The Collaborative UK ECMO Trial: Follow-up to 1 Year of Age

Abstract

Objective. To evaluate the clinical effectiveness of neonatal extracorporeal membrane oxygenation (ECMO), in terms of mortality and morbidity, in the treatment of cardiorespiratory failure in term infants. Methods. The criteria for trial entry were: an oxygenation index of >40 or arterial partial pressure of carbon dioxide (Paco2) >12 kPa for at least 3 hours; gestational age at birth of 35 completed weeks or more; a birth weight of 2 kg or more; <10 days high-pressure ventilation; an age of <28 days; and no contraindication to ECMO such as previous cardiac arrest or intraventricular hemorrhage. Eligible infants were randomized either to be transferred to one of five ECMO centers in the United Kingdom or to continue conventional treatment. The principal outcome was death or severe disability at the age of 1 year. Severe disability was defined as an overall developmental quotient of <50 using the Griffiths Mental Development Scales, or blindness or a level of function so as to make assessment using the Griffiths Scales impossible. Families of surviving children were contacted at regular intervals during the first year and at the age of 1, and an assessment of the child was performed by one of three developmental pediatricians. This included a neurologic examination, assessment of hearing and vision, developmental level, general health, and health service use. Results. Of 185 infants recruited into the trial, 93 infants were in the ECMO arm and 92 were allocated conventional treatment. The groups were comparable at trial entry. Thirty of 93 (32%) ECMO infants died before the age of 1 year and 54 of 92 (59%) of the infants in the conventional group died. Two infants were lost to follow-up, 1 from each arm of the trial. Of the remaining 99 survivors, at the age of 1 year, 2 infants (1 in each arm) were still in the hospital, and 5 (3 in the ECMO arm and 2 conventional) still required supplementary oxygen. Fifteen infants had tone changes in the limbs, 10/62 (16%) in the ECMO arm and 5/37 (13.5%) in the conventional arm. These signs were more common on the left side in both groups. One infant (in the ECMO arm) had bilateral sensorineural deafness and 1 infant (also in the ECMO arm) had low vision. Overall, 2 infants were severely disabled (1 ECMO and 1 conventional), 16 others also had evidence of functional loss (12 vs 4), and 8 had impairment without functional loss (4 vs 5). There was a trend toward proportionately greater respiratory morbidity in the conventional group. Neurologic morbidity was more common in the ECMO group, reflecting the larger number of survivors. The lower rate of adverse primary outcome (death or severe disability at 1 year) was found among infants allocated ECMO in all the predefined stratified analyses. Disease severity at trial entry and type of referral center did not appear to alter the effects of ECMO. Only 4 of 18 infants with congenital diaphragmatic hernia survived and at age 1 year only 1 of the 4 survivors was considered normal. Conclusion. These results are in accord with the earlier preliminary findings that a policy of ECMO support reduces the risk of death without a concomitant rise in severe disability. However, 1 in 4 survivors had evidence of impairment with or without disability. Further follow-up is planned at the age of 4 and 7 years.