- Email: [email protected]
Assessment of cisplatin-induced ototoxicity using derived-band ABRs
International Journal of Pediatric Otorhinolavngology, 22 (1991) 237-248 0 1991 Elsevier Science Publishers B.V. All rights reserved 016%5876/91/$03.50
237
PEDOT 00743
Assessment of cisplatin-induced ototoxicity using derived-band ABRs * Stuart G. Coupland ‘,2, Curtis W. Ponton ‘, Jos J. Eggermont Thomas J. Bowen ‘S and Ronald M. Grant 1*5
3,
Departments of ’ Pediatrics, ’ Clinical Neurosciences and Surgery, and .’ Psychology, Unil~ersityof Calgary, Alta. (Canada), ’ House Ear Institute, Los Angeles, CA (U.S.A.) and 5 Pediatric Oncology Clinic, Tom Baker Center and Alberta Children’s Hospital, Calgary, Alta. (Canada)
(Received 12 February 1991) (Revised version received 12 April 1991) (Accepted 21 April 1991)
Key words: Cisplatin; Ototoxicity; Hearing loss; Auditory brainstem response: Derived auditory brainstem response
Abstract
Ototoxicity is an adverse side effect of numerous therapeutic agents (aminoglycoside antibiotics, blood chelating agents, diuretics and oncologic drugs) used in treatment of both adult and pediatric patients. Recently, there has been increasing interest in using the auditory brainstem response (ABR) to detect both short-term effects of ototoxicity in adults and long-term effects of drug administration on neonates and children. Since click ABRs have relatively poor frequency selectivity they best approximate the pure-tone hearing threshold in the 2000-4000 Hz frequency range. Hearing loss above or below that frequency range can be present without producing significant abnormalities in the ABR waveform parameters. Frequency-specific ABRs can be obtained using the derived response technique. The purpose of this study was to investigate early cisplatin ototoxicity using both the broadband click and derived ABR and to monitor progressive hearing Ioss with repeated drug trials in 18 patients studied over a 2-year period. ABRs were obtained serially prior to and following intravenous administration of cisplatin. Derived ABRs were found to be more sensitive than broadband click ABR in
* This paper was first presented at the 4th International Evoked Potentiat Symposium. Toronto, Canada, September, 1990. Correspondence: S.G. Coupland, Sensory Physiology Laboratory, Alberta Children’s Hospital. 1820 Richmond Road SW, Calgary, Alta., Canada T2T X7.
238
detecting early high-frequency hearing loss. For click ABRs, the cumulative dosage of cisplatin at age of ABR examination was correlated with hearing loss in only those patients under 3 years of age. No significant correlation was found between cumulative cisplatin dosage when tested and degree of hearing loss in those patients over 3 years of age.
Introduction Ototoxicity is an adverse side effect of numerous therapeutic agents (aminoglycoside antibiotics, blood chelating agents, diuretics and oncologic drugs) used in treatment of both adult and pediatric patients. Aminoglycoside antibiotics are frequently used for treatment of upper respiratory infections and are known to have both nephrotoxic and ototoxic effects [ll]. Clinically, the manifestations of ototoxic damage include an initial high-frequency sensorineural hearing loss which impinges upon lower frequency range with prolonged drug exposure [12]. Unlike aminoglycoside nephrotoxicity, the auditory toxicity produced by aminoglycosides is frequently irreversible resulting in permanent hearing loss. In the pediatric population even mild to moderate hearing loss may seriously impede the normal progress of language development. Thus, the early detection and evaluation of hearing loss in infants at risk for ototoxicity is of unquestionable value. Other drugs used in pediatric medicine are considerably more ototoxic than aminoglycosides. Cisplatin (cis-diaminedichloroplatinum; DDP) is a chelated heavy metal with proven antitumor activity used in the treatment of recurrent pediatric brain tumors. In recent studies using pure-tone audiometry, significant sensorineural hearing impairment has been identified in all patients undergoing cisplatin chemotherapy [13] and even in patients having only one therapy trial [S]. These studies clearly demonstrate the highly ototoxic nature of cisplatin and it has been reported that the value of therapy must be balanced against the risk of significant and irreversible toxicity. The auditory brainstem response (ABR) consists of a series of 5-7 vertex-positive waves generated in the first 10 ms following the presentation of a brief transient acoustic stimulus such as a click or a tonepip. The ABR is easily measured and has proven clinically useful in the diagnosis of neurological [15] and audiological disease [6]. Recently, there has been increasing interest in using the ABR to detect both short-term effects of aminoglycoside ototoxicity [9] in adults as well as long-term effects of drug administration on neonates [l] and children [lo]. Generally, the ABR has been shown to be less sensitive than pure-tone audiometry in detection of early high-frequency hearing loss in these patients. Click ABRs have relatively poor frequency selectivity and best approximate the pure-tone hearing threshold in the 2000-4000 Hz frequency range [5,16]. Hearing loss above or below that frequency range can be present without producing significant abnormalities in the ABR waveform. Frequency-specific ABRs can be obtained using a derived technique [3] using broadband clicks embedded in high-pass
2.79
filtered white or pink noise. This technique has the advantage of deriving placespecific responses from the basilar membrane. Click ABRs were acquired in the presence of high-pass filtered noise which effectively masked response from the more basal portion of the basilar membrane down to the region of click activation. Consecutive ABRs obtained at one octave differences in high-pass noise cut-off frequency were subtracted to acquire the derived ABR in the narrow-frequency band. The ABR waveform at narrow-frequency bands can quite easily be obtained between 250 Hz and 12 kHz and reconstruction of the audiogram is possible using the derived ABR technique [4]. The derived ABR has been used successfully to investigate hearing loss in animals [14]; however, to our knowledge there are no derived response measurements relating to ototoxicity in humans nor are there any ABR measurements relating to cisplatin-induced hearing loss in humans. The purpose of this study was to investigate early cisplatin ototoxicity using both broadband click and derived ABR techniques and to monitor progressive hearing loss with repeated drug trials. In addition, we were concerned whether there was evidence of age-dependent susceptibility to cisplatin-induced hearing impairment.
Methods
Eighteen patients (8 male, 10 female) obtained from the Pediatric Oncology Program at Alberta Children’s Hospital were studied over a 2-year period. Patient ages at testing ranged from 6 months to 16 years and all patients were tested in the supine position. All newly diagnosed patients had a pretreatment click and derived ABR examination prior to beginning cisplatin therapy. When possible, a repeat study was conducted within a 72-h period following intravenous therapy. Most patients were followed on a monthly basis until ABR changes stabilized. ABR recordings were done either in a quiet neurophysiology laboratory (ambient noise level between 40 and 45 dB SPL) or when necessary, at the patient’s bedside. Approximately 20% of all studies were done at the patient’s bedside. Two estimates of hearing level were studied: the click ABR wave V threshold and the presence of the derived ABR wave V component at two intensity levels. Broadband click ABR recording
Standard broadband click ABR data were collected from patients using monaural condensation click stimulation at a level of 70 dB nHL using Sony MDR-E225 headphones. The clicks were presented at a rate of 11.9/s with recording electrodes located at the vertex CC,), with a reference electrode located on the earlobe ipsilateral to stimulation. The ground electrode was located on the forehead. Electrode impedances were monitored periodically and maintained below 5 kfi. Two series of 1000 artifact-free samples were obtained with an amplifier gain of 10,000 X and recording bandpass of 150-3000 Hz. Wave V threshold determination was performed at a 41.9/s rate with stimulus intensity decreasing in 5 dB steps until the wave V component was no longer detectable. Patients were
240
compared to age-matched hearing loss present.
normative
control
data to determine
the degree of
Derived ABR recording
For the high pass masking conditions, the click and white noise from a Nicolet NIC-1007A noise masking module were presented at the same level; 60 or 30 dB nHL. During the high-pass noise masking conditions, the white noise was high-pass filtered through two cascaded channels of a Krohn-Hite 3342 filter (96 dB/ octave slope) at 16, 8, 4, 2, 1, and 0.5 kHz, respectively. For each condition, the final ABR was based on the averaged response from 1000 click presentations. The complete high-pass filtered protocol took on average about 20 min/ear. The derived ABR was obtained using the method previously described [3] at center frequencies corresponding to 11.3, 5.6, 2.8, 1.4 and 0.7 kHz. The wave V component of the derived ABR was visually inspected and judged to be present or absent at each of the two intensity levels.
Results Cisplatin effects on broadband click ABR threshold
The relationship between click ABR wave V threshold, patient age and cumulative dosage of cisplatin is described in Table I and Figs. 1,3. Elevation in click ABR wave V threshold was used to estimate the amount of hearing loss. Figure 1 includes data from all children and reveals that there is extreme variability in the degree of hearing loss associated with cumulative drug levels with no apparent relationship being seen. In patients over 3 years of age (see Fig. 3) no significant correlation exists between cumulative cisplatin dosage when tested and degree of hearing loss was found. The cumulative dosage of cisplatin at age of click ABR testing, however, was highly correlated with hearing loss in those patients under 3 years (r = 0.88, F = 27.5, P < 0.0005) as seen in Table I and Fig. 3.
TABLE I Relationship between the cumulative dosage of cisplatin at ABR testing and dB hearing loss in patients grouped by age
ns = not significant. Patient group
r
All ages (n = 34 ears)
0.09
< 3 years (n = 11 ears)
0.88
> 3 years (n = 23 ears)
0.02
F-test
P
0.27
ns
27.5 0.01
0.0005 ns
241
All Children
__
f6.09 50.
0
40.
0 0
30.
0
201
0
0
100 2do 360 400 Cumulative Cisplatin (mglm2)
0 Fig. I. No significant time of ABR
relationship
(r = 0.091 was found between
testing and dB hearing loss in the entire
.
60
I
relationship
time of ABR
the cumulative children
dosage of cisplatin
at
aged 6 months to I6 years.
ChIldron Under 3 yews r-.66
t
0
t
.
0 Fig. 2. A significant
group of
600
500
25 50 75 100 125 150 Cumulative Cisplatin (mgIm2) (r = 0.88) was found between
175
200
the cumulative
testing and dB hearing loss in those children
dosage of cisplatin
al
under 3 years of age
Children Over 3 your 601 b.02 0
+ 1
0
0
100 200 300 400 Cumulative Cisplatin (mg/m2)
500
600
Fig. 3. No significant relationship (r = 0.02)was found between the cumulative dosage of cisplatin time of ABR testing and dB hearing loss in those children over 3 years of age.
at
242
I
12
I
I
3
I,,
4
66
,
,
,
7
8
91Olll21314l6
,
,
,
,
,
,
Latency (meed Fig. 4. Derived ABRs from a 3-year-old patient obtained prior to cisplatin treatment with components labelled as I through V. CM indicates the presence of the cochlear microphonic preceding wave I in these data. Derived ABR wave V components are visible at all center frequencies.
Cisplatin effects on presence of derived ABR at fixed levels
Pretreatment baseline derived ABRs were obtained on each newly diagnosed patient. Figure 4 illustrates the derived-band ABRs obtained at 60 dB nHL on a 3-year-old patient. The presence of wave V can be seen at each of the center frequencies between 0.7 and 11.3 kHz, respectively. Figure 5 illustrates the derived-band ABRs obtained from the same patient following a single cisplatin treatment with a dosage of 55 mg/m2. The wave V component is absent or ambiguous in 11.3, 5.7 and 2.8 kHz derived responses but appears to be present at 1.4 and 0.7 kHz. Interestingly, the unmasked broadband click ABR (uppermost tracing in Fig. 5) is still well preserved and within normal limits. In Fig. 6 the presence of each derived ABR wave V component (at 60 dB nHL) obtained at center frequencies between 0.7 and 11.3 kHz are indicated by symbols.
ACIitBB1616 I II
I=
0.26
m
I1IIII11I~~‘~~J 1 2 3 4
5
6
7
Latency Fig. 5. Derived-ABRs from a 3-year-old patient obtained wave V components from 1 I.3 to 2.X kHz center
8
9
10 ll
pvoltkl
12 13 14 16
(msec) after single cisplatin treatment. Derived-ABR frequencies are absent or ambiguous.
In this 15-year-old, pretreatment-derived ABRs were obtained at all center frcquencies at 60 dB nHL. For accumulated cisplatin dosage up to and including 35X mg/m2 the wave V component was present at all center frequencies. At the 443 mg/m2 level there was a loss of the derived ABR in the 11.4 kHz frequency band at 60 dB nHL. At this cumulative dosage level the broadband click ABR wave V thresholds were 40 dB nHL. Figure 7 illustrates the derived ABR results at 30 dB nHL intensity obtained from the same patient. At an accumulated dosage of 353 mg/m’ there was a loss of derived ABR at all center frequencies at the 30 dB nHL level in this 15-year-old although the component was still present at the higher intensity suggesting a moderate threshold increase. The early loss of derived ABR wave V components was more apparent in the younger children. Figure 8 illustrates the presence of wave V in the derived ABRs
244 Cisplatln
A
Inducad
hearing
A
A
loss
+ A
-
MC 15 yn. (150 dB nHL)
250
150
n
amp.
0
57w
A
117nrp.
+
1Ymg.
Cl
=m(l.
A MQ.
50
-50
10
1
.l
Cantor
Fraquoncy
4 100
(kHr.)
Fig. 6. Derived ABR presence for 60 dB nHL click obtained at center frequencies between 0.7 and 11.4 kHz as indicated by symbols. In this 15-year-old, derived ABRs were obtained at all center frequencies for accumulated cisplatin dosage less than 443 mg/m*. At that level there was a loss of the derived ABR at the 11.4 kfiz frequency band.
Clsplatln Induced hearing loss
= t-l MC
15 yrs. (30 dB nHL)
350 -
ow
250 -
150-
57~.
A
ll7mg.
+
146mg. 353mg
t
443 m.J.
l
50-
l
l
l
l
l
-50 1
1
Center Frequency
10
100
(kHs.)
Fig. 7. Derived ABR presence for 30 dB nHL click obtained at center frequencies between 0.7 and 11.4 kHz as indicated by symbols in the same subject as illustrated in Fig. 6. Arrows indicate that at an accumulated dosage of 353 mg/m2 there was a loss of derived ABR at all center frequencies.
245 Claplatin lnducod hoaring loss
UM 4.5 yrs. (60 dE nHL) 0
0
l
0
1
I.
l zanmp
Center Frequency
(kHz.)
Fig. 8. Derived ABR presence for 60 dB nHL click obtained at center frequencies between 0.7 and kHz. as indicated by symbols in an 4.5-year-old. Arrows indicate loss of the derived ABR at the kHz frequency with an accumulated dosage of 280 mg/m*.
11.4 1I .4
obtained in a 4.5year-old patient at 60 dB nHL. There was clear indication of loss of the derived ABR at the 11.3 kHz frequency with an accumulated dosage of 280 mg/m2. At this accumulated dosage level the broadband click ABR thresholds Cisplatin
induced hearing loss
TC .7 yrs. (60 dB nHL)
(j0l
n
omg
0 A
92mg 152mg
LIO0
Center
Frequency
(kHz.)
Fig. 9. Derived ABR presence for 60 dB nHL click obtained at center frequencies between 0.7 and 11.4 kHz as indicated by symbols in an &month-old infant. Arrows indicate early loss of the derived ABR at the 11.4kHz frequency with an accumulated dosage of 92 mg/m”. With increasing cisplatin levels there is a dramatic loss of derived ABR at frequencies down to 1.4 kHz.
246
were estimated at 10 dB nHL. Figure 9 illustrates the early loss of the derived ABR in an S-month-old infant after an accumulated cisplatin dosage of only 92 mg/m’. With increasing drug levels there is a dramatic loss of derived ABR at frequencies down to 1.4 kHz. Broadband click ABR threshold was 30 dB nHL at this same accumulated dosage level.
Discussion
Oncology patients exposed to cisplatin are at significant risk for the development of drug-induced sensorineural hearing loss and the majority of patients treated with cisplatin develop hearing impairment early in the course of treatment. In our group of pediatric patients, the relationship between maturational age at treatment and magnitude of hearing loss was investigated using both the broadband click ABR threshold and presence of derived ABR at fixed intensity levels to estimate degree of hearing impairment. In those children under 3 years of age, we found a significant relationship between the cumulative dosage of cisplatin at time of click ABR testing and ABR threshold increase. No significant relationship between cumulative dosage and subsequent hearing loss was established in the older children. One interpretation of these data is that younger children may be more susceptible to drug-induced hearing loss as estimated by the broadband click ABR threshold and derived ABR presence. Like other investigators, we observed initial high frequency hearing loss with subsequent spread into the lower frequency range. The absence of wave V was first apparent in the derived ABRs at higher center frequency bands, and with subsequent chemotherapy trials wave V absence was also noted in the lower frequency bands. Infants and younger children showed loss of derived responses at lower normalized cumulative cisplatin levels than older children providing further confirmation that there is significantly greater risk for hearing loss in this population. One clinical implication of these findings would appear to be the necessity for frequency-specific hearing assessment of those younger patients receiving cisplatin chemotherapy. Another consistent finding in our patients was the earlier loss of derived ABR in the high frequency range when broadband click ABR thresholds were still within normal limits. Derived-band ABR measurement is much more frequency-specific than standard broadband click ABR and is a more efficient assessment of hearing level than behavioral field testing in this age group. The need for longitudinal ABR testing cannot be overemphasized as we have observed changes in the derived ABR occurring some 12 weeks following chemotherapy treatment. It would appear that some of the effects of ototoxicity are not immediately detectable in the brainstem response or that the degree of neurotoxicity is cumulative with time. In summary, we have demonstrated the usefulness of the derived ABR in detection of early frequency-specific hearing loss due to cisplatin ototoxicity. Onset of hearing loss occurred earlier after treatment in those patients under 3 years of age. In this same group the magnitude of hearing loss was shown to be dose-de-
347
pendent. These findings stress the demand for repeated derived ABR testing of hearing status in infants and young children receiving cisplatin chemotherapy. Given that the complete high-pass filtered protocol took on average about an additional 20 min/ear, the information derived from this additional testing was clearly clinically beneficial. There were no practical difficulties encountered with derived ABR testing. These studies required some additional equipment and time to conduct but they can be performed reliably in the clinical laboratory setting or even at the patient’s bedside. Approximately 20% of all studies were done at the patient’s bedside. More importantly, the results of these electrodiagnostic investigations revealed early evidence of auditory dysfunction (see Fig. 9>, before there was any clinical evidence of hearing impairment. The earlier detection of hearing impairment with the derived ABR in conjunction with behavioral testing has led to earlier amplification in those affected patients.
Acknowledgements
This research was supported by a grant from the MS1 Foundation (to S.G.C. and J.J.E.). C.W.P. was supported by an Alberta Heritage Foundation for Medical Research Graduate Studentship. The investigators wish to thank Ms. Sandi Miki, R.N. for her technical assistance in these investigations.
References 1 Bernard,
P.A.,
treated
Pechere,
J.C.
human neonates,
2 Catlin,
F.I., Prevention
and Hebert,
R.. Altered
Arch. Otorhinolaryngol.,
of hearing impairment
objective
228 (1980)
from infection
audiometry
in aminoglycosides-
205-210. and ototoxic drugs, Arch. Otolaryngol..
111 (1985) 377-384. 3 Don,
M.
and Eggermont,
J.J., Analysis
high-pass noise masking, J. Acoust. 4 Don,
M., Eggermont,
J.J. and Brackmann,
responses and high-pass masking, Ann. 5 Eggermont.
J.J.,
applications.
The
Ann.,
6 Eggermont,
Grune
7 Eggermont,
J.J. and Don,
8 Granowetter. 9 Guerit.
L.,
D.E.,
Otol.
of click-evoked
J.M., J.W.,
Herndon, patients
auditory
J.G.
and
treated
of the audiogram Suppl. 57 (1979)
brainstem
(Ed.),
using
using brainstem l-20.
responses
Gary, with
R.J.,
Oncoi.,
in audiological
Brain-Stem
and Winkler,
ototoxic
drugs,
potentials
Sot. Am., 68 (1980)
Evoked
Enhanced
in man using high 1671-1675.
cis-platinum
neurotoxicity
in
1 (19831 293-297.
S. and Herbay,
in man, Arch.
L.B.
Bases of Auditory brainstem
J. Acoust.
Packer,
P., Houben-Giurgea, D.N.,
in man
1983, pp. 2X7-315.
M., Analysis of the click-evoked
evoked potentials
burn-wound
Laryngol.,
In E.J. Moore
New York,
II. Effect of click intensity.
Mahieu,
auditory
potentials
Sci., 388 (19821 707-709. Disorders.
Rosenstock,
brainstem
1084-1092.
Reconstruction
Rhinol.
patients with brain tumors, J. Nemo.
brainstem 10 Hall.
Acad.
& Stratton,
pass noise masking. pediatric
inadequacy
N.Y.
J.J., Audiological
Responses,
of the click-evoked
Sot. Am., 63 (1978)
S.. The influence
Otorhinolaryngol., J.B., Auditory Int.
J. Pediatr.
of ototoxic
drugs on
233 (19811 189-199. brainstem
response
Otorhinolaryngol.,
in young I7
(1986)
187-203. 11 Matz, Lerner,
G.J.
and Lerner,
G.J.
Matz,
1981. pp. 327-339.
S.A.,
Prospective
and J.E Hawkins
studies of aminoglycoside
(Eds.),
Aminoglycoside
ototoxicity
Ototoxicity.
Little.
in adults.
In S.A.
Brown,
Boston,
248 12 Rybak, L.P., Treatable Sensorineural hearing loss, Am. J. Otolaryngol., 6 (198.5) 482-489. 13 Sexauer, CL., Khan, A., Burger, P.C., Krischer, P., Van Eys, J., Vats, T. and Ragab, A.H., Cisplatin in recurrent pediatric brain tumors: a POG phase II study, Cancer, 56 (1985) 1497-1501. 14 Shepherd, R.K. and Clark, G.M., Progressive ototoxicity of neomycin monitored using derived brainstem response audiometry, Hear. Res., 18 (1985) 105-110. 15 Starr, A. and Hamilton, A., Correlation between confirmed sites of neurological lesions and abnormalities of far-field auditory brainstem responses, Electroenceph. Clin. Neurophysiol., 41 (1976) 595-608. 16 Van der Drift, J.F.C., Brocaar, M.P. and Van Zanten, G.A., The relation between pure-tone audiogram and the click auditory brainstem response threshold in cochlear hearing loss, Audiology, 26 (1987) l-10.
237
PEDOT 00743
Assessment of cisplatin-induced ototoxicity using derived-band ABRs * Stuart G. Coupland ‘,2, Curtis W. Ponton ‘, Jos J. Eggermont Thomas J. Bowen ‘S and Ronald M. Grant 1*5
3,
Departments of ’ Pediatrics, ’ Clinical Neurosciences and Surgery, and .’ Psychology, Unil~ersityof Calgary, Alta. (Canada), ’ House Ear Institute, Los Angeles, CA (U.S.A.) and 5 Pediatric Oncology Clinic, Tom Baker Center and Alberta Children’s Hospital, Calgary, Alta. (Canada)
(Received 12 February 1991) (Revised version received 12 April 1991) (Accepted 21 April 1991)
Key words: Cisplatin; Ototoxicity; Hearing loss; Auditory brainstem response: Derived auditory brainstem response
Abstract
Ototoxicity is an adverse side effect of numerous therapeutic agents (aminoglycoside antibiotics, blood chelating agents, diuretics and oncologic drugs) used in treatment of both adult and pediatric patients. Recently, there has been increasing interest in using the auditory brainstem response (ABR) to detect both short-term effects of ototoxicity in adults and long-term effects of drug administration on neonates and children. Since click ABRs have relatively poor frequency selectivity they best approximate the pure-tone hearing threshold in the 2000-4000 Hz frequency range. Hearing loss above or below that frequency range can be present without producing significant abnormalities in the ABR waveform parameters. Frequency-specific ABRs can be obtained using the derived response technique. The purpose of this study was to investigate early cisplatin ototoxicity using both the broadband click and derived ABR and to monitor progressive hearing Ioss with repeated drug trials in 18 patients studied over a 2-year period. ABRs were obtained serially prior to and following intravenous administration of cisplatin. Derived ABRs were found to be more sensitive than broadband click ABR in
* This paper was first presented at the 4th International Evoked Potentiat Symposium. Toronto, Canada, September, 1990. Correspondence: S.G. Coupland, Sensory Physiology Laboratory, Alberta Children’s Hospital. 1820 Richmond Road SW, Calgary, Alta., Canada T2T X7.
238
detecting early high-frequency hearing loss. For click ABRs, the cumulative dosage of cisplatin at age of ABR examination was correlated with hearing loss in only those patients under 3 years of age. No significant correlation was found between cumulative cisplatin dosage when tested and degree of hearing loss in those patients over 3 years of age.
Introduction Ototoxicity is an adverse side effect of numerous therapeutic agents (aminoglycoside antibiotics, blood chelating agents, diuretics and oncologic drugs) used in treatment of both adult and pediatric patients. Aminoglycoside antibiotics are frequently used for treatment of upper respiratory infections and are known to have both nephrotoxic and ototoxic effects [ll]. Clinically, the manifestations of ototoxic damage include an initial high-frequency sensorineural hearing loss which impinges upon lower frequency range with prolonged drug exposure [12]. Unlike aminoglycoside nephrotoxicity, the auditory toxicity produced by aminoglycosides is frequently irreversible resulting in permanent hearing loss. In the pediatric population even mild to moderate hearing loss may seriously impede the normal progress of language development. Thus, the early detection and evaluation of hearing loss in infants at risk for ototoxicity is of unquestionable value. Other drugs used in pediatric medicine are considerably more ototoxic than aminoglycosides. Cisplatin (cis-diaminedichloroplatinum; DDP) is a chelated heavy metal with proven antitumor activity used in the treatment of recurrent pediatric brain tumors. In recent studies using pure-tone audiometry, significant sensorineural hearing impairment has been identified in all patients undergoing cisplatin chemotherapy [13] and even in patients having only one therapy trial [S]. These studies clearly demonstrate the highly ototoxic nature of cisplatin and it has been reported that the value of therapy must be balanced against the risk of significant and irreversible toxicity. The auditory brainstem response (ABR) consists of a series of 5-7 vertex-positive waves generated in the first 10 ms following the presentation of a brief transient acoustic stimulus such as a click or a tonepip. The ABR is easily measured and has proven clinically useful in the diagnosis of neurological [15] and audiological disease [6]. Recently, there has been increasing interest in using the ABR to detect both short-term effects of aminoglycoside ototoxicity [9] in adults as well as long-term effects of drug administration on neonates [l] and children [lo]. Generally, the ABR has been shown to be less sensitive than pure-tone audiometry in detection of early high-frequency hearing loss in these patients. Click ABRs have relatively poor frequency selectivity and best approximate the pure-tone hearing threshold in the 2000-4000 Hz frequency range [5,16]. Hearing loss above or below that frequency range can be present without producing significant abnormalities in the ABR waveform. Frequency-specific ABRs can be obtained using a derived technique [3] using broadband clicks embedded in high-pass
2.79
filtered white or pink noise. This technique has the advantage of deriving placespecific responses from the basilar membrane. Click ABRs were acquired in the presence of high-pass filtered noise which effectively masked response from the more basal portion of the basilar membrane down to the region of click activation. Consecutive ABRs obtained at one octave differences in high-pass noise cut-off frequency were subtracted to acquire the derived ABR in the narrow-frequency band. The ABR waveform at narrow-frequency bands can quite easily be obtained between 250 Hz and 12 kHz and reconstruction of the audiogram is possible using the derived ABR technique [4]. The derived ABR has been used successfully to investigate hearing loss in animals [14]; however, to our knowledge there are no derived response measurements relating to ototoxicity in humans nor are there any ABR measurements relating to cisplatin-induced hearing loss in humans. The purpose of this study was to investigate early cisplatin ototoxicity using both broadband click and derived ABR techniques and to monitor progressive hearing loss with repeated drug trials. In addition, we were concerned whether there was evidence of age-dependent susceptibility to cisplatin-induced hearing impairment.
Methods
Eighteen patients (8 male, 10 female) obtained from the Pediatric Oncology Program at Alberta Children’s Hospital were studied over a 2-year period. Patient ages at testing ranged from 6 months to 16 years and all patients were tested in the supine position. All newly diagnosed patients had a pretreatment click and derived ABR examination prior to beginning cisplatin therapy. When possible, a repeat study was conducted within a 72-h period following intravenous therapy. Most patients were followed on a monthly basis until ABR changes stabilized. ABR recordings were done either in a quiet neurophysiology laboratory (ambient noise level between 40 and 45 dB SPL) or when necessary, at the patient’s bedside. Approximately 20% of all studies were done at the patient’s bedside. Two estimates of hearing level were studied: the click ABR wave V threshold and the presence of the derived ABR wave V component at two intensity levels. Broadband click ABR recording
Standard broadband click ABR data were collected from patients using monaural condensation click stimulation at a level of 70 dB nHL using Sony MDR-E225 headphones. The clicks were presented at a rate of 11.9/s with recording electrodes located at the vertex CC,), with a reference electrode located on the earlobe ipsilateral to stimulation. The ground electrode was located on the forehead. Electrode impedances were monitored periodically and maintained below 5 kfi. Two series of 1000 artifact-free samples were obtained with an amplifier gain of 10,000 X and recording bandpass of 150-3000 Hz. Wave V threshold determination was performed at a 41.9/s rate with stimulus intensity decreasing in 5 dB steps until the wave V component was no longer detectable. Patients were
240
compared to age-matched hearing loss present.
normative
control
data to determine
the degree of
Derived ABR recording
For the high pass masking conditions, the click and white noise from a Nicolet NIC-1007A noise masking module were presented at the same level; 60 or 30 dB nHL. During the high-pass noise masking conditions, the white noise was high-pass filtered through two cascaded channels of a Krohn-Hite 3342 filter (96 dB/ octave slope) at 16, 8, 4, 2, 1, and 0.5 kHz, respectively. For each condition, the final ABR was based on the averaged response from 1000 click presentations. The complete high-pass filtered protocol took on average about 20 min/ear. The derived ABR was obtained using the method previously described [3] at center frequencies corresponding to 11.3, 5.6, 2.8, 1.4 and 0.7 kHz. The wave V component of the derived ABR was visually inspected and judged to be present or absent at each of the two intensity levels.
Results Cisplatin effects on broadband click ABR threshold
The relationship between click ABR wave V threshold, patient age and cumulative dosage of cisplatin is described in Table I and Figs. 1,3. Elevation in click ABR wave V threshold was used to estimate the amount of hearing loss. Figure 1 includes data from all children and reveals that there is extreme variability in the degree of hearing loss associated with cumulative drug levels with no apparent relationship being seen. In patients over 3 years of age (see Fig. 3) no significant correlation exists between cumulative cisplatin dosage when tested and degree of hearing loss was found. The cumulative dosage of cisplatin at age of click ABR testing, however, was highly correlated with hearing loss in those patients under 3 years (r = 0.88, F = 27.5, P < 0.0005) as seen in Table I and Fig. 3.
TABLE I Relationship between the cumulative dosage of cisplatin at ABR testing and dB hearing loss in patients grouped by age
ns = not significant. Patient group
r
All ages (n = 34 ears)
0.09
< 3 years (n = 11 ears)
0.88
> 3 years (n = 23 ears)
0.02
F-test
P
0.27
ns
27.5 0.01
0.0005 ns
241
All Children
__
f6.09 50.
0
40.
0 0
30.
0
201
0
0
100 2do 360 400 Cumulative Cisplatin (mglm2)
0 Fig. I. No significant time of ABR
relationship
(r = 0.091 was found between
testing and dB hearing loss in the entire
.
60
I
relationship
time of ABR
the cumulative children
dosage of cisplatin
at
aged 6 months to I6 years.
ChIldron Under 3 yews r-.66
t
0
t
.
0 Fig. 2. A significant
group of
600
500
25 50 75 100 125 150 Cumulative Cisplatin (mgIm2) (r = 0.88) was found between
175
200
the cumulative
testing and dB hearing loss in those children
dosage of cisplatin
al
under 3 years of age
Children Over 3 your 601 b.02 0
+ 1
0
0
100 200 300 400 Cumulative Cisplatin (mg/m2)
500
600
Fig. 3. No significant relationship (r = 0.02)was found between the cumulative dosage of cisplatin time of ABR testing and dB hearing loss in those children over 3 years of age.
at
242
I
12
I
I
3
I,,
4
66
,
,
,
7
8
91Olll21314l6
,
,
,
,
,
,
Latency (meed Fig. 4. Derived ABRs from a 3-year-old patient obtained prior to cisplatin treatment with components labelled as I through V. CM indicates the presence of the cochlear microphonic preceding wave I in these data. Derived ABR wave V components are visible at all center frequencies.
Cisplatin effects on presence of derived ABR at fixed levels
Pretreatment baseline derived ABRs were obtained on each newly diagnosed patient. Figure 4 illustrates the derived-band ABRs obtained at 60 dB nHL on a 3-year-old patient. The presence of wave V can be seen at each of the center frequencies between 0.7 and 11.3 kHz, respectively. Figure 5 illustrates the derived-band ABRs obtained from the same patient following a single cisplatin treatment with a dosage of 55 mg/m2. The wave V component is absent or ambiguous in 11.3, 5.7 and 2.8 kHz derived responses but appears to be present at 1.4 and 0.7 kHz. Interestingly, the unmasked broadband click ABR (uppermost tracing in Fig. 5) is still well preserved and within normal limits. In Fig. 6 the presence of each derived ABR wave V component (at 60 dB nHL) obtained at center frequencies between 0.7 and 11.3 kHz are indicated by symbols.
ACIitBB1616 I II
I=
0.26
m
I1IIII11I~~‘~~J 1 2 3 4
5
6
7
Latency Fig. 5. Derived-ABRs from a 3-year-old patient obtained wave V components from 1 I.3 to 2.X kHz center
8
9
10 ll
pvoltkl
12 13 14 16
(msec) after single cisplatin treatment. Derived-ABR frequencies are absent or ambiguous.
In this 15-year-old, pretreatment-derived ABRs were obtained at all center frcquencies at 60 dB nHL. For accumulated cisplatin dosage up to and including 35X mg/m2 the wave V component was present at all center frequencies. At the 443 mg/m2 level there was a loss of the derived ABR in the 11.4 kHz frequency band at 60 dB nHL. At this cumulative dosage level the broadband click ABR wave V thresholds were 40 dB nHL. Figure 7 illustrates the derived ABR results at 30 dB nHL intensity obtained from the same patient. At an accumulated dosage of 353 mg/m’ there was a loss of derived ABR at all center frequencies at the 30 dB nHL level in this 15-year-old although the component was still present at the higher intensity suggesting a moderate threshold increase. The early loss of derived ABR wave V components was more apparent in the younger children. Figure 8 illustrates the presence of wave V in the derived ABRs
244 Cisplatln
A
Inducad
hearing
A
A
loss
+ A
-
MC 15 yn. (150 dB nHL)
250
150
n
amp.
0
57w
A
117nrp.
+
1Ymg.
Cl
=m(l.
A MQ.
50
-50
10
1
.l
Cantor
Fraquoncy
4 100
(kHr.)
Fig. 6. Derived ABR presence for 60 dB nHL click obtained at center frequencies between 0.7 and 11.4 kHz as indicated by symbols. In this 15-year-old, derived ABRs were obtained at all center frequencies for accumulated cisplatin dosage less than 443 mg/m*. At that level there was a loss of the derived ABR at the 11.4 kfiz frequency band.
Clsplatln Induced hearing loss
= t-l MC
15 yrs. (30 dB nHL)
350 -
ow
250 -
150-
57~.
A
ll7mg.
+
146mg. 353mg
t
443 m.J.
l
50-
l
l
l
l
l
-50 1
1
Center Frequency
10
100
(kHs.)
Fig. 7. Derived ABR presence for 30 dB nHL click obtained at center frequencies between 0.7 and 11.4 kHz as indicated by symbols in the same subject as illustrated in Fig. 6. Arrows indicate that at an accumulated dosage of 353 mg/m2 there was a loss of derived ABR at all center frequencies.
245 Claplatin lnducod hoaring loss
UM 4.5 yrs. (60 dE nHL) 0
0
l
0
1
I.
l zanmp
Center Frequency
(kHz.)
Fig. 8. Derived ABR presence for 60 dB nHL click obtained at center frequencies between 0.7 and kHz. as indicated by symbols in an 4.5-year-old. Arrows indicate loss of the derived ABR at the kHz frequency with an accumulated dosage of 280 mg/m*.
11.4 1I .4
obtained in a 4.5year-old patient at 60 dB nHL. There was clear indication of loss of the derived ABR at the 11.3 kHz frequency with an accumulated dosage of 280 mg/m2. At this accumulated dosage level the broadband click ABR thresholds Cisplatin
induced hearing loss
TC .7 yrs. (60 dB nHL)
(j0l
n
omg
0 A
92mg 152mg
LIO0
Center
Frequency
(kHz.)
Fig. 9. Derived ABR presence for 60 dB nHL click obtained at center frequencies between 0.7 and 11.4 kHz as indicated by symbols in an &month-old infant. Arrows indicate early loss of the derived ABR at the 11.4kHz frequency with an accumulated dosage of 92 mg/m”. With increasing cisplatin levels there is a dramatic loss of derived ABR at frequencies down to 1.4 kHz.
246
were estimated at 10 dB nHL. Figure 9 illustrates the early loss of the derived ABR in an S-month-old infant after an accumulated cisplatin dosage of only 92 mg/m’. With increasing drug levels there is a dramatic loss of derived ABR at frequencies down to 1.4 kHz. Broadband click ABR threshold was 30 dB nHL at this same accumulated dosage level.
Discussion
Oncology patients exposed to cisplatin are at significant risk for the development of drug-induced sensorineural hearing loss and the majority of patients treated with cisplatin develop hearing impairment early in the course of treatment. In our group of pediatric patients, the relationship between maturational age at treatment and magnitude of hearing loss was investigated using both the broadband click ABR threshold and presence of derived ABR at fixed intensity levels to estimate degree of hearing impairment. In those children under 3 years of age, we found a significant relationship between the cumulative dosage of cisplatin at time of click ABR testing and ABR threshold increase. No significant relationship between cumulative dosage and subsequent hearing loss was established in the older children. One interpretation of these data is that younger children may be more susceptible to drug-induced hearing loss as estimated by the broadband click ABR threshold and derived ABR presence. Like other investigators, we observed initial high frequency hearing loss with subsequent spread into the lower frequency range. The absence of wave V was first apparent in the derived ABRs at higher center frequency bands, and with subsequent chemotherapy trials wave V absence was also noted in the lower frequency bands. Infants and younger children showed loss of derived responses at lower normalized cumulative cisplatin levels than older children providing further confirmation that there is significantly greater risk for hearing loss in this population. One clinical implication of these findings would appear to be the necessity for frequency-specific hearing assessment of those younger patients receiving cisplatin chemotherapy. Another consistent finding in our patients was the earlier loss of derived ABR in the high frequency range when broadband click ABR thresholds were still within normal limits. Derived-band ABR measurement is much more frequency-specific than standard broadband click ABR and is a more efficient assessment of hearing level than behavioral field testing in this age group. The need for longitudinal ABR testing cannot be overemphasized as we have observed changes in the derived ABR occurring some 12 weeks following chemotherapy treatment. It would appear that some of the effects of ototoxicity are not immediately detectable in the brainstem response or that the degree of neurotoxicity is cumulative with time. In summary, we have demonstrated the usefulness of the derived ABR in detection of early frequency-specific hearing loss due to cisplatin ototoxicity. Onset of hearing loss occurred earlier after treatment in those patients under 3 years of age. In this same group the magnitude of hearing loss was shown to be dose-de-
347
pendent. These findings stress the demand for repeated derived ABR testing of hearing status in infants and young children receiving cisplatin chemotherapy. Given that the complete high-pass filtered protocol took on average about an additional 20 min/ear, the information derived from this additional testing was clearly clinically beneficial. There were no practical difficulties encountered with derived ABR testing. These studies required some additional equipment and time to conduct but they can be performed reliably in the clinical laboratory setting or even at the patient’s bedside. Approximately 20% of all studies were done at the patient’s bedside. More importantly, the results of these electrodiagnostic investigations revealed early evidence of auditory dysfunction (see Fig. 9>, before there was any clinical evidence of hearing impairment. The earlier detection of hearing impairment with the derived ABR in conjunction with behavioral testing has led to earlier amplification in those affected patients.
Acknowledgements
This research was supported by a grant from the MS1 Foundation (to S.G.C. and J.J.E.). C.W.P. was supported by an Alberta Heritage Foundation for Medical Research Graduate Studentship. The investigators wish to thank Ms. Sandi Miki, R.N. for her technical assistance in these investigations.
References 1 Bernard,
P.A.,
treated
Pechere,
J.C.
human neonates,
2 Catlin,
F.I., Prevention
and Hebert,
R.. Altered
Arch. Otorhinolaryngol.,
of hearing impairment
objective
228 (1980)
from infection
audiometry
in aminoglycosides-
205-210. and ototoxic drugs, Arch. Otolaryngol..
111 (1985) 377-384. 3 Don,
M.
and Eggermont,
J.J., Analysis
high-pass noise masking, J. Acoust. 4 Don,
M., Eggermont,
J.J. and Brackmann,
responses and high-pass masking, Ann. 5 Eggermont.
J.J.,
applications.
The
Ann.,
6 Eggermont,
Grune
7 Eggermont,
J.J. and Don,
8 Granowetter. 9 Guerit.
L.,
D.E.,
Otol.
of click-evoked
J.M., J.W.,
Herndon, patients
auditory
J.G.
and
treated
of the audiogram Suppl. 57 (1979)
brainstem
(Ed.),
using
using brainstem l-20.
responses
Gary, with
R.J.,
Oncoi.,
in audiological
Brain-Stem
and Winkler,
ototoxic
drugs,
potentials
Sot. Am., 68 (1980)
Evoked
Enhanced
in man using high 1671-1675.
cis-platinum
neurotoxicity
in
1 (19831 293-297.
S. and Herbay,
in man, Arch.
L.B.
Bases of Auditory brainstem
J. Acoust.
Packer,
P., Houben-Giurgea, D.N.,
in man
1983, pp. 2X7-315.
M., Analysis of the click-evoked
evoked potentials
burn-wound
Laryngol.,
In E.J. Moore
New York,
II. Effect of click intensity.
Mahieu,
auditory
potentials
Sci., 388 (19821 707-709. Disorders.
Rosenstock,
brainstem
1084-1092.
Reconstruction
Rhinol.
patients with brain tumors, J. Nemo.
brainstem 10 Hall.
Acad.
& Stratton,
pass noise masking. pediatric
inadequacy
N.Y.
J.J., Audiological
Responses,
of the click-evoked
Sot. Am., 63 (1978)
S.. The influence
Otorhinolaryngol., J.B., Auditory Int.
J. Pediatr.
of ototoxic
drugs on
233 (19811 189-199. brainstem
response
Otorhinolaryngol.,
in young I7
(1986)
187-203. 11 Matz, Lerner,
G.J.
and Lerner,
G.J.
Matz,
1981. pp. 327-339.
S.A.,
Prospective
and J.E Hawkins
studies of aminoglycoside
(Eds.),
Aminoglycoside
ototoxicity
Ototoxicity.
Little.
in adults.
In S.A.
Brown,
Boston,
248 12 Rybak, L.P., Treatable Sensorineural hearing loss, Am. J. Otolaryngol., 6 (198.5) 482-489. 13 Sexauer, CL., Khan, A., Burger, P.C., Krischer, P., Van Eys, J., Vats, T. and Ragab, A.H., Cisplatin in recurrent pediatric brain tumors: a POG phase II study, Cancer, 56 (1985) 1497-1501. 14 Shepherd, R.K. and Clark, G.M., Progressive ototoxicity of neomycin monitored using derived brainstem response audiometry, Hear. Res., 18 (1985) 105-110. 15 Starr, A. and Hamilton, A., Correlation between confirmed sites of neurological lesions and abnormalities of far-field auditory brainstem responses, Electroenceph. Clin. Neurophysiol., 41 (1976) 595-608. 16 Van der Drift, J.F.C., Brocaar, M.P. and Van Zanten, G.A., The relation between pure-tone audiogram and the click auditory brainstem response threshold in cochlear hearing loss, Audiology, 26 (1987) l-10.