The role of ECMO in the management of congenital diaphragmatic hernia

https://doi.org/10.1053/j.semperi.2019.07.005Get rights and content

Abstract

Congenital diaphragmatic hernia (CDH) is the most common indication for extra-corporeal membrane oxygenation (ECMO) for neonatal respiratory failure. CDH management is evolving with advanced prenatal diagnostic imaging modalities. The risk profiles of infants receiving ECMO for CDH are shifting towards higher risk. Many clinicians are developing and following clinical practice guidelines to standardize and optimize the care of CDH neonates. Despite these efforts, there are significant differences in the practice patterns among ECMO centers as to how and when they choose to initiate ECMO for CDH, when they believe repair is safe, as well as many other nuances that are based on center experience or style. The purpose of this report is to summarize our current understanding of the new and recent developments regarding management of infants with CDH managed with ECMO.

Introduction

CDH is a condition in which the natural barrier between the abdomen and thorax is missing. The resulting pulmonary parenchymal compression affects formation of both lungs at the structural level during fetal development. This compression, in addition to the early embryologic insult, results in infants born with a varying range of hypoplastic and poorly developed lungs. In utero, there is minimal consequence for the developing fetus as placental and fetal circulation allow for normal levels of gas exchange. After birth, immature lungs are not able to meet the demands of adult type circulation. This leads to hypoxia, which further exacerbates pulmonary vascular resistance and causes progressive respiratory failure. If not treated with mechanical ventilation and other interventions, this is a lethal problem for a newborn.

ECMO is a rescue therapy that maintains cardiac and pulmonary function, allowing recovery from a reversible respiratory problem. Neonates with CDH have varying degrees of pulmonary insufficiency, and in the most severe cases infants receive ECMO to support cardiopulmonary function. Up to 30% of neonates with CDH receive ECMO.1, 2, 3 As is the case with all procedures, ECMO risks and benefit must be carefully considered. As reported by the Extracorporeal Life Support Organization (ELSO) Registry, there are about 250–300 infants with CDH per year who go on to develop respiratory failure significant enough to receive ECMO.4 Roughly half of these infants do not survive.4 The remaining survivors have a high degree of long-term morbidity such as poor neurodevelopmental outcomes.5, 6 With ongoing clinical and basic science research, there is a significant effort to improve the care of these infants born with CDH. In this review, we sought to highlight those areas where there is alignment in patient care as well challenging aspects with differing opinions on the management of CDH and ECMO.

The early days of neonatal respiratory ECMO centered on the management of persistent pulmonary hypertension and a series of randomized trials followed to test the benefits of this new intervention. The first prospective randomized trial of ECMO in neonatal respiratory failure was conducted by Bartlett et al. who involved an adaptive design using a randomized “play-the winner” method. The trial only had one control patient who died; all survivors were in the ECMO group.7 There were significant commentaries that followed which criticized the trial, given there was only a single control patient. The UK collaborative conducted a larger trial and showed further benefit towards ECMO in neonates demonstrated by greater survival and lesser degrees of neurodevelopmental disability at 1 year of age.8 Globally, clinicians soon realized that ECMO regularly resulted in survival in neonatal respiratory failure and many surviving neonates grew up to have a comparable quality of life, which therefore led to wide adoption of ECMO.

The first use of ECMO for CDH was described by German in 1977.9 This report described repair of CDH on ECMO resulting in the survival of 1 out of 4 infants. There has never been a randomized trial conducted to determine whether the use of ECMO in CDH is beneficial. This is notable because CDH carries the greatest mortality of all common neonatal conditions receiving respiratory ECMO.10 To date, the UK collaborative trial had the largest number of infants with CDH, where 14/18 in the ECMO group died, and 17/17 in the conventional arm died.8 Mortality among infants with a primary diagnosis other than CDH was 21% in the same study.8 Of note, this study was conducted in the late 1980s and survival rates have improved since that time.11

In 1997, “A Tale of Two Cities” reported the comparative experience between Boston Children's Hospital and The Hospital for Sick Children, Toronto. According to this report, neither high frequency ventilation nor ECMO significantly improved outcome.12 Postnatal management strategies have evolved significantly since then, with a strong emphasis on protocolized practice guidelines, postnatal stabilization, gentle ventilation, and multi-modal treatment of pulmonary hypertension. ECMO is presently utilized for the most severe cases when pre-specified limits of medical management have been reached.

Approximately 68% of neonates with CDH are diagnosed prenatally.13 Prenatal prognostic metrics to assess the severity of pulmonary hypoplasia and pulmonary hypertension have been developed. Antenatal parameters such as Lung-Head-Ratio (LHR), observed to expected LHR (O/E LHR), total fetal lung volume (TFLV), observed to expected TFLV (O/E TFLV), percent predicted lung volume (PPLV), and liver to thorax ratio (LiTR) define the degree of pulmonary hypoplasia, while pulmonary arterial hypoplasia/hypertension can be estimated by the McGoon's index or modified McGoon's index. These prenatal factors along with herniation of liver, position of stomach, and lung/liver signal intensity ratio (LLSIR) have been shown to be associated with postnatal survival14 and are addressed elsewhere in this edition of the Seminars.

There are studies specifically examining prenatal predictors of need for ECMO in the CDH population. In a recently completed meta-analysis, Russo et al. found that LHR <1 [RR 1.65 (95%CI 1.27–2.14)], O/E LHR measured by US [pooled standardized mean difference (SMD) −0.73 (95%CI −1.07 to −0.2.14)], and O/E TLV measured by MRI [SMD −1.00 (95%CI −1.52 to −0.48)], as well as liver herniation [RR 3.04 (95%CI 2.23–4.14) significantly predicted the need for ECMO.15 By itself, intrathoracic position of the liver is also highly associated with ECMO use. Hedrick et al. showed that 80% of fetuses with liver up on prenatal imaging would go on to require ECMO, compared to 25% of fetuses with liver down.15, 16 There is heterogeneity in these measurements and ECMO utilization criteria among different institutions, which may be where some the variability stems from for these measurements. It is certainly reasonable for institutions to self-determine their own cut-off where ECMO is very likely and which infants will benefit from birth at or near the ECMO center.

In addition to lung size and liver position, fetuses with CDH must be assessed for associated structural and genetic anomalies, as the existence of these anomalies could significantly impact the decision to initiate ECMO as well as the overall outcome of the fetus. The ECMO utilization rate for CDH reported to ELSO with concomitant chromosomal abnormalities is rare,17 yet this demonstrates that it is not an absolute contraindication, and should be a multi-disciplinary institutional decision with a thoughtful informed consent process. The congenital diaphragmatic hernia composite prognostic index (CDH-PI) takes into account numerous antenatal variables and was found to be predictive of not only postnatal mortality, but also the need for ECMO.18 The CDH-PI consists of a scoring system of 10 parameters that includes karyotype abnormalities, syndromic features, presence of congenital heart disease, left ventricle/right ventricle proportion, modified McGoon's index, presence of hernia sac, liver herniation, LHR, TLV, and PPLV. The authors found that a CDH-PI score of >8 was associated with improved survival (89% vs. 38% for infants with a CDH-PI score of <8). Furthermore, decreasing CDH-PI scores were associated with increasing need for ECMO, where 63% of neonates with CDH with a score of <6 were treated with ECMO, while 75% of those with a score of <5 received ECMO therapy.

Ex utero intrapartum treatment (EXIT) to ECMO has been proposed as a treatment for severe CDH, i.e. those infants who have an extremely high mortality or ECMO risk based on prenatal imaging. This involves cannulating the infant and instituting ECMO therapy while the infant remains on placental support. The theoretical advantages of such an approach are avoidance of barotrauma, avoidance of hypoxia, minimizing the affects hemodynamic instability and end-organ damage from shock. EXIT to ECMO was initially described by Kunisaki et al., for infants with liver herniation, LHR<1.4 and predicted lung volume <15%. In this study, the EXIT procedure was performed in conjunction with a 20 min ventilation trial. Of the 14 infants in the study, 11 underwent EXIT to ECMO, where 4 died and 7 survived. The results of this study suggested EXIT to ECMO was a feasible option. However, a second analysis by Stoffan et al., didn't support these results as 5/10 infants survived without EXIT to ECMO compared to 2/6 who survived after EXIT to ECMO.19 The last study that evaluated EXIT to ECMO for CDH sought to determine if there were morbidity differences in survivors.20 In this study, 8 EXIT to ECMO survivors were compared to 9 non-EXIT survivors and all had mild motor or speech delay without any differences between the groups. These data do not support leveraging EXIT to ECMO as a treatment option for high-risk infants given equivocal outcomes. Based on the data from these studies, EXIT to ECMO is not routinely practiced nor generally recommended.

Initial management of newborn infants with CDH has significantly evolved over the last few decades and multiple collaboratives have adopted clinical practice guidelines (CPGs).21, 22 Currently, the CDH Study Group is also developing a proposed overarching CPG. Overall, the essential components of CPGs include a well-coordinated initial resuscitation in the delivery room, followed by a protocolized assessment and resuscitation within the neonatal intensive care unit. After delivery, with near term vaginal delivery accepted as safe for infants with CDH, initial resuscitation efforts should include standard newborn care, a secured airway, and alimentary tract decompression with a large orogastric tube. Once transported to the neonatal intensive care unit, a chest radiograph, head ultrasound and an echocardiogram should be obtained. Adequate venous and arterial access must be established. Gentle ventilator management, with pre-specified limits, is employed to avoid barotrauma, volutrauma, and atelectrauma. Preductal saturations and blood gases should be trended and utilized for decisions regarding escalation of support.

Neonates with CDH can progress rapidly from acceptable hemodynamic parameters to cardiopulmonary failure. For neonates with tenuous respiratory status, the use of high frequency oscillatory ventilation (HFOV) is a common next step. It is worth noting that a recent trial was not able to demonstrate a difference in mortality or development of bronchopulmonary dysplasia between conventional ventilation and HFOV.23 If preductal saturations remain <80%, inhaled nitric oxide (iNO) may be started, though strong data question its efficacy. Although studies have not shown iNO to either avert ECMO or improve survival in the CDH patient population,2, 24 it continues to be commonly used.25, 26, 27 iNO use can be considered for severe pulmonary hypertension with close assessment of patient response, or simply as a bridge to ECMO.28, 29 Based on available evidence, use of iNO in the setting of left ventricular dysfunction may have deleterious effects, such as non-response to vasodilators and occurrence of pulmonary hemorrhage.30, 31 When these collective strategies fail to adequately address severe physiologic derangements of hypoxia, acidosis, hypercarbia, and/or hypotension, ECMO should be considered (Table 1).

Of course, there is institutional variation on ECMO utilization for CDH.32 Prenatal criteria as discussed above can be useful in predicting severe disease and need for ECMO in order to marshal appropriate resources and help guide initial management. In addition, Jancelewicz et al. have described an ECMO prediction model using postnatal data.33

ECMO is used in infants with CDH experiencing deteriorating clinical status, frequently secondary to pulmonary hypertensive crisis—in theory, a reversible condition. As noted, the indications for ECMO are not uniform. Relative indications for ECMO have been studied by multiple investigators and include elevated oxygenation index (OI), persistently low oxygen saturations despite maximal ventilator assistance, hypercarbia, and elevated alveolar-arterial oxygen gradient (A-aDO2) (Table 2).

The Conventional Mechanical Ventilation Versus High-Frequency Oscillatory Ventilation for Congenital Diaphragmatic Hernia Trial (The VICI-trial) utilized the following predetermined failure criteria as possible indications for ECMO if they were met at two consecutive time points for at least 3 h: 1 - inability to maintain preductal oxygen saturation above 85% (52 mmHg or 7 kPa) or post ductal saturations above 70% (40 mmHg or 5.3 kPa); 2 - increase in PaCO2 > 65 mmHg or 8.5 kPa despite optimization of ventilatory management; 3 - peak inspiratory ventilator pressure (PIP) > 28 cm H2O; 4 - mean airway pressure (MAP) > 17 cm H2O; 5 - inadequate oxygen delivery with metabolic acidosis defined as lactate ≥ 5 mmol/L and pH < 7.20; 6 - hypotension resistant to fluid therapy and inotropic support resulting in a urine output < 0.5 ml/kg/hour; 7 - oxygenation index ≥ 40.23 Overall these indications provide a good summary of when ECMO support may be provided for neonates with CDH. However, all are subject to numerous, patient-specific factors and actual patients may meet these criteria for a variable period of time, followed by periods of stability, which can make the decision to initiate ECMO difficult. Other patients may have a more dramatic deterioration making the decision to provide ECMO support the only option.

Contraindications for ECMO support for CDH are also changing as the field is evolving.34 The generally accepted contraindications for ECMO are: 1 – significant congenital anomalies (e.g. severe cardiac lesions) and lethal chromosomal disorders, 2 - irreversible brain damage, 3 - uncontrolled bleeding; and 4 - intraventricular hemorrhage grade III or greater. Other relative contraindications for infants with CDH include: 1 - <2 kg in weight; 2 - <32–34 weeks gestational age; and 3 - high probability of a poor prognosis (Table 3).

The two methods of delivering ECMO support to the newborn with CDH are venoarterial (VA) and venovenous (VV). VA ECMO is most commonly performed by cannulation of the right carotid artery and jugular vein. VV ECMO is accomplished through placement of a double lumen (DL) cannula into the right internal jugular vein. VV cannulation is not always possible if the jugular vein is not adequate in size to accommodate the smallest VVDL cannula (currently 13 French). VV cannulation is also not an option with extremely poor cardiac function or for extracorporeal cardiopulmonary resuscitation.35 Mode of cannulation is most often dictated by institutional preference or the perceived degree of illness of the neonate.32

A number of studies have compared VV versus VA ECMO for CDH and none have reported a difference in mortality.4, 36, 37 These studies were criticized for not being able to adequately control for disease severity or selection bias.38 To address this, recently a propensity score-based analysis of the ELSO registry was carried out from the years 2000–2016. The propensity score matching identified 3304 infants (VA = 2470, VV = 834) and the odds of death were not significantly different the two groups (OR = 1.01, 95% CI: 0.86–1.18, P = 0.95).39 Neither was a difference seen in the odds of severe acute neurologic events between the matched VV and VA patients.39 Subgroup analysis of pre-ECMO CDH repair compared 175 VA with 70 VV cases. In this group, the odds of death for the VV group was more than doubled compared to the VA group (OR = 2.10, 95% CI: 1.19–3.69; P = 0.01). There was no difference in odds of severe neurologic complications within the two groups undergoing pre-ECMO CDH repair (OR = 1.48; 95% CI: 0.59–3.71; P = 0.39). Since this was a subgroup and the number of patients are much smaller, the significance of this finding in the pre-ECMO repair group remains unclear. Other subgroups based on timing of repair were also analyzed and no differences in mortality or acute severe neurologic events were noted.39

Clearly, these studies all have limitations related to the source of data which can limit matching methodologies. Therefore, conclusions drawn are not equal to a well-designed randomized controlled trial. What can be recommended based on the evidence are the following: 1 - many prefer VA ECMO and have a bias for VA ECMO when the neonate is deemed to be more critical; 2 - VA ECMO should be used in cases of sudden arrest or extremely poor cardiac function; and 3 - assuming vessel size is not the limiting factor, VV and VA have similar outcomes and the modality selected should be based on individual and intuitional experience or preference. Table 4 summarizes the advantages of these two modalities of ECMO support.

In the setting of VV ECMO, some centers routinely place an additional cephalad jugular venous catheter.40 The reasoning behind this is to reduce cerebral venous pressure and decrease the risk of intracerebral hemorrhage. Based on ELSO data, this practice has not been shown to decrease the risk of neurologic complications or survival for VV cannulation in the neonate with CDH.41 Thus, the decision to place a cephalad drainage catheter should be based on individual patient factors, equipment availability and center experience.

Due to the complex of anatomic challenges in infants with CDH, cannulation can be more difficult compared to ECMO cannulation for neonatal respiratory failure for other reasons (Fig. 1). For example, infants with CDH have been described to have hypoplastic neck vasculature which may limit the size of the cannula, rendering VV challenging, high-risk, or impossible.42 Bicaval double lumen cannulas are associated with significant challenges for accurate placement and complications during cannulation have been reported.43, 44 Right-sided CDH presents additional challenges given their unique thoracic anatomy and mediastinal shift due to liver position in the right thorax. As a result, the venous cannula may preferentially enter the ostium of an ectatic azygous vein, which can be difficult to recognize and correct.45, 46 Beyond these anatomic challenges, there is no evidence that VV or VA cannulation is superior in infants with right-sided CDH.4, 39 A recent APSA (American Pediatric Surgical Association) survey study revealed a preference for VA ECMO in right-sided CDH patients, which may be due to the perception of increased severity of right-sided CDH.32

ECMO circuits are powered by one of two types of pumps: roller or centrifugal. The roller pump works by compressing the tubing, thereby propelling blood forward via positive displacement; simultaneously, blood from the venous reservoir is pulled into the tubing due to negative pressure. The advantage of the roller pump is constant antegrade flow and reduced hemolysis at the low flow rates required in neonates. The disadvantage of the roller pump is that it will continue to rotate independent of the pressure in the circuit or the volume of blood in the venous reservoir. Thus, it requires servo-regulation mechanisms to ensure that circuit volumes and pressures do not exceed safe levels. Furthermore, wear on the tubing and even raceway tubing rupture have been reported.47

The centrifugal pump is being increasingly used in ECMO.48 This pump employs a spinning magnetic disc that creates a constrained vortex. Negative pressure is therefore generated, pulling blood into the pump and then out of the top of the vortex. The centrifugal pump eliminates repeated compression of the tubing, reduces the risk of excessive negative pressure, and decreases priming time and volume.49 The two main disadvantages of this pump are the difficulty in maintaining set flow and the increased incidence of red blood cell hemolysis from turbulent flow in the pump head vortex. Hemolysis can lead to hyperbilirubinemia, acute renal failure, and other end-organ damage.49

There is only one study examining the relationship between ECMO pump type and outcomes in the CDH population. Using the ELSO registry, Delaplain et al. compared roller pumps to centrifugal pumps with a propensity matched cohort. In their study, PS matching identified 1808 infants (centrifugal = 904, roller = 904). There was no difference in mortality or severe neurologic injury between the two groups.49 There was at least a six-fold increase in odds of hemolysis for centrifugal pumps in all groups.49 Despite these findings, many centers are using centrifugal pumps for neonatal ECMO.

The optimal time to perform surgical repair of the diaphragm after birth has long been an area of interest and considerable debate. Emergent repair was thought to be necessary for decades due to the belief that decompressing the lung by removing the herniated abdominal contents from the chest was crucial to improve respiratory distress. Several reports of successful preoperative stabilization prior to repair were reported in the 1980s, showing equivalent to improved survival.50, 51 Lung compliance was even shown to worsen after surgical repair rather than improving.52 By the 1990s, delayed repair after a period of stabilization had become a widely accepted approach.53, 54, 55, 56, 57 As described previously, a strategy of low pressure gentle mechanical ventilation with permissive hypercapnea is employed until sufficient stability for an operation has been demonstrated.58 Repair is commonly delayed for 24–72 or more hours.

Up to 30% of patients with CDH remain unstable despite optimal medical management and require ECMO cannulation.59, 60 There is no consensus on the ideal timing of surgical repair in the setting of ECMO. Arguments have been made for early repair on ECMO, late repair on ECMO, or waiting until the infant is able to be decannulated from ECMO prior to repair. As repair on ECMO entails performing the operation while the infant is fully anticoagulated, the risk of hemorrhage during or after repair is central to this discussion.61 Proponents of waiting to repair until after ECMO have argued that not only is the risk of significant bleeding less, but that repair itself has not been shown to improve pulmonary hypertension or respiratory failure. The ELSO Congenital Diaphragmatic Hernia Interest Group and the Congenital Diaphragmatic Hernia Study Group have each reported a survival benefit to delaying CDH repair until after decannulation from ECMO.62, 63 Similarly, Robertson et al. reported a survival benefit for repair after ECMO when they evaluated their institutional experience.64 In this comparison of their evolving institutional strategy, survival in patients able to be decannulated prior to repair was 94.4% compared with just 43.3% in the early repair group, despite the early repair strategy being the more recent one adopted. Moreover, they found that the early repair strategy was associated with longer ECMO duration. Partridge et al. also found that in their institutional experience, repair after ECMO was associated with increased survival (100%), lower rates of surgical bleeding and decreased duration of ECMO compared with repair on ECMO (survival 43.9%).65 Of note however, 16 patients of 77 did not survive to repair. If repair confers any benefit, it is plausible that some of these patients may have survived if repaired earlier.

Early repair appeals to the intuition. By removing herniated contents from the chest and recreating normal anatomy, the compressed lungs may expand and contribute more to gas exchange and perhaps pulmonary hypertension may also thereby be improved. If bleeding could be avoided, early repair may thus reduce duration of ECMO. Those who favor this approach have also argued that bleeding risk may be less in the period immediately after ECMO initiation compared with weeks later when tissue edema may be worsened. Dassinger et al. reported that of 34 patients repaired on ECMO in their institution a median of 1 day after cannulation, 71% survived with an 8.8% rate of significant perioperative bleeding (and none of these leading to death).66 Similarly, when Fallon et al. compared their evolving strategies of early repair (<72 h), late repair (>72 h), and post-decannulation repair, they found a survival benefit to early repair (73% vs. 50% late, and 64% post) despite those repaired early having worse prenatal factors.59 Those repaired early were also found to have a shorter duration of ECMO by 6 days. Bleeding risk was similar. A more recent study out of the CDH Study Group also found that survival in those repaired early (<72 h), survival was 87%; however, in this study the duration of ECMO was longer.67

A strategy of late repair on ECMO can be viewed as a compromise between these two approaches. Repair is performed on ECMO, typically after the patient has been proven ready to decannulate. If all goes well, the patient may be decannulated the following day. If physiologic parameters worsen as a result of perioperative stress, ECMO is already in place as a safety-net, sparing the need for a second more risky cannulation in the contralateral neck (and the eventual ligation of bilateral carotid arteries and jugular veins after decannulation). On the other hand, if hemorrhage occurs during or after surgery, the patient in theory may be decannulated, obviating the need for anticoagulation. One problem with this approach is that not all patients will reach the point of readiness to decannulate soon after repair. This strategy also begs the question: how often is repair so hemodynamically destabilizing that re-cannulation to ECMO is required? A recent study indicates this may be less of concern than previously thought: of 668 patients in the CDH Study Group database who were repaired after decannulation, only 6 (0.9%) required recannulation.68

Thus, a consensus on the best time to repair the diaphragm in relation to ECMO has not been reached and practices vary.69 More work to help identify patients who would most benefit from early repair is needed. Coming full circle to original management strategies, Kays et al. have even argued recently that urgent early repair within the first few hours of life before anticipated ECMO may offer the best chance of survival in the most severe cohort of left-sided liver-up CDH patients.70

Finally, a recent investigation from the CDH study group evaluated 1581 patients with CDH who received ECLS, performing two comparisons: 1. repair on ECLS compared to repair after ECLS and 2. repair early on ECLS compared to repair later on ECLS. These groups were matched based on propensity score and, importantly, the study accounted for non-repaired patients. CDH repair on ECLS resulted in a 46% reduction in mortality compared to repair after ECLS while early repair on ECLS was associated with a 49% reduction in mortality compared to late repair on ECLS.71 While resource intensive and optimal with appropriate experience and expertise, it appears that early repair, on ECLS, affords the best survival for all patients on ECLS.

As discussed, the primary concern with surgical repair on ECMO is bleeding due to anticoagulation. Hemorrhagic complications can be fatal. Different strategies have been used to reduce this risk. First, the level of anticoagulation should be decreased temporarily at the time of surgery.72 Use of thromboelastography (TEG) to guide anticoagulation management perioperatively in CDH patients, already described in post-cardiopulmonary bypass patients, is an emerging area of study and could be useful in the management of anticoagulation during repair.73 Maintaining adequate platelet count, typically >100,000 perioperatively, has been described as another component of a successful perioperative strategy to limit bleeding.72

Addition of anti-fibrinolytic therapy has been shown to reduce surgical bleeding events on ECMO.74 A 10-year single-institution review found that use of Aminocaproic acid (Amicar) significantly decreased the rate of surgical site bleeding.75 In this protocol, Amicar was administered as a 100 mg/kg IV bolus around cannulation and continued at a 30 mg/kg/h infusion for 72 h or beyond if continued bleeding was encountered. A specific study looking at Amicar during on-ECMO CDH repair is lacking. Use of perioperative tranexamic acid (TXA) has been specifically studied in the CDH population.76 Using a protocol of 4 mg/kg bolus 30 min to repair and infusing 1 mg/kg/h for 24 h postoperatively, authors found significantly less surgical site bleeding and need for blood transfusions.76 Anti-fibrinolytic medications can also have thrombogenic effects on the circuit, which must be balanced with the needs of the patient. The condition of the circuit should additionally be considered before these agents are used.

The surgical techniques used during on ECMO repair can also contribute significantly to decrease bleeding risk. Although there are no studies published to describe differences in surgical technique, minimizing dissection of the posterior rim of diaphragm can decrease bleeding, as can liberal use of electrocautery and even argon beam coagulation. The use of a patch should also be considered, given primary repair of certain defects may require significant dissection of the posterior rim of the diaphragm. Temporary abdominal closure is another useful method to avoid abdominal compartment syndrome secondary to bleeding in these tenuous patients; this technique can also allow for packing of any bleeding sites.77 Topical hemostatic agents such as Surgicel®, thrombin, and Gelfoam® may also be considered. The use of pledgeted sutures may help attain hemostasis at suture sites.

Survival rate of infants with CDH treated with ECMO is 50%, a figure that has minimally changed over the last few decades. However, upon closer examination, the modern CDH population demonstrates a higher risk profile than in the past.11 We postulate that improvements in medical management have shifted the application of ECMO to the highest severity infants. Whether ECMO improves survival in infants with CDH remains a critical question. As discussed earlier, there has never been a randomized controlled trial specific to CDH and ECMO. Therefore, this question cannot be answered directly. However, evidence from the literature would suggest that ECMO does indeed help those infants who may not otherwise survive.78

Postnatal determination of mortality risk is a useful means to understand the trajectory of patients. There have been several risk models developed to determine risk of mortality in infants with CDH. These prediction tools are addressed in detail in this edition of Seminars. However, none of the established risk models can adequately predict risk of mortality once ECMO is required. Thus, on-ECMO mortality risk models have been developed. One such score is the Pittsburgh Index for Pre-ECMO Risk (PIPER+) specific for venoarterial (VA) ECMO, which can be used to predict on-ECMO mortality risk for infants with CDH needing VA ECMO.79 We have also developed an ELSO-based CDH-specific risk model to predict mortality before and during ECMO.17 The on-ECMO CDH score can be used to determine risk of mortality while on ECMO and is the most accurate risk model for CDH neonates needing ECMO.

The reasons why infants do not survive after initiation of ECMO is complex and is very center specific.80 The primary determinant of ECMO-specific mortality and development of complications is the duration of ECMO.79, 81, 82 Duration of ECMO support is individualized for each CDH patient and depends on their unique risk profile as well as the specifics of the treatments they have received. There are no established guidelines as to how long ECMO support should be provided, but the survival rates start to decline significantly beyond 4 weeks to 10–20%; beyond 5 weeks it is only <5–18%.17, 83

One prenatal therapy that shows promise in decreasing the need for postnatal ECMO in the CDH population is fetoscopic endoluminal tracheal occlusion (FETO). FETO has been shown to feasible and safe in the most severe CDH patients with LHR <1 and liver herniation.84 The use of FETO is being further investigated and its use may increase in the future.85 Other potential treatment strategies may include pumpless arteriovenous ECMO and the artificial placenta.86 These therapies may individually and/or collectively, represent breakthroughs in the management of this complex patient population. Lastly, other future therapies may involve in-utero delivery of growth factors, specific cell or sub-cellular material and gene therapy. Circuits are becoming more simplified and monitoring is becoming more sophisticated. In the future ECMO cannulation for infants with CDH may be done at non-ECMO centers then transported to an ECMO center using compact ECMO pump systems such as PediMag or CentriMag.87

As new therapies evolve for CDH, the need for ECMO may decrease and could eventually become obsolete. This is similar to the decline of ECMO utilization for neonatal respiratory distress syndrome. Nevertheless, it is currently a lifesaving intervention that allows time for pulmonary vasculature stabilization, novel therapy application, surgical decompression of the pulmonary parenchyma, and cardiopulmonary recovery. Thus, it is unlikely that ECMO will be completely replaced in the treatment of CDH. What is more likely is that the risk curves will augment further, and ECMO utilization will be targeted to the sickest CDH infants with highly hypoplastic lungs, refractory pulmonary hypertension, ineffective fetal tracheal balloon occlusion, or infants with concomitant complex cardiac malformations. We may also see a shift towards increasing use of ECMO for more premature infants with improved anticoagulation management and the development of smaller cannulas, as well as pumpless AV ECMO. Ongoing research efforts to standardize care and direct the effects of our therapies must remain at the forefront. These efforts will improve and affect future applications of ECMO therapy for CDH.

References (93)

  • P.T. Delaplain et al.

    Cannulating the contraindicated: effect of low birth weight on mortality in neonates with congenital diaphragmatic hernia on extracorporeal membrane oxygenation

    J Pediatr Surg.

    (2017)
  • R.A. Dimmitt et al.

    Venoarterial versus venovenous extracorporeal membrane oxygenation in congenital diaphragmatic hernia: the extracorporeal life support organization registry, 1990-1999

    J Pediatr Surg

    (2001)
  • A. Kugelman et al.

    Venovenous versus venoarterial extracorporeal membrane oxygenation in congenital diaphragmatic hernia

    J Pediatr Surg

    (2003)
  • D.W. Kays

    ECMO in CDH: is there a role?

    Semin Pediatr Surg

    (2017)
  • Y.S. Guner et al.

    Outcomes of infants with congenital diaphragmatic hernia treated with venovenous versus venoarterial extracorporeal membrane oxygenation: a propensity score approach

    J Pediatr Surg

    (2018)
  • E.D. Skarsgard et al.

    Extracorporeal life support organization R: venovenous extracorporeal membrane oxygenation in neonatal respiratory failure: does routine, cephalad jugular drainage improve outcome?

    J Pediatr Surg

    (2004)
  • B. Frenckner et al.

    Neonates with congenital diaphragmatic hernia have smaller neck veins than other neonates-An alternative route for ECMO cannulation

    J Pediatr Surg

    (2002)
  • J.C. Fisher et al.

    Challenges to cannulation for extracorporeal support in neonates with right-sided congenital diaphragmatic hernia

    J Pediatr Surg

    (2007)
  • J.C. Langer et al.

    Timing of surgery for congenital diaphragmatic hernia: is emergency operation necessary?

    J Pediatr Surg

    (1988)
  • H. Sakai et al.

    Effect of surgical repair on respiratory mechanics in congenital diaphragmatic hernia

    J Pediatr

    (1987)
  • B. Frenckner et al.

    Improved results in patients who have congenital diaphragmatic hernia using preoperative stabilization, extracorporeal membrane oxygenation, and delayed surgery

    J Pediatr Surg

    (1997)
  • J.M. Wilson et al.

    Evolution of the technique of congenital diaphragmatic hernia repair on ECMO

    J Pediatr Surg

    (1994)
  • S.C. Fallon et al.

    Repair of congenital diaphragmatic hernias on extracorporeal membrane oxygenation (ECMO): does early repair improve patient survival?

    J. Pediatr. Surg.

    (2013)
  • J.O. Robertson et al.

    Comparison of early versus delayed strategies for repair of congenital diaphragmatic hernia on extracorporeal membrane oxygenation

    J Pediatr Surg

    (2018)
  • E.A. Partridge et al.

    Timing of repair of congenital diaphragmatic hernia in patients supported by extracorporeal membrane oxygenation (ECMO)

    J Pediatr Surg

    (2015)
  • M.S. Dassinger et al.

    Early repair of congenital diaphragmatic hernia on extracorporeal membrane oxygenation

    J Pediatr Surg

    (2010)
  • D.W. Kays et al.

    Improved survival in left liver-up congenital diaphragmatic hernia by early repair before extracorporeal membrane oxygenation: optimization of patient selection by multivariate risk modeling

    J Am Coll Surg

    (2016)
  • A.A. Desai et al.

    Optimal timing of congenital diaphragmatic hernia repair in infants on extracorporeal membrane oxygenation

    Semin Pediatr Surg

    (2015)
  • C.D. Downard et al.

    Impact of AMICAR on hemorrhagic complications of ECMO: a ten-year review

    J Pediatr Surg

    (2003)
  • F.H. van der Staak et al.

    Surgical repair of congenital diaphragmatic hernia during extracorporeal membrane oxygenation: hemorrhagic complications and the effect of tranexamic acid

    J Pediatr Surg

    (1997)
  • P. Laje et al.

    Delayed abdominal closure after congenital diaphragmatic hernia repair

    J Pediatr Surg

    (2016)
  • T.H. Keshen et al.

    Does extracorporeal membrane oxygenation benefit neonates with congenital diaphragmatic hernia? Application of a predictive equation

    J Pediatr Surg

    (1997)
  • T. Jancelewicz et al.

    Risk-stratification enables accurate single-center outcomes assessment in congenital diaphragmatic hernia (CDH)

    J Pediatr Surg

    (2019)
  • D.W. Kays et al.

    Extracorporeal life support in patients with congenital diaphragmatic hernia: how long should we treat?

    J Am Coll Surg

    (2014)
  • D.W. Kays et al.

    Extracorporeal life support in patients with congenital diaphragmatic hernia: how long should we treat?

    J Am Coll Surg

    (2014)
  • P. Dekoninck et al.

    Results of fetal endoscopic tracheal occlusion for congenital diaphragmatic hernia and the set up of the randomized controlled TOTAL trial

    Early Hum Dev

    (2011)
  • E.A. Partridge et al.

    Pumpless arteriovenous extracorporeal membrane oxygenation: a novel mode of respiratory support in a lamb model of congenital diaphragmatic hernia

    J Pediatr Surg

    (2018)
  • J. Boloker et al.

    Congenital diaphragmatic hernia in 120 infants treated consecutively with permissive hypercapnea/spontaneous respiration/elective repair

    J Pediatr Surg

    (2002)
  • L.R. Putnam et al.

    Congenital diaphragmatic hernia defect size and infant morbidity at discharge

    Pediatrics

    (2016)
  • L.R. Putnam et al.

    Evaluation of variability in inhaled nitric oxide use and pulmonary hypertension in patients with congenital diaphragmatic hernia

    JAMA Pediatr

    (2016)
  • H.F. Shieh et al.

    Does the ex utero intrapartum treatment to extracorporeal membrane oxygenation procedure change morbidity outcomes for high-risk congenital diaphragmatic hernia survivors?

    J Pediatr Surg

    (2016)
  • R.H. Bartlett et al.

    Extracorporeal circulation in neonatal respiratory failure: a prospective randomized study

    Pediatrics

    (1985)
  • UK collaborative randomised trial of neonatal extracorporeal membrane oxygenation

    Lancet

    (1996)
  • R.P. Barbaro et al.

    Development and validation of the neonatal risk estimate score for children using extracorporeal respiratory support

    J Pediatr

    (2016)
  • Y.S. Guner et al.

    Trends in mortality and risk characteristics of congenital diaphragmatic hernia treated with extracorporeal membrane oxygenation

    ASAIO J

    (2018)
  • F.M. Russo et al.

    Lung size and liver herniation predict need for extracorporeal membrane oxygenation but not pulmonary hypertension in isolated congenital diaphragmatic hernia: systematic review and meta-analysis

    Ultrasound Obstet Gynecol

    (2017)
  • Cited by (38)

    • Management of the CDH patient on ECLS

      2022, Seminars in Fetal and Neonatal Medicine
      Citation Excerpt :

      A follow-up study of the ELSO registry also showed that centrifugal pumps had higher odds of hemolysis than roller pumps [49]. However, newer centrifugal pump designs have lessened the risk of hemolysis and therefore many centers utilize centrifugal pumps in their neonatal ECLS circuits [50]. Currently, early repair on ECLS or repair after decannulation (post-ECLS) are the most accepted options for timing of repair of the diaphragmatic hernia for neonates who require ECLS.

    • Long term outcome of babies with pulmonary hypertension

      2022, Seminars in Fetal and Neonatal Medicine
      Citation Excerpt :

      Pulmonary hypertension is almost a universal finding in the acute phase of this disease [2]. In some cases pulmonary hypertension and hypoxia may be severe with 30% requiring extracorporeal support [61]. Despite the advances in neonatal intensive care, survival rates for those requiring ECMO remain approximately 50–60%.

    • Diagnosis &amp; management of pulmonary hypertension in congenital diaphragmatic hernia

      2022, Seminars in Fetal and Neonatal Medicine
      Citation Excerpt :

      Of note, severe, early and disproportionate PH combined with high lactate levels and bilious nasogastric aspirates may indicate a rare complication of bowel compromise in CDH and may be an indication for immediate exploratory surgery and repair [111]. Extra-corporeal membrane oxygenation is a rescue support for cardio-respiratory failure, used globally in around 30% of CDH cases per year [112]. However, controversies remain in relation to patient selection, timing of initiation and timing of surgical repair on ECMO.

    View all citing articles on Scopus

    Sources of support: Pediatric Subspecialty Faculty Tithe Funding, Children's Hospital of Orange County.

    View full text