Key Points
- •
Congenital diaphragmatic hernia occurs in 1 to 4 in 10,000 births. The condition is isolated in more than 50% of the cases.
- •
The main causes of mortality and morbidity are respiratory insufficiency and persistent pulmonary hypertension of the newborn.
- •
Prenatal diagnosis should be made by screening ultrasound, after which patients are referred to specialised centres.
- •
In isolated cases, the size of the lungs and the presence of liver herniation are antenatal predictors of outcome.
- •
In cases with anticipated poor outcome, a potential option is fetal treatment in the form of fetoscopic endoluminal tracheal occlusion.
Epidemiology and Background
Congenital diaphragmatic hernia (CDH) is a developmental anomaly with a prevalence ranging between 1 and 4 in 10,000 births, which means that in Europe, around 2000 children are born with this condition every year. Despite being relatively uncommon, CDH is a major clinical concern inside the realm of neonatology, with important implications for diagnosis, management and prognosis. Although medical and surgical management have improved the outcome of this condition, CDH remains associated with high mortality and significant morbidity.
A primary characteristic of CDH is that the diaphragm fails to form properly during embryogenesis. Normally, the diaphragm develops to form a continuous sheet that completely separates the thoracic and abdominal cavities before the major period of internal organ growth. In the case of CDH, a significant proportion of the diaphragm is absent. The defect is usually on the left side (85%) but can also occur on the right (12%–15%) or bilaterally. In some rare instances, a true agenesis of the hemidiaphragm is present, but in most cases, the defect is limited to the posterolateral region of the diaphragm (referred to as Bochdalek hernia). The anterior (Morgagni hernia; 25%–30%) or central regions (2%–5%) can also be affected. Less often, the diaphragm is present but thinned and devoid of muscular fibres (diaphragmatic eventration).
The diaphragmatic defect allows herniation of abdominal viscera into the thorax, where they compete for the space normally reserved to accommodate the growing lungs. When the defect is located on the left side, the thorax may contain small and large bowel, the spleen, the stomach, the left lobe of the liver and, occasionally, the kidney. Right-sided CDHs virtually always contain part of the right lobe of the liver and sometimes the bowel, kidney, or both. The loss of a continuous diaphragmatic muscle also impairs fetal breathing movements that are necessary for proper stretch-induced lung maturation.
Lungs of fetuses with CDH display variable degrees of lung hypoplasia, with impairment of both airway and vascular maturation. These changes become symptomatic immediately after birth, when neonates have variable degrees of respiratory insufficiency and persistent pulmonary hypertension (PPH), which is often resistant to inhaled nitric oxide (iNO).
In 50% to 60% of cases, the diaphragmatic defect and lung hypoplasia are the only significant anomalies. In the remaining cases, there are major nonpulmonary congenital anomalies. Cardiovascular defects such as ventricular septal defects, cardiac outflow anomalies (tetralogy of Fallot, double outlet right ventricle, transposition of the great vessels and others) and abnormal great vessels (right aortic arch, double aortic arch, truncus arteriosus, abnormal subclavian arteries and others) are the most common associated anomalies, found in about one third of patients with CDH. Left ventricular hypoplasia has also been described, yet its occurrence and clinical relevance are debated. Musculoskeletal defects such as anomalies of the limbs or of the number and shape of the vertebral bodies or ribs, neural tube defects, abdominal wall defects, craniofacial defects or urinary tract anomalies have also been reported. Associated malformations are sometimes components of Pallister-Killian and Fryns syndromes; Ghersoni-Baruch syndrome; Wilms tumour, aniridia, genitourinary anomalies, and mental retardation (WAGR); Denys-Drash; and other syndromes. Some chromosomal anomalies, such as 9p tetrasomy, have CDH as part of their spectrum. For further information we refer to the excellent review by Slavotinek and colleagues.
Finally, the presence of the intestine in the thorax during late fetal development causes malrotation, malfixation, or both, which can further complicate the disease.
Aetiology and Pathogenesis
The causes of CDH are largely unknown, although exposure to teratogens or pharmacologic agents has been suggested. In particular, phenmetrazine, thalidomide, quinine, nitrofen and vitamin A deficiency have been linked to this disease. In the classical view, the defect in CDH occurs first in the muscular part of the diaphragm. However, studies in rats have suggested that CDH is a primary lung pathology, even in humans. Keijzer and colleagues proposed the ‘dual-hit hypothesis’, that is, that two independent events cause the major features seen in CDH. These hits disturb normal lung development (first hit) and diaphragm formation (second hit). Data from animal studies confirm this hypothesis: in the toxic nitrofen model in rats, abnormalities in the ipsilateral as well as the contralateral lung are present already before the development of the diaphragm.
Another theory is based on the hypothesis that nonclosure of the pleuroperitoneal canals would be caused by a defect in the pleuroperitoneal folds (PPFs), the source of diaphragmatic cells. Of interest, the diaphragmatic defect in the nitrofen model is located more medial than could be expected from nonclosure of these canals. Therefore it has been proposed that the origin of the diaphragmatic defect lies in the amuscular mesenchymal precursor cells of the diaphragm, which are also derived from the PPFs. This theory is based on the observation that although the migration of muscular precursors is not disturbed, a defect occurs in regions of the underlying mesenchymal substratum of the PPF. This would subsequently contribute to the defective region in CDH.
Finally, an involvement of the retinoid signalling pathway is likely in CDH. Both animal and clinical studies have shown that retinol and retinol-binding proteins are decreased in newborns with this malformation. Moreover, some of the genes involved in the pathogenesis of human CDH are tightly related to retinoid signalling.
Prognosis
Congenital diaphragmatic hernia was described many years ago, but survival after repair was not achieved until the 20th century. The symptoms of insufficient gas exchange are associated with those of PPH, and unless invasive treatment is undertaken, the respiratory condition deteriorates rapidly until death.
Survival rates vary widely amongst various neonatal management centres. When only liveborn children undergoing surgery are included, survival until discharge approaches 70% for isolated CDH. If terminations of pregnancy, spontaneous abortions, stillbirths, prehospital or preoperative deaths and surgical mortality are taken into account, the mortality rate is between 50% and 60%. The gap between these numbers is usually referred to as the ‘hidden mortality’. However, significant advances in the postnatal management, with the introduction of ‘ gentle ventilation and permissive hypercarbia ’, have resulted in improved survival rates over the past 2 decades. The improved survival of very sick babies is, however, associated with a higher risk for severe morbidity persisting beyond the initial hospitalisation, especially in those treated with extracorporeal membrane oxygenation (ECMO).
The severity of lung hypoplasia and PPH are the key determinants of both mortality and morbidity and thus of quality of life. More than half of survivors are oxygen dependent at 28 days of age, and 16% require oxygen at the time of discharge for a mean duration of 14.5 months. Restrictive and obstructive lung diseases have also been reported in CDH survivors many years after operation. Diaphragmatic rigidity and thoracic deformities can play a minor role in chronic lung disease. Bronchodilators may be needed in 40% of patients in the first year of life. PPH may persist in up to 30% of patients at 2 months of age and is associated with increased risk for early death and increased morbidity. PPH also deeply affects the quality of life in CDH survivors.
Nonpulmonary diseases are also relatively frequent in CDH survivors. Gastroesophageal reflux is caused by a distorted anatomy of the diaphragmatic sling, both congenital and related to the surgical CDH repair. Also, esophageal motility and gastroesophageal sphincter function are disturbed. In addition, malrotation may delay gastric emptying, and the abnormal balance of pressures in the thorax and abdomen may facilitate retrograde passage of gastric contents to the oesophagus. Reflux can complicate the preexisting respiratory disease, and for all these reasons, a considerable proportion of these patients respond poorly to medical treatment and ultimately require antireflux surgery. Reflux and the need for antireflux surgery are dependent on the degree of pulmonary hypoplasia.
Neurodevelopmental deficits are possible in patients in whom brain oxygenation was marginal for long periods of time and particularly when ECMO required major vascular occlusion. Neurosensory deafness occurs in a small proportion of children surviving CDH. This is progressive, and it is often tied to both prolonged antibiotic treatment and a particular sensitivity or a developmental defect of the inner ear.
These sequelae, and the frequent associated malformations, require long-term follow-up for early diagnosis and proactive management. Most babies eventually lead a life very close to normal provided when managed in a multidisciplinary follow-up program.
Congenital Diaphragmatic Hernia–Related Aberrations in Lung Development
In CDH, the lung is hypoplastic not only on the side of the hernia, but the contralateral side is affected as well but to a variable extent. Peripheral airways are less developed, and there are markedly less and smaller alveoli, thickened alveolar walls and an increased amount of interstitial tissue so that there is less alveolar airspace and gas exchange surface area. Parallel to airway changes, there is a reduction in arteries, resulting in a hypoplastic vascular bed. Morphologically, a thickening of the vascular wall is determined by an increase in arterial media and adventitia and by neomuscularisation of the small pulmonary arteries, which are normally partially or nonmuscularised. The structural remodelling of the small pulmonary arteries reduces their ability to dilate to increase the vascular bed capacity and reduce the pressure in the pulmonary circulation after birth. After birth, further muscularisation of this ‘immature’ pulmonary vasculature occurs, with migration of adventitial fibroblasts into the media and smooth muscle cells into the intima. These morphologically abnormal vessels respond abnormally to mechanical and chemical stimuli, including the shear stress accompanying raised blood flow through the narrowed vessels. Increased contractility and impaired relaxation of pulmonary arteries have been demonstrated in animal models and could be responsible for the low efficacy of conventional vasodilatory therapy.
Prenatal Diagnosis and Outcome Prediction
Today, high-resolution ultrasound and advances in prenatal diagnosis and genetic testing have made it possible to diagnose CDH relatively early and to rule out a number of associated anomalies. Ideally, antenatal ultrasound screening identifies cases in utero , reportedly in more than 70% of cases, yet lower numbers have been reported as well. Intrathoracic abdominal organs are the hallmark of CDH. Left-sided CDH typically presents with a mediastinal shift to the right caused by herniation of the stomach, intestines and in some cases liver ( Fig. 31.1A ). In right-sided CDH, part of the liver is visible in the chest, with mediastinal shift to the left ( Fig. 31.1B ). Prenatal diagnosis allows in utero referral to a tertiary care centre used to manage the condition for expert assessment, counselling and perinatal management. Additional genetic and morphologic assessment using ultrasound or magnetic resonance imaging (MRI) can be used to rule out associated malformations.
For isolated cases, clinicians should individualise prognosis to counsel parents about prenatal options. Most prediction methods are based on lung size, liver herniation and pulmonary circulation, and more recently stomach position. Ultrasound measurement of the lung-to-head ratio (LHR) is most widely used. The LHR, first described by Metkus and colleagues, provides an indirect estimate of the size of the lung contralateral to the hernia normalised for the head circumference ( Fig. 31.1C ). It is a two-dimensional measure and changes over gestation as the lung area grows more rapidly than the head circumference. The observed to expected (o/e) LHR has subsequently been proposed to eliminate the effect of gestational age at assessment. The o/e LHR has been shown to be an independent predictor of postnatal survival both in left- and right-sided CDH and to some extent of short-term morbidity. Other methods to assess lung size, such as the lung-to-thorax ratio, the quantitative lung index and three-dimensional measurements of lung volume, have also been proposed, but the predictive value of these parameters has not been validated to the same extent as the o/e LHR.
Liver herniation can be determined both for left- as well as right-sided hernia, although in the latter case, the liver is nearly always herniating through the defect. For left-sided CDH, herniation of the liver into the thorax has been recognised as a predictor of poor outcome by Harrison and colleagues and seems to be independent from lung size. Although it is theoretically possible to quantify the amount of liver in the thorax, liver position on ultrasound is usually categorised binary as either ‘up’ (in the thorax), or ‘down’ (confined to the abdomen). Because the echogenicity of the liver is very similar to that of the lung, it can be difficult to assess liver position with ultrasound ( Fig. 31.2A and 
). Additional indirect signs suggestive of liver herniation are presence of hepatic vessels above the diaphragmatic edge ( Fig. 31.2B ), abnormal position of the ductus venosus or the gallbladder ( Fig. 31.2C ) or deviation (bowing) of the umbilical vein on towards the left side ( Fig. 31.2D ). Finally, evaluation of stomach position has recently been reintroduced as an indirect method to estimate severity of the disease in left-sided CDH because it has been shown to correlate with the proportion of intrathoracic liver determined by MRI.
The combination of liver herniation and o/e LHR is now a widely accepted method to stratify fetuses with left- and right-sided CDH into groups with an increasing degree of pulmonary hypoplasia and corresponding mortality rates. It is also used to select patients for fetal therapy trials ( Fig. 31.3 ).
Severity assessment by MRI theoretically has several advantages over ultrasound. Visualisation is not limited by maternal habitus, amniotic fluid volume or fetal position. With MRI, the total (left and right) lung volume can be measured, which may better predict postnatal lung function. Volumetry may also accurately quantify liver and stomach herniation. Although one study has claimed that MRI better predicts outcome than ultrasound, the numbers do not allow such a claim nor has this been proven in clinical practice.
Lung size and liver herniation also predict neonatal morbidity, such as the duration of assisted ventilation, the need for supplemental oxygen, the need for patch repair and the time it takes to full enteral feeding. The literature on prediction of PPH is more limited (systematically reviewed by Russo and colleagues ). Several candidate parameters have been suggested in single case series, including lung size, presence of visceral herniation and direct assessment of the pulmonary vasculature, which may provide additional information. However, to our knowledge, there is currently no validated antenatal predictor for PPH.
Current Neonatal Management
In an attempt to improve outcome and permit comparison of outcome data, the CDH EURO consortium has published a standardised neonatal treatment protocol (revised in 2015 ). Delivery is planned (either via induction or via caesarean section) after 39 weeks in a high-volume tertiary centre, and the newborn is immediately intubated.
Historically, high-pressure high oxygen conventional respiratory assistance was used, leading to iatrogenic pneumothorax and acute death. Then it was realised that the hypoplastic and immature CDH lung was severely damaged by excessive oxygen delivery and by excessive airway pressure. Since Wung and colleagues introduced the ‘gentle ventilation and permissive hypercarbia strategy’, improved pulmonary outcomes and mortality rates have been reported. This policy progressively gained support and it is now the standard in most developed countries. High-frequency oscillatory ventilation was also believed to improve survival and reduce long-term morbidity. However, a recent randomised controlled trial (RCT; the VICI trial: Ventilation in Infants with Congenital diaphragmatic hernia: an International randomised clinical trial.) has shown no superiority of this technique versus conventional ventilation. The latter again shows that proper studies have to be done before implementing new technologies.
The management of PPH in newborns with CDH remains one of the major concerns in neonatal intensive care units (NICUs). Early cardiac ultrasound (within the first 24 hours of life) is recommended by the international consortia as a noninvasive tool in the diagnosis of PPH. Serial cardiac ultrasonography is then recommended to guide therapeutic choices and to monitor their effect on PPH. The cornerstone of postnatal treatment of PPH is the reversal of the vasoconstrictive component to prevent the right ventricle overload and the development of irreversible vascular remodelling. Currently, iNO is the first therapeutic choice in CDH infants with PPH. However, in an RCT, although iNO significantly decreased the need for ECMO in infants with PPH, it did not reduce mortality rate, length of hospitalisation, chronic lung injury or neurodevelopmental impairment. Furthermore, although the response rate to iNO in all infants with PPH is around 60%, in infants with CDH, it is estimated to be only 30%. Other vasodilatory drugs, such as phosphodiesterase type 5 (PDE5) inhibitors, endothelin antagonists or prostacyclins, have been used alone or in association with nitric oxide in experimental series. However, their use is highly variable throughout NICUs around the world, which makes it difficult to define their exact role. Again, trials are being designed to clarify this.
Apart from pharmacologic therapy, ECMO is in some centres an adjunct in the treatment of CDH-related PPH. It is used to unload the right ventricle while putting the lungs at rest, thereby reversing PPH, which would otherwise be lethal. It prevents additional lung injury induced by barotrauma and oxidative stress in case of maximal ventilator support. As for iNO, experience with ECMO achieved the worst results precisely in patients with CDH. Therefore there is no evidence that the outcome of CDH is better with ECMO. Together with the limitations of the technique (required weight above 2000 g, need for heparinisation), all this has somewhat tempered the initial enthusiasm followed by its diminished use in many centres.
Surgical repair should be performed electively after the newborn is stabilised. The optimal surgical technique remains under debate. Minimal access surgery is gaining ground on the open approach (thoracotomy or laparotomy). This approach has aesthetic advantages but carries a higher risk for recurrence. It may also not be applicable in severe cases. For defects that are too large to be closed by primary repair, prosthetic patches may be used to close the gap, ranging from first-generation nonabsorbable synthetic materials (Gore-Tex) to biomaterials (xenografts) as well as composite materials. The ideal mesh remains elusive.
Antenatal Therapeutic Strategies
The ability to prenatally identify a future nonsurvivor offers the potential for prenatal interventions that could avoid that outcome. The concept of tracheal occlusion (‘plug the lung until it grows’) was first introduced by Wilson and associates and is inspired by clinical observations in fetuses with congenital high airway obstruction, who have a marked increased lung volume and alveolar number. Airway obstruction prevents egress of pulmonary fluid, which experimentally has been shown to prompt lung growth by a mechanism of stretch of lung parenchymal cells. This leads to increased airway branching morphogenesis, an increase in both alveolar surface area and airspace volume and stimulated alveolisation. This concept has been further explored experimentally, demonstrating that tracheal occlusion improves neonatal lung compliance and improves ventilation. The procedure was first clinically attempted by open fetal surgery and clipping of the trachea. Progress in minimally invasive fetal surgery made percutaneous fetal endoscopic tracheal occlusion (FETO) with a balloon possible. FETO is an investigational minimally invasive, percutaneous procedure that can be done under maternal local anaesthesia ( Fig. 31.4A ). The procedure is usually performed at 27 + 0 to 29 + 6 weeks in severe cases and 30 + 0 to 31 + 6 weeks in moderate cases. The fetus is anesthetised and immobilised with an intramuscular injection of a neuromuscular blocking agent, fentanyl and atropine. A flexible cannula is inserted through the skin and myometrium and targeted to the fetal nose tip under ultrasound guidance. Fetoscopic instruments specifically designed for FETO are then introduced. These consist in a 3.3-mm sheath loaded with a fiberoptic endoscope (1.3 mm; Karl Storz) and a balloon occlusion system (catheter loaded with a detachable inflatable latex balloon with integrated one-way valve Goldbal 2, Balt Extrusion, France). A stylet or forceps can also be inserted through the sheet to remove the balloon if wrongly positioned. Fetoscopic landmarks are the philtrum and upper lip, the tongue and raphe of the palate, uvula, epiglottis, and eventually the vocal cords. The endoscope is advanced into the trachea until identification of the carina, above which the balloon is positioned by inflation and detachment from the catheter ( 
). The median duration of FETO is 10 (range, 3–93) minutes, dependent on both the experience of the operator and the position of the fetus. A longer operation time is the main risk factor for membrane rupture.
