Eur J Cardiothorac Surg 2000;18:328-333
© 2000 Elsevier Science NL
Two years’ follow-up of newborn infants after extracorporeal membrane oxygenation (ECMO)
S. Jaillarda,
V. Pierratb,
P. Truffertb,
D. Métoisa,
Y. Rioub,
A. Wurtza,
P. Lequienb,
L. Stormeb
a Department of Thoracic Surgery, A. Calmette Hospital, Centre Hospitalier Régional et Universitaire de Lille, Bd. du Professeur Leclercq, 59037 Lille cedex, France
b Department of Neonatology, J. de Flandre Hospital, Lille, France
Received 22 December 1999;
received in revised form 10 May 2000;
accepted 6 June 2000.
Corresponding author
e-mail: sjaillard{at}chru-lille.fr
 |
Abstract
|
|---|
Objective: Extracorporeal membrane oxygenation (ECMO) is a technique of extracorporeal oxygenation used in newborn infants with refractory hypoxemia after failure of maximal conventional medical management, when mortality risk is higher than 80%. We retrospectively reviewed all the neonates treated by ECMO between October 1991 and September 1997 in our newborn intensive care unit. Methods: Fifty-seven patients were treated with ECMO for severe respiratory failure: congenital diaphragmatic hernia (CDH) (n=23), neonatal sepsis (NS) (n=14), meconium aspiration syndrome (MAS) (n=12), and others (n=8). Mean gestational age and birth weight were 38±2 weeks and 3200±500 g, respectively. Oxygenation index was 61±8. Both venovenous (n=28) or venoarterial ECMO (n=29) were used. The mean time at ECMO initiation was 47 h (range 8 h–2 months). The mean duration was 134±68 h. In each case of VA ECMO, carotid reconstruction was performed. Survival at 2 years was 40/57 (70%) (CDH 12/23 (52%), NS 11/14 (79%), MAS 12/12 (100%), others 5/8). Follow-up at 2 years was available in 36 survivors. Results: Neurodevelopmental outcome was not related to the initial diagnosis: normal neurologic development (n=30), cerebral palsy (n=5), and neurologic developmental delay (n=1). Two patients remained oxygen dependant at 2 years, and four required surgical treatment for severe gastroesophageal reflux. Respiratory and digestive sequelae were more frequent in the CDH group (P<0.01). Patency and flow of the repaired carotid artery was assessed in 20 infants at 1 year of age using Doppler ultrasonography: normal (n=10), <50% stenosis (n=9), and >50% stenosis (n=1). Conclusion: ECMO increased survival of newborn infants with refractory hypoxemia. However, higher a survival rate and lower morbidity were found in non-CDH infants than in congenital diaphragmatic hernia.
Key Words: Extracorporeal membrane oxygenation • Congenital diaphragmatic hernia • Neonatal respiratory failure • Venovenous bypass • Venoarterial bypass
|
1. Introduction
|
|---|
Extracorporeal membrane oxygenation (ECMO) is an effective therapy for neonates with life-threatening respiratory failure unresponsive to conventional medical management. Since the description of Bartlett in 1975 [1,2], ECMO has been used in the support of more than 12 000 neonates collected in the ECMO Registry of Extracorporeal Life Support Organization (ELSO) [3]. The overall survival rates of neonates managed on ECMO is high in selected series excluding congenital diaphragmatic hernia (CDH). Conversely, despite the advent of ECMO, overall survival of neonates with CDH remains poor. The decrease in mortality has allowed attention to be focused on morbidity. The purpose of this study was to report results of the two years follow-up of newborn infants treated by ECMO according to the diagnosis. Although neurological complications remain to be expected, we did observe many respiratory and digestive complications.
|
2. Population and methods
|
|---|
We prospectively studied all the newborn infants treated by ECMO between October 1991 and September 1997 admitted to our newborn intensive care unit (NICU), for severe respiratory failure unresponsive to conventional medical treatment including inhaled nitric oxide (for infants enrolled after 1993), exogenous surfactant, high frequency oscillatory ventilation and prostacyclin infusion. Initial diagnosis was recorded. Consent was obtained from parents.
Our ECMO entry criteria were: (1) gestational age (GA) over 34 weeks; (2) birth weight (BW) above 1800 g; (3) respiratory failure with oxygenation index (OI) calculated as:
above 40 for 4 h despite maximum ventilatory and pharmacological support; (4) acute deterioration with postductal PaO2 below 35 cmH2O; (5) no evidence of intracranial hemorrhage or severe brain injury. Electroencephalography (EEG) and cranial ultrasonography were performed before cannulation.
Both venovenous (VV) or venoarterial (VA) ECMO were used. VV ECMO was performed using a single lumen cannula surgically inserted through the right jugular vein with the tip in the mid portion of the right atrium. This type of VV ECMO uses alternative inflow and outflow. VA ECMO was performed with both surgical cannulation of the right common carotid artery (with a short cannula (tip<1.5 cm) inserted through the artery) and of the right jugular vein (in the same position as VV ECMO). VA ECMO allows continuous flow within the circuit: the vein was used for venous drainage and the artery was used for reinfusion of oxygenated blood (Fig. 1) . Systemic anticoagulation (heparin infusion in the venous line of the ECMO circuit) was set to maintain the Hemocron time between 160 and 180 s. During the course of ECMO, infants were systematically sedated (Fentanyl) and received diuretics (Lasilix). No prophylactic antibiotics were infused.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 1. Diagram showing the position of the two surgically inserted cannulae, and components of the venoarterial circuit.
|
|
In each case of VA ECMO, the common carotid artery was reconstructed at the time of decannulation. Retention sutures were removed. A short segment with the edges of arteriotomy was resected. Inflow and backflow bleeding were checked. An end-to-end anastomosis was performed with 8/0 interrupted sutures using x3.5 magnification. Hemodynamic stability, pulmonary vasomotor tone and surgically induced complications were recorded during the procedure. Systemic anticoagulation (heparin=600 UI/kg per day) was systematically continued for at least 5 days after the carotid artery reconstruction. Then, inhibitor of thrombocyte aggregation (aspirin=5 mg/kg per day) was given over the long term (at least 2 years). Patency and flow of the repaired carotid artery were assessed by using Doppler ultrasonography.
All of the children were enrolled into a follow-up study with neurodevelopmental, respiratory and digestive examination. Neurologic evaluation consisted of clinical examination (1/6 months), electroencephalography (EEG) (at 6 months, 1 year and 2 years), cranial ultrasonography (at 1 month and 6 months) and computed tomography (CT) scan (at 1 year if cranial ultrasonography was abnormal). Neurological outcome was defined as normal neurologic development, developmental delay, or cerebral palsy according to Hagberg [4]. Neurosensorial evaluation was systematically performed including ocular and fundus oculi examination, and hearing evaluation by Boël-test and brainstem auditory-evoked potential (BAEP). Respiratory evaluation was performed: bronchopulmonary dysplasia was defined as oxygen dependency after 28 days of life, and persistence of oxygen dependency was recorded with the need of tracheostomy. Digestive evaluation consisted of the search for symptomatic gastroesophageal reflux. Growth, including weight, length and head circumference, was systematically recorded, and growth retardation was defined as weight<5th percentile. Rate of feeding gastrostomy was noted. Doppler ultrasonography was performed to check the blood flow of the repaired carotid artery.
Statistical analysis was performed using 
2 or Fisher's exact test for difference of proportions, and the Mann–Whitney test for the comparison of groups of patients. Results are mean±standard deviation (SD). P-values of less than 0.05 were considered significant.
|
3. Results
|
|---|
Fifty-seven infants were enrolled in the study (mean GA=38±2 weeks; mean BW=3200±500 g; Apgar score at 5 min=7.7±2). The OI 2 h before connection to ECMO was 61±8. Respiratory failure resulted from congenital diaphragmatic hernia (CDH; n=23), neonatal sepsis (NS; n=14), meconium aspiration syndrome (MAS; n=12) and others (n=8). The mean time at onset of ECMO was 70 h (range 8 h–2 months). The mean duration was 134±68 h. Initial cranial ultrasonography examination was normal in 47 infants. Abnormalities (compatible with entry criteria) was recorded in ten infants. All neonates had EEG abnormalities prior to initiation of ECMO (from moderate (16%) to severe (42%)).
The clinical characteristics of the population are summarized in Table 1. Two groups were compared according to the initial diagnosis: CDH (n=23) and other diagnosis (n=34). No difference between the two groups could be found for gestational age, birth weight, Apgar score, oxygenation index and age at ECMO initiation. However, ECMO duration was significantly longer in the CDH group (P=0.0053).
Twenty-eight neonates were treated by VV ECMO (nine CDH (32%)) mainly before November 1993. After November 1993, 29 newborns were treated by VA ECMO (14 CDH (48%)). The switch from VV to VA ECMO in 1993 was decided because of the occurrence of four successive deaths during VV ECMO related to severe hemodynamic impairment (two in sepsis and two in CDH) due to right cardiac failure. Distention of the right ventricle clearly prevented from filling of the left ventricle. By decreasing right ventricular preload, VA ECMO may prevent from such cardiac failure. A similar number of CDH infants were treated by VV ECMO and by VA ECMO (P=0.28, Fisher's exact test). One infant was converted from VA ECMO to VV ECMO because of a technical complication with expulsion of the arterial cannula.
Survival at 2 years (Table 2) was 40/57 (70%): it was highly related to initial diagnosis: CDH=12/23 (52%), NS=11/14 (79%), MAS=12/12 (100%) and others=5/8 (62%). Ten neonates died during extracorporeal circulation (early death): eight were on VV ECMO, two on VA ECMO. Five technical complications occurred during VV ECMO: pump malfunction (n=2), air in circuit (n=1), hemorrhagic complication (n=1) and heat exchanger malfunction (n=1). Five infants died due to medical complications: one intracranial hemorrhage, one cerebral ischemia, one multisystem organ failure and two persistent respiratory insufficiency. Seven infants (all CDH) died after ECMO support was discontinued: one total superior and inferior vena cava obstruction (1 month old), one sudden death (5 months old), one of gastric inhalation (7 months old), one secondary to general anesthesia for jejunostomy (1 year old), one bronchopulmonary dysplasia (18 months old) and one hemorrhage following insertion of a central venous line (2 years old).
Follow-up at 2 years was available in 36 cases. Neurologic outcome was not related to initial diagnosis (Table 3). Clinical examination was performed: normal neurologic development (n=30), cerebral palsy (n=5), and neurologic developmental delay (n=1). Cerebral palsies were two right hemiparesis, one right monoparesis and two lower limb hypertonicity and hyperspasticity. Two of them had evidence of gaseous emboli during extracorporeal circulation. The infant with neurologic developmental delay had no associated cerebral palsy.
Neurosensorial loss was recorded. Fundus oculi examination was abnormal in four infants: three hemorrhages of the retina (one bilateral and two unilateral), and one hemorrhage of the vitreous. Two infants had clinical visual anomalies: one with unilateral amblyopia and one with a minor visual loss. Furthermore, one strabismus and one Horner's syndrome were recorded. Five alterations of the brainstem auditory-evoked potential (BAEP) were recorded. However, only one infant had severe hearing loss requiring a prosthesis.
Anomalies of cranial ultrasonography was recorded in 16 neonates before decannulation. Anomalies ranged from moderate (intraventricular hemorrhage grade I of Papile, or transient intraparenchymal hyperechogenicity) (n=10/16), to severe (grade II or III and cerebral atrophia) (n=6/26). Persistent cerebral atrophia imaging was found in six infants. CT scan at 1 year was performed on 23 infants and showed nine abnormalities without lateralization (six cerebral atrophia and three minor abnormalities).
Study of electroencephalography shows that poor neurological outcome was related to ictal discharges lasting more than 1 day and associated with severe background abnormalities.
Beyond the neonatal period (among the 2-year survivors), respiratory and nutritional sequelae were recorded. Respiratory problems were frequent: 18/40 (45%) patients remained oxygen-dependent at 28 days (CDH=10/12 (83%), others=8/28 (29%); P=0.0021); two patients (CDH) remained oxygen-dependent at 2 years. Diagnosis of bronchopulmonary dysplasia was performed for 13 infants: all CDH, with the need of a tracheostomy for four infants. Nutritional and digestive problems were growth retardation (n=12), ten of whom were CDH; gastroesophageal reflux (n=11), all CDH, four of whom required surgical treatment and needed a feeding gastrostomy These respiratory and digestive sequelae were more frequent in the CDH group (P<0.01).
There was no evidence of adverse effects observed during decannulation according to the carotid reconstruction: no hemodynamic instability or change in pulmonary vasomotor tone were observed during surgery. No short-term complications such as air emboli, thromboembolic episode, or major bleeding from the site of reconstruction were recorded. Patency and flow of repaired carotid artery (n=27; two deaths during the 29 VA ECMO performed) was assessed in 20 infants at 1 year of age using Doppler ultrasonography. Out of 20 carotid reconstructions, 19 were considered technically successful: ten children had patent carotid arteries with normal cephalad flow across the anastomotic site and nine infants had an inferior to 50% stenosis. One carotid reconstruction failure was recorded with a more than 50% carotid stenosis.
|
4. Discussion
|
|---|
Our population is characterized by a high incidence of CDH patients (40%) compared with the others’ studies, which included 14–19% CDH [3,5–7]. These CDH patients need to be distinguished from the others because of the seriousness of the initial respiratory failure, and the different prognosis due to the late respiratory and digestive sequelae encountered. Management of these CDH patients involves initial stabilization and delayed surgery. Results involving conventional treatment without ECMO are disappointing without improvement of lung function due to persistence of severe pulmonary hypertension, which leads to death. The sequelae encountered were identified as CDH ‘hidden morbidity’ [8], by extension of Harrison's original idea of ‘hidden mortality’ as reported in 1978 [9]. This morbidity exceeds the acute problem, resolved here by ECMO treatment concerning CDH survivors. In our experience, carotid artery reconstruction at the time of decannulation appeared as a safe, technically feasible alternative to permanent ligation (no local or general complications were recorded during decannulation and reconstruction).
Although a few authors reported series with a high level of OI [7,10], in our study OI was indeed higher (61±8) than the OI calculated in the majority of the other studies (calculated means from 36 to 49) [4,11,12]. These data underline the seriousness of our treated group. Similar survival rates were found in the ELSO experience (1997) [3] concerning more than 12 000 neonates treated by ECMO with a survival rate of 81% (including both VA and VV ECMO): VA ECMO mortality rates for MAS were 94%, mortality rates for sepsis were 77%, mortality rates for others (excluding CDH) were 81%, whereas mortality rates for CDH were 58%; regarding ECMO VV only, mortality rates were 90%. Experience reported in 1996 by the ‘UK Collaborative Randomized Trial of Neonatal ECMO’ [7] compared two similar groups of newborn infants with severe respiratory failure. Infants were randomized to conventional management or to ECMO treatment, and were followed up to 1 year. Death rates differed between the two groups; 30 of 93 infants (32%) allocated ECMO died compared with 54 of 92 (59%) which were allocated to conventional care (relative risk=0.55, (P=0.0005)). These data showed a clear advantage with ECMO leading to an early halt to recruitment to the trial, and confirmed the significant beneficial effect of ECMO for the management of severe respiratory failure. However, the mortality rate of the CDH patients in this trial was 17/17 deaths (100%) with conventional management and 14/18 deaths (77%) with ECMO (P=0.10, Fisher's exact test). This high mortality rate emphasizes the severity of the respiratory failure in the CDH group.
Our ECMO experience began in October 1991 using VV ECMO with a single cannula with alternating inflow and outflow. The VV ECMO technique prevents carotid artery cannulation. However, in our experience alternative flow led to technical complications (all the technical complications (n=5) were observed during VV ECMO). Furthermore, whereas VV ECMO improves oxygenation, it does not support hemodynamic failure. Currently, the use of exogenous surfactant, inhaled nitric oxide (iNO), and high-frequency oscillatory ventilation (HFOV) in the management of respiratory hypoxemia decreases the indication of ECMO. Failure of this management with persistent hypoxemia leads to the selection of severe respiratory insufficiency associated with hemodynamic instability, mainly right ventricular failure. Introduction of VA ECMO, in our experience (from November 1993 onwards) allows us to manage right ventricular failure and to provide left ventricular support. This quality of VA ECMO appeared as especially interesting in the CDH group associated with impaired cardiac function.
Usually at the time of decannulation, carotids were ligated [13]. Different reports have demonstrated that the anatomical and physiological mechanisms that control cerebral circulation appeared to be sufficient to ‘compensate’ for loss of a major inflow vessel to the brain [14–16]. However, permanent ligation of the common carotid artery may have short- and long-term consequences and remains a major objection to the use of VA ECMO [17,18]. Long-term effects of definitive carotid ligation on the incidence and pattern of occlusive vascular disease in later life are unknown [16]. In each case of VA ECMO, we decided to perform carotid reconstruction. In order to accomplish reconstruction, the cannulation procedure had to be minimally traumatic. Excision of the edges of the arteriotomy site is recommended at the time of arterial repair [16]. Our follow-up demonstrates successful carotid reconstruction in 19 patients either with normal flow or with <50% stenosis. Nine infants had <50% stenosis. Desai [19] speculates that carotid reconstruction may have augmented the cerebral circulation during the first 5 years of life after neonatal ECMO. He demonstrates that stenosis after reconstruction may improve over time, but long-term risks and benefits remain unknown. In order to avoid cerebral emboli, and with a lack of any proven data, we chose to keep all the babies on aspirin with antiaggregant dosage (5 mg/kg per day).
In this prospective study, comparison of VV ECMO with VA ECMO is not appropriate, because of the historical composition of the two groups: two different management policies were successively introduced in our experience with an improvement of the technique and the improvement in conventional management. Comparison of the data concerning the CDH group with the other infants allows us to distinguish a population with a different short-term and long-term prognosis. These prospective studies expose the difficulty of long-term follow-up (only 36 infants of the 40 survivors had a complete and available follow-up at 2 years). Our experience began in 1991, with the introduction of VA ECMO (with systematic arterial reconstruction) by the end of 1993. We still have an insufficient retrospective in order to conclude neurological development delay for the milestones of long-term development of arterial reconstruction. Is it necessary to prescribe aspirin for infants and if so, for what duration? Current thinking suggests that a small dosage leads to a small incidence of morbidity.
ECMO improves survival of newborn infants with refractory hypoxemia. Higher survival rates and lower mortality were found in neonatal sepsis and meconium aspiration syndrome than in congenital diaphragmatic hernia. Current indication of ECMO treatment is rare with the development of conventional management including early admittance and associated techniques such as exogenous surfactant, inhaled nitric oxide, and high-frequency ventilation. However, ECMO still remains indicated for patients with failure of the maximal conventional treatment.
|
Footnotes
|
|---|
Presented at the 7th European Conference on General Thoracic Surgery, Nancy, France, October 21–23, 1999.
|
References
|
|---|
- Bartlett R.H., Gazzaniga A.B., Huxtable R.F., Schippers H.C., O'Connor M.J., Jefferies M.R. Extracorporeal circulation (ECMO) in neonatal respiratory failure. J Thorac Cardiovasc Surg 1977;74:826-833.[Abstract]
- Bartlett R.H., Gazzaniga A.B., Toomasian J., Coran A.G., Roloff D., Rucker R. Extracorporeal membrane oxygenation (ECMO) in neonatal respiratory failure: 100 cases. Ann Surg 1986;204:236-245.[Medline]
- Extracorporeal Life Support Organization. ECMO Registry of extracorporeal life support organization (ELSO). Ann Arbor, MI: ELSO, 1997.
- Hagberg B., Hagberg G., Olow I. The changing panorama of cerebral palsy in Sweden 1954–1970 II. Analysis of various syndromes. Acta Paediatr Scand 1975;64:193-200.[Medline]
- Anderson H.L., Snedecor S.M., Otsu T., Bartlett R.H. Multicenter comparison of conventional venoarterial access versus venovenous double-lumen catheter access in newborn infants undergoing extracorporeal membrane oxygenation. J Pediatr Surg 1993;28:530-535.[Medline]
- Stolar C.J.H., Snedecor S.M., Bartlett R.H. Extracorporeal membrane oxygenation and neonatal respiratory failure: experience from the extracorporeal life support organization. J Pediatr Surg 1991;26:563-571.[Medline]
- UK Collaborative Randomised Trial of Neonatal ECMO. UK collaborative randomised trial of neonatal extracorporeal membrane oxygenation. Lancet 1996;348:75-82.[Medline]
- Lund D.P., Mitchell J., Kharasch V., Quigley S., Kuehn M., Wilson J.M. Congenital diaphragmatic hernia: the hidden morbidity. J Pediatr Surg 1994;29:258-264.[Medline]
- Harrison M.R., Bjordal R.I., Langmark F., Knutrud O. Congenital diaphragmatic hernia: the hidden mortality. J Pediatr Surg 1978;13:227-230.[Medline]
- Osiovich H.C., Peliowski A., Ainsworth W., Etches P.C. The Edmonton experience with venovenous extracorporeal membrane oxygenation. J Pediatr Surg 1998;33:1749-1752.[Medline]
- Delius R., Anderson H., Schumacher R., Shapiro M., Otsu T., Toft K., Hirsch J., Bartlett R. Venovenous compares favorably with venoarterial access for extracorporeal membrane oxygenation in neonatal respiratory failure. J Thorac Cardiovasc Surg 1993;106:329-338.[Abstract]
- Cornish J.D., Heiss K.F., Clark R.H., Strieper M.J., Boecler B., Kesser K. Efficacy of venovenous extracorporeal membrane oxygenation for neonates with respiratory and circulatory compromise. J Pediatr 1993;122:105-109.[Medline]
- Bartlett R.H., Gazzaniga A.B., Huxtable R.F., Schippers H.C., O'Connor M.J., Jefferies M.R. Extracorporeal circulation (ECMO) in neonatal respiratory failure. J Thorac Cardiovasc Surg 1977;74:827-833.
- Voorhies T.M., Tardo C.L., Andrea L.S., Loe W.A., Graves E.D., Falterman K.W., Arensman R.M. Evaluation of the cerebral circulation in neonates following extracorporeal membrane oxygenation. Ann Neurol 1985;18:380.
- Perlman J.M., Altman D.I., Powers W.J., Volpe J.J. Cerebral injury and regional cerebral blood flow in newborn infants undergoing extracorporeal membrane oxygenation. Ann Neurol 1987;22:421.
- Moulton S.L., Lynch F.P., Cornish J.D., Bejar R.F., Simko A.J., Krous F. Carotid artery reconstruction following neonatal extracorporeal membrane oxygenation. J Pediatr Surg 1991;26:794-799.[Medline]
- Lott I.T., McPherson D., Towne B., Johnson D., Starr A. Long term neurophysiologic outcome after neonatal extracorporeal membrane oxygenation. J Pediatr 1990;116:343-349.[Medline]
- Spector M.L., Wiznitzer M., Walsh-Sukys M.C., Stork E.K. Carotid reconstruction in the neonate following ECMO. J Pediatr Surg 1991;26:357-361.[Medline]
- Desai S.A., Stanley C., Gringlas M., Merton D.A., Wolfson P.J., Needleman L., Graziani L.J., Baumgart S. Five years follow-up of neonates with reconstructed right common arteries after extracorporeal membrane oxygenation. J Pediatr Surg 1999;134:428-433.
Top
Abstract
1. Introduction
