interquartile range, IQR, 25%, 75%) or mean (standard deviation) depending on the data distribution, as assessed by histograms, and categorical variables were summarized as frequencies. Continuous variables that were assumed to be non-normally distributed were compared between dichotomous lung function groups using the Wilcoxon rank sum test and between the diagnosis categories using the Kruskal-Wallis test. All spirometry and plethysmography result variables were normally distributed and therefore compared between the three groups using Analysis of Variance (ANOVA). A Tukey HSD post-hoc analysis was used to adjust for multiple comparisons between group comparisons of these variables. Categorical variables were compared using the chi-squared test of independence or Fisher’s exact test when the expected values in the cells were close to 5. We also used linear regression models to evaluate the association between abnormal lung function and a list of a priori chosen cardiac function variables. These models were adjusted for the potential confounding effects of biological sex, self-reported race defined as White/Non-White, age at pulmonary function testing, and use of Extracorporeal Membrane Oxygenation (ECMO) in the neonatal period. Statistical significance was defined as P < 0.05. All analyses were conducted in R version 4.2.

Results

Patient population

During the study period, the pulmonary hypoplasia program followed 626 patients, 189 of whom were between the ages of 8 and 13 years old. Of the 100 patients enrolled in the study, 97 attempted to perform pulmonary function testing. Fifteen subjects were excluded because of inadequate effort or reproducibility on pulmonary function testing. We also excluded one subject with pentalogy of Cantrell because he had both GO and CDH, and categorization of the main cause of his pulmonary hypoplasia could not be clinically determined. Ultimately, 81 subjects were included in the study. Fifty-two had CDH, 17 had EICLM, and 12 had GO. Four subjects had a small CLM in addition to CDH (n=3) or GO (n=1). We categorized these patients in the CDH or GO categories and not the EICLM category because the CLM did not require fetal intervention or resection within the first 48 hours of life. Of those included in the study, 64 (79%) subjects had valid spirometry and plethysmography studies, and 17 (21%) subjects had valid spirometry alone.

Patient demographic and clinical characteristics are shown in Table 1. There were no significant differences in age, weight, or height at the time of pulmonary function testing between the three cohorts. The median gestational age at delivery for all subjects was 38.1 weeks, with a small but statistically significant difference between the three groups, likely attributable to a lower gestational age in the GO cohort (p = 0.02). No subjects had a tracheostomy or used any form of respiratory support at the time of pulmonary function testing. Only one subject was prescribed pulmonary hypertension (PH) medications at the time of testing. Corticosteroids and bronchodilators were prescribed to a higher percentage of subjects with CDH, although this difference was not statistically significant. Twelve (14.8%) subjects had a history of extracorporeal membrane oxygenation (ECMO) support during their initial hospitalization after birth (10 in the CDH group, two in the EICLM group).

Table 1.

Baseline demographic and clinical characteristics. Continuous variables are expressed as median and interquartile range [Q1,Q3]. CDH, congenital diaphragmatic hernia; EICLM, early intervention congenital lung malformation; GO, giant omphalocele; GERD, gastroesophageal reflux disease

Variable	CDH (n=52)	EICLM (n=17)	GO (n=12)	Overall (n=81)	P-value
AGE (yrs)	10.3 [8.94, 12.2]	10.3 [9.10, 11.9]	10.8 [8.74, 11.5]	10.4 [8.94, 11.9]	0.941
WEIGHT (kg)	31.9 [24.1, 40.1]	33.3 [30.5, 47.0]	27.6 [22.6, 34.6]	32.6 [25.4, 40.6]	0.116
HEIGHT (cm)	138 [128, 148]	140 [135, 148]	134 [129, 150]	138 [131, 149]	0.49
GESTATIONAL AGE (wks)	38.4 [37.6, 39.1]	37.7 [37.0, 38.4]	37.5 [33.9, 38.2]	38.1 [37.2, 39.0]	0.02
CHEST WALL					0.317
Chest wall asymmetry	6 (11.5%)	1 (5.9%)	0 (0%)	7 (8.6%)
No chest wall deformity	35 (67.3%)	12 (70.6%)	12 (100%)	59 (72.8%)
Pectus carinatum	4 (7.7%)	0 (0%)	0 (0%)	4 (4.9%)
Pectus excavatum	7 (13.5%)	3 (17.6%)	0 (0%)	10 (12.3%)
Rib flaring	0 (0%)	1 (5.9%)	0 (0%)	1 (1.2%)
SCOLIOSIS	11 (21.2%)	2 (11.8%)	0 (0%)	13 (16.0%)	0.177
MEDICATIONS at time of study
Bronchodilators	36 (69.2%)	9 (52.9%)	5 (41.7%)	50 (61.7%)	0.148
Inhaled corticosteroids	22 (42.3%)	4 (23.5%)	3 (25.0%)	29 (35.8%)	0.305
Allergic rhinitis medications	24 (46.2%)	5 (29.4%)	5 (41.7%)	34 (42.0%)	0.478
Pulmonary vasodilators	1 (1.9%)	0 (0%)	0 (0%)	1 (1.2%)	> 0.99
GERD medications	8 (15.4%)	2 (11.8%)	1 (8.3%)	11 (13.6%)	> 0.99
Diuretics	1 (1.9%)	0 (0%)	0 (0%)	1 (1.2%)	> 0.99
HISTORY OF ECMO	10 (19.2%)	2 (11.8%)	0 (0%)	12 (14.8%)	0.27

Chest wall abnormalities and scoliosis were only observed in subjects with CDH and EICLM (Table 1). Nearly all subjects (91.3%) with CDH underwent open surgical repair, and 22 (42.3%) required a patch repair. The diaphragm defect was on the left in most cases (86.5%). The most common grade defect was A (57.7%). Intrathoracic liver position was seen in 22 (42.3%).

In the GO group, eight patients (66.7%) underwent Schuster staged closure, three (25%) underwent primary closure, and information about surgical technique was missing in one subject. No subjects had scoliosis or a chest wall deformity.

Fourteen (82.4%) patients with EICLM had CPAM, one (5.9%) had BPS, one (5.9%) had CLO, and one (5.9%) had persistent pulmonary interstitial emphysema.

Pulmonary function

Spirometry

Spirometry results are shown in Figure 1 and Table 2. The average FEV₁/FVC was lower in subjects with CDH and EICLM compared to those with GO (ANOVA p-value 0.007, multiple comparison adjusted p-values of 0.005 and 0.052, respectively –Table S1). The FEV₁ and FVC tended to be lowest in the GO group, while the trend was for FEF_25–75 to be lower in CDH and EICLM subjects compared to those with GO, although these differences were not statistically significant.

Figure 1.

Box and whisker plots depicting lung function testing results (spirometry and plethysmography). Each box extends from the 25^th to 75^th percentile of values; the line within each box represents the median value. CDH, congenital diaphragmatic hernia; EICLM, early intervention congenital lung malformation; GO, giant omphalocele

Table 2.

Spirometry results. Values are expressed as mean (standard deviation). pp, percent predicted; CDH, congenital diaphragmatic hernia; EICLM, early intervention congenital lung malformation; GO, giant omphalocele; FEV₁, forced expiratory volume in one second; FVC, forced vital capacity; FEF_25–75, forced expiratory flow between 25 and 75 percent of forced vital capacity; pp, percent predicted

Variable	CDH (n=52)	EICLM (n=17)	GO (n=12)	Overall (n=81)	P-value
FEV1 pp	85.7 (18.7)	83.3 (18.1)	82.2 (14.0)	84.6 (17.8)	0.787
FEV1 z-score	−1.07 (1.4)	−1.25 (1.3)	−1.35 (1.1)	−1.15 (1.3)	0.773
FVC pp	95.9 (17.9)	90.9 (13.9)	84.2 (16.5)	93.1 (17.3)	0.088
FVC z-score	−0.32 (1.4)	−0.69 (1.1)	−1.22 (1.3)	−0.53 (1.3)	0.089
FEV 1 /FVC	78.7 (8.9)	79.9 (8.5)	87.4 (5.1)	80.2 (8.8)	0.007
FEV 1 /FVC pp	88.8 (9.9)	90.5 (10.1)	98.0 (6.1)	90.5 (9.9)	0.014
FEV 1 /FVC z-score	−1.36 (1.1)	−1.13 (1.3)	−0.22 (1.0)	−1.14 (1.2)	0.01
FEF25–75 pp	63.8 (25.2)	64.8 (25.0)	76.8 (18.8)	66.0 (24.5)	0.255
FEF 25–75	−1.56 (1.1)	−1.44 (1.1)	−1.03 (0.9)	−1.46 (1.1)	0.31

Lung volumes

Plethysmography results are shown in Figure 1 and Table 3. The average FRC percent predicted (pp) and z-score differed between the three groups (p=0.024 and p=0.009, respectively). With multiple comparison testing, FRC was significantly increased in patients with CDH compared to those with GO whether described by percent predicted (p = 0.021) or by z-score (p = 0.007). FRC was not significantly different between subjects with CDH and those with EICLM (pp, p=0.472 and z-score, p=0.485). The average TLC percent predicted and z-score were different between the three groups (p=0.017 and p=0.012, respectively). In multiple comparison testing, the TLC was lower in subjects with GO compared to those with CDH (pp, p=0.013 and z-score, p=0.008) and EICLM z-score (pp, p=0.071 and z-score, p=0.049). Average RV and RV/TLC tended to be increased in those with CDH and EICLM compared to the GO group, although these differences were not statistically significant.

Table 3.

Plethysmography results. Values are expressed as mean (standard deviation). CDH, congenital diaphragmatic hernia. EICLM, early intervention congenital lung malformation; GO, giant omphalocele; FRC, function residual capacity; RV, residual volume; TLC, total lung capacity; pp, percent predicted.

Variable	CDH (n=42)	EICLM (n=14)	GO (n=8)	Overall	P-value
FRC pp	124 (33.0)	113 (25.6)	91.3 (31.1)	118 (32.8)	0.024
FRC z-score	0.99 (1.32)	0.52 (1.2)	−0.603 (1.6)	0.687 (1.4)	0.009
RV pp	140 (59.7)	129 (58.7)	96.2 (38.0)	132 (58.3)	0.145
RV z-score	0.68 (1.0)	0.45 (1.03)	−0.166 (0.8)	0.523 (1.0)	0.093
TLC pp	104 (17.7)	102 (11.2)	85.1 (20.4)	101 (17.8)	0.017
TLC z-score	0.28 (1.3)	0.13 (0.8)	−1.20 (1.6)	0.064 (1.3)	0.012
RV/TLC z-score	1.11 (1.1)	0.80 (1.0)	0.514 (0.6)	0.968 (1.0)	0.257
RV/TLC pp	148 (47.2)	135 (44.1)	123 (26.4)	142 (44.8)	0.317
RV/TLC	31.6 (10.2)	29.0 (9.5)	26.6 (5.6)	30.4 (9.6)	0.334

We compared differences in abnormal pulmonary function patterns between groups (Figure 2). Forty-three percent of patients with CDH demonstrated an obstructive pattern, while 10% had a pattern of lung overdistension without obstruction. Similarly, an obstructive pattern was the most common abnormal pattern in EICLM subjects (36%). The only abnormal pulmonary function pattern seen in subjects with GO was a restrictive process (50%).

Figure 2.

Pulmonary function patterns of subjects who completed valid spirometry and plethysmography. CDH, congenital diaphragmatic hernia; EICLM, early intervention congenital lung malformation; GO, giant omphalocele.

Factors associated with abnormal pulmonary function

Overall, 42 (51.9%) patients had abnormal pulmonary function, defined as either a low FEV₁/FVC, low TLC, or high RV/TLC. This included 29 (55.8%) subjects with CDH, 8 (47.1%) with EICLM, and 5 (41.7%) with GO. In the CDH group, several factors were associated with abnormal lung function (Table 4), including chest wall abnormalities (p = 0.033), patch repair (p < 0.001), grade of defect (p < 0.001), and liver position (p = 0.017). There were no variables significantly associated with abnormal pulmonary function in patients with GO or EICLM, although this may have been the result of the smaller sample size in those groups.

Table 4.

Factors associated with abnormal lung function in congenital diaphragmatic hernia (CDH). GA, gestational age. Continuous variables are expressed as median and interquartile range [Q1,Q3].

Variable	Normal Lung Function (n=23)	Abnormal Lung Function (n=29, 55.8%)	Overall (n=52)	p-value
AGE (yrs)	10.2 [9.13, 11.5]	10.4 [8.74, 12.3]	10.3 [8.94, 12.2]	0.868
GA at Birth (wks)	38.7 [37.9, 39.2]	38.3 [37.2, 39.1]	38.4 [37.6, 39.1]	0.231
WEIGHT (kg)	33.0 [26.5, 40.8]	30.8 [23.4, 35.4]	31.9 [24.1, 40.1]	0.242
HEIGHT (cm)	139 [129, 144]	136 [129, 150]	138 [128, 148]	0.761
CHEST WALL ABNORMALITY	3 (13.0%)	14 (48.3%)	17 (32.7%)	0.033
Chest wall asymmetry	2 (8.7%)	4 (13.8%)	6 (11.5%)
Pectus excavatum	1 (4.3%)	6 (20.7%)	7 (13.5%)
Pectus carinatum	0 (0%)	4 (13.8%)	4 (7.7%)
SURGICAL TECHNIQUE
Minimally invasive	1 (4.3%)	0 (0%)	1 (1.9%)
Open	21 (91.3%)	29 (100%)	50 (96.2%)
Missing	1 (4.3%)	0 (0%)	1 (1.9%)
PATCH REPAIR	3 (13.0%)	19 (65.5%)	22 (42.3%)	< 0.001
CDH GRADE				< 0.001
A	20 (87.0%)	10 (34.5%)	30 (57.7%)
B	1 (4.3%)	8 (27.6%)	9 (17.3%)
C	0 (0%)	7 (24.1%)	7 (13.5%)
D	0 (0%)	1 (3.4%)	1 (1.9%)
Missing	2 (8.7%)	3 (10.3%)	5 (9.6%)
LIVER POSITION				0.017
Down	18 (78.3%)	12 (41.4%)	30 (57.7%)
Up	5 (21.7%)	17 (58.6%)	22 (42.3%)
CDH SIDE				> 0.99
Left	20 (87.0%)	25 (86.2%)	45 (86.5%)
Right	3 (13.0%)	4 (13.8%)	7 (13.5%)

Quantitative right ventricular function

Echocardiographic measures are described in Table 5. The average TAPSEZ and RVFAC were lower in those who had abnormal lung function compared to those without, although this association did not reach statistical significance. The differences in RVGLS and RVFWS are not clinically relevant.

Table 5.

Echocardiographic measures of right ventricular function associated with abnormal lung function. (a) Summary statistics and (b) Regression model adjusted for sex, race, age, and history of ECMO. TAPSEZ, tricuspid annular plane systolic excursion Z-score; RVFAC, right ventricular fractional area change; RVGLS, right ventricular global longitudinal strain; RVFWS, right ventricular free wall strain; CI, confidence interval

VARIABLE	Normal Lung Function (N=39)	Abnormal Lung Function (N=42)	P-Value
TAPSEZ	0.19 (−0.95, 1.74)	0.10 (−1.25, 1.64)	0.631
RVFAC	48 (44.4, 51.8)	45.5 (42, 49.5)	0.146
RVGLS	25.3 (23.2, 27.6)	26.4 (23.0, 28.5)	0.751
RVFWS	27.2 (25.8, 29.9)	28.0 (25.4, 31.2)	0.747
a

	TAPSE-Z	RVFAC	RVGLS	RVFWS
Coefficient	−0.29	−1.73	0.15	0.08
95% CI	[−1.35, 0.77]	[−4.42, 0.97]	[−1.42, 1.73]	[−2.07, 2.23]
b

Discussion

In this study, we report long-term pulmonary function outcomes in children with various causes of pulmonary hypoplasia. To our knowledge, this is the first report of pulmonary function in older children with GO and EICLM, and the first study to compare pulmonary function in children with different underlying causes of pulmonary hypoplasia. We observed pulmonary function abnormalities in most patients (51.9%) independent of the underlying diagnosis, despite their being several years removed from surgical correction of their congenital defects. Additionally, patterns of abnormal pulmonary function differed among etiologies of pulmonary hypoplasia, suggesting that subsequent lung growth differs depending on the underlying cause of pulmonary hypoplasia. An obstructive pattern was the most common abnormality in patients with CDH and EICLM, whereas restrictive lung disease was the most common abnormal pattern in those with GO.

Pulmonary hypoplasia in CDH has been hypothesized to be the result of a “dual-hit hypothesis.”²⁸ Abnormal cellular signaling pathways during embryogenesis lead to reduced alveolar number and airway branching, pulmonary vascular disease, and abnormal left ventricular development, even before the expected closure of the diaphragm in utero. Then, failure of the diaphragm to close results in mechanical compression of the lungs by abdominal viscera, which in turn further impairs lung growth, and exacerbates the existing pulmonary hypoplasia.^8,28,29 Several studies have assessed pulmonary function outcomes in individuals with CDH. Panitch et al. performed infant pulmonary function testing in 98 subjects with CDH during the first three years of life and found expiratory flows were reduced but airspace overdistension without airway obstruction was common (high RV/TLC and normal FEV_0.5/FVC). The authors hypothesized that persistent pulmonary hypoplasia and overdistension of simplified alveoli, in addition to compensatory overinflation of the contralateral lung, resulted in elevated lung volumes. Further, decreased elastic recoil by alveolar simplification would contribute to reduced expiratory flows.³⁰ In a longitudinal study of 119 subjects with CDH, Dao et al. found that the probability of developing obstructive lung disease (low FEV₁/FVC) increased by 33% each year, and that by early adulthood, nearly all subjects had obstructive lung disease.³¹ Peetsold et al. also found that obstructive lung disease was common in 53 older children and adolescents (mean age 11.9 years) with a history of CDH who underwent pulmonary function testing.³² Our findings are similar to these previous studies of older children with CDH. It seems that airspace overdistension is the primary means of volume compensation for pulmonary hypoplasia early in life (rather than more rapid increase in alveolar number). With time, the number of alveoli increases, while the number of airways does not, leading to expiratory flow limitation and an obstructive pattern on pulmonary function testing.³³

Previous studies have found the grade of defect³¹ and patch closure³⁰ to be associated with worse pulmonary function in patients with CDH. Intrathoracic liver position is associated with an increased need for ECMO, increased mortality, and worse early life lung function.^30,34,35 Our study confirms that early life abnormalities in lung function associated with patch closure, increased defect size, and intrathoracic liver position persist into late childhood. Every subject with a grade C or D defect had abnormal lung function, and nearly all (91%) subjects with normal lung function had a grade A or B defect. However, it is worth mentioning that a large proportion (46%) of those with grade A or B diaphragm defects had abnormal lung function in our study. Right laterality has been associated with pulmonary morbidity in several studies, likely due to the intrathoracic liver position that occurs in right sided CDH.^31,36 While there was no association between laterality and abnormal lung function in our cohort, this is likely due to the small number (n=7) of subjects with right sided CDH and a lack of statistical power.

Large CLMs compress lung parenchyma, leading to incomplete lung expansion and in turn to hypoplasia.⁵ Both the severity of the condition and the timing of surgical intervention play important roles in subsequent lung growth. Long term pulmonary function outcomes vary in different studies,^37–44 likely because of heterogeneity of these factors in the studied population. Most studies of pulmonary function in subjects with a history of CLM demonstrate either normal pulmonary function or mild obstructive lung disease.^39,40,44 However, our patient population is unique in that all required either prenatal treatment or excision of the CLM within the first 48 hours of life due to respiratory distress, whereas those included in previous studies underwent resection later in life. Therefore, our patients probably had more significant pulmonary hypoplasia. This may account for the large proportion (47.1%) of patients with EICLM having abnormal lung function in our study.

In GO, several explanations for the development of pulmonary hypoplasia have been proposed. Because the liver is outside of the abdominal cavity, the lower thoracic cage forms abnormally. Additionally, decreased intraabdominal pressure leads to indrawing of the lower rib cage during inhalation, and the abnormal lateral attachment of rectus abdominis muscles to the ribs exerts a downward force on the rib cage during muscle contraction. These factors lead to caudal rib declination and a narrow chest wall, which in turn is presumed to limit lung expansion and impair lung development. Others have also described diaphragm dysfunction, which further impairs lung expansion and development in utero.^45,46

A restrictive process was the most common abnormality in our subjects with GO. This is consistent with a study by Danzer et al. demonstrating that young children (mean age 19.3 months) with GO had significantly decreased TLC (69% predicted) and respiratory system compliance.⁹ While the larger median TLC (79.7 % predicted) in our study indicates some improvement in restriction occurred over time, the abnormality was still present in half of our GO patients. The restrictive process in these patients is probably related to a narrow chest wall,⁴⁶ which impairs lung expansion, resulting in a low TLC. Other factors may have also contributed to the restrictive pattern. Danzer et al. found that respiratory system compliance (C_rs) was low in young children with GO, likely reflecting a parenchymal abnormality.⁹ Because we did not measure C_rs, it is unclear if this played a role in the restrictive physiology seen in our older patients. Diaphragm dysfunction in older children with GO has not been formally studied, although if present, it would also contribute to a reduced TLC.

Akinkuotu et al. found that lung volumes on pre-natal MRI (a surrogate for pulmonary hypoplasia) differed between neonates with CDH, CLM, and omphalocele. When the authors controlled for differences in prenatal lung volume, those with CDH still had a longer duration of intubation and a higher incidence of PH.⁴⁷ These findings indicate that disease-specific factors aside from the degree of lung hypoplasia influence pulmonary outcomes. However, subjects in that study differed from our population in that all patients with omphaloceles (even small ones) and milder CLM were included, limiting comparisons with our results.

Children with pulmonary hypoplasia are at risk for developing PH.⁴⁸ PH during infancy is associated with mortality⁴⁹ and abnormal pulmonary function^30,31,50 in patients with CDH. However, in those who survive, echocardiographic measures of PH often resolve during the first few years of life.⁵¹ In CLM, PH during infancy has been described in case reports and small case series.^52–54 PH is also common in GO and is associated with worse pulmonary outcomes and mortality during infancy.¹⁴ The increased RV afterload caused by PH leads to RV dysfunction.^8,55 Qualitative assessment of RV systolic function is reader-dependent and limits vigorous longitudinal follow-up.⁵⁶ However, quantitative measures of RV dysfunction are associated with important clinical outcomes in infants with developmental lung disease, including bronchopulmonary dysplasia,²⁰ GO⁵⁷, and CDH.¹⁸ Additionally, these non-invasive echocardiographic measures correlated with pulmonary artery pressure and pulmonary vascular resistance measured by cardiac catheterization in a cohort of infants with CDH.¹⁶ However, long-term RV function in survivors of CDH and other forms of pulmonary hypoplasia has not been well-described. In one small study of seven survivors of CDH, quantitative measures of RV function were abnormal in early childhood (mean age at time of study, 6.2 years) compared with healthy controls.⁵⁸ We found a non-statistically significant association between TAPSEZ and RVFAC and abnormal lung function in a larger cohort of older children with pulmonary hypoplasia. In infants with pulmonary hypoplasia, abnormal development of pulmonary vasculature and ongoing lung disease both contribute to the development of pulmonary hypertension and subsequent RV systolic dysfunction.⁸ Therefore, it is not surprising that subjects with abnormal lung function were more likely to have signs of RV dysfunction on echocardiogram. Ultimately, larger, long-term studies assessing RV function in pulmonary hypoplasia survivors are needed.

Our study has several limitations. We did not perform bronchodilator testing, which would have provided insight into the long-term management of these patients. Studies of younger children with GO and CDH have found a substantial portion to be bronchodilator responsive (46% in GO, 22.5% in CDH).^9,30 Peetsold et al. reported a 23.7% rate of bronchodilator responsiveness in older children and adolescents with a history of CDH.³² In a longitudinal study of pulmonary function in survivors of CDH, Dao et al. noted that most subjects with obstructive lung disease did not have a positive bronchodilator response, although they did not provide further details.³¹ The sample size of our subjects was small within each disease category. Furthermore, selection bias may have been present if parents of children with more severe underlying disease would have been more likely to agree to participate in this research study. On the other hand, children with more severe underlying disease may have been unable to perform pulmonary function testing because of intellectual or developmental disability. The somewhat wide age range (8 to 13 years) was necessary to maximize enrollment. However, it is certainly possible that differences even within this age range related to puberty (lung and thorax size, respiratory muscle strength, bronchial smooth muscle tone) may have impacted pulmonary function.^59–62 Data regarding several factors associated with abnormal pulmonary function (e.g., tobacco smoke exposure,⁶³ asthma, gastroesophageal reflux disease [GERD]) were incomplete. The fact that there was no association between inhaled corticosteroid or GERD medication use and abnormal lung function, however, suggests that these conditions did not impact our pulmonary function results. Finally, our definition of abnormal lung function did not include more subtle abnormalities in pulmonary function test results such as isolated reduction in FEV₁, which can be pathologic.⁶⁴

In summary, more than half of survivors of CDH, GO, and EICLM have abnormal lung function during older childhood, but patterns of lung function abnormalities differ by underlying etiology of pulmonary hypoplasia. Patients with CDH and EICLM are more likely to have obstructive lung disease, while those with GO most commonly have restrictive lung disease. Our findings suggest that the lung grows differently, depending on the initial cause of hypoplasia. As low lung function during childhood is associated with an increased risk for pulmonary morbidity in adulthood,³³ these patients stand to benefit from longitudinal care by a pulmonologist to monitor pulmonary function and identify modifiable risk factors for further lung function decline.