The objective was to determine frequency of sensorineural hearing loss (SNHL), identified by abnormal threshold in evoked potentials, absence of otoacoustic emissions and behavioral responses, auditory neuropathy (AN) (absence of evoked potentials, with preservation of otoacoustic emissions), and neurological comorbidity in infants with hyperbilirubinemia (HB) treated with exchange-transfusion (ET). From a total of 7,219 infants, ET was performed on 336 (4.6%). Inclusion criteria were fulfilled in 102; 234 children did not meet criteria (182 outside of the study period, 34 did not have complete audiological evaluation, and 18 rejected the followup). Thirty-five children (34%) were born at-term and 67 (66%) were preterm. Children had a mean age of
The auditory pathway is known as one of the most susceptible parts of the central nervous system to noxious agents. Severe neonatal hyperbilirubinemia (HB) is a common cause of sensorineural hearing loss (SNHL) and auditory neuropathy (AN) [
Some audiological studies in children with serum bilirubin levels >20 mg/dL, have reported auditory dysfunction in 17–87% of cases [
The Joint Committee on Infant Hearing of the American Academy of Pediatrics (AAP) considers ET a risk factor for SNHL [
We designed a retrospective, case-control study, with the following inclusion criteria: having been born at the National Institute of Perinatology “Dr. Isidro Espinosa de los Reyes” (INPerIER) in Mexico City, between January 1, 2000 and December 30, 2010; a history of ET secondary to severe HB after Rh hemolytic disease; ABO incompatibility or multifactorial HB, regardless of birth gestational age or associated morbidity during the neonatal period, and belonging to the pediatric followup clinic for high-risk newborns. Severe HB was defined as a bilirubin increase >0.5 mg/dL per hour in term infants, or >0.3 mg/dL for preterm infants, requiring exchange transfusion. Rh hemolytic disease was defined as different maternal-infant antigens and a positive direct Coomb’s test. ABO incompatibility was defined as an infant’s blood type A or B with a type O mother. Multifactorial HB was defined as the same maternal-infant blood type and severe HB. Our clinic’s characteristics have been described in previous publications [
Phototherapy and ET were performed according to AAP guidelines [
Two groups were defined in the followup based on their hearing status: (1) children with SNHL and (2) a control group of children, consisting of eighty-seven children (85%), who showed bilateral normal hearing (BNH) with history of exchange transfusion for severe hyperbilirubinemia. Neonatal variables and procedures were compared as follows: gestational age at birth in weeks belongs to a term (birth age between 37 and 42 weeks) or preterm (<37 weeks) infant group; birth weight, Apgar score at one and five minutes, gender, days of endotracheal ventilation, length of hospital stay, and age at time of studies. Risk variables for SNHL documented in the neonatal period were peak serum indirect bilirubin level in mg/dL at the time of ET; days of phototherapy; exposure to other potentially ototoxic drugs such as aminoglycosides [
Samples were obtained by peripheral venipuncture. All specimens were protected from light after they were drawn, and these were analyzed immediately. For the quantitative determination of serum bilirubin, we utilized a dichlorophenyl diazonium (DPD) reagent. Measurements were performed using a Beckman Synchron CX-9 equipment (Fullerton, CA, USA).
All infants included in the study underwent determination of conventional brainstem auditory evoked potentials (BAEP) at 3 and 6 months of chronological age with a Nicolet Viking Quest (Nicolet Biomedical Inc., Madison, WI, USA) computer. The test was conducted in a soundproof room reserved for this purpose within the neurophysiology unit, with the child in physiological sleep in a regular bed. BAEP determinations were performed after skin cleaning with alcohol-acetone and to apply conductive gel, using the international 10–20 system electrode placement [
To rule out middle ear pathology, children were studied with a Carl Zeiss OP-MI-9 F-125 Oto-microscope (Jenna, Germany). Afterward, we used a Grason-Stadler GSI TympStar V.2 Impedanciometer (Madison, WI), with ANSI S3.6-1996 calibration. The test tone used in tympanometry was 226 Hz, 85 at dB, pressure range = −600 to 400 deca-Pascals, compliance range of 0.1 to 5.0 mL, with an accuracy of ±5%. Children should have had a Jerger type A curve [
Children >3 years of age underwent audiometry by conditioning game technique [
Hearing was considered normal in conditioned audiometry when the threshold was ≤20 dB in the frequencies analyzed. The criteria for SNHL were considered when both the air and bone conduction thresholds were increased and overlapping with hearing thresholds ≥25 dB in at least two of the frequencies tested. All subjects were studied, diagnosed, and followed up by a certified pediatric audiologist (MCCF).
In order to document auditory neuropathy, automatic transient-evoked otoacoustic emissions (TEOAE) were performed in infants with abnormal BAEP result in both determinations. A Madsen otoacoustic-emission-analyzer AccuScreen GN Otometrics equipment (Copenhagen, Denmark) was utilized with the following technique: the study was conducted in a soundproof room, placing the probe in each of the ear canals. Stimulation was performed using clicks for each ear sweep. Equipment displays automatically a “Pass” or “Refer” result. “Pass” is equivalent to normal function of outer hair cells of the cochlea in the explored ear. The criteria for diagnosing AN consisted of two abnormal BAEP determinations (flat line or only wave V at high intensity stimulation) and “Pass” otoacoustic emissions result [
Normal binaural hearing was considered when the infant passed the first or second test of conventional BAEP study, or when the infant passed the evaluation in the audiology clinic; these children formed the control group. SNHL was identified when the infant presented two BEAP studies with thresholds >45 dB nHL and did not pass the behavioral auditory tests. Hearing loss was classified in severity stages by averaging the hearing thresholds at 500, 1,000, and 2,000 Hz frequencies after performing the audiometric measurement for each ear. Subjects with audiometric threshold between 21 and 40 dB were classified with mild hearing loss; those between 41 and 70 dB with moderate hearing loss; children with thresholds of 71–90 dB were classified with severe hearing loss and >90 dB profound hearing loss [
The presence of neurological sequelae (pathologic condition resulting from a disease, once the offending agent is removed) was documented by serial neurological examinations performed by a certified neuropediatrician with the help of brain imaging scans, neurophysiological recordings, laboratory studies, and with posterior appointments to the follow-up clinic to determine alterations such as cerebral palsy and/or epilepsy (according to International Classification of Diseases Tenth Edition, categories G80, and G40 resp.).
Continuous data were presented as means and standard deviations and were analyzed using one-way analysis of variance (ANOVA) and the Mann-Whitney
From a population of 7,219 children in the pediatric follow-up clinic for high-risk newborns, 336 (4.6%) children had undergone ET. One-hundred-two infants met inclusion criteria for this study, with a mean age of 5.5 years ±3.9 (range of 2 to 10 years), 234 children did not meet the inclusion criteria (182 were outside the study period, 34 did not have complete audiological evaluations, and 18 rejected the followup in the clinic). Causes of ET were distributed as follows: Rh isoimmunization,
Comparison of Indirect bilirubin levels in children with exchange transfusion for different causes (
Causes of Exchange transfusion | SNHL | BNH | Both groups | ||||||
---|---|---|---|---|---|---|---|---|---|
|
IB mg/dL-sd | Range |
|
IB mg/dL-sd | Range |
|
IB mg/dL-sd | Range | |
Rh hemolytic disease | 5 | 23.8 ± 7.2 | 16.2–35.6 | 43 | 18.0 ± 5.4 | 6.8–28.4 | 48 | 18.6 ± 5.8 | 6.8–35.6 |
ABO incompatibility | 2 | 21.2 ± 9.2 | 14.7–27.8 | 26 | 18.7 ± 5.5 | 7.3–35.8 | 28 | 18.9 ± 5.6 | 7.3–35.8 |
Multifactorial hyperbilirubinemia | 8 | 21.5 ± 4.5 | 16.1–29.3 | 18 | 20.2 ± 4.6 | 9.6–28.6 | 26 | 20.6 ± 4.5 | 9.6–29.3 |
|
|||||||||
Total | 15 | 22.2 ± 5.7 | 14.7–35.6 | 87 | 18.7 ± 5.3 | 6.8–35.8 | 102 | 19.2 ± 5.4 | 6.8–35.8 |
|
|||||||||
|
|
|
|
Clinical characteristics of children with SNHL and BNH with ET are presented in Table
Clinical characteristics of groups of children with SNHL and BNH with Exchange transfusion (
Variable | SNHL group ( |
BNH group ( | ||||||
---|---|---|---|---|---|---|---|---|
|
Average ± SD | Range |
|
Average ± SD | Range | U M-Wa |
|
|
Gestational age (weeks) | 15 | 33.7 ± 2.1 | 30–39.5 | 87 | 35.2 ± 3.3 | 26–40.6 | 425 | 0.03a |
Birthweight (g) | 15 | 1927 ± 581 | 910–2,975 | 87 | 2,218 ± 790 | 790–3795 | 514 | 0.19a |
1 min Apgar score | 15 | 6.4 ± 1.7 | 1–9 | 87 | 6.7 ± 2.2 | 2–9 | 497 | 0.12a |
5 min Apgar score | 15 | 8.3 ± 0.8 | 1–9 | 87 | 8.5 ± 1.0 | 4–9 | 504 | 0.08a |
Mechanical ventilation (days) | 5 | 5 ± 4 | 1–14 | 22 | 5 ± 2 | 1–20 | 46.5 | 0.58a |
Hospital stay (days) | 15 | 31 ± 26 | 7–96 | 87 | 20 ± 18 | 5–102 | 469 | 0.08a |
Phototherapy (days) | 15 | 5.5 ± 0.7 | 4–6 | 87 | 5.5 ± 1.4 | 1–10 | 637 | 0.87a |
Indirect bilirubin (mg/dL) | 15 | 22.2 ± 5.7 | 14.7–35.6 | 87 | 18.7 ± 5.3 | 6.8–35.8 | 451 | 0.05a |
Age at follow-up (years) | 15 | 6.7 ± 2.3 | 2–10 | 87 | 5 ± 3.7 | 1–10 | 566 | 0.41a |
Male | 9 | 60% | 42 | 48% | 0.40** | |||
Female | 6 | 40% | 45 | 52% | ||||
Term infant | 1 | 7% | 34 | 39% | 0.01** | |||
Premature infant | 14 | 93% | 53 | 61% |
SNHL: sensorineural hearing loss. BNH: bilateral normal hearing.
Odds ratio calculations for risk-factors in children with SNHL and BNH with exchange transfusion.
Variable | SNHL |
BNH |
OR 95% CI |
|
||
---|---|---|---|---|---|---|
Yes | Yes | |||||
Preterm infants | 14 | 93% | 53 | 61% | 8.9 (1.1–71.4) | 0.01 |
Asphyxia | 2 | 13% | 5 | 6% | 2.5 (0.4–14.3) | 0.28 |
Neonatal sepsis | 9 | 60% | 42 | 48% | 1.6 (0.5–4.9) | 0.40 |
Intraventricular hemorrhage | 2 | 13% | 2 | 2% | 6.5 (0.8–50.5) | 0.04 |
Amikacin exposure | 7 | 47% | 49 | 56% | 0.6 (0.2–2.0) | 0.48 |
Furosemide exposure | 5 | 33% | 7 | 8% | 5.7 (1.5–21.4) | 0.005 |
SNHL: sensorineural hearing loss. BNH: bilateral normal hearing.
BAEP results in children with SNHL were as follows: three infants had hearing thresholds of 80 dB nHL, two presented only wave V at 95 dB nHL, and ten had no response to >95 dB nHL stimulation. TEOAE recordings were negative in all cases of SNHL and thus, AN was not documented in the sample. Audiometric measurements showed severe SNHL in 10 cases (hearing threshold of 82 dB) with profound SNHL in 3 cases (hearing threshold of 99 dB). In all cases the auditory alteration was bilateral and symmetrical.
The higher neurological comorbidity was observed in the group of children with SNHL. An increased frequency of cerebral palsy was documented for the group with SNHL (20%) when compared with results from those of children with BNH (3%) (OR = 7.0 [1.2–38.7],
This paper demonstrated a higher frequency of SNHL (15%) in children with a history of ET treated in a 3rd level hospital in Mexico City. Hearing alteration was produced despite the cause of severe hyperbilirubinemia and was associated to preterm birth and low gestational age, level of indirect bilirubin, and exposure to furosemide.
Comparison with other studies is limited because of the differences in methodology, severity of hyperbilirubinemia, ET criteria, and SNHL classification. The basic mechanism of bilirubin neurotoxicity remains unknown. It is unclear why some infants do not develop hearing loss or neurological injury with the serum bilirubin levels that other infants do [
Some researchers studied the effect of HB on the auditory pathway during its acute phase, with a short prospective design, assessing BAEP and otoacoustic emissions before and after phototherapy or ET [
In this paper we analyzed the variable ET under usual clinical conditions present in the NICU, where newborns with severe HB usually present other associated co-morbidities, therefore being difficult to document severe HB as single disease. For example, Patra et al. [
Severe HB that requires ET for its treatment is a clinical variable that cannot be accurate or measured objectively, since it is not easy to precisely define a serum bilirubin value that indicates the need for ET or that is directly associated with neural or auditory damage. Thus, ET is a strong qualitative variable associated as a risk factor to SNHL. This paper demonstrates the need to pay special attention to the increased risk of SNHL among infants treated with ET for severe HB. However, not all cases of severe neonatal HB with ET result in hearing or neurological deficits, and the exact threshold in which bilirubin becomes dangerous is not uniform among populations.
HB is more prevalent and severe in preterm infants, and its course is more prolonged than in term infants as a result of the red blood cells, liver, and gastrointestinal system immaturity. As consequence of these facts, in this study we found an increased risk for SNHL in preterm infants.
Loop diuretics cause SNHL by inhibiting ion transport of within the stria vascularis, reducing the electrochemical gradients that create the Endocochlear potential. More important is the fact that loop diuretics enhance the rate of permanent hearing loss induced by aminoglycosides. The mechanism for interaction between aminoglycosides and loop diuretics implies alterations in the blood labyrinth barrier, which facilitates aminoglycoside entry to the endolymphatic compartment [
The incidence of AN in infants with severe HB and ET has been reported as high [
The pathophysiology of SNHL secondary to severe HB is not well defined, although its toxicity can affect cochlear hair cells and neurons of the basal nuclei and of the central auditory pathways. A recent report of 30 infants with hearing loss and exposure to severe HB suggests that damage to the outer hair cells of the cochlea is very common; twenty-six infants (87%) out of 30 had cochlear damage and in four cases (13%) an AN was documented [
Unfortunately the damage to the auditory pathway is not the only sequelae of severe HB. As we observed in our study, Ogunlesi et al. reported cerebral palsy in 86.4% of 22 infants with bilirubin encephalopathy, seizures in 40.9%, and deafness in 36.4% [
The size of the sample was small, and therefore we must have caution in the interpretation of these results. In infants with HB treated with ET we reported here a higher frequency of SNHL and neurological comorbidity; however, we were unable to find AN. This fact merits more research in the future by our work team. These results deserve a continuous long-term pediatric followup of infants with greater number of patients.
The frequency of SNHL in children with a history of ET treated at a 3rd level hospital in Mexico City was high (15%). No cases of AN were documented. The preterm newborns have higher risk for SNHL. Moreover, children with SNHL and history of HB-ET have an increased risk of cerebral palsy and epilepsy. Thus, the early diagnosis and early intervention are very important actions for a better outcome of these patients.
The authors declare that there is no conflict of interests.