Pulmonary Morbidity in Congenital Diaphragmatic Hernia Survivors Treated at a Non‐ECMO Center From 1998 to 2015: A Cross‐Sectional Study

ABSTRACT

Introduction

A main feature of CDH is lung hypoplasia and the related presentation of pulmonary hypertension and cardiac dysfunction. Multiple factors influence pulmonary status after CDH: degree of hypoplasia, ventilator‐induced injury, altered growth and development of pulmonary structures, reduced diaphragm function and chest wall abnormalities. The evolution of pulmonary sequela in this population is still unclear. We aimed to describe the pulmonary status of our population of CDH‐survivors and evaluated on risk factors.

Methods

CDH‐survivors (1998−2015) were included and performed lung function tests and chest X‐rays.

Results

Fifty‐one (51/71, 71.8%) participated. Median age was 12.2 (5.5–21.4) years, 28 (54.9%) male, 42 (82.4%) had left‐sided hernias, 10 (19.6%) needed patch‐repair and median length of stay in hospital was 28.0 (IQR 18.5–61.6) days in Table 1. Spirometry including bronchodilator response (BDR)‐test, body plethysmography, and diffusion capacity, were available for 48, 42, and 40 participants. The mean (SD) z‐score for FEV1 and FVC was −0.26 (1.70) and −0.28 (1.70). Twenty‐one (43.8%) had obstructive patterns and six had positive BDR. TLC mean (SD) z‐score was −0.18 (1.10). Four showed restricted/mixed patterns and 13 showed signs of hyperinflation. Increased RV/TLC‐ratio and reduced FEV1 was associated with longer time on mechanical ventilation. Diffusion capacity was decreased in three cases. Chest X‐ray revealed hernia recurrence (13.9%) and scoliosis (38.9%).

Conclusion

Mild obstructive impairment and hyperinflation was frequent in our CDH cohort and only a small subset of restrictive disorders were identified. We advocate follow‐up by a specialized multidisciplinary team through childhood and into adulthood.

1. Background/Introduction

Congenital diaphragmatic hernia (CDH) is a rare defect in the diaphragm allowing abdominal contents to protrude into the chest cavity of the fetus during pregnancy. This, as well as unidentified underlying mechanisms, prevents normal development of the lung parenchyma. The ipsi‐lateral lung is the most affected, but both lungs reveal reduced bronchial and vascular branching resulting in a decreased number of airways and alveoli. In the postnatal period, the infants present with a combination of lung hypoplasia, persisting pulmonary hypertension, and cardiac dysfunction [1, 2].

The severity of the clinical presentation of CDH varies, ranging from death shortly after birth, to more subtle presentations that go unnoticed through infancy (late‐presentation). Many factors contribute to the clinical picture, but the degree of pulmonary hypoplasia remains a major determinant of morbidity and outcome [3, 4].

The literature states a varying incidence of pulmonary sequela in long‐term CDH survivors and data from systematic and protocolled longitudinal follow‐up programs are still limited. Some studies indicate persisting pulmonary morbidity in 27%–48% of CDH‐survivors [5, 6]. Restrictive and obstructive patterns, as well as impaired diffusion capacity, have been reported [5, 7, 8, 9]. Improvement of lung function over time has also been reported [10], but this conflicts with more recent results indicating a decrease in lung function during childhood (from 8 to 12 years) [11]. A better understanding and management of the long‐term pulmonary morbidities associated with CDH is warranted as overall survival rates have increased.

1.1. Aim

We described the pulmonary status of our population of CDH‐survivors, evaluated by; body plethysmography, spirometry—including bronchodilator response (BDR) test, gas exchange, and chest X‐ray. We also described the baseline and demographic data of the cohort, and evaluated factors associated with pulmonary outcome.

2. Methods

2.1. Design

We conducted a cross‐sectional study on consecutively born infants with CDH, treated at a non‐ECMO center from 1998 to 2015. This cohort consisted of survivors from our original study on mortality published in 2020 [12].

2.2. Ethics

Permission to collect historical data from medical journals was granted by The Danish Health and Medicines Authority (3‐3013‐1121/1). The Southern Region Committees on Health Research Ethics (S‐20170177) and The National Ethic Committee (221297) approved the study. The Danish data‐protection agency (18/26141) allowed for data management and storage using the RedCAP database system provided by OPEN (Open Patient Data Explorative Network; Region of Southern Denmark).

All eligible participants were contacted by letter. Responders received written and oral information and informed consent was obtained from participants 18 years or older and from parents of participants younger than 18 years.

2.3. Study Population

Odense University Hospital is one of two centers in Denmark treating patients with CDH. Our center provides services to the western region of Denmark, covering a population of 3.2 million (total population 5.93 million—Statistics Denmark, Q2 23).

We included all consecutively live‐born CDH infants from the region treated from 1998 to 2015 with a history of symptomatic CDH presenting in the first 24 h of life. We excluded all non‐survivors and survivors with severe neurological impairment.

Our center is a non‐ECMO center and FETO was not considered. Postnatal management adhered to contemporary guidelines and local practice, as we have previously described and published [12]. Prenatally diagnosed cases had a planned delivery at our hospital and cases diagnosed after birth were transferred immediately from hospitals in the western region of Denmark. None of the infants received treatment outside the region [12].

2.4. Baseline Data

Baseline and demographic data, including data on surgical and postnatal management, were collected from medical notes and electronic journals. Height and weight at follow‐up was noted prospectively.

2.5. Interview

Parents and participants were asked to report on; current medication, any medication given within the last year, symptoms and treatment of gastroesophageal reflux, respiratory problems such as episodes with airway infection or wheezing/shortness‐of‐breath within the last year, school/work achievements, and sports/exercise activities.

2.6. Chest X‐Ray

A standing chest X‐ray was performed. Postero‐anterior and lateral images were generated and all images were reviewed by a radiologist specialized in pediatric imaging.

2.7. Lung Function

Specially trained nurses (four investigators) from our pediatric and adult pulmonary outpatient clinics conducted lung function testing. The equipment was maintained and calibrated according to the manufacturer's instructions. The children/adolescents were guided throughout the procedure by the investigator, using computer animation if relevant.

Lung volumes, spirometry and diffusion capacity were measured using the body plethysmography equipment (VIASYS Healthcare, Masterscreen Body with SentrySuite 3.0 software). BDR‐test was performed using inhaled salbutamol 400 mg (Buventol Easyhaler, Orion Pharma A/S). All procedures were performed according to contemporary guidelines from The European Respiratory Society and The American Thoracic Society (ERS/ATS) [13, 14, 15, 16, 17].

The following lung function measures were generated; total lung capacity (TLC), functional residual capacity (FRC), expiratory reserve volume (ERV), vital capacity (VC), residual volume (RV), and RV/TLC. We also measured forced expiratory volume in the first second (FEV1) and forced vital capacity (FVC), and the FEV1/FVC ratio was calculated. A positive BDR‐test was defined as an increase of 10% of the predicted value in FEV1 or FVC in repeated spirometry after 15 min [16]. Variables on diffusion capacity were generated: D_LCO (diffusing capacity of the lung for carbon monoxide), V_A (alveolar volume), and K_CO (transfer coefficient of the lung for carbon monoxide). A finger prick sample was used to measure hemoglobin level for correction of D_LCO.

All results were converted into z‐scores of predicted normal value and reported as mean (SD). The 95th/5th percentile (upper limit of normal [ULN] and the lower limit of normal [LLN], equivalent to z‐score = −1.645 and z‐score = 1.645), were calculated according to the references provided by the ERS/Global Lung function Initiative network (GLI), using the equations on caucasian populations; spirometry [18], lung volumes [19], and diffusion capacity [20].

Pulmonary morbidity was reported as mild (z‐score −1.65 to −2.5), moderate (z‐score −2.51 to −4.0), or severe (z‐score < −4.1) and the interpretive strategies published by ERS/ATS in 2022 was applied for further classification [16].

2.8. Statistics

Nonparametric data were summarized as median and interquartile range values (25th and 75th percentile) or as numbers and percentage. Comparison between groups were performed using Wilcoxon rank‐sum test or the Chi²‐test.

Parametric data were summarized as mean and standard deviation (SD) for continuous variables.

Factors associated with lung function parameters were evaluated using linear regression models.

All analyses were performed using STATA/IC17.0 (Stata Statistical Software: Release 17. College Station, TX: StataCorp [LLC]) and p < 0.05 was considered significant.

3. Results

From the original cohort, we identified 74 CDH‐survivors [12]. We excluded three survivors due to severe mental impairment, leaving 71 eligible children. Fifty‐one children (71.8%) and their families consented to participate, see flowchart (Figure 1). Interviews and examinations were performed from November 2018 to September 2021.

Figure 1.

Flowchart of study inclusion.

3.1. Baseline

Treatment with HFO (high‐frequency oscillation) was more frequent in participants compared to non‐participants (p = 0.005). Otherwise, we found no significant differences in baseline or demographic data (see Table 1).

Table 1.

Demographic and baseline data on participants and non‐participants.

Baseline data	Participants (51)	Non‐participants (20)	p value
Sex, male n	28 (54.9%)	10 (50.0%)	0.710
Birth weight, g	3185 (2700–3700)	3115 (2550–3390)	0.561
Gestational age, weeks	38.6 (36.7–40.0)	38.2 (36.8–39.9)	0.850
APGAR, 1 min	8 (6–9)	8 (6–10)	0.479
Prenatal diagnosis, n	25 (49.0%)	5 (25.0%)	0.065
Associated malformations, n	5 (9.8%)	1 (5.0%)	0.306
Hernia location, sin	42 (82.4%)	16 (80.0%)	0.270
Patch, n	10 (19.6%)	2 (10.0%)	0.331
Time to surgery, days	3.0 (2.2–4.0)	3.2 (2.2–5.8)	0.574
Mechanical ventilation, days	6.32 (3.9–14.2)	5.64 (1.8–14.0)	0.472
HFO, n	44 (86.3%)	11 (55.0%)	0.005
iNO, n	13 (25.5%)	5 (25.0%)	0.966
Sildenafil, n	7 (13.7%)	0 (0.0%)	0.081
Milrinone, n	5 (9.8%)	0 (0.0%)	0.142
Vasoactive therapy, n	27 (52.9%)	9 (45.0%)	0.496
Discharge PICU FiO2, %	25% (21%−30%)	25% (21%−30%)	0.791
Day 90 FiO2, %	21% (21%–21%)	21% (21%–21%)	0.572
LOS‐PICU, days	7.3 (4.8–17.9)	8.3 (2.7–15.5)	0.338
LOS‐hospital, days	28.0 (18.5–61.6)	20.3 (12.1–39.6)	0.142

Note: Continuous data; summarized as median and interquartile range values (25th and 75th percentile) and Wilcoxon rank‐sum test performed for comparison. Categorical data; summarized as numbers/percentage, and Chi²‐test performed for comparison. p < 0.05 was considered statistical significant.

Abbreviations: HFO, high‐frequency oscillation; iNO, inhaled nitric oxide; LOS‐HOSP, length of stay in hospital (days); LOS‐PICU, length of stay in pediatric intensive care unit (days); Milrinone, phosphodiesterase‐3‐inhibitor; Patch, defect requiring patch‐repair; Sildenafil, phosphodiesterase‐5‐inhibitor.

In the group of participants, 11 had birth weights of 2500 g or below (ranging 1765–2500 g) and five were born before the gestational age of 35 weeks (ranging 32.3–34.4 weeks). Six children had associated malformations, which included esophagus atresia with/without fistula, chromosomal abnormalities diagnosed in early childhood, cerebral palsy and minor cardiac, and renal malformations. One participant required supplement oxygen treatment after discharge (weaned by 18 months), and one needed a percutaneous endoscopic gastrostomy feeding tube at a later age.

3.2. Interview

Fifty‐one CDH survivors, with a median age of 12.2 (range 5.5–21.4), completed the interview.

Eight (15.7%) reported the use of inhalation steroids, and 10 (19.6%) reported the use of inhalation β2‐agonist treatment within the last year. Three participants (6.0%) reported two to three episodes of “airway wheezing and/or shortness of breath resulting in the initiation of (or increased treatment) with inhaled medication,” whereas two participants (3.9%) reported two to three episodes of “airway infection requiring treatment with antibiotics or absence from school/work” within the past year.

Participants also reported on exercise activities. Eleven participated only in school‐related physical activities but stated other interests such as music, singing, or gaming. Twenty participated in sporting activities one to two times a week and 15 more than two times a week during leisure time. Five competed in sports at elite level. Overall, 40 (40/51, 78.4%) participants engaged in voluntary sporting activities on a regular basis.

3.3. Lung Function

Lung function testing was not successful in three cases: one due to lack of cooperation and two due to circumstances related to the Covid‐19 pandemic. In all, 48 participants contributed with lung function results. Spirometry results were available for all 48 participants. Body plethysmography and diffusion capacity results were available for 42 and 40 participants. Results were missing due to either fatigue, inability to complete the full test or technical issues.

We reported all lung function parameters as mean (SD) of z‐score of predicted normal value (see Table 2).

Table 2.

Lung function parameters.

Lung function test	All participants
	n = 48
	z‐score (mean, SD)
Spirometry (n = 48)
FEV1	−0.26 (1.71)
FVC	−0.28 (1.71)
FEV1/FVC	−1.30 (0.99)
Plethysmography (n = 42)
TLC	−0.18 (1.10)
FRC	0.26 (0.89)
RV	0.73 (0.74)
RV/TLC	1.33 (1.07)
ERV	−0.78 (1.02)
VC	−1.72 (1.75)
Diffusion capacity (n = 40)
DLco	−0.27 (0.95)
Kco	0.16 (1.06)
VA	−0.53 (1.27)

Note: Results are reported as z‐scores (mean, standard deviation).

Abbreviations: Spirometry (FEV1, forced expiratory volume in first second; FVC, forced vital capacity); body plethysmography (TLC, total lung capacity; FRC, functional residual capacity; ERV, expiratory reserve volume; VC, vital capacity; RV, residual volume); diffusion capacity (single‐breath carbon monoxide uptake, DLCO, diffusing capacity of the lung for carbon monoxide; VA, alveolar volume; KCO, transfer coefficient of the lung for carbon monoxide).

3.3.1. Spirometry

Spirometry was performed before and after bronchodilator treatment. All measurements were evaluated on repeatability according to the ERS “Standardization of spirometry 2019 update” and 168 (87.5%) of the 192 measurements reached a high quality [13].

Mean z‐scores were within normal limits, but a lower FEV1/FVC ratio was noted. Based on the algorithm recommended in the ERS/ATS technical standards on interpretive strategies, the results were classified into the different categories of ventilatory impairment (see Figure 2). Twenty‐five participants had FEV1/FVC above the LLN: 18 (37.5%) had otherwise normal spirometry results and seven had FVC below LLN. These seven were classified according to their TLC: three had TLC below LLN indicating a restrictive disorder and four had TLC above LLN and classified as nonspecific patterns.

Figure 2.

Classification of ventilatory impairment according to the ERS/ATS technical standards on interpretive strategies for lung function tests [16]. FEV1, forced expiratory volume in first second; FVC, forced vital capacity; LLN, lower limit of normal; TLC, total lung capacity.

Twenty‐three had FEV1/FVC ratio below LLN and 21 (43.8%) were classified as obstructive, one with a mixed disorder and one could not be classified as TLC was not available. Severity was assessed according to z‐scores, and of the 21 participants with obstructive patterns, 16 had mild impairment and five had moderate impairment.

BDR‐test was possible in 46 cases. Twelve participants had positive results. Of these, four had normal spirometry results and six had obstructive patterns. Of the eight participants who reported the use of inhalation medication within the past year, only one had a positive BDR‐test (see Table 3).

Table 3.

Summary of results.

Spirometry category	Spirometry (n = 48)	Positive BDR test (n = 12 [%])	Inh. medication < 1 year (n = 8)
			Positive BDR and inh. med	Negative BDR and Inh. med
Normal test	18	4 (22.2)	0	0
Obstructive	21	6 (28.6)	1	4
Mixed disorder	2	1 (50.0)	0	1
Possible restriction/nonspecific pattern	7	1 (14.3)	0	2

Note: Positive BDR‐test and use of inhalation medication within the past year reported according to spirometry results. Shown as number and percent of spirometry category, and as a number of participants on inhalation medication with positive or negative BDR.

3.3.2. Body Plethysmography

Mean values of TLC, FRC, RV, and ERV were within normal range, but we noted a higher mean RV/TLC ratio and a lower mean VC.

Classification of the results showed normal lung volumes in 25 cases (59.5%). Reduced TLC, with values below LLN, was noted in four cases (9.5%) confirming a restrictive lung disorder, as indicated by the spirometry results. Further classification revealed the presence of simple restriction (two cases), complex restriction (one case), and a mixed disorder (one), with mild impairment in two cases and moderate impairment in two cases [16].

Hyperinflation was present in 13 (31.0%) cases as indicated by the higher RV/TLC ratio. Of these, eight had obstructive spirometry patterns, five had a positive BDR test, and two were described with emphysema on chest X‐ray.

3.3.3. Diffusion Capacity

We measured D_LCO, V_A, K_CO in 40 participants; results were hemoglobin corrected and calculated as z‐scores, as summarized in Table 2. Two participants had D_LCO below LLN and V_A within normal limits, as seen in pulmonary vascular disease or emphysema. One had D_LCO below LLN combined with low V_A and K_co, also indicating the presence of emphysema with loss of alveolar structure. Six had V_A below LLN, indicating reduced alveolar capacity.

We evaluated factors associated with impaired lung function, and the following variables were included: patch repair, treatment with iNO, APGAR score at 1 min, and days on mechanical ventilation, as indicators of a more severe condition. Days on mechanical ventilation were associated with lower FEV1 (p = 0.029) and both days on mechanical ventilation and the need for patch repair were associated with increased RV/TLC (p = 0.001 and p = 0.004). All variables were evaluated in an unadjusted and adjusted model (see Table 4).

Table 4.

Regression model evaluation of factors associated with low FEV1 and increased RV/TLC.

	FEV1 unadjusted	FEV1 adjusted	RV/TLC unadjusted	RV/TLC adjusted
	p (95% CI)	p (95% CI)	p (95% CI)	p (95% CI)
Patch repair	0.125 (−2.23 to 0.28)	0.299 (−2.03 to 0.64)a	0.001 (0.53 to 2.05)	0.004 (0.42 to 2.09)a
Treatment with iNO	0.291 (−1.81 to 0.55)	0.562 (−1.56 to 0.86)a	0.301 (−0.39 to 1.24)	0.612 (−0.57 to 0.96)a
APGAR score 1 min	0.212 (−0.08 to 0.34)	0.367 (−0.12 to 0.31)a	0.419 (−0.20 to 0.08)	0.972 (−0.13 to 0.14)a
Time on MV, days	0.004 (−0.13 to −0.03)	0.029 (−0.16 to −0.01)b	0.000 (0.05 to 0.11)	0.001 (0.04 to 0.13)b

Note: Factors included patch repair, iNO treatment, APGAR 1 min, and time on mechanical ventilation (MV). Data reported as p‐values and 95th conf. interval. p < 0.05 were defined as statistically significant.

Abbreviations: iNO, inhaled nitric oxide; Patch repair, defect requiring patch‐repair.

^aRegression model adjusted for patch, iNO, and APGAR, but without time on mechanical ventilation.

^bRegression model adjusted for patch, iNO, APGAR 1 min, and time on mechanical ventilation.

3.3.4. Chest X‐Ray and Clinical Observations

Only 36 of the participants completed a chest X‐ray (median age of 12, range 6–21 years). Twelve declined due to parental concern about radiation exposure.

Recurrence of the hernia was diagnosed in five (13.9%) cases, one case with a large defect involving displacement of the lever and four with subtle presentations involving the liver or bowel. Three participants had a history of scoliosis before entering the study and scoliosis was noted in an additional 11 participants. On clinical examination 18 showed chest wall asymmetries, including six with pectus excavatum (6/36, 16.7%). Otherwise, findings on chest X‐ray included displaced cor, sequlae after thoracotomy, asymmetry of skeletal structures, and five showed signs of emphysema.

4. Discussion

In this cross‐sectional study of CDH survivors, we evaluated pulmonary status and function. Our main findings were results within the normal range in half of the cohort, the presence of airway obstruction in 43.8% (the majority with mild impairment) and hyperinflation in 31.0%. We noted mild to moderate restrictive impairment in 8.3%.

A varying prevalence (4%–12%) of childhood asthma has been reported in the Danish population and cases with obstruction and positive BDR (6/21) may also represent asthma [21]. However, a fixed (chronic) pattern was observed in more than 2/3 of the obstructive cases. This, as well as the demonstrated elevated RV/TLC ratio and presence of hyperinflation, corresponds well with findings from other CDH populations [6, 10, 22, 23]. All participants had been instructed to discontinue the use of inhalation medication before testing as recommended. Of note, we found a major discrepancy between participants with positive BDR and the use of inhalation medication. Eight participants were currently using inhalation medicine, but only one presented with a positive BDR. This could, in part, be due to residual effects despite correctly discontinued medication. The remaining 11 with positive BDR tests were untreated, indicating the challenge of evaluating the pulmonary status of this population.

A distinct combination of factors affects the lung function of CDH patients, including the initial degree of hypoplasia, chest wall abnormalities, reduced diaphragmatic function, and a lack of full compensatory alveolar growth on the ipsilateral side. TLC is often within normal range, but studies have demonstrated alveolar enlargement and altered/reduced ventilation of the ipsilateral lung manifesting as the elevated RV/TLC ratio [24, 25, 26]. We demonstrated a high prevalence of hyperinflation in our cohort, whether this was due to obstruction/air trapping or enlargement of a reduced number of alveoli, remains speculative but five also showed signs of emphysema on chest X‐ray.

Even as the pulmonary sequela of CDH becomes more evident, the impact on daily life activities remains unclear. The high rate of participation in exercise and sport in our cohort suggest a limited impact on daily life and stems well with the overall mild impairment noted according to z‐score in the majority of cases. Only three stated subjective pulmonary complaints in terms of “two to three episodes of airway wheezing and/or shortness of breath” during the last year, but this may be under‐reported, as participants might not fully recognize their limitations if they have experienced respiratory impairment since infancy.

Regression analysis of our results found patch repair and time on mechanical ventilation to be associated with increased RV/TLC. This indicates an increased risk of ventilatory impairment in the more severe cases, as patch repair represents a larger defect in the diaphragm. However, time on mechanical ventilation was the only variable significantly associated with reduced FEV1. No association was found with patch repair, treatment with iNO or APGAR score, using both an adjusted and unadjusted model. Therefore, we cannot rule out a component of ventilator‐induced injury as other authors have demonstrated similar impairments in non‐CDH populations also undergoing intensive care treatment during infancy [22]. Knowledge is still lacking on the trajectory of lung function evolvement in CDH, but recent data indicate a possible decline through childhood and into adulthood [11, 22, 27]. The pathology behind the observed decline is unclear. As growth may facilitate a decrease in RV/TLC, the presence of air trapping may cause further lesions to the lungs and emerging longitudinal data indicate an increased risk of obstructive disease and development of chronic obstructive lung disease in adulthood [27].

Overall, pulmonary impairment in our cohort is classified as mild, and the high degree of preserved diffusion capacity supports this. However, cases with highly complex restrictive disorders were identified. The impact of disease severity on pulmonary sequela is not yet fully understood, but recent studies have reported an association between decreased FEV1/fixed obstruction and markers of disease severity. The relative mild impairment in our cohort could reflect a milder case mix and several circumstances could have introduced such a bias. First, unaccounted non‐survivors (death before transport or undiagnosed) and terminated pregnancies of the more severe cases. Second, the national prenatal screening program was not fully implemented until 2004, which is reflected in our cohort as a higher ratio of postnatal diagnosed cases in the first part of the study period. A direct comparison is difficult, but more severe impairment has been reported in populations described with more days on mechanical ventilation, longer PICU/hospital stay and higher rates of patch repair, than our cohort. Of note, these longitudinal studies reported an association between markers of severity and the degree of lung function impairment, but not between severity and rate of decline in FEV1/increasing fixed obstruction over time. Indicating that both mild and severely affected cases are at risk of decline over time [11, 27].

In total, 50% presented with abnormalities on chest X‐rays and the two most consequential findings were recurrence of the hernia and scoliosis. Scoliosis was the most frequent finding and occurred at all ages in the study population (range 7−20 years). Two participants, both with other comorbidities, already received treatment. Our orthopedic department evaluated all cases of scoliosis discovered in the study, no interventions were necessary at the time, but follow‐up was scheduled for participants not yet fully grown. Scoliosis is a well‐described late‐onset complication of CDH, found with increasing prevalence up‐through childhood [28, 29]. Most cases are mild and remain asymptomatic, but patch repair has been associated with increased severity and earlier onset [30].

We confirmed all recurrent hernias with MRI, none of the participants had complaints at the time of examination and further evaluation did not lead to corrective surgery in any of the cases. Re‐herniation may be asymptomatic or present with mild symptoms, but has the potential to develop into a severe and life‐threatening condition if incarceration and sepsis develop.

Of note, simple chest X‐ray is not the optimal diagnostic tool to screen for either recurrence or scoliosis (and not the aim of this study), therefore, we cannot rule out missed cases. Despite this, our results emphasize the need for education on the lifelong risks of CDH and follow‐up including surgical and orthopedic expertise.

4.1. Strengths

We managed a high rate of participation as more than 70% of the eligible children participated. Our study population originated from a cohort of consecutively born CDH infants, uniformly registered and collected from a well‐defined geographical area covering more than half of the population in Denmark. Management of these infants were consistent and in line with contemporary and local guidelines throughout the study period [12].

4.2. Limitations

Cases not accounted for could be undiagnosed cases born at a regional hospital, which remained undiagnosed or did not survive transport. Cases with severe associated congenital heart disease (CHD) diagnosed prenatally, could in theory have been referred directly to our national center for CHD. However, we find this unlikely as these cases would also warrant a second opinion evaluation at our center.

Relevant prenatal factors and markers of severity were not available due to the retrospective design. Lund‐to‐head ratio and organ position was not recorded in a structured manner until 2010, prenatal MRI is not fully implicated at our center and hernia grade was not systematically recorded until 2016. The small data set (due to the nature of the conditioner) and wide range in age also serve as limitations.

We relied on normative values for comparison using z‐scores, but including matched healthy controls would have strengthened our data.

We advocate follow‐up by a highly specialized multidisciplinary team, not only through childhood but also into adulthood, as recent data indicate that severity may not predict the degree of decline in pulmonary function. The overall aim should be to ensure: (i) specialized care to CDH survivors with complex pulmonary impairment, (ii) early detection and treatment in case of deteriorating pulmonary function, and (iii) education and prevention of harmful effects of untreated asthma, recurrent respiratory infections, smoking, progressing skeletal deformation, and detection of recurrence.

5. Conclusion

Pulmonary sequelae following CDH repair was present in our cohort. The majority with mild impairment and a high rate of physical activity, but we identified a small subset of CDH survivors presenting with restrictive disorders.

Mild obstructive patterns and hyperinflation (increased RV/TLC) were the main findings. In light of our results and recent data indicating a risk of deterioration into chronic obstructive disease, we advocate follow‐up by a highly specialized multidisciplinary team not only through childhood but also into adulthood.

Author Contributions

Collection, management, and analysis of data: Ulla Lei Larsen, Susanne Halken, and Lone Agertoft. Draft: Ulla Lei Larsen, Work design, interpretation of data, revision, and final approval: Ulla Lei Larsen, Susanne Halken, Lone Agertoft, Thomas Strøm, Anne Maria Herskind, and Palle Toft. Revision: draft: Ulla Lei Larsen. Final approval: Ulla Lei Larsen, Susanne Halken, Lone Agertoft, Thomas Strøm, Palle Toft, and Anne Maria Herskind.

Ethics Statement

This study was approved by The Danish Health and Medicines Authority (jr nr.: 3‐3013‐1121/1), The Southern Region Committees on Health Research Ethics (jr nr: S‐20170177), The National Ethic Committee (jr nr: 221297), and the Danish data‐protection agency (jr nr. 18/26141). Data is stored using REDCap (Research Electronic Data Capture) in collaboration with OPEN (Odense Patient Data Explorative Network), Odense University Hospital/Institute of Clinical Research, University of Southern Denmark.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

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