Fetal therapy for congenital diaphragmatic hernia: past, present and future

Evolution of fetal therapy for congenital diaphragmatic hernia

As neonatal management of CDH has improved, efforts have been made to rescue lung growth in utero to prevent the devastating and irreversible sequelae of pulmonary hypoplasia. Harrison et al. first proposed the innovative concept of fetal diaphragmatic hernia repair; however, this entailed a significant risk of IUFD due to umbilical vein kinking following reduction of liver herniation.44 Extrapolating from the natural history of congenital high airway obstruction syndrome,45 in which upper airway obstruction results in pulmonary hyperplasia, tracheal occlusion (TO) in lambs with a surgically created diaphragmatic hernia was proposed to promote lung growth by preventing the egress of lung fluid. ‘Plugging’ the trachea accelerated pulmonary growth, resulting in more compliant lungs, which were more efficient in gas exchange and partially reversed the abdominal visceral herniation.46 47 Cyclical occlusion and reversal was found to achieve optimal alveolar development,48 whereas sustained occlusion resulted in exaggerated lung growth at the expense of surfactant producing type II pneumocytes.48 With reversal of occlusion (‘plug-unplug’ sequence), type II pneumocyte production recovered and pulmonary artery remodeling improved.49 50 Thus, the overall impact of TO on lung development was seen to be multifactorial—dependent on the timing, duration and reversibility of occlusion.49 51–53

Fetal TO was first undertaken in human studies via maternal laparotomy and hysterotomy by externally clipping the trachea after neck dissection, by endotracheal insertion of an orally retrievable foam plug54 55 and then subsequently without hysterotomy using a Fetendo clip.56 57 Despite producing significant lung growth, any survival benefit was compromised by the complications of prematurity, tracheal dissection, and significant maternal morbidity.54–56 Simultaneously, minimally invasive approaches were developed, including a ‘Fetendo-plug’ or foam for TO 58 and fetoscopic placement of a detachable latex balloon,59 60 which was originally developed for neurovascular indications.61 Both methods promoted pulmonary hyperplasia, while avoiding fetal neck dissection. Feasibility of the latter technique has been demonstrated in vivo,62 63 and a percutaneous approach has become the standardized method for clinical use.33 62 64

Fetal endoscopic tracheal occlusion: technical aspects

Fetoscopic balloon insertion and removal can be performed as an outpatient procedure under local anesthesia and maternal intravenous sedation with periprocedural prophylactic tocolysis (indomethacin and/or nifedipine) and antibiotics. Occasionally, external cephalic version may be required to optimize access to the fetal mouth. Once the fetus is appropriately positioned, fetal medications (rocuronium, atropine and fentanyl) are administered intramuscularly or intravenously for paralysis, prevention of bradycardia and analgesia. Direct trocar entry at the level of the fetal mouth is performed using a 10 Fr disposable sheath (Cook Medical, Bloomington, Indiana) with a sharp trocar (Karl Storz, Tuttlingen, Germany) under continuous US guidance, avoiding the placenta and maternal vessels. A Seldinger technique65 may also be used for sheath placement if uterine entry is challenging. A 3.3 mm curved fetoscopic sheath and 1.3 mm fiberoptic fetoscope (Karl Storz) loaded with an inflatable latex detachable balloon with a one-way valve delivered by a microcatheter (GOLDBAL2, BALT Extrusion, Montmorency, France) is introduced into the fetal oral cavity. It is carefully advanced over the tongue, past the uvula and epiglottis and through the vocal cords. Once in the trachea, the balloon is inflated with 0.7–0.8 mL of saline. The balloon is detached above the carina and ideally just below the vocal cords (video 1). Warm isotonic saline is infused throughout the procedure with subsequent amnioreduction as necessary. Patients are required to stay close to the fetal endoscopic tracheal occlusion (FETO) center for the duration of TO, with weekly US surveillance to assess balloon position, amniotic fluid index, cervical length, and lung response (figure 1). FETO is performed at 27–28 weeks and balloons are ideally removed at 34 weeks. Removal is either fetoscopically, using grasping forceps and a stylet to puncture the balloon (video 2). Alternatively, the balloon can be punctured under US guidance using a 20-gauge needle. After successful balloon removal or deflation, patients can await spontaneous labor and deliver vaginally. In the presence of preterm labor, preterm premature rupture of membranes (PPROM), chorion-amnion separation, suboptimal fetal position, or imminent delivery, fetoscopy may not be possible. Under these circumstances, balloon retrieval/deflation is either by US-guided needle puncture or rigid bronchoscopy by pediatric otolaryngologists during an ex utero intrapartum treatment (EXIT) procedure.66

Request permission

FETO: initial experience

A randomized controlled trial (RCT) compared the outcomes following FETO via a maternal laparotomy at 22–28 weeks’ gestational age (GA) and EXIT delivery versus standard postnatal care in cases of isolated LCDH with LHR ≤1.4.67 The trial was stopped prematurely, as there was no difference in survival between groups (73% with FETO vs. 77% with expectant management, p=1.00). All FETO cases were complicated by PPROM and preterm birth was more common (average GA at delivery 30.8±2.0 weeks vs 37.0±1.5 weeks, p<0.001).67 The lack of benefit with FETO may have been related to the inclusion of less severe disease (LHR >1.0), as well as contemporaneous improvement in neonatal and surgical management of CDH.

At the same time, the European FETO consortium described favorable outcomes with a minimally invasive, percutaneous FETO technique in 21 fetuses with severe pulmonary hypoplasia (LHR <1.0) and importantly, without any major maternal complications.62 However, PPROM occurred in nearly 50% of cases with an average GA at delivery of 34 weeks. Survival increased from 30% in the initial 10 cases to >60% in the subsequent 11 cases, which was greater than the reported survival of 8% (1/12) among contemporaneous expectantly managed fetuses. This improvement in survival coincided with iterative improvements in FETO technique, specifically moving from general to regional or local anesthesia, earlier FETO insertion (i.e., second trimester vs third trimester), and prenatal rather than intrapartum balloon removal during EXIT delivery.

The European FETO consortium subsequently published a large prospective multicenter study of 210 FETO cases (175 LCDH, 34 RCDH and 1 bilateral), all with US findings predictive of severe disease (defined as intrathoracic liver herniation and o/e LHR equivalent to an LHR ≤1.0). Compared with expectantly managed historical controls, survival after FETO increased from 24% to 49% and 0% to 35% in severe LCDH and RCDH, respectively.68 Premature delivery was more common in the FETO group, with PPROM occurring in 47% of cases and a median GA at delivery of 35.3 weeks.68 The main predictors of survival included o/e LHR prior to FETO and GA at delivery, with survival of approximately 15% if delivery was <32 weeks and nearly 60% if >32 weeks’ GA. Balloon insertion was successful in 97% of cases. Balloon reversal was elective in 44% and emergent in 56%, mostly due to preterm labor or PPROM. Removal was by fetoscopy (55%), US-guided puncture (21%), EXIT (7%), postnatal tracheoscopy (11%) and postnatal needle puncture (6%).68

Ruano et al.69 also conducted an RCT evaluating FETO at 26–30 weeks in severe isolated CDH, defined as LHR <1.0 with intrathoracic liver herniation. In the intention-to-treat (ITT) analysis, survival was significantly increased in the intervention group (50% (10/20) vs. 5% (1/21), p<0.01). Multiple single-center feasibility studies of FETO have since been published reporting survival of 45%–50%,70–73 with two North American centers reporting survival rates >80% after FETO.74 75 Possible explanations for varying survival rates include differences in o/e LHR selection criteria, with inclusion of less severe cases, and differing postnatal management, including ECLS use.

Tracheal Occlusion to Accelerate Lung Growth trials

Initial studies highlighting the potential benefit of FETO were limited by small sample sizes, lack of contemporary controls, varying prenatal selection criteria and postnatal CDH care protocols. In response, the international TOTAL trials were conducted, evaluating the role of FETO in moderate and severe isolated LCDH, as defined by o/e LHR.76 77 Exclusion criteria included multiple gestations, maternal medical or anatomical conditions precluding fetal intervention, obstetric factors associated with increased risk of prematurity or psychosocial factors that could impact adherence to protocol. The primary outcome was infant survival to discharge from neonatal intensive care unit (NICU) and the moderate trial also included survival without oxygen supplementation at 6 months of age. FETO was performed at 27⁺⁰–29⁺⁶ weeks in the severe trial, compared with 30⁺⁰–31⁺⁶ weeks in the moderate trial, with the intention of reducing the risk of extreme prematurity in fetuses with less severe disease. FETO balloon removal was planned for 34⁺⁰–34⁺⁶ weeks in both trials.

The severe trial was stopped early as efficacy was demonstrated. In the ITT analysis, including 80 patients, survival was significantly better in the FETO group (40% vs. 15%; relative risk (RR)=2.67; 95% confidence interval (95% CI) 22 to 6.11, p=0.0009), which was maintained at 6 months of life. In the moderate CDH trial, FETO was not associated with a significant increase in survival (63% vs. 50%; RR=1.27; 95% CI 0.99 to 1.63) or in survival at 6 months without supplemental oxygen (54% vs. 44%; RR=1.23; 95% CI 0.93 to 1.65). In both trials, FETO was associated with a significantly increased risk of prematurity and PPROM. In the severe group, PPROM occurred in 47%, compared with 11% of controls (RR=4.51; 95% CI 1.83 to 11.9), and preterm birth occurred in 75% of FETO infants compared with 29% of those managed expectantly (RR=2.59; 95% CI 1.59 to 4.52), with a median GA at delivery of 34.6 (32.2–36.6) weeks versus 38.4 (36.5–39.1) weeks. Similar findings were seen in the moderate CDH trial, with an increased risk of PPROM (44% vs. 12%; RR=3.79; 95% CI 2.13 to 6.91) and an earlier median GA at delivery of 35.9 (34.3–37.9) weeks versus 38.1 (37–38.9) weeks, despite later FETO intervention. Although there were no obvious differences in complications related to prematurity, the trials were not powered to evaluate these secondary outcomes. Nearly 40% of balloon removals were unplanned, mostly in the context of preterm labor and/or PPROM. Most balloons were removed by fetoscopy (more than 95% of planned removals and approximately 40% and 70% for emergent retrievals in the severe and moderate trials, respectively).76 77

The authors highlight several limitations and controversies of the trials, including the decade-long study period during which neonatal and postnatal care protocols may have changed, the lack of blinding of investigators, and that the trials were not powered to evaluate neonatal morbidity or rare maternal, fetal/pediatric complications.76 77 Furthermore, as postnatal care was provided in 46 centers (some with very low patient volumes), it is likely that institutional practice and experience may have contributed to differences in outcomes, despite the utilization of a standardized postnatal care protocol.78 The high overall incidence of non-repair in the severe trial (47% and 63% in the FETO and control groups, respectively), as well as the low rate of ECLS use (5%–20%) raised questions about the true severity of disease, variability in center experience and implementation of postnatal care protocols, and perhaps an unmeasured impact of prematurity on treatment and outcome within the FETO groups.79

Fetal and maternal safety of FETO

Within the trials, five spontaneous balloon deflations and one failed insertion were reported.76 77 There were three problematic balloon removals, with two failed removals resulting in NND, one in a non-compliant mother who returned to an inexperienced center against medical advice and another following apparent balloon puncture under direct visualization, which at autopsy was found to be intact. Tracheomalacia was diagnosed in three infants postnatally.

FETO was not associated with any major maternal complications. In the report from the European consortium, maternal complications were minimal: only one mother requiring a blood transfusion and chorioamnionitis occurring in 2% of cases, all of which were successfully treated with antibiotics with no maternal sepsis.68 In the TOTAL trial, there was one NND due to a placental laceration,77 otherwise no major maternal complications were noted. Similarly, a systematic review evaluating complications in more than 9000 fetoscopies reported a risk of 6.0% for minor complications and 1.7% for major complications.80 Among 634 FETO procedures, the most common complication was chorioamnionitis affecting 1.1% of pregnancies.80 When subsequent fertility, pregnancy and gynecological outcomes were evaluated in CDH pregnancies that were either managed with FETO or expectantly,81 there were no significant differences based on self-reported survey data. However, an impact on mental health was similarly noted in both groups.

Online supplemental file 3 summarizes the clinical experience to date with FETO.

Generalizability of the TOTAL trials

The applicability of FETO in North America has been questioned, as some US centers report better survival with expectant management of fetuses with severe pulmonary hypoplasia when compared with the TOTAL trial (71% vs. 15%).79 Lack of standardization in prenatal prognostication by imaging may contribute to variability in the estimation of severity of pulmonary hypoplasia and, accordingly, postnatal survival between sites.20 42 Data from Toronto,26 which contributed the most North American cases to the TOTAL trials (n=27), reported survival in expectantly managed fetuses with severe CDH comparable to the European FETO consortium68 and the TOTAL trials.76 77 However, more recent data from 10 North American Fetal Therapy Network (NAFTNet) centers evaluating 63 fetuses with severe LCDH undergoing FETO compared with 43 expectantly managed cases found no significant difference in 6-month survival (69.8% vs. 58.1% for FETO and expectantly managed cases, respectively, p=0.30). Although FETO patients had higher rates of PPROM (54% vs. 14%, p<0.001), prematurity (35 weeks vs. 38.3 weeks at delivery, p<0.001), and lower birth weights (2487 g vs. 2857 g, p=0.001), the observation of a shorter duration of ECLS (9 days vs. 17 days, p=0.014) in the FETO group may indicate a pulmonary hypoplasia benefit.82 Notably, in this cohort, the expectantly managed group had a higher survival compared with the TOTAL trial (68% vs. 15%). Although this was a non-randomized study, the possible difference in survival may be related to a more aggressive management approach to CDH care in North America, with higher rates of ECLS utilization, and lower non-repair rates (13% of FETO patients and 21% of expectantly managed infants,82 compared with 47% and 63%, respectively, within the TOTAL trial for severe LCDH).77 A recent retrospective European study of fetuses with severe isolated LCDH (o/e TFLV ≤35% and intrathoracic liver herniation on MRI) compared 147 cases managed expectantly in a high-volume center to 47 FETO cases at three other centers.83 Expectantly managed fetuses had lower o/e TFLV (20.6%±7.5% vs. 23.7%±6.8%, p=0.013), later GA at delivery (37.4 weeks vs. 35.1 weeks, p<0.001) and required ECLS more frequently (55.8% vs. 4.3%, p<0.001) without increase in morbidity. Of note, survival to discharge and at 2 years was higher in expectantly managed cases compared with FETO cases (74.3% vs. 44.7%, p=0.001 and 72.8% vs. 42.5%, p=0.001, respectively). A systematic review looking at postnatal care setting and mortality after FETO also demonstrated a significant survival benefit for FETO performed in integrated settings (i.e., prenatal and postnatal care in the same program) (odds ratio (OR)=2.97, 95% CI 1.69 to 4.26), and the strongest determinant of postnatal survival was the availability of ECLS (OR=18.8, p=0.049).84

These studies highlight the need for establishing high-volume CDH centers with the ability to provide both FETO and high-quality postnatal standardized treatment including access to ECLS.83 84 Although the TOTAL trials have defined a new standard of care for prenatal therapy in severe CDH, this intervention should only be offered in experienced fetoscopic centers as defined within the TOTAL trials (ie, 36 fetoscopies per year with a minimum of 15 FETO cases prior to trial participation) to ensure optimal outcomes.76 77 A recent consensus statement (endorsed by NAFTNet and the International Fetal Medicine and Surgery Society) recommends that FETO be offered only in level III fetal therapy centers (ie, full range of minimally invasive and open fetal interventions) with a functional linkage to expert, multidisciplinary postnatal care in a level IV NICU or pediatric intensive care unit with ECLS availability.15 For provision of FETO, this would require expertise in prenatal prognostication and fetoscopy, 24/7 availability for emergent balloon removal including EXIT procedure and consistent use of standardized postnatal care protocols.15 85 86

Early versus late TO

When in vivo studies evaluated lung volumes in fetuses with CDH, TO induced a net increase in lung expansion of up to 1.5% per week; GA at occlusion was the only independent predictor of lung response, with earlier occlusion resulting in a better response, particularly if <30 weeks’ GA.87 These findings may explain the differences in treatment effects between the TOTAL trials, as those in the moderate trial underwent treatment at a later GA and consequently had a shorter duration of occlusion (24 days (19–28) vs. 34 days (28–39)).76 77 Not surprisingly, less lung growth was noted in the moderate group, as determined by a lower o/e LHR following FETO (32% vs. 67% in the moderate and severe groups, respectively).76 77 When data from both trials were pooled to study the heterogeneity of the treatment effect by o/e LHR and GA at balloon insertion, the effect of FETO was not dependent on o/e LHR for any endpoint, including survival to discharge.88 The authors concluded that the discrepancy in results among the trials was likely due to differences in timing of balloon insertion.88

Prematurity, neonatal and long-term morbidity associated with FETO

GA at delivery among infants with CDH has been shown in several studies to be an independent predictor of survival.18 53 68 Data from CDHSG demonstrated increasing survival with later GA at delivery: 32% if ≤28 weeks, 42% if 31–32 weeks and 73% if ≥37 weeks.89 A recent multicenter study, evaluating survival among 53 infants with CDH delivering at <32 weeks, reported an overall survival of 43%.90 The impact of prematurity on mortality appeared to be most pronounced in the subset of LCDH fetuses with o/e LHR <35%90 with lower survival rates compared with historical controls.18 The impact of prematurity on CDH mortality may be related to ECLS being relatively contraindicated in the extreme preterm population. However, the benefit of ECLS in improving CDH mortality remains controversial,10 17 with a survival advantage restricted to high-risk patients in high-volume centers.91

Among fetuses with severe CDH undergoing FETO, GA at delivery and interval between balloon removal and delivery were significantly greater in survivors compared with non-survivors.53 Furthermore, GA at delivery was predictive of early neonatal morbidity, including duration of assisted ventilation, days on supplemental oxygen, age at full enteral feeding and the need for oxygen for at least 28 days.53 In contrast, a recent study of isolated CDH undergoing FETO (n=84) demonstrated that the most important factor associated with survival was pulmonary response after FETO, whereas delivery <32 weeks was not associated with survival after controlling for lung response.91

Although the TOTAL trials could not address the impact of FETO on neonatal morbidity, observational data suggest that FETO may improve neonatal outcomes beyond survival. In a study of 90 fetuses with severe CDH undergoing FETO, neonatal morbidity was comparable to expectantly managed fetuses with less severe or moderate disease when delivered after 34 weeks.53 Other series have reported lower ECLS rates after FETO.74 82 However, more recent studies highlight the potential morbidity among FETO patients, which may be related to more severe disease, rather than FETO itself, and include prosthetic patch repairs in 75%–90% of cases,53 76 77 gastroesophageal reflux, tube feeding, dependency and need for prolonged supplemental oxygen and bronchodilators and/or corticosteroids.92 93

Predicting FETO responders

Various predictors of survival following FETO have been reported, including lung size prior to FETO as measured by LHR and o/e LHR by US and o/e TFLV by MRI,68 94 95 the pulmonary response after intervention,91 95 96 timing of intervention,87 97 duration of TO,87 96 interval between balloon removal and delivery53 and GA at delivery.53 68 Fetal pulmonary response following FETO at 26–30 weeks is demonstrable within 2 weeks and is maximal by 4 weeks, with US determined TFLV and pulmonary vascularization index at 4 weeks post-FETO performing best for prediction of survival (area under the curve 0.98).96 Similarly, MRI determined o/e TFLV pre-FETO and 3.3 weeks post-FETO were independent predictors of survival95 (figure 2). A post-FETO increase of <10% in o/e TFLV by MRI was associated with a significant increase in mortality, both at discharge and at 1 year of life, and ECLS use.98

Request permission

Right congenital diaphragmatic hernia

Due to its rarity, the data for RCDH are less robust regarding the role of FETO. In a recent large multicenter European FETO consortium study which included 214 fetuses with isolated RCDH, 128 of whom underwent FETO, a significant improvement in survival was seen following FETO (15% vs. 41% among expectant and FETO groups, respectively, p=0.014) in severe RCDH as defined by o/e LHR ≤45%.99 Despite an increased risk of PPROM (24% vs. 2%) and lower GA at delivery (34.4±2.7 weeks vs. 36.8±3 weeks) in the FETO group, neonatal morbidity was not significantly different between groups. The optimal o/e LHR threshold for prediction of survival at discharge was 50% in this series and may be used to identify candidates for FETO.

FETO technique

Improvement in instrument technology, including miniaturization of endoscopic instruments as well as advances in occlusive devices, may help reduce the risk of prematurity with FETO.85 Prolonged procedure times have been associated with an increased risk of PPROM,68 100 which can be improved by optimizing fetal positioning prior to intervention and concentrating procedures in regional centers with requisite expertise.15 Baschat et al. reported lower PPROM rates and later GA at delivery, which they attributed to modifications in FETO technique, including preferential trocar insertion in the upper uterus, FETO reversal by US-guided needle puncture whenever feasible, routine use of vaginal progesterone after procedure and a low threshold for amnioreduction for polyhydramnios.75

Unplanned balloon removal is common, with emergent balloon removal necessary in 28–56% of cases,68 ,101which introduces additional challenges for the fetal therapy team. Difficult balloon removal contributed to an NND in 5% of cases in the initial FETO experience reported by the European consortium.68 In a report of over 300 balloon removals, almost one-third were emergent.101 Furthermore, comparing outcomes between FETO and non-FETO centers, 3/9 removals in non-FETO centers were unsuccessful with resultant NNDs.101 In contrast, all balloon removals in FETO centers were successful,101 which can be attributed to the availability of an experienced emergency airway team throughout the TO period, typically consisting of otolaryngology or pediatric surgery, maternal fetal medicine (MFM) and neonatology providers. To minimize the challenges with balloon removal, the ‘smart’ tracheal occlusion (smart-TO) balloon, which incorporates a magnetic valve that can deflate spontaneously when proximate to an MRI scanner, has been introduced.102 Although this minimally invasive removal technique holds great promise, failed deflations have been reported.103 In vivo studies are ongoing to evaluate the efficacy of smart-TO91 and may improve the safety and availability of FETO by avoiding difficult retrievals and a second fetal intervention. A significant hurdle to the availability of smart-TO may be the increasingly stringent criteria for regulatory approval, which has recently affected the availability of several well-established and proven fetal devices.104

In utero therapy: what’s next?

Given the surgical risks associated with FETO, additional strategies to promote both lung development and pulmonary vascular remodeling are being explored.85 86 These less invasive novel treatments may be used in isolation or in conjunction with FETO.

Transplacental pharmacotherapy

Sildenafil, a selective inhibitor of phosphodiesterase type 5, has vasodilatory and antiremodeling effects on the pulmonary circulation and is used for treatment of pulmonary hypertension in CDH postnatally.105 However, human trials evaluating the role of antenatal sildenafil in CDH106 were stopped due to data from a clinical trial, suggesting an increased risk of pulmonary hypertension postnatally among growth-restricted infants exposed to sildenafil in utero.107 The role of treprostinil, a prostacyclin analog, in the antenatal treatment of pulmonary hypertension108 is being evaluated with ongoing preclinical studies.85

Advances in epigenetics and stem cell-based therapies

Increasing data are highlighting the importance of microRNA (miRNA), small non-coding RNAs that regulate the expression of genes, in the pathogenesis and prognostication of prenatally diagnosed CDH.109 In animal studies, miR-200b deficient mice were found to have stiffer lungs with abnormal distal airway branching and thicker alveolae, suggesting that miR-200b is necessary to regulate distal airway development.110 Prenatal miR-200b therapy has also been shown to reduce the incidence of CDH and rescue lung hypoplasia.111 Furthermore, higher levels of miR-200b have been demonstrated in tracheal fluid of CDH survivors after FETO.112 Circular RNAs, upstream regulators of miRNAs, also have an altered expression profile in hypoplastic CDH lungs compared with controls and may represent relevant biomarkers for CDH prognostication.107 Antounians et al. have demonstrated the ability of extracellular vesicles derived from amniotic fluid stem cells (AFSC-EVs) to promote lung regeneration through the release of RNA in rodent models of pulmonary hypoplasia.113 AFSC-EVs can also improve markers of pulmonary vascular development in rodent CDH models.114 Data regarding the optimal timing, dosage and route of administration and potential risks are required prior to clinical use.115