Survival rate for patients with congenital diaphragmatic hernia in 3 therapeutic eras. The survival rate for the third era was significantly higher than that for the first and second eras (P<.02 for era 3 compared with eras 1 and 2).
Management protocol and outcome (survival) for the first therapeutic era (1970-1983) of congenital diaphragmatic hernia. Overall survival was 42% (43 of 102 patients).
Management protocols and outcome (survival) for the second therapeutic era (1984-1988) of congenital diaphragmatic hernia. Overall survival was 58% (26 of 45 patients). ECMO indicates extracorporeal membrane oxygenation.
Management protocols and outcome (survival) for the third therapeutic era (1989-1997) of congenital diaphragmatic hernia. Overall survival was 79% (44 of 56 patients). EMCO indicates extracorporeal membrane oxygenation.
Weber TR, Kountzman B, Dillon PA, Silen ML. Improved Survival in Congenital Diaphragmatic Hernia With Evolving Therapeutic Strategies. Arch Surg. 1998;133(5):498–502. doi:10.1001/archsurg.133.5.498
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1998
To compare the survival rates for 3 therapeutic eras, each using different treatment strategies for the management of newborns with congenital diaphragmatic hernia (CDH).
Retrospective review of all infants with CDH from 1970 through 1997.
Tertiary care children's hospital.
A total of 203 newborns with CDH.
Extracorporeal membrane oxygenation (ECMO) was performed with arterial and venous cannulation connected to a membrane oxygenator–roller pump perfusion apparatus, using systemic heparinization. Delayed operative therapy involved operative repair 2 to 5 days after birth using preoperative ventilation support only. Since 1970, 203 newborns with CDH were managed in 3 therapeutic eras: era 1 (1970-1983, 102 patients) was immediate CDH repair with postoperative ventilator and pharmacologic support; era 2 (1984-1988, 45 patients) was immediate repair with postoperative ventilator support (18 patients), immediate ECMO with CDH repair on ECMO (4 patients), or immediate repair with postoperative ECMO (23 patients); and era 3 (1989-1997, 56 patients) was immediate ECMO with repair on ECMO (23 patients), immediate repair with postoperative ECMO (9 patients), or delayed (2-5 days) CDH repair (24 patients).
Main Outcome Measures
Survival, defined as discharge from the hospital, and morbidity.
Survival was 42% (43/102 patients) in era 1, 58% (26/45 patients) in era 2, and 79% (44/56 patients) in era 3 (P<.02 vs eras 1 and 2). In era 3, the survival for immediate ECMO with repair on ECMO was 57% (13/23 patients), 89% (8/9 patients) for immediate repair with postoperative ECMO, and 96% (23/24 patients) for delayed repair. Eight late deaths were caused by pulmonary hypertension (1 death), sudden infant death syndrome (1 death), and other causes (6 deaths). Morbidity in survivors included mild neurologic deficit (5 patients) and pulmonary disease (3 patients).
These data demonstrate a significant improvement in survival in CDH with preoperative ECMO and with delayed repair with and without ECMO support and suggest that immediate repair of CDH without the availability of ECMO support should be abandoned.
CONGENITAL diaphragmatic hernia (CDH) continues to have a high mortality rate despite advances in neonatal critical care. The traditional treatment scheme of emergency operative repair immediately after birth, if possible, resulted in mortality rates greater than 50% and has recently given way to other forms of management, such as delayed operation, extracorporeal membrane oxygenation (ECMO), inhaled nitric oxide, and partial liquid ventilation. Although analysis1 of multi-institutional registries suggest that some of these modalities may have resulted in improved survival, few series at single institutions with sufficient patient populations confirm this observation. The present review of the changing survival and morbidity with evolving management strategies during a 28-year period at a single children's hospital was undertaken to assess the effects of new therapies on outcome in CDH.
From 1970 to 1997, 203 newborns with CDH were treated at Cardinal Glennon Children's Hospital, Saint Louis University Health Sciences Center, St Louis, Mo. All newborns with CDH who arrived alive at the hospital were included in the study, while older (>48 hours) infants and children with asymptomatic or minimally symptomatic CDH and infants who died enroute to the hospital were excluded. One hundred eighty-five neonates were term or near term (>36 weeks' gestation), while the remainder were premature (28-35 weeks' gestation). Forty-five newborns had associated anomalies, including cardiac abnormalities in 36, omphalocele in 3, pulmonary sequestration in 3, intestinal atresia in 2, and biliary atresia in 1.1 Ductus arteriorosis was not considered an associated anomaly in this series. All infants were born at nonadjacent hospitals and were transported by ground or air transport, usually by a transport team consisting of a neonatal nurse, respiratory therapist, and physician.
Prospectively, the following data were kept for each patient: management protocol (immediate or delayed repair, ECMO preoperatively or postoperatively, or repair elsewhere and transfer for ECMO), ventilator management before and after repair, and arterial blood gas levels.
The ECMO, when used, was offered to all infants who did not have any of the following exclusionary criteria: gestational age less than 32 weeks, birth weight less than 1800 g, intracranial hemorrhage greater than grade 2, and other medical disorders or additional anomalies that precluded eventual survival.
Using the prerepair or pre-ECMO blood gas levels or both, a mean PO2, PCO2, and pH were calculated for each treatment group. In addition, these blood gas values and mean airway pressure (MAP) were used to calculate the alveolar-arterial difference in partial pressure of oxygen (partial pressure of oxygen in the alveoli [PAO2] − PaO2), oxygenation index (OI), and ventilatory index (VI) using the following formulae:
(1) PAO2 − PaO2 = [(Barometric Pressure − 47) ×([FIO2− PaCO2]/0.8)] − PaO2 (2) OI = ([FIO2 × MAP]/PaO2) × 100 (3) VI = MAP × Respiratory Rate,
where FIO2 indicates fraction of inspired oxygen. The techniques of newborn ECMO have been previously described2 and include carotid artery and jugular vein cannulation. Venoarterial ECMO is initiated and continued by gravity drainage of blood to a reservoir, which is then delivered by a roller pump through a membrane oxygenator and heat exchanger and returned to the patient through the carotid artery cannula. Systemic heparinization is used throughout the period of ECMO. As lung function improves and pulmonary hypertension resolves, the ECMO flow rate is gradually tapered off and conventional ventilation is resumed and advanced as tolerated.
These newborns were treated in 3 distinct therapeutic eras.
This era included infants treated for CDH before ECMO was available in our institution. Infants in this group received immediate operative repair of CDH without regard to clinical condition. Many children were in extremis, with combinations of severe hypoxia, hypercarbia, or both, metabolic and respiratory acidosis, and shock or cardiac arrest. After operative repair, postoperative management usually consisted of high-frequency, high-pressure ventilation; fluid resuscitation; inotropic support; and attempts at pharmacological reduction of pulmonary hypertension with agents such as tolazoline hydrochloride, nitroglycerin, prostaglandin inhibitors, nitroprusside, and others.
In this era, ECMO became available and was used primarily as postoperative support. Forty-one infants in this group were treated with immediate repair of the CDH. Eighteen of these patients either recovered with conventional therapy and, therefore, ECMO was not needed or the patients were not candidates for ECMO because of exclusionary criteria. The remaining 23 patients received ECMO after CDH repair because of failure to improve with the conventional management techniques outlined in era 1. The final 4 patients received ECMO immediately after arrival at our hospital, prior to repair, because of a moribund condition. All 4 of these infants had CDH repair while receiving ECMO, and the ECMO flow was tapered off, as described above, as soon as possible after surgery to limit the risk of significant hemorrhage in these infants receiving systemic heparinization.
The emphasis in this era included delayed repair and the use of ECMO for moribund patients. The patients in this group received either immediate ECMO with repair of the CDH while using ECMO (23 patients), immediate CDH repair with postoperative ECMO support (9 patients, all CDHs repaired elsewhere), or delayed CDH repair after 2 to 5 days of support with combinations of ventilator, fluid and inotrope infusion, induced respiratory alkalosis, metabolic alkalosis, or both, and inhaled nitric oxide (24 patients).
The primary outcome measured for each therapeutic era was survival, as defined by discharge from the hospital. Late deaths defined as occurring after discharge were recorded. Morbidity was assessed for both patient complications as well as ECMO apparatus malfunction. Survival for the 3 therapeutic eras was compared using χ2 analysis.
A comparison of mean blood gas levels and mean PAO2, OI, and VI for the patients in each therapeutic group is shown in Table 1. There were no significant differences between the 3 groups with regard to these parameters, suggesting that the patient population with CDH admitted to our hospital has remained relatively constant throughout the 27 years of study.
The survival for the 3 therapeutic groups is shown in Figure 1. Therapeutic era 3 had a statistically significant improvement in survival compared with both eras 1 and 2 (P<.02). Further analysis of the survival and morbidity for each group follows.
The survival and treatments are shown in Figure 2. Overall, 43 (42%) of the 102 patients treated in this period survived. There were no recurrent herniations and no late deaths in follow-up as long as 22 years. In addition, no instances of significant neurologic impairment or chronic pulmonary disease were found among the 43 survivors from this group. It would appear that surviving CDH in this era tended to select a group of basically healthy children who go on to adulthood without significant impairment.
The outcomes for children treated in this period are shown in Figure 3. Overall, 26 (58%) patients were discharged from the hospital, but 3 patients died after discharge from bronchopulmonary dysplasia (2 patients) and sudden infant death syndrome (1 patient). Thus, the long-term survival was 51%. Among the survivors, 2 children have mild neurologic delay and 1 has chronic pulmonary disease. Apparently, the addition of ECMO to the other therapies available in this period allowed survival of some critically ill newborns that previously would not have survived. This may have resulted in several survivors that have chronic (neurologic and pulmonary) diseases as well. All patients with either late death or chronic neurologic or pulmonary disease were treated with ECMO.
The outcomes for patients in this era are shown in Figure 4. Forty-four (79%) of the 56 patients were discharged from the hospital, but 5 of these patients suffered late death, giving an overall survival of 70%. The late deaths were caused by pulmonary hypertension in 1, occluded tracheostomy in 1, bronchopulmonary dysplasia in 1, liver failure in 1, and aspiration in 1. In addition, 2 survivors have chronic pulmonary disease and 1 has cerebral palsy. As in the previous therapeutic era, all patients with either late death or chronic neurologic or pulmonary disease were treated with ECMO.
Congenital diaphragmatic hernia has traditionally had a poor prognosis, in spite of advances in neonatal care and ventilator technology and an improved understanding in the role of pulmonary hypertension in the clinical spectrum of this anomaly. The introduction of extracorporeal support (ECMO) in the late 1970s seemed to improve survival, documented in several series.2- 6 The use of ECMO was subsequently extended to include immediate ventilator support with ECMO after birth and later repair of the defect either while receiving ECMO7,8 or after tapering off ECMO support.9 We have been reluctant to attempt to taper off ECMO support prior to repair in those infants who receive ECMO immediately after birth, because of the known deterioration in pulmonary function after CDH repair. This may precipitate recurrent pulmonary hypertension that would be difficult to treat, as successive ECMO treatment is rarely successful.
After several investigators10,11 noted improved survival with delayed operative therapy, this approach became standard treatment in newborns who were stable immediately after birth and could therefore tolerate 2 to 5 days of nonoperative management prior to repair. This period presumably allows for resolution of pulmonary hypertension and seems to decrease the risk of recurrent pulmonary hypertension as well. This becomes especially important because of the well-known phenomena of temporary, but significant, deterioration of pulmonary function after repair of the hernia.12
Our approach to CDH has changed as these new therapies have been introduced and advocated. Our data would suggest that these innovations have improved survival during the past 20 to 30 years to the point where the survival rate for this disorder approaches 80%. Further improvements in survival seem less likely as a small group of newborns with CDHs will have pulmonary hypoplasia severe enough to preclude survival no matter what therapy is used.
The present study is neither randomized nor controlled. In addition, there are undoubtedly other differences in management, such as design of ventilators, from one therapeutic era to another. These factors make strict comparisons of outcomes related to each therapeutic era somewhat risky. However, the introduction of ECMO is clearly the major difference in the treatment of CDH when comparing the second therapeutic period with the first, and the additional option of delayed repair is the primary difference between the second and third eras. We recognize, however, that other more subtle differences in management have also probably improved outcome and survival in these patients.
The decreased mortality with new therapies, especially ECMO, has resulted in survival of a group of infants who would not have previously survived. This has resulted in the appearance of chronic neurologic or pulmonary disorders in several of these survivors. In addition, late death has occurred in a few patients, primarily caused by pulmonary abnormalities. However, an improvement in survival is still apparent in spite of these late deaths. In addition, at least several of our late deaths were probably preventable, which has resulted in much closer follow-up for these infants. Other investigators13- 15 have emphasized the high-risk nature of this population of patients and their tendency toward neurologic and other complications.
Other investigators16,17 have noted little or no improvement in survival with either ECMO or delayed therapy. However, most of these reports emphasize that more critically ill newborns with CDH are being transported to their referral centers, both before birth (diagnosed by prenatal ultrasound) as well as postnatally. This has resulted in attempts at management in infants who previously would not have been candidates for any form of treatment. Undoubtedly, ECMO has saved a number of these extremely high-risk neonates, but at the expense of overall survival rates and an increase in long-term morbidity.
In spite of an increasing number of complications and late deaths that has accompanied the use of ECMO, the overall survival rate for CDH seems to be rising. Further close follow-up to prevent late deaths and the advent of newer modalities, such as partial liquid ventilation, should allow continued improvement in outcome for this important congenital abnormality.
Jay L. Grosfeld, MD, Indianapolis, Ind: Dr Weber has presented the evolving management of CDH during a 25-year period. The care has changed completely from urgent to delayed surgical intervention. We have also adopted delayed surgery as the basis of managing these patients. Dr Weber's most recent results are excellent. However, Dr Weber's group is heavily dependent on ECMO, while other groups are trying to get away from using ECMO. While use of nitric oxide and surfactant, etc, have not specifically improved outcome, use of oscillating ventilators and permissive hypercapnia (allowing the babies to achieve a PCO2 level of 60-70 mm Hg) to reduce barotrauma can avoid ECMO in certain instances. In patients who fail this type of treatment, we would support Dr Weber's concept of delayed surgery and placing the patient on ECMO and performing the operative repair with the patient on the ECMO circuit, and then weaning the patient off of ECMO postoperatively. Our current survival rates are somewhat similar, at about the 70th percentile.
I have a few questions for Dr Weber. Have you used permissive hypercapnia and oscillating ventilators to avoid barotrauma and avoid ECMO? Have you been able to use venovenous perfusion for the babies receiving ECMO who have good cardiac output and avoid carotid artery cannulation? Regarding morbidity, you did not mention recurrence of diaphragmatic hernia in survivors. Some of our patients had almost an absence of the diaphragm and required a prosthetic patch insertion. In those patients, the recurrence rate of hernia after survival has been significant. Now that more of these babies are surviving, we have noted that a number of them have severe esophageal dysmotility problems, including gastroesophageal reflux; I wonder if you have encountered these problems in your practice.
Finally, based on the management principles that you have suggested, would you support the premise that all babies with a CDH diagnosed by prenatal ultrasound should be delivered at a center that offers ECMO?
James R. Debord, MD, Peoria, Ill: Have you had any opportunity to study more subtle forms of late morbidity as these children become teenagers or young adults in terms of pulmonary function and growth and development?
Dr Weber: We have used permissive hypercapnia for several years, a technique that is evolving at this point. As long as the pH level is maintained relatively normal, hypercapnia does not seem to result in severe pulmonary hypertension. Clearly, this technique decreases the incidences of barotrauma, which can be a major cause of bronchopulmonary dysplasia later in infancy.
We have not used venovenous perfusion for this group of patients. Many of our patients have an element of cardiac failure as well as pulmonary failure, and we feel that venoarterial perfusion in those children would be more beneficial. The use of venovenous perfusion in these children, however, has been reported successfully by a number of institutions.
We have a small incidence of recurrence of hernia, primarily in those infants who have an absent diaphragm and in whom a prosthetic patch is used. These CDHs seem to reoccur as the child grows during the first year or two. In addition, we have seen a number of children with esophageal dysmotility. The most difficult patients are the children who have severe gastroesophageal reflux that is unresponsive to medication and requires a fundoplication. Many of these children have extremely poor esophageal peristalsis and can be extremely challenging in terms of balancing the reflux with the partial esophageal obstruction created by a fundoplication.
Whether all infants who are diagnosed prenatally with CDH should be delivered only at ECMO centers, of the 9 patients in the third therapeutic era who had their hernia repaired elsewhere and were transferred to us for ECMO, 8 survived. The techniques of transfer have improved recently, and it is probably safe to transfer for short distances those infants who have been repaired at institutions where ECMO is not available. However, ideally these infants should be delivered at centers where ECMO is immediately available should it be needed.
I do not have data on the late morbidity in teenagers. Our oldest child is 13 years old now, and she has no neurologic or respiratory compromise whatsoever. We have a number of children older than 10 years who are similarly growing and developing normally, so it would appear that the late morbidity is going to be quite minimal.
Presented at the 105th Scientific Session of the Western Surgical Association, Colorado Springs, Colo, November 17, 1997.
Reprints: Thomas R. Weber, MD, Cardinal Glennon Children's Hospital, 1465 S Grand Blvd, St Louis, MO 63104.