OBJECTIVE: To investigate the incidence of late-onset sensorineural hearing loss (SNHL) and study the association between neonatal extracorporeal membrane oxygenation (ECMO) and SNHL in survivors of neonatal ECMO between 9 and 13 years of age. STUDY DESIGN: Retrospective analysis of medical records of 212 neonatal ECMO survivors treated between 1987 and 1991 to identify children with complete hearing evaluation at 9-13 years of age. Patients were categorised into normal hearing (NH) and SNHL groups. Cox proportional-hazard regression analysis was used to take into account the variable age at follow-up and to assess the independent effect of each risk factor for SNHL. RESULTS: Forty-eight of the 212 patients had complete hearing evaluations at 9-13 years of age and eight of 48 patients were diagnosed as having SNHL. In two patients, SNHL was first detected between 9 and 13 years. A significant difference between the SNHL and NH groups was found in median 5 min Apgar scores, mean ECMO support duration, proportion of infants with pre-ECMO Paco(2) of <30 mm Hg, and pre-ECMO use of furosemide. No differences were found in the mean cumulative dose or duration of diuretic, muscle relaxant or aminoglycoside treatments. The occurrence of clinical seizures before ECMO and the duration of ECMO therapy were independently associated with SNHL. CONCLUSION: Clinical seizure activity prior to ECMO and the duration of ECMO therapy are independently associated with SNHL. These data confirm that there is an increased incidence of SNHL in neonatal ECMO survivors at 9-13 years of age and suggest that SNHL may also present later in childhood in this patient population. Upon recovery from acute respiratory failure and after discharge from the hospital, longitudinal neurodevelopmental follow-up of infants treated with ECMO during the neonatal period is essential.
OBJECTIVE: To investigate the incidence of late-onset sensorineural hearing loss (SNHL) and study the association between neonatal extracorporeal membrane oxygenation (ECMO) and SNHL in survivors of neonatal ECMO between 9 and 13 years of age. STUDY DESIGN: Retrospective analysis of medical records of 212 neonatal ECMO survivors treated between 1987 and 1991 to identify children with complete hearing evaluation at 9-13 years of age. Patients were categorised into normal hearing (NH) and SNHL groups. Cox proportional-hazard regression analysis was used to take into account the variable age at follow-up and to assess the independent effect of each risk factor for SNHL. RESULTS: Forty-eight of the 212 patients had complete hearing evaluations at 9-13 years of age and eight of 48 patients were diagnosed as having SNHL. In two patients, SNHL was first detected between 9 and 13 years. A significant difference between the SNHL and NH groups was found in median 5 min Apgar scores, mean ECMO support duration, proportion of infants with pre-ECMO Paco(2) of <30 mm Hg, and pre-ECMO use of furosemide. No differences were found in the mean cumulative dose or duration of diuretic, muscle relaxant or aminoglycoside treatments. The occurrence of clinical seizures before ECMO and the duration of ECMO therapy were independently associated with SNHL. CONCLUSION: Clinical seizure activity prior to ECMO and the duration of ECMO therapy are independently associated with SNHL. These data confirm that there is an increased incidence of SNHL in neonatal ECMO survivors at 9-13 years of age and suggest that SNHL may also present later in childhood in this patient population. Upon recovery from acute respiratory failure and after discharge from the hospital, longitudinal neurodevelopmental follow-up of infants treated with ECMO during the neonatal period is essential.