- Email: [email protected]
Changes in Brainstem Auditory Evoked Response Latencies in Term Neonates With Hyperbilirubinemia
Changes in Brainstem Auditory Evoked Response Latencies in Term Neonates With Hyperbilirubinemia Ze Dong Jiang, MD, PhD*†, Chao Chen, MD*, Tin Tin Liu, MM*, and Andrew Robert Wilkinson, MB† Ninety term neonates with hyperbilirubinemia were studied with brainstem auditory evoked response to clarify the ototoxic effect of hyperbilirubinemia, and detect any differences in ototoxic effect between different levels of total serum bilirubin. The response threshold in these neonates was significantly elevated (P < 0.001). All wave latencies and I-V interval increased significantly (P < 0.05-0.0001), and correlated weakly with total serum bilirubin (r ⴝ 0.24-0.28, all P < 0.05). Twenty-five neonates (28%) had abnormal responses, including 14 (16%) with elevated thresholds or increased wave I latency, suggesting peripheral auditory impairment, and 16 (18%) with increased I-V interval, suggesting central auditory impairment. Wave V latency and I-V interval were longer in neonates with total serum bilirubin of <20 mg/dL than in those with bilirubin 11-15 mg/dL (P < 0.05). However, there were no significant differences in response variables between neonates with total serum bilirubin 11-15 mg/dL and those with bilirubin 16-20 mg/dL, and between neonates with bilirubin 16-20 mg/dL and those with bilirubin >20 mg/dL. Thus, although the acute ototoxic effect of hyperbilirubinemia tends to be more significant at a higher rather than lower level of total serum bilirubin, auditory impairment does not increase closely with the increase in bilirubin. © 2007 by Elsevier Inc. All rights reserved.
Jiang ZD, Chen C, Liu TT, Wilkinson AR. Changes in brainstem auditory evoked response in term neonates with hyperbilirubinemia. Pediatr Neurol 2007;37:35-41.
From the *Department of Pediatrics of Children’s Hospital, Fudan University, Shanghai, China, and †Neonatal Unit, Department of Paediatrics, John Radcliffe Hospital, University of Oxford, Oxford, United Kingdom.
© 2007 by Elsevier Inc. All rights reserved. doi:10.1016/j.pediatrneurol.2007.03.006 ● 0887-8994/07/$—see front matter
Introduction Hyperbilirubinemia during the neonatal period is known to be an important risk factor for neonatal auditory impairment [1-5]. Previous research on the ototoxic effect of neonatal hyperbilirubinemia concentrated on peripheral auditory (sensory or cochlear, or sensorineural) impairment, and gave little attention to central auditory impairment [2-5]. In the last two decades, the ototoxicity of hyperbilirubinemia in the neonate has been studied mainly using brainstem auditory evoked response (BAER), the most reliable, objective method to study both peripheral and central (specifically brainstem) auditory function in neonates. In infants with hyperbilirubinemia, the majority of researchers reported that BAER is abnormal, typically with an increase in wave latencies [5-12]. Abnormalities of BAER in infants with hyperbilirubinemia were suggested by some investigators to be a marker of bilirubin ototoxicity or neurotoxicity. On the other hand, others did not find any significant abnormalities in infants with hyperbilirubinemia. For example, Soares et al. reported that in their 72 preterm and term infants with hyperbilirubinemia, waves I, III, and V in BAER were always present, and that BAER thresholds were normal in all subjects [13]. Only eight infants manifested abnormal wave latencies. There was no significant difference in I-V interpeak interval between infants with a higher level of bilirubin and those with a lower level of bilirubin. Streletz et al. found an abnormal increase in BAER wave latencies in infants with hyperbilirubinemia, but no abnormality in the I-V interval [8]. We previously studied BAER in infants with various perinatal conditions or problems (e.g., asphyxia, very low
Communications should be addressed to: Dr. Jiang; Neonatal Unit, Department of Paediatrics; John Radcliffe Hospital; Headington, Oxford OX3 9DU, United Kingdom. E-mail: [email protected] Received November 27, 2006; accepted March 23, 2007.
Jiang et al: BAER Latencies in Hyperbilirubinemia 35
birthweight, and intrauterine growth retardation), and found that BAER is a reliable test for detecting auditory impairment in the infant, particularly in the neonate [14-17]. In order to clarify further the effect of bilirubin toxicity on neonatal auditory function, we examined BAER in term neonates who had hyperbilirubinemia and required phototherapy or exchange transfusion during the first 2 weeks of life, with the following two aims: (1) to observe changes in BAER wave latencies and I-V interpeak interval in order to examine the acute effect of bilirubin toxicity on peripheral and central auditory function, respectively, and to detect the prevalence of auditory impairment both peripherally and centrally; and (2) to detect any differences in the acute ototoxic effect between different levels of total serum bilirubin (TSB) and to examine whether auditory impairment, reflected by BAER abnormalities, is closely related to the level of TSB. Materials and Methods Study Group We recruited 90 term neonates who had a TSB level greater than 10 mg/dL and who required phototherapy or exchange transfusion during the first 2 weeks after birth. Their gestational age ranged from 37-42 weeks (39.1 [mean] ⫾ 1.3 [S.D.] weeks), and their birthweight from 2240-5305 g (3,362 [mean] ⫾ 558 [S.D.] g). Of these neonates, 34 had evidence of hemolysis (rhesus hemolytic disease, anemia, a positive direct antiglobulin test, reticulocytosis, or a peripheral blood smear compatible with hemolysis), nine had an associated infection (pneumonia or sepsis), 21 had Apgar scores of 3-7 at 1 and 5 minutes, and the remaining 26 did not have any detectable perinatal problems except for hyperbilirubinemia. Neonates were excluded if they had any perinatal conditions or problems that may have significantly affected neonatal auditory function and their central nervous system. These conditions included congenital malformations, in utero infection, a family history of hearing loss, low birthweight, severe intrauterine growth retardation, bacterial meningitis, seizures, hypoxic-ischemic encephalopathy, persistent pulmonary hypertension, and ototoxic medication [2,5,8,14-17]. Based on the level of TSB, obtained just before the start of BAER recording, neonates were divided into three subgroups: neonates with TSB at 11-15 mg/dL (n ⫽ 28, 31% of subjects), those with TSB at 16-20 mg/dL (n ⫽ 28, 31%) and those with TSB greater than 20 mg/dL (n ⫽ 34, 38%).
Control Group Forty-three term neonates served as control subjects. Their gestational age ranged between 37-42 weeks (39.0 ⫾ 1.2 weeks), and birthweight between 2569-4539 g (3,460 ⫾ 483 g). All were in good health. None had any of the perinatal conditions or problems listed above in Study Group.
University (Shanghai, China). The informed consent of parents and the pediatrician in charge was obtained before subjects entered the study. The recording of BAER began shortly after obtaining a blood sample for measurement of TSB. Subjects lay supine on a cot. Prior to BAER recording, the auditory meatus was inspected to avoid collapsing the meatal lumen, and cleaned of any blockage by vernix and wax. The left ear was tested in all subjects. The right ear was also tested in some subjects. To maintain consistency in analyzing the data and presenting the results, only data from the left ear are described here. The recording of BAER commenced soon after subjects fell asleep naturally, often after a feed. No sedation was used. As described previously, a Bravo Portable Evoked Potential System (Nicolet Biomedical, Inc., Madison, WI) was used to record and analyze BAER [15-17]. Three gold-plated disk electrodes were placed on the middle forehead (positive), the ipsilateral earlobe (negative), and the contralateral earlobe (ground). Interelectrode impedances were kept below 5 k⍀. The acoustic stimuli were rarefaction clicks, generated by rectangular pulses 100 sec in duration, and delivered monaurally to TDH 39 earphones. The clicks were presented at a repetition rate of 21/second and at an intensity of 60 dB normal hearing level in all subjects. Brain responses to 2,048 clicks were preamplified and bandpassed between 100-3000 Hz. An automatic artifact rejection was used to reduce the inclusion of high-amplitude muscular activity in the averaged responses. The ongoing filtered electroencephalogram and the running averaged BAER were carefully monitored while averaging. Sampling was discontinued whenever artifacts were visible on the monitoring oscilloscope. The analysis time of the screen was 12 ms. Duplicate recordings were made for each stimulus condition to check reproducibility. The BAER threshold, determined as the lowest intensity of clicks which produced a reliable wave V, was obtained by reducing the intensity of clicks from 60 dB normal hearing level by 5–20-dB steps until no clear and reproducible wave V occurred. If we found a lack of clear BAER waveforms at 60 dB normal hearing level, the intensity was increased at 5–10-dB steps up to a maximum of 90 dB normal hearing level.
BAER Analysis The latencies of BAER waves I, III, and V were measured without knowledge of the medical history and clinical data of each subject. The I-V interpeak interval was calculated. Measurements of two replicable recordings in each stimulus condition were averaged for data analyses. Correlation analysis was performed for the relationship between BAER variables and the level of TSB. One-way analysis of variance was used to compare the mean and standard deviation of each BAER variable between different groups of subjects.
Results Measurements of BAER variables (mean ⫾ SD) in neonates with hyperbilirubinemia and the age-matched normal controls are given in Table 1. Figures 1-4 show box plots of measurements of wave I, III, and V latencies, and the I-V interpeak interval, respectively, in neonates with different levels of TSB and normal control subjects. Comparison of BAER Between Neonates With Hyperbilirubinemia and Normal Control Subjects
BAER Testing The recording of BAER was made during the first 2 weeks after birth, when clinical signs of jaundice and a TSB level greater than 10 mg/dL occurred. There was no difference in postconceptional age at which BAER was recorded between the study group (40.0 ⫾ 1.3 weeks) and the control group (39.9 ⫾ 1.5 weeks). The study protocols and procedures were approved by the Children’s Hospital Ethics Committee of Fudan
36
PEDIATRIC NEUROLOGY
Vol. 37 No. 1
The BAER threshold in neonates with hyperbilirubinemia was significantly higher than in control subjects (analysis of variance, P ⬍ 0.0001, Table 1). The latencies of BAER waves I, III, and V were all significantly longer than in control subjects (P ⬍ 0.05-0.0001, Table 1). The difference between the two groups was more significant
Table 1.
Measurements of BAER variables (mean ⴞ standard deviation) in neonates with hyperbilirubinemia and normal controls
BAER Variables
Threshold (dB nHL) I latency (ms) III latency (ms) V latency (ms) I-V interval (ms)
Subjects
Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects
BAER Mean
S.D.
16.9 11.4 2.41 2.32 5.26 5.07 7.52 7.26 5.12 4.98
7.2 5.5 0.18 0.17 0.27 0.19 0.32 0.20 0.24 0.14
F
P
19.40
⬍0.0001
4.59
⬍0.05
12.36
⬍0.001
17.21
⬍0.0001
10.99
⬍0.001
Abbreviations: BAER ⫽ Brainstem auditory evoked response ms ⫽ Mini second nHL ⫽ Normal hearing level S.D. ⫽ Standard deviation
for later waves, i.e., those with a longer latency, than for earlier waves, i.e., those with a shorter latency. The I-V interval in neonates with hyperbilirubinemia was also significantly longer (P ⬍ 0.01, Table 1). Of 90 neonates with hyperbilirubinemia, nine (10%) had a BAER threshold greater than 20 dB normal hearing level, with seven at 25-35 dB, one at 50 dB, and one at 55 dB. A significant or abnormal increase in wave latency (greater than 2.5 standard deviations of the mean value in normal control subjects) was seen in eight (9%) neonates with hyperbilirubinemia for wave I, and in 18 (20%) for wave V. Three of nine neonates who had a threshold elevation had an increase in wave I latency. Thus, in total, 14 (16%) neonates had an elevation in their BAER threshold or an increase in wave I latency, suggesting peripheral auditory impairment. All nine neonates with a threshold elevation had an increase in wave V latency.
An abnormal increase in I-V interval was seen in 16 (18%) neonates, suggesting central auditory impairment. Five of these neonates also had an association with an elevation in BAER threshold or an increase in wave I latency. Thus, in total, 25 (28%) neonates had BAER abnormalities, suggesting auditory impairment peripherally or centrally. The BAER abnormalities tended to occur more frequently in neonates with a higher level of TSB than in those with a lower level of TSB. Comparison of BAER Between Neonates With Different Levels of TSB and Normal Control Subjects No statistically significant difference in wave I latency was found between neonates with a TSB of 11-15 mg/dL and control subjects (Fig 1). Wave III and V latencies in these neonates were significantly longer than those in the
Figure 1. Box plot of wave I latency (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
Jiang et al: BAER Latencies in Hyperbilirubinemia 37
Figure 2. Box plot of wave III latency (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
controls (P ⬍ 0.05 and 0.05, Figs 2, 3). The I-V interval, though slightly longer, was not significantly different from that in control subjects (P ⬍ 0.05, Fig 4). In neonates with TSB at 16-20 mg/dL, III and V latencies were significantly longer than those in the controls (P ⬍ 0.05 and 0.01, Figs 2-4). The I-V interval was also significantly longer (P ⬍ 0.05, Fig 4). In neonate with TSB greater than 20 mg/dL, all wave latencies were significantly longer than those in the controls (P ⬍ 0.01, 0.001, and 0.0001 for waves I, III, and V, respectively; Figs 1-3). Similarly, the I-V interval was significantly longer (P ⬍ 0.0001, Fig 4). Correlation of BAER Variables With Level of TSB In neonates with hyperbilirubinemia, the BAER threshold tended to increase with the rise in the level of TSB.
Figure 3. Box plot of wave V latency (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
38
PEDIATRIC NEUROLOGY
Vol. 37 No. 1
However, the threshold did not correlate significantly with the level of TSB. All latencies of waves I, III, and V correlated weakly, though with statistical significance, with the level of TSB (r ⫽ 0.26-0.28, all P ⬍ 0.05). The I-V interval also correlated weakly with the level of TSB (r ⫽ 0.24, P ⬍ 0.05). Comparison of BAER Between Neonates With Different Levels of TSB In general, as the level of TSB increased, the differences in BAER variables between neonates with hyperbilirubinemia and normal control subjects tended to be greater. The latencies of waves I, III, and V, and the I-V interval, were all slightly longer in neonates with TSB at 16-20 mg/dL than in those with TSB at 11-15 mg/dL (Figs 1-4).
Figure 4. Box plot of I-V interval (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
However, none of these BAER variables differed significantly between the two subgroups of neonates. Similarly, all wave latencies and the I-V interval in neonates with TSB greater than 20 mg/dL tended to be longer than in those with TSB at 16-20 mg/dL, but no statistically significant difference was found in any of these variables between the two subgroups (Figs 1-4). The wave latencies and I-V interval in neonates with TSB greater than 20/dL were all longer than in those with TSB at 11-15 mg/dL, with a statistical significance in wave V latency and I-V interval (P ⬍ 0.05 and 0.05, respectively, Figs 1-4). Discussion Bilirubin is known to inhibit mitochondrial enzymes and affect DNA synthesis and ion exchange. It disturbs neuro-excitatory signals, and impairs nerve conduction [1,18]. Some authors reported that the site of a lesion in auditory impairment caused by hyperbilirubinemia may be retrocochlear, with the cochlea unaffected [19]. Others reported that lesions causing auditory impairment after hyperbilirubinemia may include the organ of Corti, especially at the outer hair cells and the cochlear nerve [20]. In any case, it is presumed that neonatal BAER, which reflects functional integrity and the development of the auditory pathway up to the brainstem level, can be affected by hyperbilirubinemia. Our neonates who had hyperbilirubinemia showed an abnormal increase in the latencies of wave I, III, and V, suggesting that neonatal auditory function is impaired after hyperbilirubinemia. Twenty percent of 90 neonates with hyperbilirubinemia had an abnormal increase in wave V latency, and half of the 20% were associated with an elevation of BAER threshold. This abnormality appears to be somewhat similar to the prevalence (22%) of auditory impairment previously reported by some authors [9].
However, further analysis of the details of our BAER results revealed that the prevalence of auditory impairment is as high as 28%. The impairment could be peripheral or central. Previous authors reported no elevation in BAER threshold after neonatal hyperbilirubinemia [13]. In the present study, the mean value of the BAER threshold in neonates with hyperbilirubinemia was greater than in normal control subjects. Ten percent of the neonates had an elevated threshold, with seven slightly elevated, and two significantly elevated. Nine percent of the neonates had an increase in wave I latency. Some neonates had both an elevation in BAER threshold and an increase in wave I latency. These conditions led to a total of 16% of our neonates exhibiting BAER abnormalities (elevation in BAER threshold or an increase in wave I latency) that suggest peripheral auditory impairment. To date, few reports have described central auditory impairment after neonatal hyperbilirubinemia. In BAER, some authors found no abnormality in the I-V interval [8]. The present study revealed that the increase in BAER latency in the neonates with hyperbilirubinemia was more significant for the later waves than for the earlier waves, and that the I-V interval increased significantly. These results suggest that, in addition to peripheral auditory impairment, there is also central auditory impairment after neonatal hyperbilirubinemia. In the present study, central impairment was found in 18% of neonates with hyperbilirubinemia. Such impairment may have important clinical implications. Therefore, researchers will need to differentiate between peripheral and central auditory impairment, and to pay more attention to central auditory impairment after neonatal hyperbilirubinemia. In addition to hyperbilirubinemia, neonatal auditory function can be impaired by some pre-, peri-, and postnatal conditions or problems, as described in Materials and
Jiang et al: BAER Latencies in Hyperbilirubinemia 39
Methods [2-5,8,14-17]. In the present study, any neonates who had evidence of these risk factors had been excluded. Thus, our results were unlikely to be significantly confounded by any of these risk factors. Some of our neonates with hyperbilirubinemia had pneumonia, sepsis, and a low Apgar score. However, none of these neonates showed any clinical signs of significant neurological abnormalities. Our previous study revealed that a low Apgar score at 1 or 5 minutes alone, with no clinical signs of hypoxicischemic encephalopathy, does not have any significant effect on BAER [21]. Thus, these factors are unlikely to have any significant effects on the BAER results of our neonates with hyperbilirubinemia, and cannot be responsible for the BAER abnormalities seen in these neonates. Although we cannot completely rule out some subtle confounding effects of neonatal conditions associated with hyperbilirubinemia and conductive ear disorders in some of the neonates, the BAER abnormalities must be attributed mainly to bilirubin ototoxicity. Thus, our results from the BAER data, collected shortly after obtaining TSB, indicate that hyperbilirubinemia has an acute toxic effect on neonatal auditory function, both peripherally and centrally. Some of our subjects were tested repeatedly with BAER several days and weeks after the initial BAER recording. These preliminary data indicated that after phototherapy or exchange transfusion, the BAER abnormalities in neonates with hyperbilirubinemia recovered quickly, suggesting that the auditory impairment is largely transient. This is similar to findings by other investigators [22]. Infants or children who had hyperbilirubinemia were often found not to be associated with adverse neurodevelopmental outcomes after prompt treatment [23]. The toxic effect of hyperbilirubinemia on visual evoked potentials was also found to be transient after prompt treatment [24]. Our study suggests that a TSB level greater than 10 mg/dL has a toxic effect on neonatal auditory function. As the level of TSB increased, the latencies of waves I, III, and V and the I-V interval tended to be longer, and the differences in these BAER variables between neonates with hyperbilirubinemia and normal neonates tended to be greater. Thus, the acute toxic effect of hyperbilirubinemia on peripheral and central auditory function tends to be more significant at a higher level of TSB than at a lower level. However, all these variables only correlated weakly with the level of TSB. Our study further revealed that there were no major differences in BAER variables between the neonates with different levels of TSB, except for a significant increase in wave V latency and I-V interval in the neonates with TSB greater than 20 mg/dL when compared with those with TSB at 11-15 mg/dL. The BAER abnormalities occurred slightly more frequently in neonates with a higher level of TSB than in those with a lower level of TSB. It appears that auditory impairment, reflected by BAER abnormalities, after neonatal hyperbilirubinemia does not increase closely with the increase in
40
PEDIATRIC NEUROLOGY
Vol. 37 No. 1
the level of TSB. This finding may be partly related to individual differences in sensitivity to the ototoxic effect of hyperbilirubinemia and, in particular, in perinatal conditions that are associated with hyperbilirubinemia, e.g., acidosis and hypoxia, which are known to facilitate bilirubin toxicity [25].
We thank the nurses and doctors at the Neonatal Unit, Children’s Hospital, Fudan University (Shanghai, China), for their assistance in recruiting subjects. This research was supported by grants from the Children’s Hospital of Fudan University, and from Defeating Deafness, UK.
References [1] Shapiro SM. Bilirubin toxicity in the developing nervous system. Pediatr Neurol 2003;29:410-21. [2] Joint Committee on Infant Hearing. Year 2000 position statement: Principles and guidelines for early hearing detection and intervention programs. Pediatrics 2000;106:798-817. [3] Newton V. Adverse perinatal conditions and the inner ear. Semin Neonatol 2001;6:543-51. [4] Kountakis SE, Skoulas I, Phillips D, Chang CYJ. Risk factors for hearing loss in neonates: A prospective study. Am J Otolaryngol 2002;23:133-7. [5] Wilkinson AR, Jiang ZD. Brainstem auditory evoked response in neonatal neurology. Semin Fet Neonatol Med 2006;11:444-51. [6] Hung KL. Auditory brainstem responses in patients with neonatal hyperbilirubinaemia and bilirubin encephalopathy. Brain Dev 1989; 11;297-301. [7] Gupta AK, Mann SB. Is auditory brainstem response a bilirubin neurotoxicity marker? Am J Otolaryngol 1998;19:232-6. [8] Streletz LJ, Graziani LJ, Branca PA, Desai HJ, Travis SF, Mikaelian DO. Brainstem auditory evoked potentials in fullterm and preterm newborns with hyperbilirubinemia and hypoxemia. Neuropediatrics 1986;17:66-71. [9] Boo NY, Oakes M, Lye MS, Said H. Risk factors associated with hearing loss in term neonates with hyperbilirubinaemia. J Trop Pediatr 1994;40:194-7. [10] Funato M, Tamai H, Shimada S, Nakamura H. Vigintiphobia, unbound bilirubin, and auditory brainstem responses. Pediatrics 1994;93: 50-3. [11] Amin SB, Ahlfors C, Orlando MS, Dalzell LE, Merle KS, Guillet R. Bilirubin and serial auditory brainstem responses in premature infants. Pediatrics 2001;107:664-70. [12] Smith CM, Barnes GP, Jacobson CA, Oelberg DG. Auditory brainstem response detects early bilirubin neurotoxicity at low indirect bilirubin values. J Perinatol 2004;24:730-2. [13] Soares I, Collet L, Delorme C, Salle B, Morgon A. Are click-evoked BAEPs useful in case of neonate hyperbilirubinaemia? Int J Pediatr Otorhinolaryngol 1989;117:231-7. [14] Jiang ZD. Maturation of peripheral and brainstem auditory function in the first year following perinatal asphyxia—a longitudinal study. J Speech Lang Hear Res 1998;41:83-93. [15] Jiang ZD, Brosi DM, Wilkinson AR. Hearing impairment in preterm very low birth weight babies at term revealed by brainstem auditory evoked responses. Acta Paediatr 2001;90:1411-5. [16] Jiang ZD, Brosi DM, Wang J, Wilkinson AR. Brainstem auditory-evoked responses to different rates of clicks in small-forgestational age preterm infants at term. Acta Paediatr 2004;93:76-81. [17] Jiang ZD, Brosi DM, Wang J, Wilkinson AR. One-third of term babies after perinatal hypoxia-ischaemia have transient hearing impairment: Dynamic change in hearing threshold during the neonatal period. Acta Paediatr 2004;93:82-7. [18] Dennery PA, Seidman DS, Stevenson DK. Neonatal hyperbilirubinemia. N Engl J Med 2001;344:581-90.
[19] Sano M, Kaga K, Kitazumi E, Kodama K. Sensorineural hearing loss in patients with cerebral palsy after asphyxia and hyperbilirubinemia. Int J Pediatr Otorhinolaryngol 2005;69:1211-7. [20] Rhee CK, Park HM, Jang YJ. Audiologic evaluation of neonates with severe hyperbilirubinemia using transiently evoked otoacoustic emissions and auditory brainstem responses. Laryngoscope 1999;109: 2005-8. [21] Jiang ZD, Shao XM, Wilkinson AR. Brainstem auditory evoked responses in term neonates with temporary low Apgar scores. Acta Otolaryngol (Stockh) 2005;125;163-8.
[22] Wong V, Chen WX, Wong KY. Short- and long-term outcome of severe neonatal nonhemolytic hyperbilirubinemia. J Child Neurol 2006;21:309-15. [23] Newman TB, Liljestrand P, Jeremy RJ, et al. Outcomes among newborns with total serum bilirubin levels of 25 mg per deciliter or more. N Engl J Med 2006;354:1889-900. [24] Chen WX, Wong V. Visual evoked potentials in neonatal hyperbilirubinemia. J Child Neurol 2006;21:58-62. [25] Gustafson PA, Boyle DW. Bilirubin index: A new standard for intervention? Med Hypotheses 1995;45:409-16.
Jiang et al: BAER Latencies in Hyperbilirubinemia 41
Jiang ZD, Chen C, Liu TT, Wilkinson AR. Changes in brainstem auditory evoked response in term neonates with hyperbilirubinemia. Pediatr Neurol 2007;37:35-41.
From the *Department of Pediatrics of Children’s Hospital, Fudan University, Shanghai, China, and †Neonatal Unit, Department of Paediatrics, John Radcliffe Hospital, University of Oxford, Oxford, United Kingdom.
© 2007 by Elsevier Inc. All rights reserved. doi:10.1016/j.pediatrneurol.2007.03.006 ● 0887-8994/07/$—see front matter
Introduction Hyperbilirubinemia during the neonatal period is known to be an important risk factor for neonatal auditory impairment [1-5]. Previous research on the ototoxic effect of neonatal hyperbilirubinemia concentrated on peripheral auditory (sensory or cochlear, or sensorineural) impairment, and gave little attention to central auditory impairment [2-5]. In the last two decades, the ototoxicity of hyperbilirubinemia in the neonate has been studied mainly using brainstem auditory evoked response (BAER), the most reliable, objective method to study both peripheral and central (specifically brainstem) auditory function in neonates. In infants with hyperbilirubinemia, the majority of researchers reported that BAER is abnormal, typically with an increase in wave latencies [5-12]. Abnormalities of BAER in infants with hyperbilirubinemia were suggested by some investigators to be a marker of bilirubin ototoxicity or neurotoxicity. On the other hand, others did not find any significant abnormalities in infants with hyperbilirubinemia. For example, Soares et al. reported that in their 72 preterm and term infants with hyperbilirubinemia, waves I, III, and V in BAER were always present, and that BAER thresholds were normal in all subjects [13]. Only eight infants manifested abnormal wave latencies. There was no significant difference in I-V interpeak interval between infants with a higher level of bilirubin and those with a lower level of bilirubin. Streletz et al. found an abnormal increase in BAER wave latencies in infants with hyperbilirubinemia, but no abnormality in the I-V interval [8]. We previously studied BAER in infants with various perinatal conditions or problems (e.g., asphyxia, very low
Communications should be addressed to: Dr. Jiang; Neonatal Unit, Department of Paediatrics; John Radcliffe Hospital; Headington, Oxford OX3 9DU, United Kingdom. E-mail: [email protected] Received November 27, 2006; accepted March 23, 2007.
Jiang et al: BAER Latencies in Hyperbilirubinemia 35
birthweight, and intrauterine growth retardation), and found that BAER is a reliable test for detecting auditory impairment in the infant, particularly in the neonate [14-17]. In order to clarify further the effect of bilirubin toxicity on neonatal auditory function, we examined BAER in term neonates who had hyperbilirubinemia and required phototherapy or exchange transfusion during the first 2 weeks of life, with the following two aims: (1) to observe changes in BAER wave latencies and I-V interpeak interval in order to examine the acute effect of bilirubin toxicity on peripheral and central auditory function, respectively, and to detect the prevalence of auditory impairment both peripherally and centrally; and (2) to detect any differences in the acute ototoxic effect between different levels of total serum bilirubin (TSB) and to examine whether auditory impairment, reflected by BAER abnormalities, is closely related to the level of TSB. Materials and Methods Study Group We recruited 90 term neonates who had a TSB level greater than 10 mg/dL and who required phototherapy or exchange transfusion during the first 2 weeks after birth. Their gestational age ranged from 37-42 weeks (39.1 [mean] ⫾ 1.3 [S.D.] weeks), and their birthweight from 2240-5305 g (3,362 [mean] ⫾ 558 [S.D.] g). Of these neonates, 34 had evidence of hemolysis (rhesus hemolytic disease, anemia, a positive direct antiglobulin test, reticulocytosis, or a peripheral blood smear compatible with hemolysis), nine had an associated infection (pneumonia or sepsis), 21 had Apgar scores of 3-7 at 1 and 5 minutes, and the remaining 26 did not have any detectable perinatal problems except for hyperbilirubinemia. Neonates were excluded if they had any perinatal conditions or problems that may have significantly affected neonatal auditory function and their central nervous system. These conditions included congenital malformations, in utero infection, a family history of hearing loss, low birthweight, severe intrauterine growth retardation, bacterial meningitis, seizures, hypoxic-ischemic encephalopathy, persistent pulmonary hypertension, and ototoxic medication [2,5,8,14-17]. Based on the level of TSB, obtained just before the start of BAER recording, neonates were divided into three subgroups: neonates with TSB at 11-15 mg/dL (n ⫽ 28, 31% of subjects), those with TSB at 16-20 mg/dL (n ⫽ 28, 31%) and those with TSB greater than 20 mg/dL (n ⫽ 34, 38%).
Control Group Forty-three term neonates served as control subjects. Their gestational age ranged between 37-42 weeks (39.0 ⫾ 1.2 weeks), and birthweight between 2569-4539 g (3,460 ⫾ 483 g). All were in good health. None had any of the perinatal conditions or problems listed above in Study Group.
University (Shanghai, China). The informed consent of parents and the pediatrician in charge was obtained before subjects entered the study. The recording of BAER began shortly after obtaining a blood sample for measurement of TSB. Subjects lay supine on a cot. Prior to BAER recording, the auditory meatus was inspected to avoid collapsing the meatal lumen, and cleaned of any blockage by vernix and wax. The left ear was tested in all subjects. The right ear was also tested in some subjects. To maintain consistency in analyzing the data and presenting the results, only data from the left ear are described here. The recording of BAER commenced soon after subjects fell asleep naturally, often after a feed. No sedation was used. As described previously, a Bravo Portable Evoked Potential System (Nicolet Biomedical, Inc., Madison, WI) was used to record and analyze BAER [15-17]. Three gold-plated disk electrodes were placed on the middle forehead (positive), the ipsilateral earlobe (negative), and the contralateral earlobe (ground). Interelectrode impedances were kept below 5 k⍀. The acoustic stimuli were rarefaction clicks, generated by rectangular pulses 100 sec in duration, and delivered monaurally to TDH 39 earphones. The clicks were presented at a repetition rate of 21/second and at an intensity of 60 dB normal hearing level in all subjects. Brain responses to 2,048 clicks were preamplified and bandpassed between 100-3000 Hz. An automatic artifact rejection was used to reduce the inclusion of high-amplitude muscular activity in the averaged responses. The ongoing filtered electroencephalogram and the running averaged BAER were carefully monitored while averaging. Sampling was discontinued whenever artifacts were visible on the monitoring oscilloscope. The analysis time of the screen was 12 ms. Duplicate recordings were made for each stimulus condition to check reproducibility. The BAER threshold, determined as the lowest intensity of clicks which produced a reliable wave V, was obtained by reducing the intensity of clicks from 60 dB normal hearing level by 5–20-dB steps until no clear and reproducible wave V occurred. If we found a lack of clear BAER waveforms at 60 dB normal hearing level, the intensity was increased at 5–10-dB steps up to a maximum of 90 dB normal hearing level.
BAER Analysis The latencies of BAER waves I, III, and V were measured without knowledge of the medical history and clinical data of each subject. The I-V interpeak interval was calculated. Measurements of two replicable recordings in each stimulus condition were averaged for data analyses. Correlation analysis was performed for the relationship between BAER variables and the level of TSB. One-way analysis of variance was used to compare the mean and standard deviation of each BAER variable between different groups of subjects.
Results Measurements of BAER variables (mean ⫾ SD) in neonates with hyperbilirubinemia and the age-matched normal controls are given in Table 1. Figures 1-4 show box plots of measurements of wave I, III, and V latencies, and the I-V interpeak interval, respectively, in neonates with different levels of TSB and normal control subjects. Comparison of BAER Between Neonates With Hyperbilirubinemia and Normal Control Subjects
BAER Testing The recording of BAER was made during the first 2 weeks after birth, when clinical signs of jaundice and a TSB level greater than 10 mg/dL occurred. There was no difference in postconceptional age at which BAER was recorded between the study group (40.0 ⫾ 1.3 weeks) and the control group (39.9 ⫾ 1.5 weeks). The study protocols and procedures were approved by the Children’s Hospital Ethics Committee of Fudan
36
PEDIATRIC NEUROLOGY
Vol. 37 No. 1
The BAER threshold in neonates with hyperbilirubinemia was significantly higher than in control subjects (analysis of variance, P ⬍ 0.0001, Table 1). The latencies of BAER waves I, III, and V were all significantly longer than in control subjects (P ⬍ 0.05-0.0001, Table 1). The difference between the two groups was more significant
Table 1.
Measurements of BAER variables (mean ⴞ standard deviation) in neonates with hyperbilirubinemia and normal controls
BAER Variables
Threshold (dB nHL) I latency (ms) III latency (ms) V latency (ms) I-V interval (ms)
Subjects
Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects Hyperbilirubinemia Normal control subjects
BAER Mean
S.D.
16.9 11.4 2.41 2.32 5.26 5.07 7.52 7.26 5.12 4.98
7.2 5.5 0.18 0.17 0.27 0.19 0.32 0.20 0.24 0.14
F
P
19.40
⬍0.0001
4.59
⬍0.05
12.36
⬍0.001
17.21
⬍0.0001
10.99
⬍0.001
Abbreviations: BAER ⫽ Brainstem auditory evoked response ms ⫽ Mini second nHL ⫽ Normal hearing level S.D. ⫽ Standard deviation
for later waves, i.e., those with a longer latency, than for earlier waves, i.e., those with a shorter latency. The I-V interval in neonates with hyperbilirubinemia was also significantly longer (P ⬍ 0.01, Table 1). Of 90 neonates with hyperbilirubinemia, nine (10%) had a BAER threshold greater than 20 dB normal hearing level, with seven at 25-35 dB, one at 50 dB, and one at 55 dB. A significant or abnormal increase in wave latency (greater than 2.5 standard deviations of the mean value in normal control subjects) was seen in eight (9%) neonates with hyperbilirubinemia for wave I, and in 18 (20%) for wave V. Three of nine neonates who had a threshold elevation had an increase in wave I latency. Thus, in total, 14 (16%) neonates had an elevation in their BAER threshold or an increase in wave I latency, suggesting peripheral auditory impairment. All nine neonates with a threshold elevation had an increase in wave V latency.
An abnormal increase in I-V interval was seen in 16 (18%) neonates, suggesting central auditory impairment. Five of these neonates also had an association with an elevation in BAER threshold or an increase in wave I latency. Thus, in total, 25 (28%) neonates had BAER abnormalities, suggesting auditory impairment peripherally or centrally. The BAER abnormalities tended to occur more frequently in neonates with a higher level of TSB than in those with a lower level of TSB. Comparison of BAER Between Neonates With Different Levels of TSB and Normal Control Subjects No statistically significant difference in wave I latency was found between neonates with a TSB of 11-15 mg/dL and control subjects (Fig 1). Wave III and V latencies in these neonates were significantly longer than those in the
Figure 1. Box plot of wave I latency (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
Jiang et al: BAER Latencies in Hyperbilirubinemia 37
Figure 2. Box plot of wave III latency (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
controls (P ⬍ 0.05 and 0.05, Figs 2, 3). The I-V interval, though slightly longer, was not significantly different from that in control subjects (P ⬍ 0.05, Fig 4). In neonates with TSB at 16-20 mg/dL, III and V latencies were significantly longer than those in the controls (P ⬍ 0.05 and 0.01, Figs 2-4). The I-V interval was also significantly longer (P ⬍ 0.05, Fig 4). In neonate with TSB greater than 20 mg/dL, all wave latencies were significantly longer than those in the controls (P ⬍ 0.01, 0.001, and 0.0001 for waves I, III, and V, respectively; Figs 1-3). Similarly, the I-V interval was significantly longer (P ⬍ 0.0001, Fig 4). Correlation of BAER Variables With Level of TSB In neonates with hyperbilirubinemia, the BAER threshold tended to increase with the rise in the level of TSB.
Figure 3. Box plot of wave V latency (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
38
PEDIATRIC NEUROLOGY
Vol. 37 No. 1
However, the threshold did not correlate significantly with the level of TSB. All latencies of waves I, III, and V correlated weakly, though with statistical significance, with the level of TSB (r ⫽ 0.26-0.28, all P ⬍ 0.05). The I-V interval also correlated weakly with the level of TSB (r ⫽ 0.24, P ⬍ 0.05). Comparison of BAER Between Neonates With Different Levels of TSB In general, as the level of TSB increased, the differences in BAER variables between neonates with hyperbilirubinemia and normal control subjects tended to be greater. The latencies of waves I, III, and V, and the I-V interval, were all slightly longer in neonates with TSB at 16-20 mg/dL than in those with TSB at 11-15 mg/dL (Figs 1-4).
Figure 4. Box plot of I-V interval (bold line across box, median; box, 25th and 75th centiles; extensions, largest and smallest values) in various groups of neonates. See text for significance of statistical comparison between different groups.
However, none of these BAER variables differed significantly between the two subgroups of neonates. Similarly, all wave latencies and the I-V interval in neonates with TSB greater than 20 mg/dL tended to be longer than in those with TSB at 16-20 mg/dL, but no statistically significant difference was found in any of these variables between the two subgroups (Figs 1-4). The wave latencies and I-V interval in neonates with TSB greater than 20/dL were all longer than in those with TSB at 11-15 mg/dL, with a statistical significance in wave V latency and I-V interval (P ⬍ 0.05 and 0.05, respectively, Figs 1-4). Discussion Bilirubin is known to inhibit mitochondrial enzymes and affect DNA synthesis and ion exchange. It disturbs neuro-excitatory signals, and impairs nerve conduction [1,18]. Some authors reported that the site of a lesion in auditory impairment caused by hyperbilirubinemia may be retrocochlear, with the cochlea unaffected [19]. Others reported that lesions causing auditory impairment after hyperbilirubinemia may include the organ of Corti, especially at the outer hair cells and the cochlear nerve [20]. In any case, it is presumed that neonatal BAER, which reflects functional integrity and the development of the auditory pathway up to the brainstem level, can be affected by hyperbilirubinemia. Our neonates who had hyperbilirubinemia showed an abnormal increase in the latencies of wave I, III, and V, suggesting that neonatal auditory function is impaired after hyperbilirubinemia. Twenty percent of 90 neonates with hyperbilirubinemia had an abnormal increase in wave V latency, and half of the 20% were associated with an elevation of BAER threshold. This abnormality appears to be somewhat similar to the prevalence (22%) of auditory impairment previously reported by some authors [9].
However, further analysis of the details of our BAER results revealed that the prevalence of auditory impairment is as high as 28%. The impairment could be peripheral or central. Previous authors reported no elevation in BAER threshold after neonatal hyperbilirubinemia [13]. In the present study, the mean value of the BAER threshold in neonates with hyperbilirubinemia was greater than in normal control subjects. Ten percent of the neonates had an elevated threshold, with seven slightly elevated, and two significantly elevated. Nine percent of the neonates had an increase in wave I latency. Some neonates had both an elevation in BAER threshold and an increase in wave I latency. These conditions led to a total of 16% of our neonates exhibiting BAER abnormalities (elevation in BAER threshold or an increase in wave I latency) that suggest peripheral auditory impairment. To date, few reports have described central auditory impairment after neonatal hyperbilirubinemia. In BAER, some authors found no abnormality in the I-V interval [8]. The present study revealed that the increase in BAER latency in the neonates with hyperbilirubinemia was more significant for the later waves than for the earlier waves, and that the I-V interval increased significantly. These results suggest that, in addition to peripheral auditory impairment, there is also central auditory impairment after neonatal hyperbilirubinemia. In the present study, central impairment was found in 18% of neonates with hyperbilirubinemia. Such impairment may have important clinical implications. Therefore, researchers will need to differentiate between peripheral and central auditory impairment, and to pay more attention to central auditory impairment after neonatal hyperbilirubinemia. In addition to hyperbilirubinemia, neonatal auditory function can be impaired by some pre-, peri-, and postnatal conditions or problems, as described in Materials and
Jiang et al: BAER Latencies in Hyperbilirubinemia 39
Methods [2-5,8,14-17]. In the present study, any neonates who had evidence of these risk factors had been excluded. Thus, our results were unlikely to be significantly confounded by any of these risk factors. Some of our neonates with hyperbilirubinemia had pneumonia, sepsis, and a low Apgar score. However, none of these neonates showed any clinical signs of significant neurological abnormalities. Our previous study revealed that a low Apgar score at 1 or 5 minutes alone, with no clinical signs of hypoxicischemic encephalopathy, does not have any significant effect on BAER [21]. Thus, these factors are unlikely to have any significant effects on the BAER results of our neonates with hyperbilirubinemia, and cannot be responsible for the BAER abnormalities seen in these neonates. Although we cannot completely rule out some subtle confounding effects of neonatal conditions associated with hyperbilirubinemia and conductive ear disorders in some of the neonates, the BAER abnormalities must be attributed mainly to bilirubin ototoxicity. Thus, our results from the BAER data, collected shortly after obtaining TSB, indicate that hyperbilirubinemia has an acute toxic effect on neonatal auditory function, both peripherally and centrally. Some of our subjects were tested repeatedly with BAER several days and weeks after the initial BAER recording. These preliminary data indicated that after phototherapy or exchange transfusion, the BAER abnormalities in neonates with hyperbilirubinemia recovered quickly, suggesting that the auditory impairment is largely transient. This is similar to findings by other investigators [22]. Infants or children who had hyperbilirubinemia were often found not to be associated with adverse neurodevelopmental outcomes after prompt treatment [23]. The toxic effect of hyperbilirubinemia on visual evoked potentials was also found to be transient after prompt treatment [24]. Our study suggests that a TSB level greater than 10 mg/dL has a toxic effect on neonatal auditory function. As the level of TSB increased, the latencies of waves I, III, and V and the I-V interval tended to be longer, and the differences in these BAER variables between neonates with hyperbilirubinemia and normal neonates tended to be greater. Thus, the acute toxic effect of hyperbilirubinemia on peripheral and central auditory function tends to be more significant at a higher level of TSB than at a lower level. However, all these variables only correlated weakly with the level of TSB. Our study further revealed that there were no major differences in BAER variables between the neonates with different levels of TSB, except for a significant increase in wave V latency and I-V interval in the neonates with TSB greater than 20 mg/dL when compared with those with TSB at 11-15 mg/dL. The BAER abnormalities occurred slightly more frequently in neonates with a higher level of TSB than in those with a lower level of TSB. It appears that auditory impairment, reflected by BAER abnormalities, after neonatal hyperbilirubinemia does not increase closely with the increase in
40
PEDIATRIC NEUROLOGY
Vol. 37 No. 1
the level of TSB. This finding may be partly related to individual differences in sensitivity to the ototoxic effect of hyperbilirubinemia and, in particular, in perinatal conditions that are associated with hyperbilirubinemia, e.g., acidosis and hypoxia, which are known to facilitate bilirubin toxicity [25].
We thank the nurses and doctors at the Neonatal Unit, Children’s Hospital, Fudan University (Shanghai, China), for their assistance in recruiting subjects. This research was supported by grants from the Children’s Hospital of Fudan University, and from Defeating Deafness, UK.
References [1] Shapiro SM. Bilirubin toxicity in the developing nervous system. Pediatr Neurol 2003;29:410-21. [2] Joint Committee on Infant Hearing. Year 2000 position statement: Principles and guidelines for early hearing detection and intervention programs. Pediatrics 2000;106:798-817. [3] Newton V. Adverse perinatal conditions and the inner ear. Semin Neonatol 2001;6:543-51. [4] Kountakis SE, Skoulas I, Phillips D, Chang CYJ. Risk factors for hearing loss in neonates: A prospective study. Am J Otolaryngol 2002;23:133-7. [5] Wilkinson AR, Jiang ZD. Brainstem auditory evoked response in neonatal neurology. Semin Fet Neonatol Med 2006;11:444-51. [6] Hung KL. Auditory brainstem responses in patients with neonatal hyperbilirubinaemia and bilirubin encephalopathy. Brain Dev 1989; 11;297-301. [7] Gupta AK, Mann SB. Is auditory brainstem response a bilirubin neurotoxicity marker? Am J Otolaryngol 1998;19:232-6. [8] Streletz LJ, Graziani LJ, Branca PA, Desai HJ, Travis SF, Mikaelian DO. Brainstem auditory evoked potentials in fullterm and preterm newborns with hyperbilirubinemia and hypoxemia. Neuropediatrics 1986;17:66-71. [9] Boo NY, Oakes M, Lye MS, Said H. Risk factors associated with hearing loss in term neonates with hyperbilirubinaemia. J Trop Pediatr 1994;40:194-7. [10] Funato M, Tamai H, Shimada S, Nakamura H. Vigintiphobia, unbound bilirubin, and auditory brainstem responses. Pediatrics 1994;93: 50-3. [11] Amin SB, Ahlfors C, Orlando MS, Dalzell LE, Merle KS, Guillet R. Bilirubin and serial auditory brainstem responses in premature infants. Pediatrics 2001;107:664-70. [12] Smith CM, Barnes GP, Jacobson CA, Oelberg DG. Auditory brainstem response detects early bilirubin neurotoxicity at low indirect bilirubin values. J Perinatol 2004;24:730-2. [13] Soares I, Collet L, Delorme C, Salle B, Morgon A. Are click-evoked BAEPs useful in case of neonate hyperbilirubinaemia? Int J Pediatr Otorhinolaryngol 1989;117:231-7. [14] Jiang ZD. Maturation of peripheral and brainstem auditory function in the first year following perinatal asphyxia—a longitudinal study. J Speech Lang Hear Res 1998;41:83-93. [15] Jiang ZD, Brosi DM, Wilkinson AR. Hearing impairment in preterm very low birth weight babies at term revealed by brainstem auditory evoked responses. Acta Paediatr 2001;90:1411-5. [16] Jiang ZD, Brosi DM, Wang J, Wilkinson AR. Brainstem auditory-evoked responses to different rates of clicks in small-forgestational age preterm infants at term. Acta Paediatr 2004;93:76-81. [17] Jiang ZD, Brosi DM, Wang J, Wilkinson AR. One-third of term babies after perinatal hypoxia-ischaemia have transient hearing impairment: Dynamic change in hearing threshold during the neonatal period. Acta Paediatr 2004;93:82-7. [18] Dennery PA, Seidman DS, Stevenson DK. Neonatal hyperbilirubinemia. N Engl J Med 2001;344:581-90.
[19] Sano M, Kaga K, Kitazumi E, Kodama K. Sensorineural hearing loss in patients with cerebral palsy after asphyxia and hyperbilirubinemia. Int J Pediatr Otorhinolaryngol 2005;69:1211-7. [20] Rhee CK, Park HM, Jang YJ. Audiologic evaluation of neonates with severe hyperbilirubinemia using transiently evoked otoacoustic emissions and auditory brainstem responses. Laryngoscope 1999;109: 2005-8. [21] Jiang ZD, Shao XM, Wilkinson AR. Brainstem auditory evoked responses in term neonates with temporary low Apgar scores. Acta Otolaryngol (Stockh) 2005;125;163-8.
[22] Wong V, Chen WX, Wong KY. Short- and long-term outcome of severe neonatal nonhemolytic hyperbilirubinemia. J Child Neurol 2006;21:309-15. [23] Newman TB, Liljestrand P, Jeremy RJ, et al. Outcomes among newborns with total serum bilirubin levels of 25 mg per deciliter or more. N Engl J Med 2006;354:1889-900. [24] Chen WX, Wong V. Visual evoked potentials in neonatal hyperbilirubinemia. J Child Neurol 2006;21:58-62. [25] Gustafson PA, Boyle DW. Bilirubin index: A new standard for intervention? Med Hypotheses 1995;45:409-16.
Jiang et al: BAER Latencies in Hyperbilirubinemia 41