Congenital Diaphragmatic Hernia (Genetic Causes and Approach)

Isolated and complex congenital diaphragmatic hernia: Fryns syndrome (PIGN), diaphragmatic hernias 1 to 5 (15q26, 8p23.1, ZFPM2, ALDH1A2, PLS3), Donnai-Barrow syndrome (LRP2), PDAC/Matthew-Wood syndrome (STRA6, RARB), MYRF-related cardiac-urogenital syndrome, WT1 disorder, chromosomal anomalies and other syndromic causes.

Literature and regulatory information updated: 6 November 2026.

0. Initial Reference Table

Congenital diaphragmatic hernia (CDH) is a malformation and not a monogenic disease. The table brings together the malformation in general and the genetic entities in which CDH is a core or frequent feature; the remaining syndromic and chromosomal causes appear in the tables of section 2 and in the differential diagnosis of section 6.

1. Introduction

Definition and anatomical classification

CDH is a defect of diaphragm formation, present at birth in most cases, resulting from incomplete formation or muscularisation, with absence or deficiency of the muscle, or from eventration, that is, elevation of a thinned portion of the diaphragm owing to incomplete muscularisation [1]. The diaphragm is normally formed by nine weeks of gestation; the hernia arises when one of its four constituent components (septum transversum, pleuroperitoneal membranes, dorsal mesentery of the oesophagus and body wall) fails to grow or fuse by the eighth week after conception [21]. The opening allows the stomach, bowel, spleen, kidney and liver to ascend into the thorax, and lung development is compromised, with bilateral pulmonary hypoplasia and a reduced number of pulmonary vessels, leading to pulmonary arterial hypertension (PAH) [3].

The posterolateral Bochdalek hernia accounts for 80–90% of all CDH in the GeneReviews review chapter and 70–90% in the OMIM synopsis; about 85% are left-sided, 10% right-sided and 5% bilateral [1, 14]. The French protocol (PNDS, Protocole National de Diagnostic et de Soins, of the Haute Autorité de Santé) places 75% of cases in the posterolateral region of the left dome [3]. Diaphragmatic agenesis represents the severe end of the Bochdalek spectrum [1, 14]. The Morgagni (or Morgagni-Larrey) hernia, anterior, retrosternal or parasternal, accounts for about 2% of cases, usually has a sac and is often asymptomatic in the newborn; the central hernia, which involves the central tendon with a complete muscular rim, accounts for 1–2%; and the anterior hernias of the pentalogy of Cantrell accompany defects of the supraumbilical abdominal wall, lower sternum, diaphragmatic pericardium and heart [1, 14]. The presence of a sac has been associated with a better prognosis, but an eventration with a redundant membranous diaphragm may be described as a "sac", so the association should not be assumed [1].

The international surgical classification of the CDH Study Group distinguishes four defect sizes: A, small, usually intramuscular, with more than 90% of the hemidiaphragm present and less than 10% of the chest wall circumference involved; B, with 50–75% of the hemidiaphragm present; C, with less than 50% present; and D, formerly termed agenesis, with less than 10% present and mandatory need for a prosthesis or flap [22, 23]. The PNDS describes the same scale in simplified form: A with all muscular margins, B with less than 50% of the diaphragm involved and absence of the posterolateral margin, C between 50% and 100%, and D with complete absence [3].

Isolated CDH and complex CDH

CDH is isolated in 50–60% of probands and complex in 40–50%, when associated with other congenital anomalies [1]. In the hospital series of Pober and colleagues, with 203 unrelated Bochdalek hernias observed over 28 years, 112 (55%) were isolated and 91 (45%) accompanied other malformations [14, 24]. The complex form is divided into syndromic, when there is a recognisable syndrome or a chromosomal anomaly, and complex non-syndromic, when there are craniofacial, ocular, cardiovascular, central nervous system (CNS), limb or genitourinary malformations without a recognisable pattern [1]. Patients with syndromic CDH are more often diagnosed prenatally, have an affected first-degree relative in 10% of cases, are born preterm and with low birth weight in more than 50% and have a worse prognosis [1]. The Oxford Desk Reference states that 30–40% of newborns with CDH have additional major anomalies, mainly cardiovascular, CNS, musculoskeletal and genitourinary [21].

Historical overview and genetic milestones

Familial forms of unilateral diaphragmatic agenesis were described by Passarge and colleagues in 1968, and the reviews of Wolff (1980) and of Norio and colleagues (1984, 53 familial cases) favoured multifactorial inheritance, with a sibling recurrence risk of about 2% [14]. In 1979, Fryns and colleagues described two stillborn sisters with coarse facies, cloudy corneas, diaphragmatic defects, absent lung lobulation and distal limb deformities, the first description of the syndrome that bears his name [9, 25]. Donnai and Barrow described in 1993 siblings with diaphragmatic hernia, exomphalos, agenesis of the corpus callosum, hypertelorism, myopia and sensorineural deafness; the overlap with the facio-oculo-acoustico-renal syndrome of Holmes and Schepens (1972) was confirmed by the identification of LRP2 variants in 2007 [16, 26, 27]. In 2005, Ackerman and colleagues identified a de novo nonsense variant in FOG2 (ZFPM2) in a deceased child with a diaphragmatic defect and severe primary pulmonary hypoplasia, the first gene implicated in non-syndromic CDH in humans [28]. Klaassens and colleagues defined in 2005 a minimal region of about 5 Mb at 15q26.1-q26.2, with NR2F2 as the main candidate gene [14]. In 2007, STRA6 variants were associated with Matthew-Wood syndrome, and in 2013 RARB was associated with the same constellation [29, 30]. Biallelic PIGN variants were described in Fryns syndrome in 2014 and 2016 [7, 31], and trio exome sequencing identified MYRF in 2018 as the cause of a new syndrome [32]. Exome and genome studies over the last decade have shown an excess of deleterious de novo variants in both complex and isolated CDH and, in 2024, a polygenic contribution from common variants [33, 34].

Nomenclature and reclassifications

OMIM organises CDH into a phenotypic series: DIH1 (142340), associated with the 15q26 region; DIH2 (222400), with the 8p23.1 deletion; DIH3 (610187), with ZFPM2; DIH4 (620025), with ALDH1A2; and DIH5 (306950), with PLS3 at Xq23 [14]. Donnai-Barrow syndrome and FOAR syndrome are now the same entity, and the terms Holmes-Schepens syndrome, diaphragmatic hernia-exomphalos-hypertelorism syndrome and diaphragmatic hernia-hypertelorism-myopia-deafness syndrome are no longer recommended [16]. The designations Denys-Drash, Frasier and Meacham are now understood as points on a phenotypic continuum of WT1 disorder and are no longer useful clinical diagnoses [20]. Matthew-Wood syndrome is also termed PDAC syndrome (pulmonary hypoplasia/agenesis, diaphragmatic hernia/eventration, anophthalmia/microphthalmia, cardiac defect), Spear syndrome and syndromic microphthalmia 9 [29]. The GeneReviews general review chapter on CDH was retired on 17 July 2025 because the phenotype was too broad, remaining as a historical archived version, while the Fryns syndrome chapter was updated on 4 December 2025 [1, 7].

2. Molecular Genetics and Pathophysiology

Development of the diaphragm and lung

The pulmonary hypoplasia of CDH has a primary component, independent of the diaphragmatic defect, demonstrated mainly in animal models in which the hypoplasia precedes visceral herniation, and a secondary component, resulting from competition for space in the thorax, which mainly affects the ipsilateral lung [1]. The pulmonary vascular bed is reduced and the vessels show thickening of the media and adventitia, premature differentiation of smooth muscle cells into a contractile phenotype and failure of relaxation at birth owing to abnormal expression of vasoconstrictors and vasodilators [1]. Animal models support the contribution of several genes: mice with tissue-specific inactivation of Nr2f2 develop dorsolateral Bochdalek hernias similar to those of patients with the 15q26 deletion; Slit3-deficient mice have a central hernia; Pdgfra-deficient mice have posterolateral defects; and Fog2 (Zfpm2) is required for normal development of the diaphragm and lung [14, 28]. In animal models, the hernia with a sac appears to result from abnormal connective tissue, which forms a localised amuscular region, rather than from a primary defect of the myogenic cells [1]. The retinoic acid pathway is a common axis for several monogenic causes: STRA6 is the receptor that takes up retinol bound to retinol-binding protein (RBP4), RARB is a nuclear retinoic acid receptor, ALDH1A2 synthesises retinoic acid and NR2F2 (COUP-TFII) is a transcription factor activated by retinoic acid [35]. Nitrofen, used in the experimental rodent model of CDH, is one of the exposures to investigate in the clinical setting, together with phenmetrazine, thalidomide, quinine, cadmium and lead [1].

Genetic architecture

Older data from classical and molecular cytogenetics identified a genetic cause in 15–20% of patients with CDH, a figure the authors consider an underestimate [1]. Cytogenetically visible chromosomal anomalies were historically described in about 10% of patients, and OMIM reports about 15% chromosomal anomalies [1, 14]. In the series of 200 cases of Klaassens and colleagues, 7% had numerical anomalies and 5% structural anomalies, the 15q deletion being the most frequent structural anomaly [14]. Exome sequencing in 87 trios, analysed together with 39 published trios, showed an excess of de novo likely gene-disrupting variants (LGD: nonsense, frameshift and canonical splicing) and of deleterious missense variants, contributing to 21% of complex CDH and 12% of isolated CDH, with a higher burden in genes expressed in the developing diaphragm and heart [33]. In a cohort of 1469 individuals, including 1064 trios, and 6133 controls, Qiao and colleagues identified 15 candidate genes, eight of them new, through deleterious de novo variants, and two risk loci from common variants in regulatory regions of developmental patterning genes, with similar effects in Europeans and Latinos; the heritability estimated from common variants was about 19%, and polygenic risk scores did not differ between isolated and complex CDH or between carriers and non-carriers of de novo variants, supporting a polygenic model as part of the genetic architecture [34].

Monozygotic twins are discordant for CDH in most consecutive series; in the series of Pober and colleagues, all five monozygotic pairs were discordant, and small cohorts of discordant twins showed no differences in CNVs (copy number variants) or sequence variants [1, 14]. In families with syndromic and non-syndromic CDH, autosomal dominant (AD), autosomal recessive (AR) and X-linked (XL) patterns have been described, and AD pedigrees show reduced penetrance and variable expressivity [1].

Monogenic causes

GeneReviews classified the frequency of CDH in each syndrome at three levels: core feature, with CDH in more than 40% of cases (+++); variable feature, with CDH in about 10–40% (++); and rare feature, with CDH in less than 10% (+) [1]. The following table brings together the monogenic causes from that chapter, supplemented by the 2025 Fryns syndrome chapter and by OMIM.

AVSD, atrioventricular septal defect; DSD, difference of sex development; ID, intellectual disability; DD, developmental delay.

Chromosomal anomalies and copy number variants

Fryns syndrome and the PIGN gene

PIGN, at 18q21.33, encodes GPI ethanolamine phosphate transferase 1, one of the enzymes of GPI anchor biosynthesis, which attaches proteins to the outer leaflet of the cell membrane; the enzyme transfers phosphoethanolamine from phosphatidylethanolamine to the second position of the first α-4-linked mannose [7]. Defective GPI-anchored proteins are abnormally localised and disrupt developmental pathways such as Wnt, Hedgehog and BMP signalling [7]. The mechanism of Fryns syndrome is loss of function, with intragenic deletions, splicing, frameshift and nonsense variants [7]. Biallelic truncating variants are more associated with Fryns syndrome than with the allelic phenotypes and confer an increased relative risk of polyhydramnios, congenital anomalies, distal digital hypoplasia and facial anomalies, as well as higher prenatal and neonatal mortality: of 16 individuals with biallelic variants, two died in utero, six pregnancies were terminated and four died in the first two months [7, 38]. Biallelic PIGN variants cause two other phenotypes: multiple congenital anomalies-hypotonia-seizures syndrome type 1 (MCAHS1, OMIM 614080), almost always due to biallelic missense variants or a missense combined with a loss-of-function variant, without CDH and less severe; and a neurological form with developmental delay, hypotonia and seizures, without CDH or visceral malformations [7, 38].

The PIGN variants cited in GeneReviews, on the MANE transcript NM_176787.5 (validated in VariantValidator), include c.2679C>G p.(Ser893Arg), in the last exon, which does not restore surface expression of CD59 or CD55 (DAF) or binding of fluorescent proaerolysin (FLAER) and was classified as null; c.932T>G p.(Leu311Trp), with partial function; and the splicing variants c.548_549+6del and c.2180+1G>T, associated with abnormal transcripts [7, 39]. An intragenic deletion of 5.07 kb, described as c.329_549+1907del5064 p.(Ser110ArgfsTer15), is probably a founder variant in patients with ancestry from La Réunion and other Indian Ocean islands [7, 40]. Genetic heterogeneity is very likely, because patients with a clinical diagnosis of Fryns have no PIGN variants, and microdeletions at 15q26.2, 8p23.1 and 1q41-q42 have been identified in patients previously diagnosed as Fryns [7, 9].

DIH3 and the ZFPM2 gene

ZFPM2 (FOG2), at 8q23.1, encodes a zinc-finger cofactor of the transcription factors GATA4, GATA5 and GATA6 [13, 35]. The de novo nonsense variant described by Ackerman and colleagues, R112X in historical nomenclature, corresponds to NM_012082.4:c.334C>T p.(Arg112Ter) on the MANE transcript [28, 39]. Longoni and colleagues estimated that deleterious ZFPM2 variants account for almost 5% of 275 CDH exomes, mainly in isolated CDH, and identified an intragenic deletion in a multigenerational family with an estimated penetrance of 37.5% [12, 13]. Deletions of 1.04 Mb and 711 kb encompassing ZFPM2 were inherited from unaffected parents, and Brady and colleagues described the same R112X variant in two sibling fetuses with isolated CDH and in a third sister with total anomalous pulmonary venous return and right anterior eventration, all inherited from an asymptomatic mother [13]. The 46,XY gonadal dysgenesis and conotruncal heart defects associated with ZFPM2 have inconsistent evidence gradings among curators [35].

DIH2, 8p23.1 deletion and GATA4

The 8p23.1 deletion has been associated with CDH since 1990; in a series of 256 CDH trios, Yu and colleagues found two 8p23.1 deletions and defined a 3.8 Mb critical region containing GATA4 and SOX7 [11]. In one family, the GATA4 variant described as c.754C>T p.(Arg252Trp) on NM_002052.5 was inherited from the father and paternal grandfather, both asymptomatic but with small left Bochdalek hernias on MRI, and a paternal uncle died on day two with left-sided CDH [11, 41]. A second, de novo variant, c.848G>A p.(Arg283His), was found in a patient with CDH, atrial septal defect (ASD) and ventricular septal defect (VSD) [41]. On the GATA4 MANE transcript, NM_001308093.3, these variants correspond to c.757C>T p.(Arg253Trp) and c.851G>A p.(Arg284His) (NC_000008.11:g.11749056C>T and g.11750175G>A), which shows the importance of stating the transcript when comparing reports [39]. OMIM keeps the association of GATA4 with CDH as pending confirmation [11]. Carriers of GATA4 variants or deletions may have neonatal or childhood diabetes, sometimes with pancreatic hypoplasia or agenesis and exocrine insufficiency [35].

DIH1, the 15q26 region and NR2F2

The minimal region at 15q26.1-q26.2, of about 5 Mb, contains NR2F2 and CDH2; the region is gene-poor and its haploinsufficiency may not be fully penetrant, because a patient with a 15q26.1-q26.2 deletion without CDH has been described [14]. Heterozygous NR2F2 variants (MANE transcript NM_021005.4), often de novo, cause heart defects, with significant enrichment in non-syndromic atrioventricular septal defect (five rare missense variants, two de novo, in 13 trios and 112 probands), Bochdalek CDH and, in 46,XX SRY-negative children, testicular or ovotesticular development with virilisation [35, 39]. A de novo frameshift variant was described in a patient with CDH, and fetal exome sequencing in 22 trios strengthened the association of NR2F2, PTPN11 and WT1 with CDH [35, 42].

MYRF, WT1 and other causes with genitourinary expression

MYRF (MANE transcript NM_001127392.3) encodes a transcription factor bound to the endoplasmic reticulum membrane, which self-cleaves; the N-terminal fragment, containing the DNA-binding domain, migrates to the nucleus [19]. Qi and colleagues identified, in 362 trios, four deleterious de novo MYRF variants (P = 5.3×10⁻⁸) and eight more in other cohorts; the missense variants cluster in two functional domains, and the transcriptome of diaphragmatic fibroblasts suggests loss of transcriptional activity [32]. Inheritance is AD, with many de novo variants and highly variable expressivity within families; variants that truncate or alter the C-terminus are associated mainly with high hypermetropia and nanophthalmos, with reduced penetrance and severity of the other features [19]. WT1 causes CDH in fewer than ten infants described; the phenotype termed Meacham syndrome, with CDH, pulmonary dysplasia, complex heart defects and genitourinary anomalies, was lethal in early infancy in most reports, and none of the patients with a confirmed WT1 variant and a diaphragmatic defect had a complex heart defect, which led to the hypothesis of a multigenic cause [20, 43]. One patient with a clinical diagnosis of Meacham in fact had a MYRF variant [19].

STRA6, RARB and LRP2

STRA6 (MANE transcript NM_022369.4) mediates cellular uptake of retinol from RBP4; biallelic loss of function reduces intracellular retinol and retinoic acid signalling during organogenesis [35, 39]. Truncating variants cause the lethal form of Matthew-Wood syndrome, missense variants are associated with milder phenotypes and longer survival, and at the mild end there is isolated colobomatous microphthalmia [29, 35, 44]. RARB has two opposing mechanisms: biallelic truncating variants without transcriptional activity (loss of function, AR) and de novo missense variants in the ligand-binding domain, such as p.Arg387Cys and p.Arg387Ser, which increase retinoic acid-induced activity two- to threefold (gain of function, AD) [30, 35]. LRP2 (MANE transcript NM_004525.3), at 2q31.1, encodes megalin, a transmembrane endocytic receptor of the proximal tubule that reabsorbs lipoproteins, sterols, vitamin-binding proteins and hormones; its absence explains the low-molecular-weight proteinuria present in 100% of patients with Donnai-Barrow syndrome [16, 26]. LRP2 is one of the longest genes in the genome and has many private benign variants, which calls for caution in interpretation; one patient was homozygous through paternal uniparental isodisomy of chromosome 2 [16].

Genotype–outcome correlation

In a prospective cohort of 647 newborns with CDH, with genome or exome sequencing in 462 trios, complex forms had higher mortality than isolated forms (P = 8×10⁻⁶), and patients with isolated CDH carrying a likely deleterious de novo variant had mortality similar to that of complex forms and much higher than that of isolated cases without such a variant (P = 3×10⁻³); the trend was similar for pulmonary hypertension at one month, and carriers had neurodevelopmental scores 12 to 17 points lower at two years [45]. Identifying a genetic cause thus modifies prognostic counselling even when CDH appears isolated on prenatal imaging.

3. Clinical Spectrum and Manifestations by System

Prenatal presentation

CDH is diagnosed prenatally in most cases: in France, about 80% of cases are detected on a screening ultrasound, usually the second-trimester scan, by the presence of the stomach and bowel loops in the thorax, by mediastinal and cardiac shift and, sometimes, by excess amniotic fluid [3]. The overall prenatal detection rate by ultrasound is about 60% in the Oxford Desk Reference, and GeneReviews reports detection of more than 80% of left-sided CDH and more than 60% of right-sided CDH by two-dimensional ultrasound and/or MRI in the second trimester [1, 21]. Right-sided hernias are more difficult to identify, because the stomach rarely ascends and the lung and liver have similar echogenicity [1, 3]. Polyhydramnios is frequent, especially in syndromic forms [1]. In a French series of 31 prenatally diagnosed cases (Rennes, 1995–2000), 10 (32%) had associated anomalies, four of them chromosomal (trisomy 18 and 22q11 deletion) and four Fryns syndromes [21].

Neonatal period

The newborn presents with the triad of respiratory distress, displacement of heart sounds to the contralateral side and a scaphoid abdomen; the chest and abdominal radiograph shows gas-filled bowel loops in the thorax with mediastinal shift [1, 3]. Respiratory distress may appear after a period of 24–48 hours of stability [1]. Almost all patients have some degree of pulmonary hypoplasia, and pulmonary hypertension results from the reduced vascular bed and failure of vascular relaxation at birth; the therapies effective in usual persistent pulmonary hypertension of the newborn (PPHN) are of little efficacy in CDH [1]. Early left ventricular dysfunction contributes to pulmonary hypertension [3]. Respiratory complications include tracheobronchomalacia, pneumothorax and secondary infections, mainly viral, which may precipitate terminal respiratory failure months after surgery [1].

In the CDH Study Group registry, among 2183 newborns discharged from hospital, 23.5% had no morbidity and 76.5% had pulmonary, neurological or gastrointestinal morbidity at discharge; in the overall cohort, 62.2% had a prenatal diagnosis, 17.6% were preterm, 3.8% had a major heart defect, 3.4% a chromosomal anomaly, 85.1% a left-sided defect, 19.9% had been treated with extracorporeal membrane oxygenation (ECMO) and 45.0% with a patch; defects A, B, C and D accounted for 17.1%, 45.1%, 29.7% and 8.2% [22]. Patients without morbidity had ECMO in 2.7%, against 40.7% of those with pulmonary morbidity [22].

Late presentation

Presentation after the neonatal period occurs in 5–10% of patients, with respiratory distress (for example, pleural effusion from intestinal incarceration) or gastrointestinal symptoms, such as abdominal pain from chronic or intermittent obstruction; about 1% are asymptomatic and the hernia is an imaging finding [1]. Intestinal malrotation is frequent and may present as obstruction or volvulus [1]. Late forms may present as asthma, gastro-oesophageal reflux (GOR) or obstruction, and even an incidentally discovered hernia should be operated on within a short time to prevent complications [3].

Long-term respiratory system

CDH-associated pulmonary hypertension persists beyond the neonatal period in up to 20% of patients and is associated with higher mortality, although most improve over time [22]. At 30 days of life, 10–50% of survivors have pulmonary morbidity (oxygen, bronchodilators, inhaled corticosteroids or diuretics), a strong predictor of long-term morbidity; at least 4% undergo tracheostomy, and 22% in severe CDH [22]. The data summarised by the American Academy of Pediatrics (AAP) include bronchopulmonary dysplasia in 41%, readmission for respiratory infections in 33% in the first two years, wheezing in 26%, airway obstruction in 28% and abnormalities of spirometry, lung volumes and exercise tolerance in 48%; at 8–12 years, 30% have abnormal FEV1 and forced vital capacity, with a tendency to worsen [22]. The PNDS gives ranges of 22–52% for bronchopulmonary dysplasia, asthma in 15% at five years and lung disease in 7–33% [3]. Ventilation-perfusion scintigraphy shows reduced perfusion mainly in the ipsilateral lung, and an early severe mismatch (perfusion below 30–40%) is associated with pneumonia and poor weight gain [1, 22]. Hospitalisations for respiratory syncytial virus (RSV) affect 23.8% of survivors [22].

Digestive system and nutrition

Difficulty in achieving caloric intake is the most frequent problem in survivors; it results from oral aversion, dysphagia, tachypnoea, GOR, gastric dysmotility and, above all, high metabolic requirements [22]. GOR is present early in at least 40% (45–90% in the PNDS) and rarely appears after two years of age if it was not present in infancy, but it may be silent; oesophagitis affects 50% of adults and, in some series, more than half of adults have oesophageal disease, including Barrett oesophagus, with oesophageal carcinoma described in adult survivors [3, 22]. Growth impairment affects more than half in the first months and 50% at one year, and 30–40% of infants require a nasogastric tube or gastrostomy (more than 50% of those treated with ECMO), a proportion that falls to less than 10% at five years [3, 22]. In a single-centre cohort of 1247 patients treated between 2000 and 2022, the 835 survivors without major comorbidity gained weight more slowly, stabilising at a reduced level after two years; height normalised by six years in those born at term, head circumference normalised by one year and body mass index remained low; defect size and gestational age, but not ECMO, were independent predictors of growth restriction [46]. Adhesive intestinal obstruction occurs in 10–30% over the lifetime (20% in the PNDS), 80–90% of cases require surgery and there is a risk of short bowel [3, 22].

Nervous system, hearing and development

The main independent predictor of intellectual disability described in GeneReviews is ECMO, and mild (44%) or severe (13%) delay in at least one domain was reported at one year [1]. The AAP considers that the extent and duration of critical illness, reflected for example in the need for vasoactive drugs, is the most important predictor, more so than ECMO or the defect size itself [22]. Even in isolated CDH, 10–25% of survivors have moderate to severe neurodevelopmental impairment, more motor than cognitive; in a European programme with assessments at 5, 8 and 12 years, motor function was below normal in low-risk patients only at five years, and at all ages in high-risk patients; more than half of survivors had behavioural and cognitive deficits at eight years, and attention deficit hyperactivity disorder and autism spectrum behaviours are frequent (14%) [22]. In patients treated with ECMO, half had intracranial haemorrhage on MRI [22]. Sensorineural deafness, often late-onset and progressive and therefore not detected by newborn screening, has been described in 25% of patients and in up to 100% of those treated with ECMO in some older series; the AAP reports ranges from less than 5% to more than 50% and identifies ECMO, aminoglycosides (the most relevant iatrogenic factor) and prolonged ventilation as risk factors [1, 22]. Patients with multiple anomalies or genetic syndromes have more severe motor and cognitive deficits [22].

Musculoskeletal and genitourinary systems

Chest asymmetry affects up to half of patients, and pectus (mainly excavatum) and scoliosis affect about 25% in GeneReviews; the AAP reports chest deformity in more than 50% after repair with a patch or muscle flap and in less than 10% after primary repair, and thoracic scoliosis in 15%, also after primary repair, which suggests it is part of the disease spectrum [1, 22]. The PNDS indicates chest deformity in 21–48% and scoliosis in 5–27% [3]. Cryptorchidism requiring surgical correction affects 18–36% of boys, usually on the side of the hernia [22].

Hernia recurrence

At least 10% of patients have a recurrence after the initial repair, with a higher risk when a patch was used (up to 50% in some series); recurrence is more frequent in the first two years, but may occur in adolescence or adulthood, be asymptomatic and also happen after primary or muscle flap repair [1, 22]. Pregnancy in a woman who survived CDH may increase the risk of recurrence [22].

Phenotypes of the main genetic entities

Fryns syndrome. In the Smith series, diaphragmatic defects were present in 89%, genital anomalies in 86% and CNS malformations (Dandy-Walker, hypoplasia of the optic and olfactory tracts, arhinencephaly, heterotopias, agenesis of the corpus callosum) in 50% [47]. The facies is coarse, with widely spaced eyes, a broad and depressed nasal bridge, a broad nasal tip, long philtrum, low-set and abnormal ears, tented upper vermilion, macrostomia and a small mandible; the distal phalanges of the hands and feet are short and the nails small [7]. Among the 17 patients with biallelic PIGN variants and a Fryns phenotype, the frequency was 88% for polyhydramnios or cystic hygroma/increased nuchal fold, 71% for dysmorphism, cleft lip or palate and heart defect, 65% for CDH and genitourinary anomalies, 59% for distal digital hypoplasia, 47% for brain and gastrointestinal malformations, 12% for ocular anomalies and 100% for prenatal or neonatal lethality [7]. CDH is present in more than 90% of cases with a clinical diagnosis, but in 50–65% of cases with PIGN, and may be less frequent in patients from La Réunion [7]. VSD is the most frequent heart defect in clinical diagnoses; in cases with PIGN, tetralogy of Fallot, VSD, patent ductus arteriosus, overriding aorta, hypoplastic pulmonary trunk and aberrant right subclavian artery have been described [7]. Among 112 patients reviewed by Lin and colleagues, about half had a cardiovascular malformation [9]. Survivors, about 14% of cases published up to 1995, had less CDH, milder pulmonary hypoplasia, no complex heart defect and severe neurological impairment; the oldest survivor in Smith walked at six years and did not speak at nine, and there are reports of mild delay in children aged one and two years [9, 47]. Seizures in survivors with PIGN may be refractory [7].

Donnai-Barrow syndrome. The facies, with a large anterior fontanelle, wide metopic suture, widow's peak, marked hypertelorism, enlarged eyeballs, downslanting palpebral fissures, posteriorly rotated ears and a short nose, is consistent in patients with confirmed LRP2 and persists in adults [16]. High myopia (−6 dioptres or worse) is universal, retinal detachment affects 30%, sensorineural deafness is present in 100% with onset in infancy, intellectual disability is mild to moderate, and CDH (or eventration) and omphalocele (or umbilical hernia) each occur in about 40%, rarely together [16]. Low-molecular-weight proteinuria, with loss of retinol-binding protein and vitamin D-binding protein, may cause hypovitaminoses, and nephrocalcinosis, hypercalciuria, focal segmental glomerulosclerosis and progression to end-stage kidney disease have been described [16, 26]. Only two adults over 50 years with confirmed LRP2 have been published, and renal deterioration is the main long-term concern [16].

Matthew-Wood syndrome (PDAC). Most pregnancies end in miscarriage, fetal death or neonatal death from severe pulmonary or cardiac involvement; about a third of patients survive (between three months and 40 years), with severe to profound intellectual disability, hypotonia and spastic tetraparesis, although one long-term survivor had normal intelligence [29]. Findings include micro- or anophthalmia, pulmonary agenesis or hypoplasia, alveolar capillary dysplasia, tracheal atresia, CDH or eventration, anomalies of the pulmonary arteries and veins, renal dysplasia and pancreatic anomalies [29]. In the ten neonatal survivors with de novo RARB variants, all had severe global delay with progressive motor deterioration from spasticity and/or dystonia, and Chiari type I malformation was present in most [35].

MYRF-CUGS. Genital anomalies (80–89%) range from unilateral cryptorchidism to female-appearing genitalia in 46,XY individuals, and the six 46,XX individuals described included three with absent uterus and fallopian tubes, small or absent ovaries and an absent or blind-ending vagina; ocular changes (84%) are mainly high hypermetropia and nanophthalmos; heart defects (68–90%) range from isolated dextrocardia to hypoplastic left heart syndrome, with scimitar syndrome as the most frequent specific lesion; CDH occurs in 31–59%, and 30% of patients without CDH have primary pulmonary hypoplasia [19]. Patients diagnosed postnatally, especially at older ages, have far fewer severe malformations such as CDH [19]. There may be an increased risk of gonadoblastoma in 46,XY individuals, with very limited data [19].

DIH3 (ZFPM2) and DIH2 (8p23.1). In ZFPM2, isolated posterolateral CDH predominates, with a conotruncal heart defect in one patient of the Longoni series and eventration and anomalous pulmonary venous return in other relatives [13]. In DIH2, patients have pulmonary hypoplasia, respiratory failure and, in some cases, atrioventricular, atrial and ventricular septal defects [11].

Differences between left- and right-sided CDH. In a population-based study from California, the 38 patients (26%) with right-sided CDH, compared with the 108 (74%) with left-sided CDH, had fewer ASDs (1/38 vs 20/108; p = 0.015), more bilateral pulmonary hypoplasia (15/38 vs 22/108; p = 0.029), fewer abnormalities of skull or face shape (1/38 vs 17/108) and no digital or limb anomalies (0/38 vs 13/108 and 18/108) [14].

4. Clinical Recognition and Warning Signs

Prenatal period

CDH should be suspected when ultrasound shows the stomach or bowel loops in the thorax, absence of the gastric bubble below the diaphragm, mediastinal or cardiac shift and polyhydramnios; colour Doppler identifies the abnormal position of the umbilical and portal veins when the liver is herniated and helps to recognise right-sided hernias [1, 3]. About 12% (14/121) of prenatal CDH diagnoses at a fetal medicine centre in the West Midlands were false positives, and the final diagnoses included cystic adenomatoid malformation, bronchogenic cyst, hydrothorax, pleural effusion, pulmonary sequestration, duodenal atresia and malrotation; fetal MRI helps in difficult cases [21]. Signs pointing to syndromic CDH are the presence of other anomalies on the reference ultrasound (about 40% of cases), polyhydramnios, growth restriction or, conversely, macrosomia, rhizomelic limb shortening, cystic hygroma or increased nuchal translucency, hypoplasia of the distal phalanges, facial clefts, microphthalmia and genitourinary anomalies [1, 3, 7]. CDH, a heart defect and rhizomelic shortening suggest Pallister-Killian syndrome, whose most frequent ultrasound finding is the combination of polyhydramnios, macrosomia and limb shortening [1]. CDH with upper limb anomalies suggests Cornelia de Lange syndrome [21].

Newborn

Respiratory distress at birth, displaced heart sounds, reduced breath sounds on the side of the hernia and a scaphoid abdomen call for immediate intubation and transfer to a surgical neonatal intensive care unit [1, 3]. Bag-mask ventilation is contraindicated, because it distends the herniated stomach and bowel and worsens lung compression [3]. Examination by a clinical geneticist looks for major malformations and minor anomalies (craniofacial, ocular, cardiovascular, CNS, limb and genitourinary) [1].

Child, adolescent and adult

In a child or adult, CDH may present with chronic cough, recurrent lung infections, pleural effusions, pneumonia, dysphagia, abdominal pain, intestinal obstruction or volvulus [1]. In a patient who has already undergone surgery, acute respiratory distress should prompt consideration of hernia recurrence, a PAH crisis, decompensation of chronic respiratory failure, viral infection or aspiration pneumonia, and asthma; an obstructive picture should suggest adhesive obstruction or recurrence; and an episode of malaise may correspond to a PAH crisis, vagal hypertonia from GOR, hypoglycaemia or dehydration [3]. A child with repaired CDH and growth faltering, wheezing, nocturnal or exertional cough, otitis media with effusion or language delay should be referred to the centre that follows them [3].

Signs pointing to a genetic cause

An affected sibling, parental consanguinity or perinatal death of other children suggest an AR disorder, such as Fryns syndrome, Donnai-Barrow syndrome or Matthew-Wood syndrome [1, 7]. CDH in a parent, even asymptomatic, may reveal an AD form with incomplete penetrance, as in the GATA4 family in which the father and grandfather had small hernias visible only on MRI [11]. The association of CDH with genital anomalies in a 46,XY individual suggests MYRF, WT1, ZFPM2 or GATA4; with microphthalmia it suggests STRA6, RARB or the genes of microphthalmia with linear skin defects; with omphalocele, hypertelorism and low-molecular-weight proteinuria it suggests LRP2; with hypoplastic nails and short distal phalanges it suggests Fryns syndrome, a GPI anchor defect or Pallister-Killian syndrome; and with overgrowth it suggests Simpson-Golabi-Behmel or Beckwith-Wiedemann syndrome [1, 7, 16, 19]. There is no population newborn screening for CDH.

5. Diagnosis: Criteria, Tests and Algorithms

Anatomical confirmation and prenatal prognostic assessment

The PNDS recommends, when ultrasound raises suspicion, a reference ultrasound at a multidisciplinary prenatal diagnosis centre, ideally in the second trimester, to confirm the diagnosis, determine the side and look for associated malformations [3]. The prognostic assessment relies on the observed-to-expected lung-to-head ratio (o/e LHR), measured on the four-chamber view; on liver position by the Cordier technique, which relates the position of the stomach to the plane of the atrioventricular valves; on fetal echocardiography by a paediatric cardiologist; and on fetal MRI, ideally between 26 and 28 weeks of gestation, which measures the observed-to-expected total lung volume and better visualises the ipsilateral lung, without the need for a routine second MRI [3]. The prognostic factors in the literature were assessed after 21 weeks, and diagnosis in the first trimester is associated with a worse overall prognosis [3]. The Canadian consensus, in its 2023 update, recommends measuring the o/e LHR by the manual tracing (trace) method between 22 and 32 weeks, in conjunction with a regional fetal medicine programme, and using thresholds below 25% for left-sided CDH and 50% for right-sided CDH to predict poor outcome, replacing the 45% threshold for right-sided CDH of the 2018 version [48, 49]. With these thresholds, severe hypoplasia (o/e LHR ≤25% on the left, ≤50% on the right) corresponds to an estimated survival of 30% or less on the left and 20% on the right, and moderate hypoplasia in left-sided CDH is defined by an o/e LHR of 26–34% [48]. Stomach position has been proposed as a surrogate for liver herniation, which is difficult to recognise on ultrasound, and correlates with neonatal mortality and morbidity, but lacks prospective validation [48]. Fetal MRI should be considered in all fetuses and is strongly recommended in moderate and severe forms: an o/e total lung volume below 35% and liver herniation are significant predictors of mortality, MRI is more reproducible than ultrasound, does not depend on maternal habitus or fetal position, has greater sensitivity and specificity for survival and, according to the TOTAL protocol, is ideally performed at about 26 weeks; ultrasound remains the basis of assessment because of its availability, and the two techniques should be combined in high-risk fetuses [48]. Since 2023, the consensus has also recommended a detailed anatomical assessment and fetal echocardiography at a tertiary fetal medicine centre in all pregnancies with CDH, because 30–40% of prenatally diagnosed cases have other structural or genetic anomalies, mainly cardiovascular [48]. CDH EURO recommends that severity be assessed at an experienced centre by the o/e LHR and liver position, reports associated chromosomal or genitourinary anomalies in about 25% and cardiac anomalies in about 20% of cases and considers lung volumetry, 3D ultrasound and Doppler of the pulmonary vasculature to be research tools, despite the prognostic value described in individual series [50]. The LHR increases with gestational age, unlike the observed-to-expected lung volume on MRI, which is an independent predictor of mortality at six months; liver herniation is associated with a worse prognosis [1]. A 2025 review emphasises that these markers do not predict pulmonary hypertension or cardiac dysfunction and that the newer parameters lack validation and vary between centres [51]. In a Chinese cohort of 130 fetuses, the o/e LHR, o/e total lung volume, percentage predicted lung volume and liver herniation were associated with outcomes, and a multivariable model for neonatal outcome had an area under the curve of 0.929 [52].

Postnatal diagnosis

Postnatal diagnosis is based on clinical findings and the chest and abdominal radiograph, followed by assessment for associated malformations with echocardiography, abdominal and renal ultrasound, transfontanellar ultrasound and skeletal radiographs, and by a genetic study if one was not performed prenatally [3]. In late cases, CT and MRI confirm the defect [1, 7].

Genetic testing strategy

Genetics is part of the prenatal assessment in all the guidelines consulted, but with different wording. The PNDS requires, at a minimum, array comparative genomic hybridisation (array CGH), preferably on amniotic fluid, with testing for tetrasomy 12p by direct examination; if only a karyotype on chorionic villi was performed in the first trimester for another indication, the CGH should be repeated on amniotic fluid [3]. GeneReviews, in the archived version, described karyotype as the first-line test in many institutions, CMA in fetuses with additional anomalies or a suggestive family history and even in apparently isolated CDH, and exome sequencing, at the time, only in a research setting [1]. The 2023 Canadian consensus recommends offering invasive genetic testing, ideally with CMA, to all pregnancies with CDH, because isolated CDH has 10–13% CMA abnormalities, and anticipates that extended genomic analysis (exome, RNA analysis) will increase this yield; CDH EURO includes invasive sampling for high-resolution genetic testing in the complete prenatal assessment that precedes multidisciplinary counselling [48, 50]. Scott and colleagues consider CMA (with or without karyotype) the optimal test for fetal chromosomal anomalies and CNVs, and exome sequencing the most comprehensive strategy for sequence variants, with genome sequencing accompanied by CNV analysis as the potentially most efficient test when available prenatally [4]. The ACMG and the AAP recommend exome or genome sequencing as a first- or second-tier test in children with developmental delay, intellectual disability or multiple congenital anomalies [7]. The AAP recommends a genetics consultation for all patients with CDH before discharge or shortly afterwards, if not already done [22].

The discrepancy between published yields reflects different populations (fetuses or children; isolated or complex CDH; clinical reference laboratories or research cohorts) and different definitions of a positive result. In the series of 560 probands sequenced in a commercial laboratory, a "reportable finding" was described in 61% (64% in complex forms and 56% in isolated forms), but that category is not equivalent to a confirmed molecular diagnosis [55]. By contrast, a Chinese fetal cohort of 51 cases had an overall yield of 5.9%, with 22.2% in non-isolated forms and 2.4% in isolated forms [15]. The proportion of patients with a detectable variant is therefore not a measure of the clinical sensitivity of the test. Trio sequencing is preferred because the rate of de novo variants is high in both isolated and complex CDH [1].

Diagnostic algorithm

The algorithm that emerges from the sources consulted is as follows. For prenatal CDH: reference ultrasound and fetal echocardiography; fetal MRI between 26 and 28 weeks; amniocentesis (preferred) with CMA and direct examination for i(12p), or chorionic villus sampling with CMA and, if normal and suspicion of Pallister-Killian persists, amniocentesis; trio exome when CMA is normal, especially in complex forms, but also in isolated forms; and pre-test counselling [1, 3, 4, 53]. After birth or in the event of death: examination by a geneticist, photographs, skeletal radiographs and autopsy or fetal pathology, strongly recommended by the PNDS, with collection of a sample for genetic study and, if possible, fibroblasts [1, 3]. When Pallister-Killian syndrome is suspected after birth, the karyotype should be performed on skin fibroblasts, because peripheral blood is almost always normal, although CMA on blood may detect the mosaicism when the proportion of tetrasomic cells is high [1, 7]. Prenatal exposures to phenmetrazine, thalidomide, quinine, cadmium, lead and nitrofen should be investigated, and mycophenolate mofetil embryopathy includes CDH [1, 56].

Criteria for Fryns syndrome

The criteria of Lin and colleagues (2005), adopted by GeneReviews, include six elements: diaphragmatic defect (hernia at any location, eventration, significant hypoplasia or agenesis); characteristic facies; hypoplasia of the distal phalanges of the fingers and toes; significant pulmonary hypoplasia; at least one characteristic associated anomaly (polyhydramnios, orofacial cleft, cardiovascular, genitourinary, brain or gastrointestinal malformation, cloudy corneas or microphthalmia); and a family history compatible with AR inheritance [7]. The clinical diagnosis requires four of the six criteria and the absence of a CDH-associated CNV [7]. In the review of 112 patients, 82 met the strict criteria (four or more of six), 30 the broad criteria (three of six) and 12 lacked core features or had atypical features [9]. The molecular diagnosis requires biallelic pathogenic or likely pathogenic PIGN variants; the identification of two VUS (variants of uncertain significance), or of one pathogenic variant and one VUS, neither establishes nor excludes the diagnosis [7]. PIGN sequencing detects sequence variants, and if only one variant is found, deletion/duplication analysis or exome/genome sequencing follows; a 5.07 kb intragenic deletion has been described, and no pathogenic deep intronic variants have been described [7]. The limits of the clinical criteria are recognised: patients with ectrodactyly, radial aplasia, limb shortening or multiple pterygia were classified as Fryns by some authors and not by others [7].

Report content and interpretation

The molecular report should state the gene, the reference transcript with version, the HGVS nomenclature (c. and p.), the zygosity, the class according to the ACMG/AMP criteria and, in AR disorders, confirmation of phase by parental testing [7]. In the case of GATA4, the same variant is described as p.(Arg252Trp) on NM_002052.5 and as p.(Arg253Trp) on the MANE transcript NM_001308093.3, so comparison with the literature requires reference to the transcript [39, 41]. In a gene with incomplete penetrance such as ZFPM2, inheritance from a healthy parent does not exclude pathogenicity [12]. In LRP2, the size of the gene and the number of private benign variants call for careful interpretation [16]. Apparent homozygosity for a PIGN variant may result from an undetected deletion on the other allele or from uniparental isodisomy [7].

Prenatal and preimplantation diagnosis

When the causal variant is known in a relative, targeted prenatal diagnosis and preimplantation genetic testing are possible, with counselling on the reduced penetrance of AD forms [1]. In a pregnancy at high risk because of a sibling with Fryns syndrome of known molecular cause, prenatal or preimplantation diagnosis is offered; without a molecular diagnosis, two- and three-dimensional ultrasound and MRI look for polyhydramnios, increased nuchal translucency or hygroma, hyperechogenic bowel, echogenic kidneys, heart defects, renal cysts, hydroureter, ventriculomegaly, agenesis of the corpus callosum and Dandy-Walker malformation, but the absence of these findings does not exclude the disease [7]. In Donnai-Barrow syndrome, fetal ultrasound and MRI may identify agenesis of the corpus callosum, CDH, omphalocele and hypertelorism, but their absence also does not exclude an affected fetus [16].

6. Differential Diagnosis

Anatomical differential diagnosis of a fetal or neonatal thoracic mass

The imaging differential includes congenital cystic adenomatoid malformation of the lung (CCAM or CPAM), bronchogenic cysts, pulmonary sequestration (the extrapulmonary form is often associated with CDH), cystic teratomas of the anterior mediastinum, neurogenic tumours of the posterior mediastinum, paraoesophageal hernia (a type of hiatus hernia, rare in childhood), pulmonary agenesis, hydrothorax and pleural effusions, duodenal atresia and malrotation [1, 21]. Eventration of the diaphragm may coexist with or be mistaken for Bochdalek hernia, and both occur in patients with variants in the same gene [1].

Aetiological differential diagnosis of complex CDH

Fryns syndrome and Pallister-Killian syndrome may be clinically indistinguishable, and in some patients only the chromosomes or the inheritance pattern separate them; initial diagnoses of Fryns were corrected when i(12p) was found in fibroblasts or kidney [7, 9]. Sparse hair is characteristic of Pallister-Killian syndrome, whereas the sisters described by Fryns had a low hairline and hypertrichosis [7]. MCAHS1, allelic at PIGN, is distinguished from Fryns syndrome by the absence of CDH and by a picture with fewer anomalies [7]. Matthew-Wood syndrome is not associated with small nails or short digits [7]. Pallister-Killian, Cornelia de Lange, Kabuki, Simpson-Golabi-Behmel, CHARGE, Fraser, craniofrontonasal, Coffin-Siris, White-Sutton, Beckwith-Wiedemann, Wolf-Hirschhorn and 22q11.2 deletion syndromes, the trisomies, Wilms tumour predisposition and GPI anchor defects are part of the syndromic differential diagnosis.

7. Treatment and Risk Management Strategies

Principles and organisation of care

There is no treatment targeting the genetic cause of any form of CDH; treatment addresses the malformation and its pulmonary, vascular, digestive and developmental consequences, adjusted to the associated anomalies [7, 35]. The PNDS sets as objectives treating the difficulties of cardiorespiratory adaptation while minimising barotrauma and volutrauma, preventing and treating chronic PAH, feeding and oral-motor disorders, undernutrition, GOR and respiratory complications, ensuring continuity from the fetal period to adulthood, supporting breastfeeding and providing psychological and social support to the family [3]. Management requires paediatric anaesthetists, paediatric surgeons and neonatal intensive care in a level III maternity unit, ideally at a reference or competence centre [3]. Syndromes with a lethal prognosis, such as most forms of Fryns syndrome with CDH and severe pulmonary hypoplasia, call for a discussion with the family about the intensity of care, including the option of palliative care, which the PNDS also envisages in the most severe isolated forms in which resuscitation is withdrawn [3, 7].

Reference guidelines and divergences

The sources are the PNDS [3], the 2015 CDH EURO Consortium consensus, which brings together 22 European centres and grades the evidence by the SIGN system, with almost all recommendations grade D [50], the 2018 Canadian consensus and its 2023 update, produced by GRADE methodology from 103 articles published between January 2017 and August 2022 and endorsed by a modified Delphi process with a consensus threshold above 80%, which changed 20 recommendations and created 15, mainly on analgesia and neuromuscular blockade [48, 49], and the 2026 AAP clinical report [22]. After implementation of the 2010 European protocol, survival rose from 67% to 88% in two centres, a before-and-after comparison without a control group [50]. The main divergences concern saturation and PaCO2 targets, more permissive in CDH EURO; the place of inhaled NO, which the PNDS reserves for failure of ventilation and haemodynamic correction, whereas CDH EURO proposes it with oxygenation index and saturation difference criteria and the Canadian consensus indicates it in suprasystemic PAH without left ventricular dysfunction, all with discontinuation in the absence of response; milrinone, which the Canadian consensus recommends, the PNDS does not routinely recommend and CDH EURO does not grade; repair on ECMO, which CDH EURO accepts, the Canadian consensus advises against before decannulation (except early repair in patients with a low probability of weaning) and the PNDS considers controversial; and GOR prevention, for which CDH EURO proposes preventive antireflux therapy with enteral feeding, whereas the AAP does not recommend empirical therapy and the Canadian consensus rejects preventive fundoplication [3, 22, 48–50]. The minimally invasive approach, which the 2018 Canadian consensus advised against in the newborn, may be considered since 2023 in selected patients, which brings the Canadian position closer to that of the PNDS, which accepts it in forms with a favourable prognosis, and to that of CDH EURO, which accepts it in left-sided CDH with an intra-abdominal liver; all recognise the higher risk of recurrence, and the AAP describes the trade-off between more recurrences and fewer adhesive obstructions [3, 22, 48–50]. The CDH EURO recommendation not to perform FETO outside a trial predates the publication of the TOTAL trials, which the 2023 Canadian consensus already incorporates [48, 50].

Obstetric management

The pregnancy requires no special surveillance except in the case of polyhydramnios, which calls for regular surveillance to decide on amnioreduction and limit the risk of preterm delivery; antenatal corticosteroids should be given before 34 weeks and discussed up to 37 weeks in threatened preterm labour [3]. Delivery should take place in a level III maternity unit with a trained paediatric medical-surgical team; there is no argument for bringing delivery forward, and delivery near term should be the rule [3]. In France, termination of pregnancy for medical reasons may be requested in very severe forms under the 2011 bioethics law, after complete information that includes consultations with the neonatologist and the surgeon, and fetal pathology is strongly recommended [3]. CDH EURO recommends antenatal corticosteroids if delivery is expected before 34 weeks and planned delivery after 39 weeks, by induction or caesarean section and according to maternal indications, at a high-volume tertiary centre (at least six CDH cases per year); it relies on a study in which neonatal and infant mortality fell from 25% and 36% at 37 weeks to 17% and 20% at 40 weeks, although another Canadian study found no difference in mortality between gestational age categories or between vaginal delivery and caesarean section [50]. The 2023 Canadian consensus recommends considering delivery at about 39 weeks (between 38 and 39), owing to better survival at 28 days with term delivery, at a tertiary centre with neonatal and paediatric intensive care and paediatric surgery experienced in CDH, because outborn delivery is a significant predictor of mortality, and deciding the mode of delivery by the usual obstetric indications [48]. In the meta-analysis of Stege and colleagues, about a quarter of prenatally diagnosed fetuses underwent elective termination [1, 59].

Fetal tracheal occlusion

Fetal laryngeal obstruction causes lung distension through fluid retention, the rationale for fetoscopic endoluminal tracheal occlusion (FETO), which consists of placing a balloon in the fetal trachea by fetoscopy and removing it before delivery [1, 3]. An early American trial of fetal tracheal occlusion showed no benefit and had many preterm deliveries [1]. The two TOTAL (Tracheal Occlusion To Accelerate Lung growth) trials, open-label, randomised 1:1, in singleton fetuses with isolated left-sided CDH and standardised postnatal care, defined the current evidence [5, 60].

PROM, prelabour rupture of membranes; RR, relative risk; aOR, adjusted odds ratio; ITT, intention to treat.

The severe-form trial showed a survival benefit maintained at six months; the moderate-form trial showed no significant benefit, and both increased the risk of prelabour rupture of membranes and preterm delivery [5, 60]. The pooled reanalysis of individual data from the two trials included the 287 randomised pregnant women who did not withdraw consent: 196 with moderate hypoplasia and 91 with severe hypoplasia, the latter comprising the 80 included up to the third interim analysis and 11 already randomised when the decision was taken to stop the trial; recruitment was concentrated in Leuven (27%), Paris-Antoine Béclère (23%), London (16%) and Barcelona (12%) [61]. The analysis, by intention to treat, used logistic regression for binary outcomes and linear regression for gestational age at live birth, both with penalised maximum likelihood estimation, chosen to mitigate the overestimation of effect typical of a trial stopped early for benefit and the confounding between o/e LHR, liver herniation and timing of balloon placement. The covariates were the intervention, which estimates the effect of FETO with late placement (30+0–31+6 weeks), early placement (27+0–29+6 weeks), which estimates its additional effect, o/e LHR with a non-linear term, liver herniation and the trial of origin; the interaction between intervention and o/e LHR was tested with a threshold of 0.157, corresponding to the Akaike criterion [61]. Survival to discharge from the neonatal unit was 54% (79/145) with FETO and 39% (55/142) with expectant management; by trial, 63% vs 50% in the moderate form and, with the additional participants, 36% (17/47) vs 14% (6/44) in the severe form. The adjusted odds ratio (aOR) for FETO with late placement was 1.78 (95% CI 1.05–3.01; p = 0.031); the additional effect of early placement was very uncertain (aOR 1.53; 0.60–3.91; p = 0.37), and combining the two terms gave an aOR of 2.73 (1.15–6.49) for FETO with early placement. A higher o/e LHR was associated with higher survival and liver herniation with lower survival (aOR 0.50; 0.27–0.93), but the interaction between intervention and o/e LHR was negligible (p = 0.50), so there was no evidence that the benefit of FETO varied with the severity of hypoplasia. Survival at six months had almost identical results (aOR 1.72; 1.02–2.89 with late placement and 2.69; 1.14–6.35 with early placement). Survival at six months without oxygen was 43% (63/145) vs 32% (46/142) overall, 54% vs 44% in the moderate form and 21% (10/47) vs 7% (3/44) in the severe form, with estimates in the same direction but more uncertain (aOR 1.63; 0.96–2.76 with late placement and 2.21; 0.83–5.86 with early placement), and liver herniation had its most marked unfavourable effect here (aOR 0.31; 0.17–0.59) [61]. The cost is prematurity. Median gestational age at live birth, in the 279 live births, was 35.6 weeks with FETO and 38.3 weeks with expectant management (34.6 vs 38.4 weeks in the severe form; 35.9 vs 38.1 in the moderate form); in the adjusted model, birth occurred on average 1.7 weeks earlier (95% CI 1.1–2.3) with late placement and 3.2 weeks earlier (2.3–4.1) with early placement, unrelated to o/e LHR or liver position. In non-prespecified analyses, about half of the effect of FETO on gestational age (52%; 95% CI 31–84%) was mediated by prelabour rupture of membranes before 37 weeks, and, among the 138 pregnant women in whom FETO was performed or attempted, each week of delay in balloon placement postponed birth by an average of 4 days (1–7), although timing of placement explained only 11% of the variability in gestational age at birth [61]. Across the trials, o/e LHR increased on average by 67% over 34 days of occlusion started at 27–29 weeks and by 32% over 24 days of occlusion started at 30–31 weeks, which is consistent with animal models and with MRI volumetry, in which gestational age at occlusion was the only independent predictor of the lung volume achieved [61]. The authors conclude that FETO increases survival in moderate and severe forms and that the discrepancy between the two trials is more likely due to the later balloon placement in the moderate form than to a dependence of the effect on severity. They acknowledge, however, that the effect of timing of placement could not be estimated robustly, because the sample is small and timing coincides by design with severity (there was no early placement in the moderate form and no late placement in the severe form), that the severe-form trial, stopped for benefit, may overestimate the effect, and that recruitment by fetal centre and by neonatal centre was unbalanced between the trials [61]. They propose studying placement at 27–29 weeks in moderate forms as well, accepting the increase in prematurity and recording secondary outcomes that capture its consequences, preferably in a randomised trial or, as a second option, in prospective registries of consecutive cases that include centre experience and volume and right-sided forms; they advise against prolonging occlusion until 37 weeks, because many pregnant women go into labour before the scheduled removal and emergency removal is more complex and risky than elective removal [61]. The reanalysis presents no data on neonatal morbidity beyond oxygen dependence at six months, nor longer-term outcomes. On the basis of these data, the 2023 Canadian consensus recommends that FETO be considered and discussed with the parents in all cases of severe CDH and accepts it as an option in moderately severe forms, in a shared decision that weighs maternal risks, prolonged relocation to one of the few centres with fetoscopic experience, disruption of the family support structure and loss of income, and notes the lack of studies on the effect of prematurity on long-term morbidity after FETO [48]. CDH EURO, in 2015, before the results, recommended not performing FETO outside a trial because of the risk of prematurity: in the largest cohort then published, with 17.1% of patients born before 32 weeks, survival was 49.4% in isolated left-sided CDH and 37.9% in right-sided CDH [50]. Prenatal transplacental sildenafil partially reversed changes in the vascular bed and airway morphometry in animal models, and studies of transplacental transfer and safety in humans are under way that may precede a randomised trial [48]. The results apply to isolated left-sided CDH; selection for FETO generally excludes chromosomal anomalies and syndromic forms, and the Fryns syndrome GeneReviews reports that a controlled trial of occlusion by open hysterotomy did not improve survival [1, 7]. The PNDS, from 2020, described the criteria of the TOTAL trials then ongoing: moderate left-sided forms with o/e LHR of 25–34.9% with or without herniated liver or of 35–44.9% with herniated liver, with balloon at 30+0–31+6 weeks; severe forms with o/e LHR below 25%, with balloon at 27+0–29+6 weeks; removal at 34–34+6 weeks in both [3]. At a London centre, survival of newborns with severe CDH treated with FETO rose from 44% to 71% between two eras, with later repair after echocardiography-guided stabilisation, staged abdominal closure in 77.8% of recent repairs and a fall in recurrences from 33% to 0% [62]. The AAP states that long-term data after FETO are limited, that growth failure appears more frequent than pulmonary complications and that persistent tracheomegaly has no demonstrated clinical consequences [22].

Availability of the device in Europe is a current regulatory problem: the entry into force of the 2017 European Medical Device Regulation required recertification of the Goldbal balloon and the Baltacci catheter, and the manufacturer stopped distributing them in Europe; in a survey of 50 centres (response rate 95% among 39 users), all European balloons were expired, centres obtain them only with a derogation from national authorities, and no alternative marketed device was expected before 2026 [63]. An alternative balloon with magnetic release (Smart-TO) was studied in a trial completed in 2024 with 23 participants (NCT04931212), with no published results in the sources consulted.

Delivery room resuscitation and stabilisation

At birth, the newborn is intubated with a 3.5 mm tube (or a 3 mm cuffed tube) and ventilated with 100% oxygen; bag-mask ventilation is contraindicated; the ventilator is preset with a peak pressure of 18–24 cmH2O, PEEP of 2–5 cmH2O and 40–60 breaths per minute; a gastric tube is placed to decompress the stomach; heart and respiratory rate, pre-ductal and post-ductal saturation (right hand and foot), transcutaneous PaCO2 and blood pressure are monitored; and a peripheral or umbilical venous line is placed for analgesia and sedation [3]. Routine neuromuscular blockade is not recommended, and ventilation is adjusted for a pre-ductal SpO2 of 85–95% and a PaCO2 of 45–60 mmHg [3]. The Canadian consensus recommends a T-piece or a ventilator so as not to exceed an inspiratory pressure of 25 cmH2O and sedation of all ventilated newborns, with deep sedation and neuromuscular blockade reserved for cases with the greatest requirements [49]. CDH EURO recommends routine intubation without bag-mask ventilation, with confirmation of tube position by expired CO2; premedication before intubation whenever there is venous access, because intubation without premedication raises blood pressure and intracranial pressure and lowers heart rate and saturation; avoidance of neuromuscular blockers in the delivery room, because their administration immediately after intubation worsened lung compliance; the lowest possible peak inspiratory pressure, preferably 25 cmH2O or less; an oro- or nasogastric tube with continuous or intermittent suction; an arterial catheter preferably in the right radial artery (pre-ductal), with the umbilical arterial catheter (post-ductal) as an alternative; and a pre-ductal SpO2 of 80–95%, accepting values down to 70% in the first two hours if they are improving without ventilator changes, with pH >7.2 and PaCO2 <65 mmHg, and reducing FiO2 when SpO2 exceeds 95% [50]. Unlike the 100% oxygen of the PNDS, CDH EURO accepts starting with an FiO2 below 1.0, by extrapolation from data on resuscitation of term newborns with room air [3, 50]. In newborns expected to have good lung development (left-sided defect, o/e LHR >50% and intra-abdominal liver), CDH EURO accepts spontaneous breathing; in a small retrospective audit cited by the 2023 Canadian consensus, this strategy succeeded in 40% of mild CDH with o/e LHR >50%, half of whom required non-invasive ventilation, with a risk of insufflation of the hollow viscera, and survival and duration of postoperative ventilation were the same, so the Canadian recommendation of immediate intubation was not changed [48, 50]. Intact-cord resuscitation may be considered by trained teams within the French national protocol (CHIC study, NCT04429750, recruiting) [3]. The PinC trial, which compares physiology-based cord clamping, after lung aeration, with immediate clamping in newborns with isolated left-sided CDH born at 35 weeks or more, has pulmonary hypertension in the first 24 hours as its primary outcome and a minimum size of 140 newborns; it was completed in January 2026 (NCT04373902), with no published results in the sources consulted [64].

High-frequency oscillatory ventilation (HFOV) is indicated in persistent hypercapnia (PaCO2 >60 mmHg) and/or hypoxaemia (pre-ductal SpO2 <85% with FiO2 of 100%), and the available studies favour conventional ventilation as first line with low PEEP [3]. Exogenous surfactant should not be given routinely [3]. In the CDH EURO VICI trial, initial conventional ventilation and HFOV did not differ in the primary outcome (death or bronchopulmonary dysplasia at day 28), but conventional ventilation was associated with shorter duration of ventilation, less need for inhaled NO or sildenafil, shorter duration of vasoactive drugs and less ECMO; the analysis included 171 of the 356 planned patients [48, 50]. CDH EURO therefore recommends pressure-controlled conventional ventilation with PIP <25 cmH2O, PEEP of 3–5 cmH2O and 40–60 breaths per minute, adjusted for a PaCO2 of 50–70 mmHg (6.9–9.3 kPa), first reducing peak pressure during weaning while pCO2 is below 50 mmHg; HFOV serves as rescue, and the need for a PIP above 28 cmH2O to reach targets should prompt consideration of HFOV or ECMO [50]. After the first two hours, the pre-ductal target is preferably 85–95%, accepting down to 80% with good perfusion (pH >7.2, lactate <5 mmol/L, urine output >1 mL/kg/h) and post-ductal SpO2 above 70% [50]. Two later retrospective studies, one with 25 years of experience at a centre without ECLS and another from the Japanese study group with 250 patients on HFOV and 77 on conventional ventilation, showed no differences in survival, NO use, duration of ventilation or oxygen at discharge, and suggested that high-frequency ventilation was chosen in the sickest patients; high-frequency positive pressure ventilation, neurally adjusted ventilatory assist (NAVA) and heliox-oxygen mixture are promising in single-centre series [48]. CDH EURO sees no rationale for surfactant, because the amount of surfactant appears adequate for the lung size, and asks for a chest radiograph as early as possible [50].

Hypoperfusion (capillary refill >3 s, lactate >3 mmol/L, urine output <1 mL/kg/h) and blood pressure below the norm are treated, according to the Canadian consensus, with very judicious crystalloid administration, generally not exceeding 20 mL/kg, and with inotropes such as dopamine, adrenaline or noradrenaline, with hydrocortisone in hypotension poorly responsive to volume and vasopressors and echocardiographic assessment of cardiac function if hypoperfusion persists [48, 49]. The 2023 update justifies caution with volume by ventricular dysfunction, which contributes to persistent hypotension and worsens with fluid excess; it maintains dopamine, adrenaline and noradrenaline as first line, noting that high doses of adrenaline may cause tachyarrhythmia, hyperglycaemia and lactic acidosis and that noradrenaline increases afterload and perhaps pulmonary arterial resistance; and it describes vasopressin as promising in catecholamine-resistant shock, with no effect on pulmonary pressures, in a retrospective series of 13 patients [48]. Cardiovascular phenotypes vary between patients and evolve in the acute phase, and no trial has compared strategies by phenotype (acute PH, right, left or biventricular dysfunction), so treatment is adjusted to the physiology assessed at the bedside with echocardiography [48]. Acute kidney injury, defined and staged by the modified neonatal KDIGO creatinine criteria, is frequent: in retrospective series, survival of patients with acute kidney injury was 37–47% and decreased with stage, with associations with lower prenatal lung volume, liver herniation, vancomycin, corticosteroids, diuretics, abdominal closure, hypotension and elevated plasma free haemoglobin, and an additional risk of fluid overload on ECLS [48]. CDH EURO recommends maintaining blood pressure normal for gestational age, without seeking supranormal values if pre-ductal SpO2 remains above 80%; in hypotension or poor perfusion, 0.9% NaCl 10–20 mL/kg at most twice (or Ringer's lactate, in the first two hours, if there is hypovolaemia), followed by inotropes or vasopressors (dopamine, dobutamine, noradrenaline or adrenaline, according to local practice) and hydrocortisone after failure; it dispenses with inotropic support if heart rate is normal, urine output above 1 mL/kg/h and lactate below 3 mmol/L without other signs of poor perfusion; it asks for echocardiography if poor perfusion or hypotension is associated with pre-ductal SpO2 below 80%; and it recommends fluid restriction to 40 mL/kg/day, including drugs, in the first 24 hours, diuretics in persistent positive balance without hypovolaemia (urine output target >1 mL/kg/h) and only parenteral nutrition before repair [50].

The 2023 Canadian consensus recommends at least two standardised echocardiograms, one in the first 24–48 hours (or before surgery) and another at two to three weeks, to assess PH and cardiac function, with additional examinations according to the clinical picture, for example before surgery or discharge [48]. The first may be deferred beyond 24 hours to avoid excessive handling during pulmonary vascular adaptation, but should be brought forward in severe cardiorespiratory instability or when ECLS is anticipated; in one study, a care bundle that deferred echocardiography was associated with fewer inotropes, less ECLS, earlier repair and better survival [48]. An echocardiogram at 5–7 days of life may be indicated, especially on ECLS, because the ratio of pulmonary artery acceleration time to ejection time (PAAT/ET) is lower on the initial echocardiogram of patients who go on to need ECLS and, at 5–7 days, in those who do not survive on ECLS [48]. Brain natriuretic peptide (BNP) or NT-proBNP may serve as ancillary markers of cardiac strain, and their rise was associated with death or respiratory support at 56 days, but availability varies and there are no data showing improved outcomes with surveillance based on these markers alone [48]. CDH EURO asks for an echocardiogram in the first 24 hours to exclude structural heart disease, assess right ventricular function and classify PH as below or above two-thirds of systemic pressure [50].

Analgesia, sedation and neuromuscular blockade

The 2023 Canadian consensus created a new area on pain: all ventilated newborns should have individualised analgesia and sedation, titrated by a validated pain or sedation scale; the intravenous opioid (morphine or fentanyl) should be given as a continuous infusion, combined with a short-acting benzodiazepine, which may reduce the opioid dose; routine neuromuscular blockade should be avoided during preoperative stabilisation and considered only if PH worsens or ventilatory targets are difficult to achieve; and intravenous paracetamol should be considered postoperatively to reduce opioids [48]. These recommendations rely on guidelines for ventilated term and near-term newborns, which prefer continuous fentanyl infusion to morphine in hypotension or renal failure and recommend rotating opioids in the face of tolerance; in a neonatal trial, continuous fentanyl infusion had favourable pharmacokinetics and pain scores equivalent to those of intermittent boluses [48]. Paracetamol was used in 48% of repaired patients in the CDH database of the Children's Hospitals Neonatal Consortium, reduced opioids in a Cochrane review and in a trial in non-cardiac surgery, and a quality improvement protocol with intravenous paracetamol, education and standardised pain handover reduced opioids and duration of intubation in CDH [48]. As for neuromuscular blockade, lung compliance decreased after pancuronium in a prospective study of 15 ventilated newborns, prolonged sedation or paralysis was associated with longer admissions and higher mortality in a multicentre registry, and blockade before repair was almost twice as frequent in non-survivors as in survivors (87% vs 48%), which appears to reflect its use in the sickest patients who do not stabilise [48]. CDH EURO recommends sedation throughout mechanical ventilation until the start of weaning, with opioids (morphine or fentanyl), blood pressure monitoring after sedation, especially with midazolam, assessment by validated scales such as the COMFORT behavior scale, comfort measures (containment, swaddling, reduction of noise, light and nociceptive stimuli) and avoidance of neuromuscular blockers, because of the risk of hypoxaemia [50]. The guidelines do not give doses of opioids or benzodiazepines; for intravenous paracetamol, the US label indicates, in newborns born at 32 weeks or more up to 28 days of age, 12.5 mg/kg every 6 hours, maximum 50 mg/kg/day, and, from 29 days to 2 years, 15 mg/kg every 6 hours, maximum 60 mg/kg/day, as a 15-minute infusion and counting all routes of administration in the daily maximum [65].

Acute pulmonary hypertension

The priority is not to normalise post-ductal SpO2, but to optimise circulatory status, with regular clinical and echocardiographic monitoring of the shunts across the foramen ovale and ductus arteriosus, tricuspid regurgitation, pulmonary pressures and right and left ventricular function, adapting treatment to the mechanism; aggravating factors (stress, pain, lung overdistension) should be limited, and hyperventilation is not recommended [3].

Inhaled NO is not recommended as first line by the PNDS and should be reserved for severe persistent PAH after failure of conventional ventilation and haemodynamic correction, because its routine use is suspected of increasing mortality [3]. In a CDH Study Group cohort of 1777 infants treated between 2015 and 2020, the 863 (48.6%) who received NO in the first three days, before any ECMO, had, after adjustment for severity, higher mortality (adjusted OR 2.06; 95% CI 1.05–4.03) and greater use of ECMO (adjusted OR 3.44; 95% CI 2.11–5.60), with no echocardiographic subgroup showing benefit [66]. In 48 US children's hospitals (2016–2023, 2565 newborns), NO use fell from 51.4% to 42.9%, varied between 24% and 100% by hospital, and hospitals with high use had higher mortality (19.9% vs 12.4%) with no difference in ECMO [67]. The US label for INOmax® expressly states that clinical trials did not demonstrate a reduction in the need for ECMO with inhaled NO in newborns with CDH [68]. A phase 4 de-implementation trial of inhaled NO in CDH is recruiting, with 900 planned participants (NCT05213676). These data are observational and subject to confounding by indication, but converge with the position of the PNDS. CDH EURO proposes inhaled NO as first choice when pre-ductal SpO2 falls below 85% or there are signs of poor perfusion, at a dose of 10–20 ppm for at least one hour, if there is an extrapulmonary right-to-left shunt with an oxygenation index above 20 and/or a pre- and post-ductal saturation difference greater than 10%, and requires it to be stopped in non-responders; response is defined by a 10–20% fall in the saturation difference, a 10–20% rise in PaO2 or haemodynamic improvement (10% rise in mean arterial pressure or fall in lactate) [50]. The European consensus notes that no dose-dependent effect has been demonstrated and that, in one study, those treated needed ECMO more often, and finds no evidence in favour of increasing systemic vascular resistance to treat the right-to-left shunt [50]. The 2023 Canadian update maintained NO only with demonstrable echocardiographic and clinical improvement, to be stopped if this is lacking [48].

Continuous infusion of prostaglandin E1 (alprostadil) is recommended by the PNDS, off-label, when circulatory failure is associated with a closing ductus arteriosus and suprasystemic pulmonary hypertension, in an approach similar to that of functional hypoplasia of the left heart [3]. The 2023 Canadian consensus broadened the indications, based on two retrospective studies that showed improvement in pre- and post-ductal saturations, ductal flow, FiO2, echocardiographic indices and BNP (57 patients in one of them): PGE1 should be used if pulmonary or systemic flow depends on the ductus arteriosus or there is an anatomical cardiac lesion; it may be considered with suprasystemic right ventricular pressures, right ventricular failure or a predominantly right-to-left ductal shunt; and it should be considered to maintain ductal patency in left ventricular dysfunction or functional aortic atresia with systemic right ventricular or pulmonary pressures [48]. CDH EURO recommends considering it intravenously in suprasystemic pulmonary arterial pressure with right-to-left shunt across the foramen ovale, to protect the right ventricle from overload [50]. Milrinone is not routinely recommended by the PNDS, despite the recognised role of left ventricular dysfunction; the Canadian consensus recommends it in cardiac dysfunction, especially associated with pulmonary hypertension, for its lusitropic and pulmonary vasodilator properties, by extrapolation from infants with heart disease and PH, with caution in hypotension and bearing in mind renal elimination; CDH EURO cites only a series of six patients with improved right ventricular function and oxygenation index [3, 48–50]. A phase 2 trial of milrinone in CDH (NCT02951130, 66 participants) was completed in May 2025, with no published results in the sources consulted, and a phase 2 trial of a phosphodiesterase 5 inhibitor in the newborn (TOP-CDH, NCT05201144) is recruiting. In the absence of response to NO, CDH EURO proposes intravenous prostacyclin, intravenous sildenafil or an endothelin pathway drug and recommends considering intravenous sildenafil in severe PH, noting that the intravenous form was not then FDA-approved for this use; the Canadian consensus considers sildenafil in PH refractory to NO or as an adjunct during NO weaning and added in 2023 prostacyclin (treprostinil, epoprostenol) and vasopressin as rescue therapy before, during or after ECLS in severe refractory PH, without delaying ECLS in patients who already meet the criteria [48–50].

Extracorporeal membrane oxygenation

Venoarterial ECMO provides oxygenation and CO2 removal, limits barotrauma and volutrauma and reduces preload and right ventricular failure; it is indicated when hypoxaemia persists despite optimal medical treatment, including haemodynamic optimisation, reopening of the ductus arteriosus and pulmonary vasodilators [3]. The usual PNDS criteria are pre-ductal SpO2 below 85% with inspiratory pressure above 28 cmH2O (or mean pressure above 16 cmH2O on HFOV) and signs of circulatory failure, an oxygenation index above 40 for more than six hours, or intercurrent refractory septic shock, with gestational age above 34 weeks and weight above 2 kg for technical reasons; for some teams, persistent hypercapnia or hypoxaemia despite optimal ventilation reflects severe pulmonary hypoplasia and contraindicates ECMO [3]. Before starting ECMO, coexisting anomalies (especially lethal ones), gestational age, weight, uncontrollable bleeding, intracranial haemorrhage and the prospects of medical treatment are considered [1]. Complications include air embolism, intracranial haemorrhage, cerebral infarction, seizures, bleeding at the cannulation site, disseminated intravascular coagulation, left-to-right shunt through the ductus arteriosus, renal failure, systemic hypertension and infection [1]. The Canadian consensus recommends that the possibility of ECMO be discussed during prenatal counselling, explaining that the available evidence does not suggest a survival benefit [49]. CDH EURO, which considers the benefit of ECMO uncertain (survival of 51% in the 2014 ELSO registry and no difference in survival between those born at centres with and without ECMO in the VICI trial), defines as criteria the inability to maintain pre-ductal SpO2 >85% or post-ductal SpO2 >70%; rising PaCO2 with respiratory acidosis of pH <7.15 despite optimised ventilation; the need for PIP >28 cmH2O or mean pressure >17 cmH2O to achieve SpO2 >85%; inadequate oxygen delivery with metabolic acidosis (lactate ≥5 mmol/L and pH <7.15); systemic hypotension resistant to fluids and inotropes with urine output <0.5 mL/kg/h for at least 12–24 hours; and an oxygenation index ≥40 for at least three hours [50]. The oxygenation index threshold is therefore similar in the two European guidelines, but the PNDS requires more than six hours and CDH EURO at least three [3, 50]. The 2023 Canadian update endorses the ELSO indications (hypoxaemic or hypercapnic respiratory failure, circulatory failure or acute clinical deterioration) and accepts considering ECLS, in special circumstances, in patients with the traditional contraindications of size, age or comorbidity: in the most recent ELSO registry, overall survival was 50% (n = 7564), 44% below 34 weeks and 29% below 2 kg, and in a systematic review survival of preterm infants with CDH treated with ECLS was similar to that of preterm infants without ECLS; given the high risk of death and neurological injury, such use should remain experimental and limited to high-volume ECLS centres [48]. A large retrospective cohort showed higher overall mortality with ECLS, with an advantage only in a high-risk subgroup at high-volume centres; two studies on costs and social implications argue against runs of indefinite duration; and two series, with 31 patients in total, support a second ECLS run with the same cannulation criteria, without it being known whether this aggravates the poorer neurodevelopment already associated with ECLS [48].

Surgery

The consensus is to operate as soon as haemodynamic and respiratory stabilisation is achieved, defined in the PNDS by mean arterial pressure normal for age, pre-ductal SpO2 of 85–95%, urine output above 1 mL/kg/h, lactate below 3 mmol/L and non-suprasystemic pulmonary pressures, with a case-by-case decision if these criteria are not met by 48 hours of life or if there is renewed deterioration [3]. The Canadian consensus adds FiO2 below 0.5 and proposes considering repair or a palliative approach if the criteria are not met within two weeks [49]. The 2023 update adds that there is no benefit in delaying repair once physiological stability has been achieved: in a retrospective study of 158 newborns, an oxygenation index <9.4 was associated with survival, and any delay in repair after reaching this value increased ventilator days and delayed discharge; in a prospective Chinese study of 30 patients with mild to moderate CDH (LHR >1), delaying thoracoscopy beyond 48 hours brought no benefit; and centres with fewer unrepaired patients had higher survival, which suggests that higher-risk patients, when repaired, may survive [48]. CDH EURO defines preoperative stabilisation by mean arterial pressure normal for gestational age, pre-ductal SpO2 of 85–95% with FiO2 below 0.5, lactate below 3 mmol/L and urine output above 1 mL/kg/h, recommends elective surgery without a routine chest drain (except for symptomatic effusion, such as chylothorax) and asks for the defect size to be recorded in all operations so that series can be compared [50]. Repair consists of reducing the herniated organs and closing the defect, directly or with a patch [3]. Laparotomy is the most widely used approach, considered the gold standard, and allows all forms to be repaired; thoracotomy does not allow control of intra-abdominal arrangement; thoracoscopy reduces adhesive obstruction, pain and cosmetic impact, but has a higher risk of recurrence, acidosis and hypercapnia and lacks control of bowel arrangement, being beneficial in forms with a favourable prognosis; laparoscopy is not indicated in the newborn [3]. The 2023 Canadian consensus accepts minimally invasive repair in patients who easily reach preoperative ventilatory targets, with subsystemic pulmonary pressures and normal cardiac function, if the surgical team has technical proficiency and the anaesthetic team can continuously monitor and correct intraoperative hypercapnia and acidosis; in five recent cohorts (137 patients), recurrence after thoracoscopy was 7–21%, a preoperative oxygenation index above 3 predicted failure (conversion or severe postoperative complication) in a multicentre study of 37 infants, and low-quality evidence suggests that a biological reinforcing mesh reduces recurrence and adhesive obstruction [48]. CDH EURO accepts minimal-access surgery in left-sided CDH with an intra-abdominal liver, with a higher risk of recurrence and longer operating times, but survival and patch use similar to those of open surgery in a meta-analysis, and asks for minimal CO2 insufflation pressures because of CO2 absorption [50]. Large defects are repaired mainly with polytetrafluoroethylene patches (Gore-Tex®, Gore DualMesh®), whose recurrence reaches 40% in some series; the tension-free, oversized cone- or dome-shaped technique reduces recurrences to about 5% and abdominal hypertension, and absorbable or biological patches should not be used; muscle flaps (internal oblique and transversus, or transversus) may be used as first line, depending on the team's experience, or after patch complications [3]. The Canadian consensus maintains oversized, tension-free polytetrafluoroethylene patches for defects that do not allow direct closure and accepts since 2023 the oblique muscle flap when technical expertise exists: in two studies with 97 newborns with large defects, recurrence at five years was 3% and 3.5%, repair on ECLS was similar to that of the patch group (39% vs 31%), as were complications, including bleeding on ECLS; another 50 flap repairs in earlier series had three recurrences (6%), and scoliosis and chest deformity were equivalent with patch and flap [48]. A routine chest drain is not recommended and, if placed, should not be put on suction (or only between −4 and −6 cmH2O), because of the risk of mediastinal shift and death [3]. Pulmonary sequestration and malrotation or defective fixation of the colon should be looked for, possibly with a Ladd procedure, and, if abdominal closure causes raised abdominal pressure, a patch or silo with delayed closure is used, monitoring intravesical pressure [3, 21]. In WT1 disorder with an indication for peritoneal dialysis, a small diaphragmatic hernia should be looked for and repaired before dialysis is started [20]. In Donnai-Barrow syndrome, repair of CDH and omphalocele does not appear to carry more risk than in other syndromes [16].

Repair on ECLS is the divergence with the most data. Survival to discharge of patients with CDH who need ECLS is about 50%, reaching nearly 70% in some centres; complications of repair on ECLS are metabolic, circuit-related or haemorrhagic (including the surgical site), the latter in 25% of cases, and about 15% of patients on ECLS are never repaired [48]. In a propensity-matched CDH Study Group study, repair on ECLS was associated with higher survival (HR 0.54; 95% CI 0.38–0.77), with over-representation of high-volume centres, but the benefit was reversed when unrepaired patients were excluded; in an ELSO registry study with more than 2200 matched patients, without unrepaired patients, repair on ECLS tripled mortality and increased, with run duration, the risk of severe neurological injury; at a centre in Ann Arbor, the highest survival (94%) was in those repaired after decannulation [48]. As for timing, two CDH Study Group studies and a single-centre study of 33 patients favoured early repair (less than 72 hours, or less than 24 hours vs 24–72 hours), but another single-centre study showed that an early repair protocol (≤5 days) predicted mortality (HR 3.48; 95% CI 1.28–9.45); at two years, those repaired on ECLS had lower cognitive and motor scores than those repaired after ECLS [48]. The relationship between survival and timing of repair depends on whether the mortality of unrepaired patients is excluded or attributed to the group repaired after ECLS; the Canadian consensus therefore maintains that surgery should wait for decannulation, but accepts early repair in patients with a low probability of survival by prenatal predictors or by severity at cannulation, who risk not being weaned [48]. CDH EURO accepts repair on ECMO, citing one study with higher survival, less bleeding and shorter ECMO with repair after ECMO and another, with 46 patients, in which repair in the first 72 hours of ECMO shortened the run, reduced circuit complications and showed a trend towards higher survival [50].

Postoperative period

After repair, the newborn remains in intensive care for early weaning from invasive ventilation, possibly with non-invasive ventilation and nasal oxygen; echocardiographic assessment of PAH with treatment if needed; early enteral feeding, between 24 and 48 hours, ideally with breast milk; optimised parenteral nutrition while required; and surveillance for abdominal compartment syndrome, persistent pleural effusion, chylothorax, obstruction and recurrence [3]. Nutritional intake should be optimised from the first day, with protein at 2 g/kg on day 0 and 3 g/kg in the first days, intravenous lipids from day 1 up to 2–3 g/kg/day and glucose adapted to tolerance up to 14–16 g/kg/day within a few days [3]. The AAP states that an intake of at least 120 kcal/kg/day is needed to maintain adequate weight gain, that breast milk and formulas are often insufficient without fortification and that a transpyloric tube speeds the achievement of full feeds, although early tube dependence may worsen oral aversion [22]. Gastrostomy is discussed in severe and persistent oral feeding disorders, and parenteral nutrition in persistent undernutrition despite maximal enrichment; later removal of the gastrostomy does not compromise linear growth [3, 22]. Breastfeeding should be explained during pregnancy without being presented as mandatory, with support from a lactation consultant, manual expression of colostrum in the first six hours, use of colostrum for oral care, a double electric pump and alternative methods such as the Morton technique, straw and cup [3]. Casein hydrolysate is not used routinely, but cow's milk protein allergy should be considered in the presence of compatible symptoms [3].

GOR is treated with positional measures, thickened feeds and proton pump inhibitors (PPIs), omeprazole and esomeprazole, off-label before one year of age (indicated above one year and 10 kg), with antireflux surgery if medical treatment fails [3]. The 2023 Canadian consensus does not recommend preventive fundoplication at repair: reflux disease persists in more than 60% of infants beyond one year of age on impedance monitoring, but, in a French prospective multicentre study, preventive fundoplication (27 patients, 11%) in high-risk CDH repaired with a patch had more redo fundoplications than later fundoplications for refractory GOR in the group without fundoplication, longer admissions, more oral aversion and more tube feeding beyond six months; fundoplication is reserved for failure of medical treatment [48]. CDH EURO recommends starting preventive antireflux therapy with enteral feeding and cites a study with no benefit beyond the first year after a prophylactic Thal procedure at primary repair [50]. The AAP, citing the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN), does not recommend empirical therapy, proposes a 4- to 8-week PPI trial in evident or suspected GOR and notes that there is no consistent recommendation on duration, given the immunological risks of prolonged use; prophylactic neonatal fundoplication in high-risk CDH was associated with more nutritional failure and oral aversion, without reducing redo fundoplications [22]. Respiratory complications are treated with continuous oxygen as needed, bronchodilators and inhaled or oral corticosteroids as needed, maintenance inhaled corticosteroids on specialist indication and antibiotics for superinfections [3].

Residual or chronic pulmonary hypertension

Prolonged PAH is associated with excess mortality, and a benefit of active treatment has been described, but, in the absence of comparative studies, the PNDS does not recommend a specific class and proposes, based on published cases and expert opinion, the following order: phosphodiesterase 5 inhibitor (sildenafil), prostacyclin analogues and endothelin receptor antagonist (bosentan), with inhaled NO in intensive care [3]. The French medicines agency emphasised in 2012 that none of these drugs had an indication for CDH-associated PAH in children [3].

The alprostadil dose is that of the label for duct-dependent congenital heart disease, a declared extrapolation to CDH, which the PNDS treats as an off-label indication [3, 71]. The sources consulted do not include the Portuguese SmPC or the Brazilian package insert for these drugs. The milrinone dose in the table is that of the US adult label, a declared extrapolation, because none of the guidelines consulted gives a dose for CDH and the label does not establish paediatric use [72]. Off-label prescribing in France requires, according to the PNDS, the absence of a suitable alternative, informing the patient of the off-label nature and the risks, a mention on the prescription and justification in the medical record [3].

During PH treatment, the AAP recommends echocardiography every two to four weeks after initiation, then every four to six months once stabilised, and only in the event of a clinical change or new radiographic finding after normalisation of pressures and discontinuation of drugs [22]. The PNDS proposes stopping routine checks after two normal echocardiograms one month apart without treatment, at least one performed by a paediatric cardiologist, and repeating them before surgery or an examination with ventilation, in the event of respiratory decompensation, while there is oxygen dependence and at the transition to adulthood [3].

Prevention of respiratory infections and vaccination

The PNDS recommends preventing nosocomial infections (early extubation, isolation), vaccinating parents, siblings, grandparents and carers against influenza and pertussis, vaccinating the child against pneumococcus and against influenza from six months, arranging suitable childcare avoiding nurseries, avoiding high-risk environments, combating passive smoking and avoiding allergens in atopic families [3]. In 2020, the PNDS indicated RSV immunoprophylaxis with palivizumab in patients with severe respiratory involvement or PAH in the first and possibly the second winter [3]. The AAP, in 2026, considers CDH survivors at high risk of severe RSV (RSV hospitalisation up to 23.8%) and recommends prophylaxis in all, high and low risk, with nirsevimab before or shortly after discharge (five months of protection) and a second dose before the second season, between 8 and 19 months, even if the mother was vaccinated, because maternal antibodies are lost with ECMO [22]. The difference reflects the evolution of RSV prophylaxis between 2020 and 2025 and not a contradiction about the risk. In the EU, nirsevimab (Beyfortus®) has been authorised since 31 October 2022 for newborns and infants in their first RSV season, as a single intramuscular dose of 50 mg below 5 kg and 100 mg from 5 kg, and since 1 August 2024 for children up to 24 months who remain vulnerable to severe disease in their second season, with 200 mg (2 × 100 mg); after cardiac surgery with cardiopulmonary bypass, an additional dose is given [73]. Clesrovimab (Enflonsia®), as a fixed intramuscular dose of 105 mg, was approved by the FDA on 9 June 2025 and authorised in the EU on 15 April 2026, for the first season only [74, 75]. Palivizumab (Synagis®), 15 mg/kg intramuscularly monthly during the season, remains indicated in children born at 35 weeks or less and aged under 6 months at the start of the season and in children under 2 years with bronchopulmonary dysplasia treated in the previous 6 months or with haemodynamically significant congenital heart disease; CDH without these conditions is not among its indications [76]. Patients with lung disease who need oxygen or medication are at risk of severe hypoxaemia during air travel, and the hypoxic challenge test identifies those who need in-flight oxygen [22].

Treatment of the syndromic entities

In Fryns syndrome there are no published clinical guidelines; after diagnosis, GeneReviews recommends chest and abdominal radiographs, ENT and craniofacial team assessment, echocardiography, renal ultrasound, transfontanellar ultrasound and brain MRI, EEG if seizures are suspected, upper gastrointestinal contrast study to exclude malrotation, ophthalmological examination with fundoscopy, developmental assessment, genetic counselling and social support [7]. Epilepsy is treated with antiseizure medication according to standard practice, because none has shown specific efficacy, and the myoclonus of one survivor was controlled with valproate [7, 9]. Clefts, heart defects, genitourinary and gastrointestinal malformations and cataracts are treated by the respective specialists, and the need for palliative care and home nursing should be assessed on an ongoing basis [7]. In Donnai-Barrow syndrome, treatment includes corrective lenses for myopia, prophylactic peripheral laser photocoagulation to reduce the risk of retinal detachment, hearing aids or cochlear implant, standard antiseizure medication and vitamin A and D supplementation with serum level monitoring (25-hydroxyvitamin D) [16]. In MYRF-CUGS, treatment follows standard practice for differences of sex development (hormone therapy, psychosocial support, assessment of gender identity, orchidopexy and hypospadias repair), for the thyroid if there is hypothyroidism, for refractive errors and nanophthalmos, for heart defects and for diaphragmatic defects; prolonged oral androgens (methyltestosterone, fluoxymesterone) should be avoided because of hepatic toxicity, as should hormone therapy in hormone-dependent cancers [19]. In WT1 disorder, steroid-resistant nephrotic syndrome should not be treated with immunosuppressants, and prophylactic gonadectomy prevents gonadoblastoma in disorders of testicular development [20].

Investigational therapies and developments in the last 12 months

There is no gene therapy or cause-directed therapy for any of the genetic forms of CDH, and the sources consulted do not describe gene therapy clinical trials in this malformation [7, 35]. Research is focused on fetal and neonatal physiology. Expression of vascular endothelial growth factor (VEGF) is reduced in the distal lung epithelium of human fetuses with CDH, and exogenous VEGF administration is described as promising in preclinical studies, without use in patients with Fryns syndrome or other genetic forms [7]. The level of evidence of the developments identified is as follows: the AAP clinical report on follow-up after discharge, published in February 2026, is a society recommendation [22]; the improved survival after FETO at a London centre is a retrospective observational study [62]; the shortage of FETO balloons in Europe stems from a regulatory rule and not from a device safety problem [63]; the PinC trial was completed in January 2026 without published results; the inhaled NO de-implementation trial began recruiting in November 2025; and the GeneReviews general review chapter was retired in July 2025, whereas the Fryns syndrome chapter was updated in December 2025 [1, 7]. In genetics, the 2025–2026 cohorts confirmed fetal exome yields of between 2% and 12% after normal CMA, higher in complex forms, and identified a PLS3 variant in a family with X-linked recurrence [15, 52, 53]. The absence of other trials in the sources consulted does not mean that none exist.

Agents and circumstances to avoid

Bag-mask ventilation in the delivery room, hyperventilation, routine neuromuscular blockade, active suction of a chest drain, routine surfactant, absorbable or biological patches and routine use of inhaled NO should be avoided, according to the PNDS [3]. Prolonged exposure to aminoglycosides is the iatrogenic factor most associated with sensorineural deafness [22]. Prolonged PPIs without a clear indication carry immunological and developmental risks [22]. Passive smoking should be avoided [3]. The PNDS and French emergency documents recommend respecting drug contraindications and carrying an emergency card, available in French and English [3].

Pregnancy in a survivor, perioperative care and transition

The AAP draws attention to the risk of hernia recurrence during pregnancy in a woman who survived CDH, which justifies counselling and surveillance [22]. Any surgery or examination with ventilation requires prior echocardiographic assessment of PAH in patients with a history of PH [3]. Transition to adult medicine should be prepared, with an adult provider familiar with the condition and continued access to a centre of excellence; the PNDS recommends follow-up throughout childhood and probably into adulthood, acknowledging the lack of such long-term data [3, 22]. Endoscopic surveillance of the oesophagus is advocated in high-risk adults because of the prevalence of oesophagitis and Barrett oesophagus [22].

Regulatory status

The sources consulted do not describe the marketing or reimbursement of these drugs for this indication in Portugal and Brazil.

8. Reproductive Options and Family Counselling

CDH with an identified cause

When a syndrome, a monogenic disorder or a chromosomal anomaly is identified, the recurrence risk is that of the entity and mechanism [1]. In AR disorders, such as PIGN-related Fryns syndrome, Donnai-Barrow syndrome and STRA6-related Matthew-Wood syndrome, each sibling has a 25% risk of being affected, 50% of being an asymptomatic carrier and 25% of being neither affected nor a carrier; the parents should be tested to confirm heterozygosity, and, if only one is a carrier, a de novo variant in the proband or parental mosaicism should be considered [7, 16]. The proband's aunts and uncles have a 50% probability of being carriers, and testing of at-risk relatives is possible once the molecular diagnosis has been established [7]. In Donnai-Barrow syndrome, the presence of CDH or omphalocele in one sibling does not predict the presence of either defect in another affected sibling [16]. Patients with Fryns syndrome are not known to have had offspring [7]. If the partner of a PIGN variant carrier has ancestry from La Réunion or other Indian Ocean islands, screening for the founder intragenic deletion may be offered [7]. In MYRF-CUGS, with AD inheritance, many cases are de novo, each child of an affected person has a 50% probability of inheriting the variant, and manifestations vary greatly within the same family, with offspring sometimes more or less severely affected than the proband [19]. In ZFPM2, the estimated penetrance of 37.5% in one family means that a healthy carrier parent is to be expected and that the risk of CDH recurrence in a carrier's child is below 50% [12, 35]. In GATA4, clinically asymptomatic carriers may have small diaphragmatic defects visible only on MRI [11]. In de novo deletions (8p23.1, 15q26, 1q41-q42), the recurrence risk is low, without excluding germline mosaicism, and in aneuploidies it follows their general rule [1]. In Pallister-Killian syndrome, the cause is usually a sporadic chromosomal aberration, which distinguishes counselling from that for Fryns syndrome [9].

CDH without an identified cause

Most patients with isolated CDH are the only case in the family [1]. The sibling recurrence risk for an isolated Bochdalek CDH is estimated at less than 2% in GeneReviews and at about 1% in the Oxford Desk Reference; in the series of Pober and colleagues, only one of 203 probands had a previously affected sibling (0.9%), but 4% of siblings had another major malformation, and the Finnish review by Norio and colleagues estimated about 2% [1, 14, 21]. These estimates are based mainly on previous occurrence in siblings and not on consecutive recurrence series [1]. In complex CDH without a diagnosis, the average risk is low, but it results from averaging a negligible risk in most families and a risk that may reach 25–50% in a minority (unrecognised AR or AD disorders), so counselling resembles that for multiple congenital anomalies of unknown cause [1]. In multiplex families, concordance of defect type and side is very high, although size varies, and one sibling may have eventration and another a hernia, or one a unilateral and another a bilateral hernia [1]. There are no systematic studies of the risk in the children of survivors; theoretically it may reach 50% if the survivor has a fully penetrant de novo dominant variant [1].

Prenatal and preimplantation diagnosis and DNA banking

When the familial variant is known, prenatal diagnosis and preimplantation genetic testing are possible, and the ideal time to assess the risk and discuss these options is before pregnancy [1, 7]. In AD forms, counselling should include reduced penetrance [1]. When a molecular diagnosis has not been obtained, DNA banking from the proband should be considered, and autopsy, photographs, radiographs and samples for genetic study and fibroblast culture from deceased fetuses or newborns are essential for future counselling [1, 7]. In a subsequent pregnancy of a couple with a child with possible Fryns syndrome without a molecular diagnosis, fetal karyotype and CMA and a detailed ultrasound with echocardiography are indicated, bearing in mind that a normal examination does not exclude the disease [7]. Pregnancies at high risk because of family history should be followed at a high-risk obstetric centre [1].

Pre-test counselling and information for parents

Prenatal pre-test counselling should explain the possible genetic causes, the testing modalities and the uncertainty that may persist after a negative result or a VUS [4]. Identification of a likely deleterious variant worsens the prognosis for mortality, pulmonary hypertension and neurodevelopment, even in forms that appear isolated, and may change eligibility for fetal therapy [4, 45]. In isolated left-sided CDH eligible for FETO, information for parents can draw on the pooled data from the TOTAL trials: survival to discharge of 54% with FETO vs 39% with expectant management and survival at six months without oxygen of 43% vs 32%, at the cost of birth on average 1.7 to 3.2 weeks earlier, depending on the timing of balloon placement, mainly from prelabour rupture of membranes [61]; the 2023 Canadian consensus frames this decision as shared, weighing maternal risks, prolonged relocation to a fetal centre and the family and financial consequences [48]. The PNDS recommends that parents receive the most complete information possible about the child's future from the paediatric medical-surgical team, be informed about patient associations and breastfeeding and receive psychological support from the prenatal period to adolescence [3]. In MYRF-CUGS, GeneReviews considers it appropriate to clarify the status of apparently asymptomatic younger and older relatives in order to start surveillance and treatment [19].

9. Special Considerations: Population Diversity and Epidemiology

Prevalence

The prevalence of CDH is 3–3.6 per 10,000 live births according to five studies cited in GeneReviews, and OMIM reports 1 in 2500 to 1 in 4000 newborns [1, 14]. Orphanet gives a birth prevalence of 30/100,000 worldwide, 21.2/100,000 in Europe (EUROCAT network, 2010) and 33/100,000 in the USA [2]. EUROCAT national figures, which measure birth prevalence in the regions covered by the registries and include different methods of counting terminations, range from 5.9/100,000 in Germany to 49.8/100,000 in Malta, with 18.9/100,000 in Portugal, 9.4 in Spain, 23.6 in France, 22.6 in Italy, 29.2 in the Netherlands, 19.7 in the United Kingdom, 38.9 in Switzerland and 39.3 in Austria [2]. In France, the PNDS estimates 1 in 3500 live births, about 350 new cases per year [3]. A five-year US state study estimated 1 in 3700 live births [21]. The sources consulted provide no data from Brazil. Differences between registries reflect coverage, inclusion of terminations and fetal deaths and the era, and cannot be read as true population differences.

Population mortality and selection bias

Series from reference centres, which mainly include isolated left-sided CDH, report survival of 80–90%, whereas population-based studies including all prenatal diagnoses report mortality of at least 50% when terminations are included [1]. In the meta-analysis of Stege and colleagues, 31% of live-born infants with a prenatal diagnosis died, most in the first 24 hours, in addition to terminations, miscarriages (3%) and fetal deaths (3%), which requires counselling to specify the population to which each figure refers [1, 59].

Genetic entities: frequency and founder populations

Fryns syndrome was estimated at 0.7 per 10,000 births (six cases in 112,276 consecutive births, including perinatal deaths) in a French population in 1989, before current genetic methods; in a cohort of 1833 patients with CDH observed over six years, 23 (1.3%) had Fryns syndrome, and older studies estimated 4–10% [7, 9, 77]. PIGN-related Fryns syndrome has a probable founder intragenic deletion in patients with ancestry from La Réunion and other Indian Ocean islands [7, 40]. Donnai-Barrow syndrome has been described in 49 individuals with sufficient data, many from large consanguineous families, in northern and central Europeans, Middle Eastern populations, Americans of European descent and African Americans, with no ethnic predominance [16]. Families with CDH and consanguinity have been described in Arab, Kuwaiti, Egyptian, Pakistani and Finnish populations [14]. A STRA6 variant was described in two Hmong families with contractures and camptodactyly [29]. The most frequent monogenic causes in CDH cohorts are ZFPM2 (about 5% of 275 exomes) and Fryns syndrome [7, 12].

Inequalities and special populations

The AAP reports that health inequalities contribute to higher mortality and greater use of ECMO in patients from racially and ethnically minoritised groups, who therefore face more barriers to follow-up [22]. Preterm and low-birth-weight newborns, more frequent in syndromic forms, have poorer growth and more morbidity, and ECMO has technical limits below 34 weeks and 2 kg [1, 3, 46].

European reference networks and patient associations

The relevant European reference network is ERNICA (rare inherited and congenital digestive and gastrointestinal diseases), which covers CDH [6]. In France, the Rare Disease Reference Centre for CDH has sites in Lille, Clamart and Marseille, with the FIMATHO network (fimatho.fr) and the website hernie-diaphragmatique.fr [3]. Patient associations include CDH International (cdhi.org), CDH UK (cdhuk.org.uk), CHERUBS (cherubs-cdh.org), APEHDia (apehdia.org), the Léonore association (association-leonore.com) and Groupe Miam-Miam for oral feeding disorders [1, 3, 21]. For bereaved families, GeneReviews lists The Compassionate Friends and HAND (Helping After Neonatal Death) [7].

10. Illustrative Clinical Case

Nicholas and colleagues described, as part of a rapid genome sequencing project in newborns in Utah, a family in which CDH and atrial septal defect co-segregated in a mother and her daughter [54]. In a 21-year-old pregnant woman, gravida 2 and nulliparous, fetal ultrasound showed a left-sided CDH with liver, stomach and bowel in the thorax and an estimated left lung volume of 25.3%, and fetal echocardiography suggested a hypoplastic aortic arch, possible coarctation and a hypoplastic left ventricle; there was also intermittent urinary tract dilatation [54]. The pregnancy was complicated by diet-controlled gestational diabetes and prelabour rupture of membranes, with induction of labour and caesarean section at 34 weeks and 6 days; the Apgar score was 1 and 7, and the newborn girl was small for gestational age (weight at the 9th centile, length at the 2nd centile, head circumference at the 18th centile), with a scaphoid abdomen, hypoplasia of the distal phalanx of the fifth finger and positional adduction of the thumb [54]. The mother had been born with left-sided CDH, ASD and scoliosis, all operated on, and had residual pulmonary hypoplasia and pulmonary hypertension, with no facial dysmorphism, no consanguinity and no other affected relatives [54]. After birth, echocardiography excluded coarctation, but confirmed a hypoplastic arch, pulmonary hypertension and a small ostium secundum ASD; the hernia was repaired with a muscle flap on day three, with abdominal closure five days later, and the recurrence was operated on again at four months, followed by tracheostomy and gastrostomy at six months; the course was complicated by chylothorax, inferior vena cava thrombus and severe pulmonary hypertension treated with several therapies [54]. Rapid trio genome sequencing, with a median coverage above 45-fold and three structural variant calling algorithms, identified a complex 4.37 Mb rearrangement at 8q22.3-q23.1, with a 4.36 Mb inversion encompassing 18 genes, flanked by two 3.7 kb deletions, whose 3' breakpoint in intron 2 of ZFPM2 displaces and inverts the first two exons of the gene; the rearrangement was present in the mother and absent in the father, the karyotype was normal 46,XX and interphase FISH confirmed the inversion [54]. The authors concluded that the rearrangement would not be detected by SNP array, exome or targeted panel and that the deletions, entirely non-coding, would not be interpreted as pathogenic in isolation; the diagnosis limited further investigation and informed family planning [54]. The figures of the article show the scheme of the rearrangement in relation to the ZFPM2 exons (https://pmc.ncbi.nlm.nih.gov/articles/PMC9000945/figure/mgg31888-fig-0001/), confirmation by interphase FISH (https://pmc.ncbi.nlm.nih.gov/articles/PMC9000945/figure/mgg31888-fig-0002/) and amplification of the breakpoints (https://pmc.ncbi.nlm.nih.gov/articles/PMC9000945/figure/mgg31888-fig-0003/).

The case illustrates three points of this article: AD inheritance with variable expression of ZFPM2, here with CDH in mother and daughter, after other families with healthy carriers [12, 13]; recurrence after repair of a large defect, more frequent in the first two years [22]; and the technical limit of karyotype, CMA and exome in the face of complex balanced rearrangements [54]. Rapid genome sequencing with research structural variant callers, used in this case, does not correspond to the routine clinical tests described in the guidelines consulted.

11. Prognosis and Long-Term Follow-up

Survival and prognostic factors

CDH mortality ranges from 20% to 60% depending on the population and data collection methods [1]. Overall neonatal mortality is 30–40% in the PNDS and survival, across all forms, 60–70%; more than half of children lead a normal life, and morbidity does not persist beyond childhood in the great majority, with quality of life in adolescence apparently preserved [3]. Isolated CDH treated at a tertiary centre has survival above 50% in the Oxford Desk Reference and 80–90% in reference series of isolated left-sided CDH, whereas population-based series with all prenatal diagnoses have mortality of at least 50% [1, 21]. Factors of worse prognosis are the complex form (chromosomal anomaly, monogenic disorder, major malformations, heart defect), defect size, the degree of pulmonary hypoplasia (o/e LHR, o/e total lung volume, liver herniation), the severity of perinatal pulmonary hypertension, early diagnosis and bilateral CDH; right-sided CDH has more adverse ultrasound predictors but, in recent studies, survival equivalent to left-sided CDH [1, 21]. A likely deleterious de novo variant is associated, even in isolated CDH, with mortality similar to that of complex forms and with poorer neurodevelopment at two years [45]. Fryns syndrome with PIGN variants was lethal in all 17 cases published up to 2025, and in Matthew-Wood syndrome about a third survive [7, 29]. After FETO in severe left-sided CDH, survival to discharge was 40% in the TOTAL trial and 71% in the most recent era at a London centre [5, 62]. In the pooled reanalysis of the two TOTAL trials, o/e LHR and liver herniation were independent predictors of survival to discharge (aOR for liver herniation 0.50; 0.27–0.93) and of survival at six months without oxygen (aOR 0.31; 0.17–0.59), but not of gestational age at birth, which depended on FETO and timing of balloon placement [61].

Long-term morbidity

Almost half of survivors will have sequelae, mainly respiratory and digestive [3]. The PNDS summarises the frequencies of bronchopulmonary dysplasia (22–52%), asthma (15% at five years), lung disease (7–33%), GOR (45–90%), oesophagitis (50% of adults), oral feeding disorders (25%), gastrostomy (33% at one year and 7% at two years), growth delay (50% at one year), adhesive obstruction (20%), recurrence (8–50%), neurological abnormalities (12%), chest deformity (21–48%) and scoliosis (5–27%), with no data on persistent PAH or deafness [3]. The AAP adds more recent data: persistent pulmonary hypertension in up to 20%, tracheostomy in at least 4%, abnormal FEV1 and forced vital capacity in 30% at 8–12 years, moderate to severe neurodevelopmental impairment in 10–25%, autism spectrum and hyperactivity behaviours in 14%, intestinal obstruction in 10–30%, recurrence in about 10% after open repair and operated cryptorchidism in 18–36% [22]. Health-related quality of life is, in most recent studies, similar to that of the general population, but is lower with ECMO, home oxygen, reoperations, special education or low socioeconomic status; in a longitudinal study of 133 survivors, health status declined between 8 and 12 years without a reduction in quality of life [22]. The AAP defines high risk as defect C or D, ECMO, patch, significant morbidity at discharge or concomitant diagnoses, and the Canadian consensus, in the 2023 version, as ECLS, repair with patch or muscle flap, or respiratory support at 30 days of life [22, 48].

Organisation of follow-up

The PNDS proposes visits every three months in the first year, every six months in the second and yearly until transition, each with a paediatrician and a surgeon, preceded by a multidisciplinary meeting before discharge with the primary care physician and local professionals [3]. The AAP proposes different schedules for high and low risk; an intermediate patient, for example with a B defect repaired with a patch, starts on the high-risk schedule and moves to the low-risk schedule if there are no adverse effects in the first months after discharge [22]. The two documents are compatible in content, but differ in the frequency of radiographs and the timing of pulmonary function tests. The 2023 Canadian consensus recommends standardised multidisciplinary follow-up, adjusted to risk level, ideally with general and developmental paediatrics, nutrition, paediatric surgery, respiratory medicine and cardiology, and other specialties as needed, refers to the AAP schedules and notes the lack of knowledge about transition to adult care [48].

No laboratory test is routine in the follow-up of isolated CDH [3]. The report of each visit should be sent to the primary care physician, who monitors growth and development, advises on diet, treats intercurrent illnesses, detects complications and checks vaccinations [3]. In low-resource settings, the AAP proposes a primary care physician close to the family, in collaboration with a centre of excellence, with remote consultations and screening tools to identify patients who need advanced care [22].

12. Summary and Final Considerations

CDH affects about 3 in every 10,000 births and is a heterogeneous malformation, isolated in 50–60% of patients and associated with other anomalies in the rest, in which pulmonary hypoplasia and pulmonary hypertension determine survival. The genetic cause is identifiable in a growing proportion of cases: chromosomal anomalies and CNVs (Pallister-Killian syndrome, trisomies, 15q26, 8p23.1, 1q41-q42 and 4p16.3 deletions) are detected by karyotype, CMA and FISH on appropriate tissues, and monogenic disorders, notably Fryns syndrome, ZFPM2, MYRF, LRP2, STRA6, RARB and NR2F2, by exome or genome sequencing, which find a causal variant in 2–12% of fetuses with normal CMA and in larger proportions in complex forms. A positive result modifies the prognosis, even in apparently isolated CDH, and counselling, with recurrence risks of 25% in AR disorders, up to 50% in fully penetrant AD disorders and generally below 2% in isolated CDH without an identified cause. Treatment is not directed at the genetic cause: it relies on delivery at a level III centre near term, immediate intubation without mask ventilation, gentle ventilation with saturation and PaCO2 targets that differ between the French PNDS, CDH EURO and the Canadian consensus, restricted use of inhaled NO and repair after stabilisation, with fetal tracheal occlusion reserved for severe isolated left-sided CDH, where it increased survival to discharge from 15% to 40% at the cost of more preterm deliveries. Survivors need multidisciplinary follow-up into adulthood, stratified by risk, for pulmonary hypertension, respiratory function, nutrition, GOR and the oesophagus, hearing, neurodevelopment, scoliosis, chest deformity, cryptorchidism and hernia recurrence, with RSV prophylaxis for all. Genetics consultation, autopsy in fatal cases and DNA banking when a diagnosis is not obtained are part of the approach and allow the diagnosis to be revisited as testing evolves.

13. References

1. Longoni M, Pober BR, High FA. Congenital Diaphragmatic Hernia Overview (retired chapter, for historical reference only). In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2006 [updated 2019 Mar 28; revised 2020 Nov 5; retired 2025 Jul 17]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1359/
2. Orphanet. Congenital diaphragmatic hernia (ORPHA:2140). https://www.orpha.net/en/disease/detail/2140
3. Centre de Référence Maladies Rares Hernie de Coupole Diaphragmatique (Storme L, Benachi A, Boubnova J, Panait N, Mur S). Protocole National de Diagnostic et de Soins (PNDS): Hernie de coupole diaphragmatique. Saint-Denis La Plaine: Haute Autorité de Santé; 2020 [published 2020-10-27]. https://www.has-sante.fr/jcms/p_3214228
4. Scott DA, Gofin Y, Berry AM, Adams AD. Underlying genetic etiologies of congenital diaphragmatic hernia. Prenat Diagn. 2022;42(3):373-386. doi:10.1002/pd.6099
5. Deprest JA, Nicolaides KH, Benachi A, Gratacos E, Ryan G, Persico N, et al. Randomized trial of fetal surgery for severe left diaphragmatic hernia. N Engl J Med. 2021;385(2):107-118. doi:10.1056/NEJMoa2027030
6. ERNICA — European Reference Network for rare Inherited and Congenital (digestive and gastrointestinal) Anomalies. https://ern-ernica.eu
7. Slavotinek A. Fryns Syndrome. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2007 [updated 2025 Dec 4]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1459/
8. Orphanet. Fryns syndrome (ORPHA:2059). https://www.orpha.net/en/disease/detail/2059
9. OMIM. %229850 Fryns syndrome; FRNS. Baltimore: Johns Hopkins University; edit 2025-12-18. https://www.omim.org/entry/229850
10. Orphanet. 8p23.1 microdeletion syndrome (ORPHA:251071). https://www.orpha.net/en/disease/detail/251071
11. OMIM. %222400 Diaphragmatic hernia 2; DIH2. Baltimore: Johns Hopkins University; edit 2025-12-01. https://www.omim.org/entry/222400
12. Longoni M, Russell MK, High FA, Darvishi K, Maalouf FI, Kashani A, et al. Prevalence and penetrance of ZFPM2 mutations and deletions causing congenital diaphragmatic hernia. Clin Genet. 2015;87(4):362-367. doi:10.1111/cge.12395
13. OMIM. #610187 Diaphragmatic hernia 3; DIH3. Baltimore: Johns Hopkins University; edit 2025-06-17. https://www.omim.org/entry/610187
14. OMIM. %142340 Diaphragmatic hernia, congenital (DIH1). Baltimore: Johns Hopkins University; edit 2023-11-14. https://www.omim.org/entry/142340
15. Lü Y, Yu Y, Chang J, Li M, Yang X, Zhou X, et al. Genetic diagnosis and clinical features of fetuses with congenital diaphragmatic hernia. Prenat Diagn. 2025;45(12):1651-1659. doi:10.1002/pd.6727
16. Longoni M, Kantarci S, Donnai D, Pober BR. Donnai-Barrow Syndrome. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2008 [updated 2018 Nov 21]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1878/
17. Orphanet. Donnai-Barrow syndrome (ORPHA:2143). https://www.orpha.net/en/disease/detail/2143
18. Orphanet. Matthew-Wood syndrome (ORPHA:2470). https://www.orpha.net/en/disease/detail/2470
19. Kaplan JD, Stewart B, Prasov L, Pyle TLC. MYRF-Related Cardiac Urogenital Syndrome. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2022 [updated 2025 Jul 31]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK586169/
20. Lipska-Ziętkiewicz BS. WT1 Disorder. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2020 [updated 2025 May 15]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK556455/
21. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 714–715 (Congenital diaphragmatic hernia).
22. Jancelewicz T, Lucke A, Guillory C; Section on Surgery; Committee on Fetus and Newborn. Postdischarge follow-up of infants with congenital diaphragmatic hernia: clinical report. Pediatrics. 2026;157(2):e2025074114. doi:10.1542/peds.2025-074114
23. Lally KP, Lasky RE, Lally PA, Bagolan P, Davis CF, Frenckner BP, et al. Standardized reporting for congenital diaphragmatic hernia—an international consensus. J Pediatr Surg. 2013;48(12):2408-2415. doi:10.1016/j.jpedsurg.2013.08.014
24. Pober BR, Lin A, Russell M, Ackerman KG, Chakravorty S, Strauss B, et al. Infants with Bochdalek diaphragmatic hernia: sibling precurrence and monozygotic twin discordance in a hospital-based malformation surveillance program. Am J Med Genet A. 2005;138A(2):81-88. doi:10.1002/ajmg.a.30904
25. Fryns JP, Moerman F, Goddeeris P, Bossuyt C, Van den Berghe H. A new lethal syndrome with cloudy corneae, diaphragmatic defects and distal limb deformities. Hum Genet. 1979;50(1):65-70. doi:10.1007/BF00295591
26. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 354 (Donnai-Barrow syndrome).
27. Kantarci S, Al-Gazali L, Hill RS, Donnai D, Black GC, Bieth E, et al. Mutations in LRP2, which encodes the multiligand receptor megalin, cause Donnai-Barrow and facio-oculo-acoustico-renal syndromes. Nat Genet. 2007;39(8):957-959. doi:10.1038/ng2063
28. Ackerman KG, Herron BJ, Vargas SO, Huang H, Tevosian SG, Kochilas L, et al. Fog2 is required for normal diaphragm and lung development in mice and humans. PLoS Genet. 2005;1(1):58-65. doi:10.1371/journal.pgen.0010010
29. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 294–295 (Matthew-Wood syndrome).
30. Srour M, Chitayat D, Caron V, Chassaing N, Bitoun P, Patry L, et al. Recessive and dominant mutations in retinoic acid receptor beta in cases with microphthalmia and diaphragmatic hernia. Am J Hum Genet. 2013;93(4):765-772. doi:10.1016/j.ajhg.2013.08.014
31. McInerney-Leo AM, Harris JE, Gattas M, Peach EE, Sinnott S, Dudding-Byth T, et al. Fryns syndrome associated with recessive mutations in PIGN in two separate families. Hum Mutat. 2016;37(7):695-702. doi:10.1002/humu.22994
32. Qi H, Yu L, Zhou X, Wynn J, Zhao H, Guo Y, et al. De novo variants in congenital diaphragmatic hernia identify MYRF as a new syndrome and reveal genetic overlaps with other developmental disorders. PLoS Genet. 2018;14(12):e1007822. doi:10.1371/journal.pgen.1007822
33. Longoni M, High FA, Qi H, Joy MP, Hila R, Coletti CM, et al. Genome-wide enrichment of damaging de novo variants in patients with isolated and complex congenital diaphragmatic hernia. Hum Genet. 2017;136(6):679-691. doi:10.1007/s00439-017-1774-y
34. Qiao L, Welch CL, Hernan R, Wynn J, Krishnan US, Zalieckas JM, et al. Common variants increase risk for congenital diaphragmatic hernia within the context of de novo variants. Am J Hum Genet. 2024;111(11):2362-2381. doi:10.1016/j.ajhg.2024.08.024
35. GeneGuidance (local curated database) — entries ZFPM2, NR2F2, GATA4, PIGN, MYRF, LRP2, STRA6, RARB, WT1; accessed 2026-11-06.
36. Scott TM, Campbell IM, Hernandez-Garcia A, Lalani SR, Liu P, Shaw CA, et al. Clinical exome sequencing data reveal high diagnostic yields for congenital diaphragmatic hernia plus (CDH+) and new phenotypic expansions involving CDH. J Med Genet. 2022;59(3):270-278. doi:10.1136/jmedgenet-2020-107317
37. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 90 (15q24 microdeletion syndrome).
38. Loong L, Tardivo A, Knaus A, Hashim M, Pagnamenta AT, Alt K, et al. Biallelic variants in PIGN cause Fryns syndrome, multiple congenital anomalies-hypotonia-seizures syndrome, and neurologic phenotypes: a genotype-phenotype correlation study. Genet Med. 2023;25(1):37-48. doi:10.1016/j.gim.2022.09.007
39. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348
40. Alessandri JL, Gordon CT, Jacquemont ML, Gruchy N, Ajeawung NF, Benoist G, et al. Recessive loss of function PIGN alleles, including an intragenic deletion with founder effect in La Réunion Island, in patients with Fryns syndrome. Eur J Hum Genet. 2018;26(3):340-349. doi:10.1038/s41431-017-0087-x
41. Yu L, Wynn J, Cheung YH, Shen Y, Mychaliska GB, Crombleholme TM, et al. Variants in GATA4 are a rare cause of familial and sporadic congenital diaphragmatic hernia. Hum Genet. 2013;132(3):285-292. doi:10.1007/s00439-012-1249-0
42. Schwab ME, Dong S, Lianoglou BR, Aguilar Lucero AF, Schwartz GB, Norton ME, et al. Exome sequencing of fetuses with congenital diaphragmatic hernia supports a causal role for NR2F2, PTPN11, and WT1 variants. Am J Surg. 2022;223(1):182-186. doi:10.1016/j.amjsurg.2021.07.016
43. Suri M, Kelehan P, O'Neill D, Vadeyar S, Grant J, Ahmed SF, et al. WT1 mutations in Meacham syndrome suggest a coelomic mesothelial origin of the cardiac and diaphragmatic malformations. Am J Med Genet A. 2007;143A(19):2312-2320. doi:10.1002/ajmg.a.31924
44. Chassaing N, Golzio C, Odent S, Lequeux L, Vigouroux A, Martinovic-Bouriel J, et al. Phenotypic spectrum of STRA6 mutations: from Matthew-Wood syndrome to non-lethal anophthalmia. Hum Mutat. 2009;30(5):E673-E681. doi:10.1002/humu.21023
45. Qiao L, Wynn J, Yu L, Hernan R, Zhou X, Duron V, et al. Likely damaging de novo variants in congenital diaphragmatic hernia patients are associated with worse clinical outcomes. Genet Med. 2020;22(12):2020-2028. doi:10.1038/s41436-020-0908-0
46. Jordan AJ, Mohr C, Martel R, Zahn K, Schaible T, Adam R, et al. Growth challenges and recovery in 1247 children with congenital diaphragmatic hernia: a 10-year follow-up. Eur J Pediatr. 2025;184(12):738. doi:10.1007/s00431-025-06479-w
47. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 292–293 (Fryns syndrome).
48. Puligandla P, Skarsgard E, Baird R, Guadagno E, Dimmer A, Ganescu O, et al. Diagnosis and management of congenital diaphragmatic hernia: a 2023 update from the Canadian Congenital Diaphragmatic Hernia Collaborative. Arch Dis Child Fetal Neonatal Ed. 2024;109(3):F239-F252. doi:10.1136/archdischild-2023-325865. PMID: 37879884
49. Canadian Congenital Diaphragmatic Hernia Collaborative; Puligandla PS, Skarsgard ED, Offringa M, et al. Diagnosis and management of congenital diaphragmatic hernia: a clinical practice guideline. CMAJ. 2018;190(4):E103-E112. doi:10.1503/cmaj.170206. PMID: 29378870
50. Snoek KG, Reiss IK, Greenough A, Capolupo I, Urlesberger B, Wessel L, et al. Standardized postnatal management of infants with congenital diaphragmatic hernia in Europe: the CDH EURO Consortium Consensus - 2015 update. Neonatology. 2016;110(1):66-74. doi:10.1159/000444210. PMID: 27077664
51. Abgral M, Russo FM, Deprest J, Benachi A. Prenatal assessment and prediction in CDH: for what purpose? Semin Fetal Neonatal Med. 2025;30(3):101654. doi:10.1016/j.siny.2025.101654
52. Huang R, Fu F, Mei S, Liu L, Zhong W, Han J, et al. Genetic burden and multidimensional predictors in prenatal diagnosis of fetal congenital diaphragmatic hernia. Hum Genet. 2025;144(9-10):1035-1050. doi:10.1007/s00439-025-02777-3
53. Weller K, Westra D, Peters NCJ, Wilke M, Van Opstal D, Feenstra I, et al. Exome sequencing in fetuses with congenital diaphragmatic hernia in a nationwide cohort. Prenat Diagn. 2024;44(11):1288-1295. doi:10.1002/pd.6622
54. Nicholas TJ, Al-Sweel N, Farrell A, Mao R, Bayrak-Toydemir P, Miller CE, et al. Comprehensive variant calling from whole-genome sequencing identifies a complex inversion that disrupts ZFPM2 in familial congenital diaphragmatic hernia. Mol Genet Genomic Med. 2022;10(4):e1888. doi:10.1002/mgg3.1888
55. Blair J, Carratu K, Rintoul NE, Hedrick HL, Peranteau WH, Avitabile CM, et al. Genomic contributors to congenital diaphragmatic hernia: results of exome sequencing in 560 probands and cross reference of findings in an independent cohort. Am J Med Genet A. 2026;200(8):1821-1831. doi:10.1002/ajmg.a.70145
56. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 804 (Mycophenolate mofetil embryopathy).
57. Orphanet. PAGOD syndrome (ORPHA:991). https://www.orpha.net/en/disease/detail/991
58. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 224 (Simpson-Golabi-Behmel syndrome).
59. Stege G, Fenton A, Jaffray B. Nihilism in the 1990s: the true mortality of congenital diaphragmatic hernia. Pediatrics. 2003;112(3 Pt 1):532-535. doi:10.1542/peds.112.3.532
60. Deprest JA, Benachi A, Gratacos E, Nicolaides KH, Berg C, Persico N, et al. Randomized trial of fetal surgery for moderate left diaphragmatic hernia. N Engl J Med. 2021;385(2):119-129. doi:10.1056/NEJMoa2026983
61. Van Calster B, Benachi A, Nicolaides KH, Gratacos E, Berg C, Persico N, et al. The randomized Tracheal Occlusion To Accelerate Lung growth (TOTAL)-trials on fetal surgery for congenital diaphragmatic hernia: reanalysis using pooled data. Am J Obstet Gynecol. 2022;226(4):560.e1-560.e24. doi:10.1016/j.ajog.2021.11.1351. PMID: 34808130
62. Jeffreys E, Jenkinson A, Kronfli R, Billington J, Bhat R, Nanjundappa M, et al. Changes in management of FETO-treated infants with severe CDH. Eur J Pediatr. 2026;185(11):888. doi:10.1007/s00431-026-07526-w
63. Russo F, Benachi A, Meijer F, Cauvet F, Berrué-Gaillard H, Power B, Deprest J. The fall out of the 2017 European Medical Device Regulation for tracheal occlusion. Prenat Diagn. 2025;45(4):539-543. doi:10.1002/pd.6763
64. Horn-Oudshoorn EJJ, Vermeulen MJ, Knol R, Bout-Rebel R, Te Pas AB, Hooper SB, et al. Multicentre, randomised controlled trial of physiological-based cord clamping versus immediate cord clamping in infants with a congenital diaphragmatic hernia (PinC): statistical analysis plan. Trials. 2024;25(1):198. doi:10.1186/s13063-024-08027-7
65. U.S. Food and Drug Administration. Acetaminophen injection (Baxter Healthcare, ANDA 214331) — prescribing information; revised 2024-05. DailyMed setid 0f67c1bd-915b-4681-88dd-22f5154a80b3.
66. Noh CY, Chock VY, Bhombal S, Danzer E, Patel N, Dahlen A, et al. Early nitric oxide is not associated with improved outcomes in congenital diaphragmatic hernia. Pediatr Res. 2023;93(7):1899-1906. doi:10.1038/s41390-023-02491-8
67. Sescleifer AM, Chidiac C, Hellmann ZJ, McDermott KM, Garcia AV, Rhee DS, et al. Revisiting inhaled nitric oxide utilization in congenital diaphragmatic hernia: a national cohort study of major children's hospitals in the United States. J Pediatr Surg. 2025;60(9):162423. doi:10.1016/j.jpedsurg.2025.162423
68. U.S. Food and Drug Administration. INOmax (nitric oxide) gas for inhalation — prescribing information (NDA 020845); revised 2025-12.
69. U.S. Food and Drug Administration. REVATIO (sildenafil) tablets, oral suspension, injection — prescribing information (NDA 021845/022473/203109); revised 2024-12. Drugs@FDA/DailyMed via openFDA.
70. U.S. Food and Drug Administration. TRACLEER (bosentan) — prescribing information; revised 2025-11.
71. U.S. Food and Drug Administration. PROSTIN VR PEDIATRIC (alprostadil) injection — prescribing information; revised 2026-08.
72. U.S. Food and Drug Administration. Milrinone lactate injection (Mullan Pharmaceutical, ANDA 216373) — prescribing information; revised 2026-07. DailyMed setid 07dddead-22ed-004c-e063-6294a90a76fc.
73. European Medicines Agency. Beyfortus (nirsevimab) — EPAR, product information (Annex I, Summary of Product Characteristics). Initial authorisation 31/10/2022; extension to the 2nd season (II/0005) 01/08/2024. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/beyfortus [accessed 2026-11-07].
74. Merck Sharp & Dohme LLC. ENFLONSIA (clesrovimab) injection: prescribing information. BLA 761432; approval 2025 Jun 9 [accessed 7 Nov 2026]. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/label/2025/761432s000lbledt.pdf
75. European Medicines Agency. Enflonsia (clesrovimab): EPAR — product information. EMEA/H/C/006497; CHMP opinion 18/09/2025; authorisation 15/04/2026 [accessed 7 Nov 2026]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/enflonsia
76. European Medicines Agency. Synagis (palivizumab) — Summary of Product Characteristics. Initial authorisation on 13 Aug 1999. Available from: https://www.ema.europa.eu/en/documents/product-information/synagis-epar-product-information_en.pdf. Accessed on 7 Nov 2026.
77. Neville HL, Jaksic T, Wilson JM, Lally PA, Hardin WD, Hirschl RB, et al. Fryns syndrome in children with congenital diaphragmatic hernia. J Pediatr Surg. 2002;37(12):1685-1687. doi:10.1053/jpsu.2002.36695