Language
Development
in a Group
of Very
Low-Birth-Weight
Children
Whose
Postauricular
Myogenic
Response
Was
Tested
in Infancy
Nan Holmes, BSc, HonsMPhil, M. J. Conway, MSc,
L. Flood, FRCS, J. G. Fraser, FRCS, and
Ann Stewart, MBChB DCH
From the Royal Ear Hospital and Departments of Paediatrics and of Medical Physics and Bioengineering, School of Medicine, University College, London
ABSTRACT. A new instrument for the detection of the
postauricular myogenic (PAM) response was used to test
the hearing of 106 infants weighing 1,500 g when they
were aged 1 to 21 months. Eighty-eight infants showed a
positive response at 60 dB hearing level (HL) (normal).
The other 18 did not respond; four were found to have
sensory neural hearing loss and another six had conduc-tive loss due to secretory otitis media. Of the 106 children,
90 aged 2 years or more (mean 27 months) were living in
the United Kingdom, and their language development
was assessed. It was normal in 67/75 children whose PAM
response had been normal in infancy. The remaining
eight children with normal PAM responses in infancy,
had language delay. All eight children had problems that
were thought to account for the delay, including three
with mental retardation, three with cerebral palsy, and
two whose families did not speak English. Language
development was normal in 11/15 children tested whose
PAM responses had been found to be abnormal, including all six whose secretory otitis media had been diagnosed
and treated at the time of the PAM test. Delay in
lan-guage development was found in 3/4 children with sen-sory neural hearing loss who were available for testing and in one child with overall developmental delay. It is
concluded that a positive PAM response at 60 dB HL in
infancy indicated hearing adequate for the development
of normal speech in otherwise normal children among a
group of infants at high risk of hearing loss. Pediatrics 1983;71:257-261, hearing loss, language, lou’ birth weight.
In 1963 Kiang et al’ described the postauricular
myogenic (PAM) response. The PAM response is a
Received for publication Nov 24, 1981; accepted May 6, 1982. Reprint requests to (AS.) Department of Paediatrics, School of
Medicine, University College London, The Rayne Institute, 5 University St, London WC1, England.
PEDIATRICS (ISSN 0031 4005). Copyright © 1983 by the
American Academy of Pediatrics.
motor reflex that follows an acoustic stimulus. The
central connections are uncertain; however, the
ef-ferent pathway is via the facial nerve (VII) and the
posterior auricular muscle. The possible application
of this finding in the diagnosis of hearing loss was
further explored ten years later by Douek et al.2 In
1978 Fraser et al3 introduced a portable machine
designed to detect the PAM response in infants and
young children, using an averaging technique. They
proposed that the presence of a satisfactory PAM
response could be used to test infants for hearing
loss; and they suggested that a positive PAM
re-sponse to a wide-band click at 60 dB hearing level
(HL),
with a latency of 14 to 19 ms indicated intacthearing that would allow the development of
nor-mal speech and language.
At University College Hospital London (UCH)
we have, since 1976, used the PAM response as a
screening test for hearing among young children
aged 2 years or less. To test the hypothesis that a
positive PAM response to a 60 dB HL click in
infancy indicates adequate hearing for the
devel-opment of normal speech and language, we have
studied prospectively, the members of an ongoing
follow-up study of very low-birth-weight infants
(1,500 g or less) who were born in 1976 and 1977. In
the course of neurodevelopmental follow-up, all
these infants had PAM tests during infancy.
Lan-guage and other psychological assessments were
made later, when the children became old enough.
METHODS
Infants Studied
Of the 122 infants of birth weight 1,500 g or less
unit of UCH during the years 1976 and 1977, 106
had PAM tests in infancy. There were 48 boys and
58 girls. Sixteen infants were not tested as they did
not attend UCH for follow-up in infancy, usually
because they were living abroad.
PAM Testing
The PAM test was performed in the audiology
department of the Royal Ear Hospital by one of us
(M.J.C.),
as described by Fraser et al.3 With thechild sitting on the mother’s knee, electrodes were
applied to the preauricular and postauricular skin
of one ear and an earth electrode was applied,
usually to the back of the neck. A loudspeaker in
the machine then presented free-field click stimuli
to the baby at a rate of 10/s. The electrical response
to the clicks was averaged and tested for
signifi-cance by the machine scoring system in the
instru-ment.3 A positive response at 60 dB HL, with a
latency of between 14 to 19 ms was regarded as
normal. If no response was obtained at 60 dB HL,
the test was repeated at 80 dB HL. A negative
response at both 60 and 80 dB HL was regarded as
abnormal, and a negative response at 60 dB HL
followed by a positive response at 80 dB HL was
considered dubious. The whole test took five to ten
minutes.
Children whose PAM tests were negative or
du-bious were retested approximately 1 month later. If
a positive response at 60 dB HL was not obtained
on this second occasion, the child was referred to a
special audiology clinic in the Royal Ear Hospital
for detailed investigation by both a consultant
neuro-otologist and a teacher for the hearing
im-paired. Assessment included distraction testing
throughout the frequency ranges, otologic
exami-nation, electroacoustic impedance testing and, if
necessary, electrocochleography.
Throughout the study, the ages of the children
were corrected for preterm birth by subtracting the
number of weeks that the child was born before the
expected date of delivery from the chronologic age.
This correction was used in the calculation of all
scores.
PAM
testing was usually performed as near aspossible to 6 months of corrected age. This age was
chosen because early experience with the test
mdi-cated that by 6 months, consistent and reliable
responses could be expected. It allowed further investigation of suspected hearing loss to be
com-pleted by 9 months, the earliest age that aiding was
considered to be feasible when the study was
planned. A few children were tested earlier because
of clinical concern.
Follow-up
Although the children were attending a special
follow-up clinic at UCH for regular
neurodevelop-mental and psychological assessments,4 an
addi-tional assessment at a mean corrected age of 37
months (range 26 to 53 months) was arranged once
for each child during the period from November
1979 to August 1980 for the purposes of this study.
These assessments were conducted by one of us
(N.H.) who had no knowledge of the original PAM
result. The children were assessed with the Reynell
Developmental Language Scales,5 a test that
mea-sures expressive language and verbal
comprehen-sion independently. Scores of 1.5 or more standard
deviations below the mean were considered to
in-dicate significant delay in language development.
The Symbolic Play Test6 was also given to children
aged 3 years or less. This test evaluates early
con-cept formation and symbolization and was used to
augment information on the children’s overall
de-velopment obtained from developmental
assess-ments7 performed in the first 18 months of life.
Children aged 3.5 years or more were assessed with
a standard test of cognitive functioning, usually the
Stanford-Binet Intelligence Scale,8 or for very shy
children or those who spoke no English, the
Merrill-Palmer Scale.9
Otologic examinations and, when necessary,
tym-panometry were performed on all the children,
usu-ally on the same day as the psychological
assess-ment. Any child who was considered to be abnormal
was referred to the Royal Ear Hospital for full
investigation.
RESULTS
Infancy
At a mean corrected age of 7.5 months (range
-1 to 21 months), 88/106 (83%) children had normal
PAM responses, eight (7.5%) had dubious
re-sponses, and ten (9%) children failed the PAM test
at both 60 and 80 dB HL on two successive
occa-sions (Table 1).
Among the eight children with dubious PAM
.tests, four had normal responses to distraction
test-ing throughout the frequency ranges, had normal
tympanograms, and were considered to have
nor-mal hearing; one had dubious responses to
high-frequency sounds but, because the parents failed to
keep appointments, a diagnosis of sensory neural
hearing loss (SNHL) complicated by secretory otitis
media (SOM) was only made later from a pure-tone
audiogram; one had SNHL of 65 dB HL at 4 kHz
sur-TABLE 1. Postauricular Myogenic
Mean Age of 7.5 Months* PAM Response
(PAM) Testing at
TABLE 2. Scores for Reynell Expressive Scale,
Ac-cording to Results of Postauricular Myogenic (PAM)
Testing in Infancy*
* Eighty children were tested at a mean corrected age of 37 months.
* Ninety children were tested at a mean corrected age of 37 months.
No. of Children
Normal (positive at 60 dB HL)
“Dubious” (positive at 80 dB HL)
SNHL
SNHL and SOM
Conductive loss (SOM)
Normal hearing
Abnormal (negative at 60 and 80
dB HL)
SNHL
SNHL and SOM
Conductive loss (SOM)
Normal hearing 88 (83.0%) 8 (7.5%) [1] [1] [1] [5] 10 (9.5%) [1] [1] [5] [3] Total 106(100.0%)
* Abbreviations used are: SNHL, sensory neural hearing loss; SOM, secretory otitis media.
gical intervention. The eighth child moved shortly
after the initial PAM test and before keeping her
appointment in the audiologic clinic. She was
in-vestigated elsewhere, and at the age of 4 years was
considered to have normal hearing, confirmed on
audiogram.
Of the ten children who failed the PAM test at
60 and 80 dB HL seven were found to have hearing
losses. Five had conductive losses due to SOM,
including two children who required surgical
inter-vention at the time of the investigation; one had a
high-tone SNHL complicated by SOM and one
child had profound SNHL. Electrocochleography
on this child at the age of 6 months indicated losses
of more than 80 dB HL on both sides and these
findings have since been confirmed on audiogram.
Thus, among the 18 children who either failed the
PAM test or who had dubious PAM responses,
problems of hearing were diagnosed in infancy in
nine (50%), including two with serious sensory
neural hearing losses, one with a high-tone loss and
SOM, and six with SOM. A tenth child was
sus-pected of having high-tone loss at audiologic
as-sessment in infancy but this could not be confirmed
as the parents failed to keep appointments. Hearing
loss was eventually confirmed from a pure-tone
audiogram.
Childhood
By the time the children were old enough for
language testing, 13/106 members of the group had
left the United Kingdom and were living abroad
with their families, including one of the children
with SNHL. Of the remaining 93 children, 90 had
language assessments, including four children living
a long distance from UCH who were assessed by
colleagues living locally. All 90 children completed
the comprehension scale of the Reynell
Develop-mental Language Test, but only 80 completed the
expressive scale due to shyness or foreign origins.
The parents of the remaining three children failed
to keep their appointments. Of the 86 children
assessed in UCH, 84 had otologic examinations.
Language Development
Distribution of standard scores for the expressive
and comprehension scales of the Reynell test
ac-cording to the results of the original PAM test are shown in Tables 2 and 3. Among the 75 children
whose PAM test was normal in infancy, there were
eight children with standard scores of 1.5 or more
below the mean on one or both scales. Three of
these children (who had low scores on both scales)
were mentally retarded with IQ scores of 2 or more
standard deviations below the test mean, including
one with a recurrent conductive loss due to SOM.
Two additional children had ataxic cerebral palsy
with specific problems of articulation; one child
with cerebral palsy had a conductive loss due to
SOM at the time of language testing; and the other
two children did not speak English. The language
development of all the other children whose PAM
responses were normal in infancy was considered to
be acceptable for their age, and there was no
evi-dence from these assessments, of children with
sig-nificant hearing loss.
Among the 15 children tested, whose original
PAM result was dubious or negative, four children
had standard scores of 1.5 or more below the mean
SD PAM Negative (n=7) PAM Dubious (n=7) PAM Normal (n=66)
-1.5 2 (29%) 2 (29%) 6 (9%)
0 to -1.4 2 (29%) 2 (29%) 18 (27%)
0 to +1.4 3 (42%) 2 (29%) 35 (53%)
+1.5 ... 1 (13%) 7 (11%)
TABLE 3. Scores for Reynell Comprehension Scale,
According to Results of Postauricular Myogenic (PAM)
Testing in Infancy*
SD PAM Negative (n=8) PAM Dubious (n=7) PAM Normal (n=75)
-1.5 1 (12%) 2 (29%) 5 (7%)
0to -1.4 4 (50%) 2 (29%) 19 (25%)
0 to +1.4 3 (38%) 1 (13%) 49 (65%)
on one or both scales. These children included three
with SNHL (one high tone only) and one child with
overall developmental delay and an IQ of 77. Otologic Examinations
Nine children, aged 26 to 53 months, whose
orig-inal PAM tests in infancy were normal, were
iden-tffied as having SOM during the course of the
investigations for this study. Surgical intervention
was considered necessary in six. Only 2/9 had low
scores on the Reynell Scales. As already indicated,
both these children had other neurodevelopmental
abnormalities, which also may have contributed to
their poor language development. In contrast,
treat-ment of SOM diagnosed as a result of the original
PAM testing had been effective clinically, and the
children’s language development was within normal
limits.
DISCUSSION
Intact hearing is vital to the development of
speech and language, yet 0.1% of all infants have
profound hearing losses,’0 and there are some
groups of infants, such as those of very low birth
weight (<1,500 g), in whom the risk of sensory
neural loss may be as high as 10%.h112 there is
evidence that immature infants treated with
me-chanical ventilation through endotracheal tubes
have an increased risk of SOM during the first year
of life.’3 In general, conductive losses due to SOM
are temporary. However, if they are present during
the first year of life and early in the second year,
they can affect early language development;
there-fore it is as important to identify these children as
those with sensory neural losses, to ensure that
medical or surgical management and facilities for
appropriate training and education are made
avail-able, both to the children and their families.
Although screening procedures in the very young
have been developed, the early ones were frequently
unreliable and always controversial.’4”5
Electroa-coustic impedance bridge measurement will only
detect a profound loss, and electrococbieography in
young children must be performed under a general
anaesthetic.’6 The reliability of cortical evoked
re-spouse audiometry in young children is not yet
established. Thus, all these methods are
inappro-priate as screening procedures in infancy.
Recently, a method of screening for moderate to
severe SNHL in infancy (“Crib-o-gram”), which
depends on electronic observation of motor
behav-ior responses has been introduced. Although the
method is noninvasive, it requires observation of
the infant in the hospital for a minimum period of
six hours in term infants and 24 hours in intensive
care units. It is only applicable in the early weeks
of life. In contrast, PAM testing is a simple,
nonin-vasive method that gives objective evidence of
in-tact hearing in only a few minutes. It can be used
at any age, although it is not satisfactory when the
baby is asleep or drowsy.3 The equipment is
porta-ble and, as ideal acoustic conditions are
unneces-sary, it is suitable for use in outpatient and infant
clinics.
In this prospective study of a group of 106 very
low-birth-weight (<1,500 g) infants at high risk for
SNHL and conductive losses due to SOM, all those
infants who had positive PAM responses at 60 dB
HL either had normal language development or, if
delayed, there were other contributory factors
in-cluding mental retardation, cerebral palsy, or
for-eign origins. None had any evidence of sensory
neural hearing loss. In contrast, PAM testing
iden-tifled the only four children in the group with
sensory neural hearing losses affecting the speech
frequencies. In addition, six infants were found with
conductive losses due to SOM at the time of PAM
testing, which was treated. Although the proportion
of hearing losses was greater (7/10 infants) among
the infants with negative PAM responses compared
with the children whose PAM responses were
du-bious (3/8), both groups included infants with
SNHL.
It will not be possible to citaw final conclusions
on the accuracy of PAM testing until the children
have had audiograms. These investigations, which
are being performed as the children become old
enough, are part of another study, the results of
which will be published separately (J. G. Fraser,
personal communication, 1982). So far, 71/90 (79%)
children concerned in this report have been tested,
including 55/74 (74%) children whose PAM test in
infancy was normal and 16/18 (89%) children with
abnormal responses. No cases of SNHL have been
identified among the children with normal PAM
responses and no other children have been found
with SNHL among those with abnormal responses.
From the data reported here, it appears that a
positive PAM response at 60 dB HL accurately
identifies infants whose hearing is adequate for the
learning of speech and that children with negative
or dubious responses warrant careful investigation
and follow-up during the early years of life, as at
least half are likely to have sensory neural or
con-ductive losses at the time of testing. It is as yet not
clear whether the eight children in the study, who
had negative or dubious PAM responses and in
whom no defmite hearing loss was diagnosed in
infancy by clinical testing, were genuine
false-neg-atives. Audiograms so far have not revealed SNHL
but, of course, we cannot rule out the possibility
conduc-tive losses due to SOM at the time of the PAM test,
especially as two thirds of them were mechanically ventilated. ‘
Even if the results from these eight infants were
false-negatives, the overall rate of 8% is the same as
the false-negative rates for the general newborn
population (8%) noted in trials with the Crib-o-gram
and much lower than the rate of 20% obtained with
the Crib-o-gram in infants in neonatal intensive
care units.’7 Thus PAM testing appears to be an
efficient and acceptable screening test for both
SNHL and conductive losses. It is of particular
value for high-risk infants such as those oflow birth
weight who were treated with mechanical
ventila-tion. Although PAM testing was originally designed
to identify SNHL, from the results reported here,
PAM testing may also have a role as a screening
test for conductive losses due to SOM.
ACKNOWLEDGMENTS
This work was supported by a grant from the
Depart-ment of Health and Social Security.
We thank Jonathan Hazell, John Graham, Liz
Meadows, and the staff of the Audiology Clinic, Royal
Ear Hospital, for the detailed hearing assessments on the
children with abnormal postauricular myogenic
re-sponses; and Grace Rawlings for advice on the
psycholog-ical aspects of the study.
REFERENCES
1. Kiang NYS, Crist AH, French MA, et al: in quarterly
prog-rem report, Laboratory ofElectronics, Massachusetts Insti-tute of Technology 1963;68:218
2. Douek E, Gibson W, Humphries K: The crossed acoustic
response. JLaryngol Otol 1973;87:711-726
3. Fraser JG, Conway MJ, Keene MH, et al: The post-auricular
myogenic response: A new instrument which simplifies its
detection by machine scoring. J Laryngol Otol
1978;92:293-303
4. Stewart AL, Turcan D, Rawlings G, et al: Outcome for infants at high risk for major handicap. Ciba Found Symp 1978;50
5. Reynell J: Manual for the Reynell Developmental Lan-guage Scales (revised edition). NFER, 1977
6. Lowe M, Costello A: The Symbolic Play Test. (experimental edition). NFER, 1976
7. Knobloch H, Pasamanick B, Sherard ES: A developmen-t,al screening inventory for infants. Pediatrics 1966;
30(suppl):1095-1108
8. Terman LM, Merrill MA: Stanford-Binet Intelligence Scale (Form LM). London, Han-up, 1961
9. Stutsman R: Mental Measurement of Pre-school Children. New York, Harcourt, Brace and World, 1931
10. Ruben RI: in Barnett HL, Einhom AH, Paediatrics, ed 15. (eds): London, Butterworths, 1972
11. Abramovich SJ, Gregory 5, Slemick M, et al: Hearing loss in very low birthweight infants treated with neonatal intensive
care. Arch Di.s Child 1979;54:421-426
12. Stennert E, Schulte FJ, Vollrath M: Incubator noise and
hearing loss. Early Hum Dev 1977;1:113-115
13. Balkany TJ, Berman SA, Simmons MA, et al: Middle ear
effusions in neonates. Laryngoscope 1978;88:398-405 14. Bench J, Boscak N: Some applications of signal detection
theory to paedo-audiology. Sound 1970;4:58-61
15. Fisch L: The probability of response to test sounds in young
children. Sound 1971;5:7-10
16. Taylor IG: The deaf child, in Ballantyne J, Grooves J (eds):
The Ear, ed 4. London, Butterworths, 1979, vol 2
17. Simmons FB, McFarland WF, Jones FR: An automated hearing screening technique for newborns. Acta Otolaryngol
