Van Meurs et aL
33. Bayley N. Bayley Scales of Infant Development: birth to two years. New York: Psychological Corp., 1969. 34. MuUen EM. Mullen Scales of Early Learning. Circle Pines, Minnesota: American Guidance Service, 1993. 34a. Graziani L J, Streletz L J, Baumgart S, et al. Prognosis and neonatal EEG studies of infants treated with ECMO. Pediatr Res 1993;33:371A. 34b. Lioy J, Claney RR, Polin RA, et al. Electroencephalography during ECMO. Pediatr Res 1993;33:373A. 35. Alexander A, Mitchell DG, Merton DA, et al. Cannulainduced vertebral steal in neonates during extracorporeal membrane oxygenation: detection with color Doppler US. Radiology 1992;182:527-30. 36. Krupski WC, Effeney D J, Ehrenfeld WK. In: Bergan JJ, Yao JST, eds. Cerebrovascular insufficiency. Grune Stratton, New York, 1983.
The Journal of Pediatrics August 1994
37. Liekweg WG, Greenfield LJ. Management of penetrating carotid arterial injury. Ann Surg 1978;188:587-92. 38. Cho J, Smullens SN, Streletz L J, Fariello RG. The value of intraoperative EEG monitoring during carotid endarterectomy. Ann Neurol 1986;20:508-12. 39. Holmes GL, Rowe J, Hafford J, Schmidt R, Testa M, Zimmerman A. Prognostic value of the electroencephalogram in neonatal asphyxia. Electroencephalogr Clin Neurophysiol 1982;53:60-72. 40. Scher MS, Klesh KW, Murphy TF, Guthrie RD. Seizures and infarction in neonates with persistent pulmonary hypertension. Pediatr Neurol 1986;2:332-9. 41. Korinthenberg R, Kachel W, Koelfen W, Schultze C, Varnholt V. Neurological findings in newborn infants after extracorporeal membrane oxygenation, with special reference to the EEG. Dev Med Child Neurol 1993;35:249-57.
Clinical and laboratory observations Intracranial abnormalities and neurodevelopmental status after venovenous extracorporeal membrane oxygenation Krisa P. Van Meurs, MD, Hanh T. Nguyen, RN, MSN,William D. Rhine, MD, Michael P. Marks, MD, Barry E. Fleisher, MD, and William E. Benitz, MD From the Department of Pediatrics, Lucile Salter Packard Children's Hospital, and the Department of Radiology, Stanford University Medical Center, Stanford, California
Computed tomography scans of the head and early neurodevelopmental assessment (Bayley Scales of Infant Development) were recorded for 24 surviving in~fants who received venovenous exfracorporeal membrane oxygenation and were compared with those of infants treated with venoarterial bypass matched by diagnosis and oxygenation index before extracorporeal membrane oxygenation. A comparable neuroradiographic and early neurodevelopmental outcome was documented for survivors of venoarterial and venovenous extracorporeal membrane oxygenation. (J PEDIATR4994;425:304-7)
Supported by Human Health Services grant No. MO1-RR00070, General Clinical Research Centers, National Center for Research Resources, National Institutes of Health, and a grant from the National Institutes of Health. Ms. Nguyen received funding as a Keck Scholar under the Stanford University Medical Student Scholars Program.
Submitted for publication Dec. 7, 1993; accepted Feb. 21, 1994. Reprint requests: Krisa P. Van Meurs, MD, Department of Pediatrics, Stanford University School of Medicine, 750 Welch Rd,, Suite 315, Palo Alto, CA 94304. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$3.00 + 0 9/24/55367
The Journal of Pediatrics Volume 125, Number 2
Extracorporeal membrane oxygenation has become an accepted and successful therapy for newborn respiratory failure unresponsive to conventional medical management. Despite encouraging results in this critically ill population, ECMO is not without significant risk. Until recently, the technique has necessitated ligation of both the right common carotid artery and right jugular vein. A double-lumen venovenous catheter was designed by Dr. Robert Bartlett and fabricated by Kendall Med-West (Salt Lake City, Utah), which allows single-site cannulation of the internal jugular vein, thereby sparing the carotid artery from ligation. 1 In this article we compare the neuroradiographic findings and early neurodevelopmental outcomes of 24 survivors of venovenous ECMO with those of a matched group of venoarterial ECMO survivors. METHODS Study population. From July 1990 to July 1992, a total of 63 infants with severe respiratory failure unresponsive to maximal conventional therapy met institutional criteria and were treated with ECMO. The causes of respiratory failure in these patients were as follows: meconium aspiration syndrome (21 patients), sepsis/pneumonia (16), persistent pulmonary hypertension of the newborn (6), congenital diaphragmatic hernia (7), respiratory distress syndrome (3), congenital heart disease (8), and other (2). Twenty-eight of these infants received venovenous ECMO after a cardiac ultrasound study showed a normal shortening fraction and CT ECMO OI
Computed tomography Extracorporeal membrane oxygenation Oxygenation index
the infant's weight and vessel size were found to be appropriate. A number of infants were excluded because of small vessel size despite a weight within the catbeter's specified range. Early in our use of venovenous bypass, infants were excluded because of other factors such as prior cardiac arrest or dependence on multiple high doses of vasopressors. With greater experience, fewer infants were excluded for these reasons. Two children required conversion to venoarterial bypass and are not included in this study; one was converted emergently because of a cardiac arrest after a platelet transfusion, and the second was electively converted to receive a lung transplant on bypass. Two infants died while receiving venovenous ECMO. In neither case was the discontinuation of ECMO related to intracranial hemorrhage: Twenty-four infants (92%) who received venovenous ECMO survived to discharge from the hospital and were available for neuroradiographic and neurodevelopmental evaluation. Comparison group. To compare relative outcomes, for
Van Meurs et al.
Table I, Comparison of clinical data for patients receiving venoarterial and venovenous ECMO Venoarterial V e n o v e n o u s ECMO ECMO
OI before ECMO Duration of bypass (hr) Time to extubation (hr) NS, Not significant.
67 _+ 36
121 _+ 47
146 + 141
each of the 24 surviving infants receiving venovenous ECMO who had computed tomography scans of the head or for whom developmental follow-up was available, an infant treated with venoarterial ECMO was matched by diagnosis and severity of illness as measured by the last oxygenation index before ECMO. Birth weight, gestational age, and Apgar scores were also considered to minimize the clinical differences between the two populations. Selection of the matched patients for inclusion in the venoarterial comparison group was performed by a researcher (H.T.N.) unaware of the neuroradiographic and neurodevelopmental outcomes. Procedure. Maximal medical therapy is very variable among institutions, but in this study included induction of systemic alkalosis with either hyperventilation or use of bicarbonate to maintain the blood pH >7.5, muscle paralysis with pancuronium, sedation with fentanyl or morphine sulfate, and infusion of dopamine, dobutamine, or both to maintain the mean arterial pressure >50 mm Hg. The Stanford ECMO selection criteria are as follows: weight ~ 2 . 0 kg, gestational age >--34 weeks, no intraventricular hemorrhage greater than grade 1, no lethal congenital anomalies, and a calculated OI >45 on two successive arterial blood gas measurements at least 30 minutes apart. The mean last OI before ECMO was 53 _+ 25 for the venovenous group and 67 + 35 for the venoarterial group (p = 0.20). Venovenous ECMO was performed with the 14F double-lumen Kendall catheter as described by Anderson et al.1 Cephalad jugular cannulation was not performed during the period of this study. The venoarterial and venovenous comparison groups had no significant differences in last OI before ECMO, mean duration of ECMO, and mean time to extubation after ECMO (Table I). The survival rate for the venovenous group was 92% (24/26). Of the initial group of 63 patients who received ECMO during this period, 35 patients were treated with venoarterial bypass. Nine children were either cardiac or pediatric patients with clear contraindications to venovenous bypass. The survival rate for the remaining 26 venoarterial bypass patients was 85%. The
Van Meurs et al.
The Journal of Pediatrics August 1994
Table II. Bayley Scales of Infant Development for
dence for contingency tables were performed according to methods of Sokol and Rohlf. 3
survivors of venoarterial and venovenous ECMO Venoarterial Venovenous ECMO ECMO p"
Mental Developmental Index Psychomotor Developmental Index
112 + 11
111 +_ 16
102 _+ 15
111 +_ 16
*Unpaired t test.
causes of death for the patients in the venoarterial group were congenital diaphragmatic hernia (2 patients), septic shock (1), and severe pulmonary hypoplasia (1).
Follow-up procedure Neuroradiography. Of the 24 children who received venovenous ECMO and survived until discharge, 22 had a CT scan of the head after ECMO. Computed tomography scans of the head for 22 of the 24 matched venoarterial bypass infants were also collected. The CT scans were reviewed by a neuroradiologist (M.M.) unaware of the clinical history and were scored according to the neuroimaging scoring system of Taylor et al., 2 which divides the intracranial abnormalities into four major categories: ventricular dilation, hemorrhage, parenchymal lesions, and extraaxial fluid collection. Descriptive labels for each category further identify the location and severity of the lesions. Developmental status. Of the 24 children who received venovenous bypass and survived until discharge home, 20 were seen for developmental testing; evaluations were performed with the Bayley Scales of Infant Development, which includes a Mental scale, a Motor scale, and an Infant Behavior record (findings result in a Mental Development index and a Psychomotor Development Index). Three children were lost to follow-up. One infant is not included because of a diagnosis of trisomy 14 mosaicism. The patients' ages at the time of the most recent assessment ranged from 4 to 24 months (mean, 8 months). The most recent evaluation was the one used for this study and was compared with results of the Bayley Scales of Infant Development for the control infant at the same chronologic age who received venoarterial bypass. The neurodevelopmental outcome was stratified by one of the following descriptive labels: "normal," "suspect," "delayed," or "significantly delayed." Developmental outcome was assessed as "normal" if both the Mental and Motor indexes of the Bayley Scales were >90, as "suspect" if one index was <90, or as "delayed" if both the Mental and Motor indexes were <90. "Significantly delay" and "profound delay" were present if the Bayley scores were both <70 or <50, respectively. Statistical analysis. The data are presented as the mean _+ SD. Unpaired t tests and the G test of indepen-
Neuroradiographie findings. Intracranial abnormalities were present in 2 (9%) of 22 infants receiving venovenous bypass. One infant had mild ventricular dilation and the other a small area of infarction in the region of the left basal ganglia. Daily ultrasound studies of the head were normal in all 24 infants who received venovenous bypass. Intracranial abnormalities were present in 3 (fourteen percent) of 22 infants who received venoarterial bypass; one had left hemisphere atrophy and periventricular leukomalacia, one had severe bilateral ventricular dilation and bilateral intraventricular hemorrhage, and one had left subependymal hemorrhage and left subdural hematoma (p = 0.65 by G test for a 2 × 2 contingency table)P Head ultrasound and CT findings were in agreement in 20 infants who received venoarterial bypass. The infant with atrophy and periventricular leukomalacia had normal results of daily ultrasound studies of the head; three infants with small intraventricular hemorrhages on ultrasonography later were found to have normal results of CT scans of the head. The neur0imaging scoring system of Taylor et al. was used to compare the head CT scans of the two groups. The two infants in the venovenous group who had abnormal results of CT scans of the head had a total score of 2.5, and the three infants in the venoarterial group had a total score of 7. Neurodevelopmental findings. The Bayley Mental Development Index scores were 111 _+ 16 for children in the venovenous group and 112 + 11 for children in the venoarterial group; Psychomotor Development Index scores were 111 + 16 for children in thevenovenous group and 102 + 15 for children in the venoarterial group (unpaired t test; p = 0.67 and 0.07, respectively) (Table II). Stratification of the Bayley results yielded 19 normal outcomes and 1 delayed outcome for the venovenous group and outcomes of 16 normal, 3 suspect, and 1 delayed for the venoarterial group (p = 0.16 by G test for a 2 × 3 contingency table). 3 DISCUSSION The relative risks of venoarterial and venovenous ECMO were compared by reviewing the neuroradiographic findings and developmental outcome in the two groups of survivors. A matched comparison group was selected from our venoarterial bypass population. Limitations were imposed by the relatively small pool of venoarterial ECMO survivors with both results of CT scans of the head and neurodevelopmental follow-up data available. In addition, any attempt to match patients by a fixed number of variables ignores the multitude of known and unknown variables affecting neuroradiographic and developmental outcome. Nevertheless,
The Journal of Pediatrics Volume 125, Number 2
we believe that this process minimized differences between the venoarterial and venovenous ECMO survivors and allowed for a relevant comparison of the two modes of therapy. The length of follow-up is relatively short and Bayley Scales scores are a reliable assessment of current status, but not necessarily predictive of long-term outcome. Our results reveal a relatively low incidence of abnormalities demonstrated by CT scans of the head and a good developmental outcome for both venoarterial and venovenous ECMO survivors. Taylor et al. 4 reported a 46% incidence of abnormalities demonstrated by CT scans in a review of 207 infants treated with venoarterial ECMO during the period 1984 to 1988. Schumacher et al. 5 reviewed CT scans of the head for 22 infants who received venovenous bypass and found abnormalities in 73%. Developmental outcome as measured by the Bayley Scales at 12 months of age was reported by Glass et al. 6 for 42 infants receiving venoarterial ECMO between 1984 and 1986; the authors found 59% normal, 20% with suspected delay, and 21% with delay. Of infants who received venoarterial ECMO, we found 80% normal, 15% with suspected delay, and 5% with delay; of infants who received venovenous ECMO, 95% were normal, none had suspected delay, and 5% had delay. The differences in neuroradiographic outcome may be the result of refinement of ECMO technical practices during the period between these studies. More important, both neuroradiographic and developmental outcome may be affected by changes in the tolerance of adverse clinical events before the start of ECMO therapy. Physicians may be more inclined to use ECMO earlier in the treatment of the critically ill infant, rather than as a last resort after or during a resuscitation.
Van Meurs et al.
The results of this preliminary study demonstrate comparable neuroradiographic and neurodevelopmental outcome for venoarterial and venovenous ECMO survivors evaluated in the first year of life. Recent reports of venoarterial ECMO survivors evaluated at age 5 years have identified behavioral, neuroperceptual, and learning difficulties in a group of children previously thought to be normal] Whether these findings will also be present in the venovenous population will be important to determine. REFERENCES
1. Anderson HL, Otsu T, Chapman RA, Bartlett RH. Venovenous extracorporeal membrane life support in neonates using a double lumen catheter. ASAIO Transactions 1989;35: 650-3. 2. Taylor GA, Glass P, Fitz CR, Miller MK. Neurologic status in infants treated with extracorporeal membrane oxygenation: correlation of imaging findings with developmentaloutcome. Radiology 1987;165:679-82. 3. Sokol RR, Rohlf FJ. Biometry. New York: WH Freeman; 1981:745-6. 4. Taylor GA, Short BL, Fitz CR. Imaging of cerebrovascular injury in infants treated with extracorporeal membrane oxygenation. J PEmATR1989;114:635-9. 5. Schumacher RE, Kewitz G, Brunberg J. CT brain scan findings post venovenous(VV) ECMO. Abstract presented at the 7th Annual Children's National Medical Center ECMO Symposium, Breckenridge, Colorado, Feb. 24-28, 1991. 6. Glass P, Miller M, Short B. Morbidity for survivorsof extracorporeal membrane oxygenation: neurodevelopmental outcome at one year of age. Pediatrics 1989;83:72-8. 7. Wagner A, Glass P, Papero P, Short BL. Neuropsychological outcome and adjustment to kindergarden in ECMO-treated neonates. Abstract presented at the 9th Annual Children's National Medical Center ECMO Symposium,Keystone, Colorado, Feb. 28-March 5, 1993.