SENSITIVITY AND SPECIFICITY OF THE NEONATAL BRAIN-STEM AUDITORY-EVOKED POTENTIAL FOR HEARING AND LANGUAGE DEFICITS IN SURVIVORS OF EXTRACORPOREAL MEMBRANE-OXYGENATION
S. Desai et al., SENSITIVITY AND SPECIFICITY OF THE NEONATAL BRAIN-STEM AUDITORY-EVOKED POTENTIAL FOR HEARING AND LANGUAGE DEFICITS IN SURVIVORS OF EXTRACORPOREAL MEMBRANE-OXYGENATION, The Journal of pediatrics, 131(2), 1997, pp. 233-239
Objective: We determined the sensitivity and specificity of neonatal b
rain-stem auditory evoked potentials (BAEP) as markers for subsequent
hearing impairment and for developmental problems found later in infan
cy and childhood. Methods: BAEP studies were performed before discharg
e in infants treated with extracorporeal membrane oxygenation (ECMO),
and two specific abnormalities were analyzed: elevated threshold and d
elayed central auditory conduction. Behavioral audiometry was repeated
during periodic follow-up until reliable responses were obtained for
all frequencies, and standardized developmental testing was also condu
cted. The sensitivity and specificity of an elevated threshold on the
neonatal BAEP for detecting subsequent hearing loss, and the relations
hip of any neonatal BAEP abnormality to language or developmental: dis
orders in infancy, were calculated. Results: Test results for 46 ECMO-
treated infants (57.5%) were normal, and those for 34 infants (42.5%)
were abnormal, with either elevated wave V threshold, prolonged wave I
-V interval, or both on neonatal BAEP recordings. Most significantly,
7 (58%) of the 12 children with subsequent sensorineural hearing loss
had left the hospital after showing normal results on threshold tests.
There was no significant difference in the frequency of hearing loss
between subjects with abnormal (5/21, or 24%) and those with normal BA
EP thresholds (7/59, or 12%; Fisher Exact Test, p = 0.28). Therefore t
he sensitivity of neonatal BAEP testing for predicting subsequent hear
ing loss was only 42%. Neonatal BAEP specificity for excluding subsequ
ent hearing loss was 76%. In contrast, on language development testing
, 19 children demonstrated receptive language delay. Of these children
, 12 (63%) had abnormal neonatal BAEP recordings and 7 (37%) had a nor
mal BAEP threshold, normal central auditory conduction test results, o
r both (p = 0.04). Conclusions: Neonatal BAEP threshold recordings wer
e of limited value for predicting subsequent hearing loss common in EC
MO-treated survivors. However, an abnormal neonatal BAEP significantly
increased the probability of finding a receptive language delay durin
g early childhood, even in those with subsequently normal audiometry f
indings. Because neonatal ECMO is associated with a high risk of heari
ng and receptive language disorders, parents should be counseled that
audiologic and developmental follow-up evaluations in surviving childr
en are essential regardless of the results of neonatal BAEP testing.