Distortion-product otoacoustic emissions testing in neonates treated with an aminoglycoside in a neonatal intensive care unit.
We evaluated the ototoxic effect of aminoglycosides on the outer hair cells of newborns in a neonatal intensive care unit (NICU) by means of distortion-product otoacoustic emissions (DPOAE) testing. Our study population was made up of 164 newborns who were divided into three groups: group A consisted of 105 infants who were given aminoglycoside therapy (either gentamicin or amikacin, or a combination of the two) as treatment for suspected or proven bacterial infection and septic states; group B included 30 newborns who were not given an antibiotic or who were given an antibiotic other than an aminoglycoside; group C, a control group, was made up of 29 healthy neonates who were hospitalized in the well-baby nursery. All the neonates underwent DPOAE testing in both ears (the [f.sub.2] primary tone was presented at 2.0, 2.5, 3.2, and 4.0 kHz). We found that 41 patients in group A (39.0%) and 13 in group B (43.3%) failed the DPOAE test in one or both ears; the difference between these two groups was not statistically significant (p = 0.673). In group C, the DPOAE fail rate was 13.8% (4 newborns). In group A, there was no statistically significant association between the pass/fail rate and the specific aminoglycoside that was administered, or in the duration of antibiotic treatment, the number of doses, and the size of the mean daily dose and the mean total dose. In clinical practice, DPOAE testing is a sensitive method of evaluating the integrity of the outer hair cells in the basal turn of the cochlea after exposure to ototoxic drugs such as aminoglycosides. However, our study did not demonstrate that the aminoglycosides had any ototoxic effect on the hearing of neonates in the NICU.
Aminoglycosides are widely used in clinical practice for the treatment of tuberculosis and gram-negative aerobic bacterial infections. Inner ear toxicity caused by aminoglycosides has been extensively described in numerous studies. (1-4) The classic audiometric pattern consists of a high-frequency downsloping sensorineural hearing loss. The main site of damage is the outer hair cells in the basal turn of the cochlea, where aminoglycosides are received from the lysosomes of the cells' cytoplasm, which leads to the local formation of toxic free radicals and finally to hair cell apoptosis. (5)
Aminoglycosides are also commonly used to treat suspected or confirmed bacterial infection and septic states in newborns in a neonatal intensive care unit (NICU). According to the Joint Committee on Infant Hearing (JCIH) position statement of 1994, aminoglycosides were among the risk factors implicated in the appearance of newborn hearing loss. (6) However, subsequent research led the JCIH to remove aminoglycosides as a risk factor because of the low percentage of newborns who were being found to be affected by them. (7) Nevertheless, their effect on otoacoustic emissions has not been adequately tested, especially with respect to the duration of treatment and the dosage.
Distortion-product otoacoustic emissions (DPOAE) are the sounds emitted by the outer hair cells of the cochlea in response to a pair of simultaneously presented pure-tone stimuli. (8) Testing for these emissions is used to screen neonates for hearing loss. (9-12) DPOAE testing has been shown to produce reliable results in high frequencies up to 8.0 kHz, (13) the same range that is related to ototoxicity.
There is a lack of controlled studies on the possible ototoxic effects of aminoglycosides on the hearing of neonates. It is important that more of these studies be conducted since aminoglycosides are so widely used for the treatment of neonates in the NICU. We designed the present study to evaluate the effect of aminoglycoside therapy on the outer hair cells of neonates.
Patients and methods
We conducted a prospective study of 164 consecutively presenting newborns who were born at Ahepa Hospital in Thessaloniki, Greece, between December 1, 2008, and November 30, 2009. The enrolled infants were divided into three groups:
* Group A consisted of 105 infants--60 boys and 45 girls, aged 12 to 126 days (mean: 17.66 [+ or -] 16.38). Their mean weight was 2,316 [+ or -] 785 g. These infants were administered aminoglycoside therapy (either gentamicin or amikacin, or a combination of the two) for the treatment of suspected or proven bacterial infection and septic states. They all received standard intravenous dosing. The duration of aminoglycoside exposure ranged from 2 to 16 days (mean: 5.21).
* Group B included 30 newborns--10 boys and 20 girls, aged 2 to 92 days (mean: 13.63 [+ or -] 21.10). Their mean weight was 2,254 [+ or -] 692 g. These infants were either not given an antibiotic for their treatment or they were given one other than an aminoglycoside (i.e., a nonototoxic antibiotic) during their stay in the NICU.
* Group C was comprised of 29 healthy, full-term, neonates--8 boys and 21 girls, aged 2 to 5 days (mean: 3.22 [+ or -] 0.51). Their mean weight was 2,012 [+ or -] 53 g. They were being cared for in the well-baby nursery of the maternity ward.
We tested consecutive newborns in groups A and B until the number of patients in both groups reached a minimum of 30. Given the sizes of these two groups, we determined that a difference of 25% in the fail rate (e.g., 60 vs. 35%) could be assessed with a power of 0.80.
The risk factors for hearing loss in the two NICU groups are shown in table 1.
Testing protocol. Before the neonates underwent DPOAE testing (Madsen AccuScreen; GN Otometrics; Schaumburg, 111.), they all were examined by pneumatic otoscopy and otologic microscopy to exclude middle ear effusion. The external ear canal was also checked for vernix caseosa and debris. Calibration of the DPOAE equipment was performed, and a correct fit of the probe in the infant's external auditory meatus was checked before each evaluation.
The newborns were examined just after their afternoon feeding to ensure a sleep state without body activity. Testing of newborns who were awake, moving, or crying was postponed until a later time. All tests were done at the crib side in a quiet room to avoid background noise. The left and right ears were tested separately. In those newborns who received an antibiotic, all measurements were made immediately after the drug was administered.
Otoacoustic emissions were elicited by delivering two pure tones ([f.sub.1] and [f.sub.2] primary tones; [f.sub.2]/[f.sub.1] = 1.2), with primary levels of 59 and 50 dB sound pressure level (SPL) for [f.sub.1] and [f.sub.2], respectively. The [f.sub.2] primary tone was presented at four frequencies: 2.0, 2.5, 3.2, and 4.0 kHz. A pass result was given when the DPOAE testing result was within normal limits in at least three of the four frequencies. A fail result was given when the result was abnormal in at least two frequencies, as well as in the event of excessive background noise or bad probe fitting. A DPOAE response with an amplitude of at least 6 dB SPL above the noise floor at each frequency was classified as a pass. A newborn was considered to have failed when a fail result occurred in one or both ears; a pass result indicated that both ears had pass results.
In group A, we analyzed the association between the results of DPOAE testing and the specific type of aminoglycoside administered, the duration of treatment, the number of doses, and the size of the mean daily dose and the mean total dose.
Statistical analysis. The accumulated data were encoded, processed, and analyzed with the Statistical Package for the Social Sciences software (v. 16.0; SPSS; Chicago). The Student t test was used to compare quantitative variables, and the Fisher exact test was used for categorical variables in groups A and B. Multivariate logistic regression models were specified to evaluate associations between DPOAE fail rates and the reported variables, using data from all three study groups. The imbalance of the sample sizes was controlled for by logistic regression analysis, where all recorded risk factors were accounted for (introduced as covariates).
The area under the receiver operating characteristic (ROC) curve was initially calculated to determine the association between the duration of aminoglycoside administration and DPOAE fail rates.
Ethical considerations. Medical Ethics Committee approval was granted for the study protocol, and parental consent was obtained for each newborns participation.
Pass/fail results. Analysis revealed that 41 newborns in group A (39.0%) and 13 in group B (43.3%) failed DPOAE testing in either one or both ears; the difference between the two groups was not statistically significant (p = 0.673) (table 2).
Failure in one ear was seen in 23 of the group A neonates (21.9%) and 12 of the group B neonates (40.0%); failure in both ears was seen in 18 patients in group A (17.1%) and in 1 patient in group B (3.3%) (table 2).
The fail rate (unilateral and bilateral) for groups A and B combined was 40.0% (54 of 135), and the fail rate for the three groups combined was 35.4% (58 of 164). Among all three groups, 23.2% failed in one ear and 12.2% failed in both ears (table 2).
The figure displays the fail rates for each of the three groups at each of the tested frequencies. The fail rate appeared to increase toward the higher [f.sub.2] frequencies (3.2 and 4 kHz) in the two NICU groups.
Variables. In group A, we found no statistically significant association between pass/fail rates and the specific aminoglycoside administered, the duration of antibiotic treatment, the number of doses, and the size of the mean daily dose and the mean total dose (table 3). Although the duration of treatment had no significant influence on the fail rate in group A (p = 0.720), the ROC curve analysis revealed that the cutoffvalue that best predicted DPOAE failure was 6 days. This means that a course of aminoglycoside treatment of 6 days or more might be related to a tendency toward ototoxicity.
Multivariate logistic regression analysis of the three groups demonstrated that only prematurity (gestational age: <37 wk) was a significant predictor of DPOAE failure at the tested frequencies, while respiratory distress was significant in DPOAE failure only at 2.0 kHz (table 4).
It is known that the outer hair cells in the basal turn of the cochlea are the cochlear components most susceptible to injury from ototoxic drugs such as aminoglycosides. In our study, we used DPOAE testing, which reflects the functional level of the outer hair cells better than other hearing tests such as auditory brainstem responses, which may not be feasible in the NICU because they are time-consuming and expensive. In our study, screening of newborns with DPOAE testing revealed no increase in the occurrence of hearing loss in the aminoglycoside-exposed newborns compared with other newborns in the same NICU (fail rates of 39.0% in group A and 43.3% in group B; p = 0.673). According to the 1994 JCIH statement, aminoglycosides represented a risk factor for hearing loss in neonates. (6) By 2000, however, the JCIH (7) recognized that although as many as 45% of neonates were treated in the NICU with an aminoglycoside, (14) the prevalence of hearing loss among them was only 1.5 to 2.0%. (15) Nothing since then has altered the JCIH s opinion. In addition, other studies have not confirmed any relationship between hearing loss and aminoglycoside exposure in the NICU. (10,12,15-20)
In our study, the combined DPOAE failure rate in groups A and B (40.0%) is within the range of what has been found in other studies (1.3 to 49%). (10-12-15-26) Likewise, our total pass rate (60.0%) was similar to those described in larger series (51% (11) and 63.2% (23)).
It has been reported that otoacoustic emissions testing is more sensitive than traditional pure-tone audiometry in detecting early signs of aminoglycoside-induced injury to the peripheral auditory system. (27) Ototoxic damage reportedly starts at frequencies higher than 8.0 kHz (i.e., up to 20 kHz), and then progresses to the lower frequencies. (28) In our study, the fail rate tended to increase in the higher [f.sub.2] frequencies (i.e., at 3.2 and 4.0 kHz), which corresponds with findings on conventional high-frequency pure-tone audiometry. (29) Notably, these fail rates were seen in both group A and group B, which indicates that they were not related to aminoglycoside administration.
While high total doses and prolonged treatment have been reported to be the main causes of aminoglycoside-induced ototoxicity,3 our findings revealed no correlation between hearing impairment and the duration of treatment, the number of doses, or the size of the mean daily dose or the mean total dose. Our statistical analysis did, however, indicate that a duration of aminoglycoside treatment of 6 days or more might cause damage to the outer hair cells. Similarly, ConeWesson et al reported that the most prevalent risk factor for hearing loss in an NICU was a duration of aminoglycoside therapy for 5 or more days. (12) The authors of other studies have claimed that a course of aminoglycoside treatment that exceeds 10 days was a potential risk factor for hearing impairment, but their samples were too small to allow for firm conclusions to be drawn. (10,30)
We observed a relatively high incidence of respiratory distress, low Apgar scores, and the need for prolonged (>24 hr) respiratory mechanical ventilation in both groups A and B. These factors, which have been implicated in hypoxia conditions, have been most common in the range of 0.5 to 2.0 kHz [f.sub.2] frequencies at 1 month after birth. (31) Therefore, it is possible that the higher [f.sub.2] frequencies tested in our study were not sensitive to hypoxia as a cause of injury.
In our study, prematurity was the primary factor that predisposed our patients to hearing loss that affected DPOAE results. In other studies, DPOAE fail rates have been estimated to range from 25% (gestational age: <33 wk) to 44% (gestational age: <37 wk). (11,32) The difference has been attributed to the fact that the outer hair cell properties are considered to be mature at 32 weeks of gestation, (33) while it has been proposed that maturation of the medial olivocochlear efferent innervation of the outer hair cells occurs at 37 weeks. (34) It is possible that in our study, the relevant cochlear immaturity influenced the generation of otoacoustic emissions among the premature neonates (gestational age: <37 wk) in both groups A and B. Even so, this did not statistically preclude us from showing that the aminoglycosides did not have an effect on the hearing of neonates at the high frequencies.
In conclusion, our study did not demonstrate that aminoglycosides had any ototoxic effect on the outer hair cells ofneonates who were being treated for various conditions in a NICU. We recognize that the number of infants in our three groups was limited, and we recommend that larger, controlled studies be conducted. However, based on our results, we feel confident in asserting that a short course of aminoglycoside treatment (<5 days) does not appear to increase the risk of ototoxicity in the outer hair cells of neonates treated in a NICU.
We thank Ms. Esther Eshkol for her editorial assistance, and Mr. Giannis Giaglis for his statistical advice in this study.
(1.) Matz GJ. Aminoglycoside cochlear ototoxicity. Otolaryngol Clin North Am 1993;26(5):705-12.
(2.) Minor LB. Gentamicin-induced bilateral vestibular hypofunction. JAMA 1998;279(7):541-4.
(3.) Forge A, Schacht J. Aminoglycoside antibiotics. Audiol Neurootol 2000;5(l):3-22.
(4.) Roland PS. Characteristics of systemic and topical agents implicated in toxicity of the middle and inner ear. Ear Nose Throat J 2003;82 (Suppl l):3-8.
(5.) Rizzi MD, Hirose K. Aminoglycoside ototoxicity. Curr Opin Otolaryngol Head Neck Surg 2007;15(5):352-7.
(6.) Joint Committee on Infant Hearing. 1994 position statement. ASHA 1994;36(12):38-41.
(7.) Joint Committee on Infant Hearing. Year 2000 position statement: Principles and guidelines for early hearing detection and intervention programs. Joint Committee on Infant Hearing, American Academy of Audiology American Academy of Pediatrics, American Speech-Language-Hearing Association, and Directors of Speech and Hearing Programs in State Health Welfare Agencies. Pediatrics 2000; 106(4):798-817.
(8.) Kemp DT, Brown AM. Ear canal acoustic and round window electrical correlates of 2fl-f2 distortion generated in the cochlea. Hear Res 1984;13(1):39-46.
(9.) Hatzopoulos S, Pelosi G, Petruccelli J, et al. Efficient otoacoustic emissions protocols employed in a hospital-based neonatal screening program. Acta Otolaryngol 2001;121(2):269-73.
(10.) de Hoog M, van Zanten GA, Hoeve LJ, et al. A pilot case control follow-up study on hearing in children treated with tobramycin in the newborn period. Int J Pediatr Otorhinolaryngol 2002;65(3):225-32.
(11.) Chiong CM, Dv Llanes EG, Tirana-Remulla AN, et al. Neonatal hearing screening in a neonatal intensive care unit using distortion-product otoacoustic emissions. Acta Otolaryngol 2003;123(2):215-18.
(12.) Cone-Wesson B, Vohr BR, Sininger YS, et al. Identification of neonatal hearing impairment: Infants with hearing loss. Ear Hear 2000;21 (5):488-507.
(13.) Franklin DJ, McCoy M J, Martin GK, Lonsbury-Martin BL. Test/retest reliability of distortion-product and transiently evoked otoacoustic emissions. Ear Hear 1992;13(6):417-29.
(14.) Vohr BR, Widen JE, Cone-Wesson B, et al. Identification of neonatal hearing impairment: Characteristics of infants in the neonatal intensive care unit and well-baby nursery. Ear Hear 2000;21(5):373-82.
(15.) Finitzo-Hieber T, McCracken GH Jr., Brown KC. Prospective controlled evaluation of auditory function in neonates given netilmicin or amikacin. J Pediatr 1985;106(l):129-36.
(16.) Hille ET, van Straaten HI, Verkerk PH; Dutch NICU Neonatal Hearing Screening Working Group. Prevalence and independent risk factors for hearing loss in NICU infants. Acta Pediatr 2007;96 (8):1155-8.
(17.) Yoshikawa S, Ikeda K, Kudo T, Kobayashi T. The effects of hypoxia, premature birth, infection, ototoxic drugs, circulatory system and congenital disease on neonatal hearing loss. Auris Nasus Larynx 2004;31(4):361-8.
(18.) Meyer C, Witte J, Hildmann A, et al. Neonatal screening for hearing disorders in infants at risk: Incidence, risk factors, and follow-up. Pediatrics 1999;104(4 Pt l):900-4.
(19.) Hess M, Finckh-Kramer U, Bartsch M, et al. Hearing screening in at-risk neonate cohort. Int J Pediatr Otorhinolaryngol 1998;46(1-2): 81-9.
(20.) Kountakis SE, Psifidis A, Chang CJ, Stiernberg CM. Risk factors associated with hearing loss in neonates. Am J Otolaryngol 1997;18 (2):90-3.
(21.) Korres S, Nikolopoulos TP, Komkotou V, et al. Newborn hearing screening: Effectiveness, importance of high-risk factors, and characteristics of infants in the neonatal intensive care unit and well-baby nursery. Otol Neurotol 2005;26(6):1186-90.
(22.) White KR, Vohr BR, Maxon AB, et al. Screening all newborns for hearing loss using transient evoked otoacoustic emissions. Int J Pediatr Otorhinolaryngol 1994;29(3):203-17.
(23.) Ohl C, Dornier L, Czajka C, et al. Newborn hearing screening on infants at risk. Int J Pediatr Otorhinolaryngol 2009;73(12):1691-5.
(24.) Yoon PJ, Price M, Gallagher K, et al. The need for long-term audiologic follow-up of neonatal intensive care unit (NICU) graduates. Int J Pediatr Otorhinolaryngol 2003;67(4):353-7.
(25.) De Capua B, De Felice C, Costantini D, et al. Newborn hearing screening by transient evoked otoacoustic emissions: Analysis of response as a function of risk factors. Acta Otorhinolaryngol Ital 2003;23(l):16-20.
(26.) Finitzo-Hieber T, McCracken GH Jr., Roeser RJ, et al. Ototoxicity in neonates treated with gentamicin and kanamycin: Results of a four-year controlled follow-up study. Pediatrics 1979;63(3):443-50.
(27.) Stavroulaki P, Apostolopoulos N, Dinopoulou D, et al. Otoacoustic emissions--an approach for monitoring aminoglycoside induced ototoxicity in children. Int J Pediatr Otorhinolaryngol 1999;50(3): 177-84.
(28.) Fausti SA, Larson VD, Noffsinger D, et al. High-frequency audiometric monitoring strategies for early detection of ototoxicity. Ear Hear 1994;15(3):232-9.
(29.) Kastanioudakis I, Ziavra N, Anastasopoulos D, Skevas A. Measuring of distortion product otoacoustic emissions using multiple tone pairs. Eur Arch Otorhinolaryngol 2003;260(7):395-400.
(30.) Hotz MA, Harris FP, Probst R. Otoacoustic emissions: An approach for monitoring aminoglycoside-induced ototoxicity. Laryngoscope 1994;104(9):1130-4.
(31.) Zang Z, Wilkinson AR, Jiang ZD. Distortion product otoacoustic emissions at 6 months in term infants after perinatal hypoxiaischaemia or with alow Apgar score. Eur J Pediatr 2008;167(5):575-8.
(32.) Smurzynski J, Jung MD, Lafreniere D, et al. Distortion-product and click-evoked otoacoustic emissions of preterm and full-term infants. Ear Hear 1993;14(4):258-74.
(33.) Bonfils P, Francois M, Avan P, et al. Spontaneous and evoked otoacoustic emissions in preterm neonates. Laryngoscope 1992;102(2): 182-6.
(34.) Chabert R, Guitton MJ, Amram D, et al. Early maturation of evoked otoacoustic emissions and medial olivocochlear reflex in preterm neonates. Pediatr Res 2006;59(2):305-8.
From the 1st Department of Otorhinolaryngology-Head and Neck Surgery, Ahepa Hospital, Aristotle University, Thessaloniki, Greece (Dr. Vital, Dr. Psillas, and Dr. Kekes); the 2nd NICU and Neonatology Department (Dr. Nikolaides) and the 2nd Department of Otorhinolaryngology-Head and Neck Surgery (Dr. Constantinidis), Papageorgiou Hospital, Aristotle University, Thessaloniki; and the Department of Audiology, University of Ferrara, Ferrara, Italy (Dr. Hatzopoulos). The study described in this article was conducted at Ahepa Hospital.
Corresponding author: Dr. Iosif Vital, 1st Department of Otorhinolaryngology-Head and Neck Surgery, Ahepa Hospital, Aristotle University of Thessaloniki School of Medicine, 1 Stilponos Kyriakidi St., 54006, Thessaloniki, Greece. Email: email@example.com
Table 1. Risk factors for hearing loss in group A and group B n (%) Group A Group B Risk factor (n = 105) (n = 30) Respiratory distress 75 (71.4) 8 (26.7) Gestational age <37 wk 64 (61.0) 23 (76.7) Respiratory mechanical ventilation >24 hr 24 (22.9) 5(16.7) Hyperbilirubinemia 18(17.1) 5(16.7) Birth weight <1,500 g 16 (15.2) 3(10.0) Apgar score <7 at 1 min 14(13.3) 1 (3.3) Apgar score <7 at 5 min 2(1.9) 1 (3.3) Convulsions 7 (6.7) 0 Hypotonia 2(1.9) 0 Meningitis 1 (1.0) 0 Table 2. Pass/fail results of DPOAE testing in the three groups Pass Fail, n (%) Group n (%) One ear Both ears Total Group A (n = 105) 64 (61.0) 23 (21.9) 18 (17.1) 41 (39.0) * Group B (n = 30) 17 (56.7) 12 (40.0) 1 (3.3) 13 (43.3) * Group C (n = 29) 25 (86.2) 3 (10.3) 1 (3.4) 4 (13.8) Total (N = 164) 106 (64.6) 38 (23.2) 20 (12.2) 58 (35.4) * The difference in the fail rate between group A and group B was not statistically significant (p = 0.673). Table 3. Association between the results of DPOAE testing and selected variables in group A (n = 105) Variable Fail Pass Any aminoglycoside, 40 (38.1) 65 (61.9) n (%) Gentamicin (n = 65), 25 (38.5) 40 (61.5) n (%) * Amikacin (n = 36), 12 (33.3) 24 (66.7) n (%) Duration of 5.35 [+ or -] 2.67 5.14 [+ or -] 3.12 aminoglycoside treatment, days, mean [+ or -] SD No. doses, mean 4.55 [+ or -] 2.66 4.29 [+ or -] 2.87 [+ or -] SD Daily dose,1 mg, 8.22 [+ or -] 3.27 8.41 [+ or -] 3.6 mean [+ or -] SD Total dose,1 mg, 44.79 [+ or -] 33.22 41.93 [+ or -] 28.7 mean [+ or -] SD Variable p Value Odds ratio (95% CI) Any aminoglycoside, 0.605 0.81 (0.35 to 1.83) n (%) Gentamicin (n = 65), 0.609 1.25 (0.53 to 2.94) n (%) * Amikacin (n = 36), 0.609 0.80 (0.34 to 1.88) n (%) Duration of 0.720 1.03 (0.90 to 1.17) aminoglycoside treatment, days, mean [+ or -] SD No. doses, mean 0.644 1.04 (0.89 to 1.19) [+ or -] SD Daily dose,1 mg, 0.782 0.98 (0.88 to 1.10) mean [+ or -] SD Total dose,1 mg, 0.639 1.00 (0.99 to 1.02) mean [+ or -] SD * Compared with amikacin. ([dagger]) Converted to gentamicin-equivalent doses when amikacin was used. Table 4. Predominant risk factors for DPOAE test failure according to [f.sub.2] frequency Frequency Risk factor p Value * Odds Ratio (95% CI) 2.0 kHz Prematurity 0.034 2.38 (1.07 to 5.30) 2.0 kHz Respiratory distress 0.034 2.74 (1.08 to 6.96) 2.5 kHz Prematurity 0.015 2.23 (1.00 to 5.02) 3.2 kHz Prematurity 0.05 2.66 (1.21 to 5.83) 4.0 kHz Prematurity 0.05 3.11 (1.40 to 6.89) * All values are statistically significant.