Hearing screening in at-risk neonate cohort

International Journal of Pediatric Otorhinolaryngology 46 (1998) 81 – 89

Hearing screening in at-risk neonate cohort M. Hess a,*, U. Finckh-Kra¨mer a, M. Bartsch b, G. Kewitz b, H. Versmold b, M. Gross a a

Department of Audiology and Phoniatrics, Benjamin Franklin Medical Center, Freie Uni6ersita¨t Berlin, Fabeckstrasse 62, D-14195 Berlin, Germany b Department of Pediatrics and Neonatal Care, Benjamin Franklin Medical Center, Freie Uni6ersita¨t Berlin, Hindenburgdamm 30, 12200 Berlin, Germany Received 17 March 1998; received in revised form 20 September 1998; accepted 21 September 1998

Abstract Objective: this prospective study reports on the prevalence of hearing impairment in an at-risk neonatal intensive care unit (NICU) population. Design: from 1990 to 1997, 942 neonates were screened with transient evoked otoacoustic emissions (TEOAE) and brainstem evoked response audiometry (BERA). Results: 835 infants passed the primary screen for both ears, 57 for one ear, adding up to 94.7%. Seventeen infants (1.9%) were lost to follow-up. In thirteen infants (1.4%), bilateral hearing impairment above 30 dB was confirmed. While all children with hearing impairment belonged to the group of 820 children receiving aminoglycosides, only one presented no other risk factors. In 11 of the hearing impaired children other anamnestic factors, i.e. dysmorphism, prenatal rubella or cytomegaly, family history of hearing loss or severe peri- and postnatal complications seem to be more probable causes of the identified hearing loss. Conclusions: from our data, aminoglycosides seem not to be an important risk factor for communication related hearing impairment, when serum levels are continuously monitored, as occurred in our cohort. After adjustment for other risk factors, birth weight between 1000 and 1500 g and a gestational age between 29 and 31 weeks were no predictive markers for hearing impairment. It might be speculated that the improved medical treatment in a Neonatal Intensive Care Unit (NICU) reduces the probability of hearing impairment for those two groups. Conductive hearing loss as a possible additional cause for hearing impairment was not studied in detail, but the high percentage of malformations detected (four out of 13 hearing impaired infants) demands further monitoring, close follow-up, counselling and adequate treatment. © 1998 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Transient evoked otoacoustic emissions; Brainstem evoked response audiometry; Hearing impairment; Hearing screening; Risk factors

* Corresponding author. Tel.: +49 30 84452435; fax: + 49 30 84456855; e-mail: [email protected] 0165-5876/98/$ - see front matter © 1998 Published by Elsevier Science Ireland Ltd. All rights reserved. PII S0165-5876(98)00151-7

82

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

Fig. 1. Screening results for 942 at risk neonates.

1. Introduction Permanent hearing impairment occurs with a prevalence of about one to three per 1000 live births [1–4]. In neonates at risk, the prevalence is considerably higher (1:50) [1]. About 5 – 12% of infants are treated in a NICU, but these newborns account for 40–70% of all cases of early onset sensorineural hearing loss [5]. There is a general agreement that early detection of hearing impairTable 1 Risk factors for hearing impairment [7] Preterm birth (B32 weeks gestation) Birth weight B1500 gr Pre-, peri- or postnatal asphyxia (APGAR-score after 5 or 10 min 56 or Umbilical artery pH57.1) Prolonged mechanical ventilation for \10 days Pre-, peri- or postnatal infection known or suspected to be associated with Sensorineural hearing loss (e.g. toxoplasmosis, herpes, rubella, cytomegaly infection) Ototoxic medications including but not limited to aminoglycosides Dysmorphia (e.g. craniofacial anomalies or stigmata or other findings associated with a syndrome known to include hearing loss) Hyperbilirubinemia at a level exceeding indication for exchange transfusion Family history of congenital or delayed onset childhood sensorineural impairment

ment is warranted and should lead to early treatment with hearing aids, appropriate counselling and rehabilitation [6]. Screening for hearing impairment in neonates is considered to be an effective procedure in early detection of hearing impairment in infants [1,5,7–15]. Since 1990, we conducted a prospective study on screening for hearing impairment in neonates at-risk Fig. 1. To achieve most benefit from screening, a selection based in the knowledge of risks associated with possible hearing impairment was choosen. Our criteria were derived from high risk registers [7]. High risk for hearing impairment was assumed when there was at least one of the following criteria, as stated in Table 1.

2. Methods Between December 1990 and November 1997, 1060 neonates at risk for hearing impairment [7] were identified at the NICU (out of : 2700 children admitted to the NICU during this period). One-hundred and eighteen neonates could not be screened for various reasons, resulting in 942 otherwise non-selected neonates at-risk for hearing impairment, which were screened by the staff of the Department of Audiology and Phoniatrics. Screening was conducted for most children for an

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

audiological opinion at the NICU in the same building, whereas at follow-up infants were mostly examined in a sound proof chamber. For children of extreme preterm birth, screening before 35 weeks gestation was avoided and delayed for later testing, if possible.

2.1. Hearing assessment All children were tested bilaterally either with brainstem evoked response audiometry (BERA) using click-stimuli or transient evoked otoacoustic emissions (TEOAE). Both methods have been found to be highly valuable in hearing assessment [1]. TEOAE measurement is considered a very good screening technique because it correctly identifies cochlear integrity with very high sensitivity in the frequency range of 1 – 4 kHz. Its assessment is rapid, non-invasive, and easy to manage. However, with TEOAE hearing thresholds cannot be detected. BERA is considered the gold standard in auditory threshold identification in infants. Skill and experience with this complex technology are required for interpretation of the results and threshold identification [11]. Equipment for BERA (EVOPORT, Pilot Electronics, Germany; Screening-BERA ALGO 1+ ®, Natus Medical, Foster City, CA) and measurement of TEOAE (ILO88, Otodynamics, London, UK) were both available since 1993. During the first 2 years, BERA was the only test carried out. In 1993, some children were screened by both methods for evaluation. Usually the TEOAE screening took place without any sedation while the children were sleeping. Whenever ‘pass’ criteria were not achieved, a follow-up examination was done with determination of bilateral hearing thresholds with Click-BERA in a sound proof chamber. Usually the children were under sedation with promethazine (Atosil®) or seldom under general anesthesia.

2.2. ‘Pass’ criteria for hearing screening For screening purposes, adequate hearing was assumed when in at least one ear either: (i) ALGO-1 + achieved the ‘pass’ criterium; (ii) TEOAE correlation threshold reached \ 60%

83

and other factors were considered valid [11,12]; (iii) reproducible waveforms in Click-BERA down to 30 dBHL were present [16]. When first screening failed the test, it was repeated within 48 h. If this was not possible, short term audiologic follow-ups within 1–4 weeks at our department were offered to ensure high compliance, as a previous study reported [17]. All children who achieved the ‘pass’ criterium during this process were added up as ‘passed’. However, some children who did not achieve ‘pass’ criteria during the NICU period were lost to the enforced short term follow-up.

2.3. Data base Patient identification, risk factors, audiologic assessment data, and related information were gathered continuously at our department for prospective evaluation in a dBase3+ ® data base. Furthermore, patient data from the NICU were also compiled in a data base at the pediatric department. After selecting infants with high risk factors for this study, both data bases were merged, and missing data were completed. All data were checked, sorted, and edited manually to fit items properly for statistical computer analysis, which was accomplished with ACCESS® software.

3. Results Nine-hundred and forty-two cases were collected in 7 years. Based on the abovementioned criteria, hearing impairment was excluded bilaterally in 835 neonates. For 57 children, audiometric ‘pass’ criteria were obtained in one ear. Both groups add up to 892, which represents 94.7% of the investigated cohort. Fifty cases (5.3%) failed the screening testing at this first examination period for various reasons. At subsequent audiologic examination, 13 out of those 50 children had bilateral hearing loss (1.4% of the 942 infant cohort). Due to loss to follow-up, no immediate information could be collected in 37 children. Verification on hearing abilities in these 37 ‘lost’ neonates was attempted. Our investigations (including calling up parents) disclosed that in 19

84

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

Fig. 2. Risk factors in 942 infants.

neonates hearing impairment has been excluded elsewhere. Hence, data from 18 out of the above mentioned 37 infants were not available (Fig. 1). Two of those 18 neonates died, in one other case hearing loss of moderate or higher degree could

be excluded, but mild hearing impairment is still suspected, and in two cases no valid information could be obtained from audiologic assessment. Comparison between pass/fail percentages of tests for bilateral hearing (ALGO1+ , ClickBERA and TEOAE) shows that, after the first screening cycle, pass criteria were obtained with: (a) ALGO1+ in 138 out of 152 measurements (90.8%); (b) Click-BERA in 41 out of 57 measurements (71.9%); and (c) TEOAE in 460 out of 626 measurements (73.5%). Distribution of risk factors are given in Fig. 2. Potentially ototoxic drug medication was administered in 847 cases (89.9%), including 820 cases with aminoglycoside. Most commonly, neonates had one risk factor (n= 415), but more than half of the cases presented two or more risk factors (Fig. 3).

3.1. Clinical cases

Fig. 3. Number of risk factors per child.

In 13 neonates, hearing impairment was ascertained through measurement of hearing thresholds with BERA. Four infants (cases 1–4) showed craniofacial malformations (combined, in

m

f

f

m m m

m w

m

f

f m

2

3

4

5 6 7

8

9

10

11

12

13

36

41

39

39

27

26

28 34 25

35

30

40

31

Gestation age (weeks)

2600

4160

2370

3300

1330

920

1090 1380 430

2270

1370

2390

1740

Birth weight (g)

Submucosal cleft palate, neurologic deficiencies, other dysmorphic signs (Gordon syndrome?), prolonged mechanical ventilation for 15 days, consanguine parents (cousins) Multiple dysmorphic changes (choanal stenosis, cleft soft palate, dysplastic pinna) syndrome? OME? Dysmorphic middle ear? (loss to follow up) Asphyxia, cerebral palsy, twisting of the bowel, micrognathia, ear dysplasia and facial asymmetry Prenatal rubella Prenatal cytomegalie AIS; antibiotic treatment and repeated infections for 66 days (including AG), resuscitation for cardiac failure, mechanical ventilation for 60 days, BNS-epileptic attacks; mental retardation; OME Mechanical ventilation for 15 days, several infections, treatment with AG for 14 days AIS; antibiotic treatment and repeated infections for (including AG), severe birth asphyxia, resuscitation for cardiac failure, mechanical ventilation for 41 days Severe birth asphyxia due to placental bleeding (insertio velamentosa), APGAR 1/1/3 after 1/5/ 10 min, subsequent renal failure, mechanical ventilation for three weeks, multiple blood transfusions; AG immediately after birth, changed to vancomycin on day 2 (antibiotics for 39 days) AIS, treatment with AG for 10 days, polyerythemia, two days of mechanical ventilation because of perinatal asphyxia; only slight HL at lower frequencies? AIS, treatment with AG for 10 days (serum levels for gentamycin 2h after infusion 11,1mg/l) Family history of hearing impairment (both parents, and other siblings).

CHARGE-association

Clinical synopsis

55

100

50

45

100

60

n.p. 90 30*; 95–100**; 80***

60–70

40

n.p.

Not determinable

BERA right ear (dB)

50

100

50

45

100

70

100 40 50*; 95–100**; 35***

40

40

n.p.

60

BERA left ear (dB)

Fitted with HA

Fitted with HA

Fitted with HA

Lost to followup Fitted with HA

HA HA to be fitted

HA

Lost to followup

Lost to followup Dead

Outcome

Abbreviations: f, female; m, male; AIS, amnion infection syndrome; AG, aminoglycosides; HA, hearing aids; n.p., no click-stimuli related potentials. * First test; ** second test; *** third test.

f

Sex

1

Case No.

Table 2 Clinical synopsis of 13 neonates with hearing impairment

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89 85

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

86

Table 3 Risk factors in 13 out of 942 neonates with hearing impairment (suspected=hearing impairment was shown in extended audiologic examination, but not confirmed due to loss in follow-up) Risk factors

Proved hearing impairment

Suspected

Craniofacial anomaly Preterm birth B29 weeks gestation Prenatal rubella/CMV infection Mature infants with periand postnatal risk factors Family history of hearing impairment

1 3

3

2 3

1

part, with low birth weight under 1500 g or history of ototoxic drug administration). In cases 1 and 3, extended hearing evaluation was not accomplished due to loss to follow-up, so only our last evaluated hearing level(s) are presented. Table 2 lists cases 1 – 13 in a synoptical manner. Tables 3 and 4 summarize our findings concerning risk factors in those thirteen children with ascertained or high likelihood of hearing impairment. Fig. 4 displays the distribution of risk factors, respectively. Four out of thirteen neonates with hearing disorders suffered from craniofacial malformation, and these infants also showed associated anomalies in other parts of the body or neurological disorders. Only one infant had no other risk factor than ototoxic drug administration. Table 4 Degree of hearing impairment (better hearing ear) in 13 neonates with proved and suspected hearing loss (suspected = hearing impairment was shown in extended audiologic examination, but not confirmed due to loss in follow-up) Degree of hearing impairment (dB)

Proved hearing impairment

Mild (B40) Moderate (40–69) Severe (70–94) Profound (\94)

1 6 1 2

Suspected

2 1

4. Discussion Infants with one or more of the abovementioned risk factors are prone to develop hearing impairment, compared to infants from non-risk factor cohorts [1]. As screening capacities are limited, hearing assessment for neonates at risk should have priority [7]. In our study we analyzed 942 neonates who were admitted in the NICU. Several other studies on risk screening data have been published [5,8,9,14], representing data from 417, 229, 322, and 284 at-risk neonates, respectively. It should be mentioned, however, that only in the last years the first studies about general newborn screening programs have been published [2–4]. Out of 942 at risk infants, 13 (1.4%) were confirmed to have considerable hearing impairment requiring medical treatment, hearing aids and further rehabilition. Twelve of those infants had more than one risk factor for hearing impairment. Three out of 13 infants with hearing loss were not entirely assessed due to death or loss to follow-up. Those three had multiple handicaps. Conductive hearing loss due to OME might have been underlying in at least one of these cases (case 3). Usually, OME seems to be prevalent in more than 10% of newborns [18], requiring repeated medical monitoring and treatment. In the dysmorphic infant subgroup, middle ear disorders due to possible dysmorphic changes could not be ascertained nor excluded at this age. Possibly, additional conductive hearing loss may have contributed to the identified (air conduction) hearing level. Thus, for dysmorphic neonates thorough audiometric assessment including otoscopy and tympanometry and close follow-up is highly advisable. All neonates in our study that received aminoglycosides (820) were monitored for aminoglycoside peak and trough (lowest level) serum levels. If treatment with ototoxic drugs like aminoglycosides theoretically would have been excluded from the risk factor catalogue, one infant with hearing loss would have been missed (case 12). On the other hand, 807 infants received aminoglycosides without developing hearing impairment. Therefore, it is likely that this infant has hearing

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

87

Fig. 4. Risk factors— children with hearing aids.

impairment of unknown etiology, as prevalence for hearing impairment in non-risk cohorts can be assumed to be one in 2000 [5], or up to one to three per 1000 live births [1 – 4]. However, we continue to strongly recommend proper monitoring of aminoglycoside serum levels. It also should be kept in mind that aminoglycoside therapy might have impaired high frequency hearing, which could not be detected with the methods used in this study. From 942 neonates, 306 had a birth weight below 1500 g and 346 were born preterm (5 31 weeks gestation). From the 13 neonates with hearing impairment seven belonged to these risk factor groups (five had both risk factors, one only was preterm, one only had low birth weight; see Table 2). Prevalence of hearing loss in very low birth weight cohorts is reported to be between 0.8 and 7.8% [19–25]. Two (low birthweight), respectively three (born preterm) of these infants belonged to the dysmorphia group. If risk factor definitions would have been restricted to even lower birth weight than 1500 g, i.e. B1000 g, and to lower gestational age of B30 weeks, all hearing impaired infants in our study would have been detected as well. In other words, no hearing

impaired infant without other risk factors was found in the birth weight group between 1000 and 1500 g or pre-term group between 29 and 31 weeks gestation, respectively. It may be argued that with improved intensive care nowadays the cut-off level possibly may be set at lower thresholds, which has to be proven in further studies. The break down across tests revealed that—for bilateral hearing screening—ALGO1+ measurements had a better pass/fail percentage compared with TEOAE measurements. Neonates passed bilaterally in 90.8% with ALGO1+ , compared with 73.5% in TEOAE tests. In view of cost effectiveness of screening, this difference is noteworthy. High risk factor catalogues are important in neonatal screening. However, : 50% of preschool children with permanent hearing impairment (besides OME) are not detected in neonatal screening programs for neonatal at risk [5]. Other studies show that postnatally acquired hearing impairment and progressive hearing loss of unknown etiology, including genetically related hearing disorders, are around half of the cases [9,10]. For this reason, other epidemiological and health care programs should be applied in addition to neonatal screening [1,15].

88

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

Screening per se does not detect mild hearing impairment (B 30 dB), resulting in labeling these ‘passed’ infants as ‘normal’ for their hearing abilities. However, infants with mild bilateral hearing loss for more than 3 months are at risk for developmental disorders including central auditory deficits [26]. In our cohort, four of 13 infants (and three ‘lost’ infants with not yet completed audiometric assessment due to loss to follow-up) had ‘craniofacial anomalia’ as risk factor and had possible OME-induced hearing loss at the time of screening. As stated before, more than 50% of infants may develop OME after screening as well, unrelated to neonatal history. This points out the importance of ongoing control for OME-related hearing impairment. Parents of infants with adequate unilateral hearing were advised to retest hearing abilities after 6 months. Furthermore, possible progressive hearing loss with potential onset in all age groups was mentioned. Our counselling always called attention to further observation of speech and language development of the child. For this issue, we handed out an information leaflet containing a list of milestones for speech and language development in children. Although our data of at-risk neonates assume a lower percentage for at-risk hearing loss than in other study groups, neonatal hearing screening still is necessary regarding common public health purposes (such as quality of life improvement following early intervention and rehabilitation, intellectual benefits, and socio-economic cost-efficiency calculations [15]). Prevalence of hearing impairment in the overall population still demands and necessitates screening of all neonates, as mentioned in the Editorial of the British Medical Journal in January 1998 [27], and required by law in several states in USA [28].

5. Conclusions In our study, 13 of 942 screened neonates atrisk (1.4%) suffered from permanent hearing impairment. Eighteen infants (1.9%) were lost in follow-up. Craniofacial anomalia was associated

in about one third of cases (four out of 13) with conductive or sensorineural hearing impairment, making it advisable to point out the importance of assessing middle ear function and sensorineural hearing abilities in this subgroup. We detected only one hearing impaired infant with the single risk factor ‘aminoglycoside treatment’ out of 820 treated infants nevertheless serum levels did not exceed the therapeutic range (keeping in mind that high frequencies can not be tested with TEOAE and BERA). This ratio suggests that this neonate (statistically) might as well be considered an infant with non-risk-factor-related hearing impairment, since prevalence of hearing impairment is stated to be  1:1000–6:1000 cases for general newborn populations (compared with 1:50 for at-risk populations). From our data, aminoglycoside treatment is not an important risk factor for hearing impairment if serum levels are continuously monitored, as in all neonates in our cohort. Birth weights between 1000 and 1500 g and a gestational age between 29 and 31 weeks were no predictive markers for hearing impairment when no other risk factors were found. It might be speculated that in those children improved medical care statistically reduces the probability of hearing impairment to levels of children without risk factors. Newborn screening programs cannot detect children who will develop hearing impairment after the screening period [1]. Early identification of these children is difficult, and little is known about prevalences.

Acknowledgements We are indebted to the audiometricians M. Bautz, D. Golle, M. Petzold, B. Ruddigkeit, and I. Ufert for their clinical support. The authors thank Waltraud Finckh, MD, who provided many helpful comments on an earlier version of this paper. Supported in part by a grant from the ‘Verein zur Fo¨rderung ho¨r-, sprach- und stimmgesto¨rter Patienten an der FU Berlin e.V.’ and the ‘German Register for Hearing Loss in Children’ (nonprofit organizations).

M. Hess et al. / Int. J. Pediatr. Otorhinolaryngol. 46 (1998) 81–89

References [1] Early identification of hearing impairment in infants and young children. NIH Consens. Statement 11 (1) (1993) 1– 24. [2] P.M. Watkin, Neonatal otoacoustic emission screening and the identification of deafness, Arch. Dis. Child. 74 (1996) F16 – 25. [3] M.P. Downs, Universal newborn hearing screening The Colorado story, Int. J. Pediatr. Otorhinolaryngol. 32 (1995) 257 – 259. [4] A.B. Maxon, K.R. White, B.R. Vohr, T.R. Behrens, Using transient evoked otoacoustic emissions for neonatal hearing screening. Br. J. Audiol. (1993) 149–153 [5] D.R. Watson, R.J. McClelland, D. Adams, Auditory brainstem response screening for hearing loss in high risk neonates, Int. J. Pediatr. Otorhinolaryngol. 36 (1996) 147 – 183. [6] M.P. Downs, The rationale for neonatal hearing screening, in: E.T. Swigart (Ed.), 1986: Neonatal Hearing Screening, Taylor and Francis, London, 1986. [7] Joint Committee on Infant Hearing 1990 position statement. Asha 33 (Suppl. 5) (1991) 3–6. [8] U. Sitka, C. Rasinski, V. Gall, Ergebnisse eines Neugeborenen-Ho¨rscreenings mittels transitorisch evozierter otoakustischer Emissionen, Z. Geburtsh. Neonatol. 199 (1995) 71 – 77. [9] P.M. Watkin, M. Baldwin, G. McEnery, Neonatal at risk screening and the identification of deafness, Arch. of Dis. Child. 66 (1991) 1130–1135. [10] R.J. McClelland, D.R. Watson, V. Lawless, H.G. Houston, D. Adams, Reliability and effectiveness of screening for hearing loss in high risk neonates, Br. Med. J. 304 (1992) 806 – 809. [11] M. Ptok, A. Ptok, R. Scho¨nweiler, Audiometrie im Sa¨uglings- und Kindesalter, HNO aktuell. 4 (1996) 209– 216. [12] A. Lamprecht-Dinnesen, Otoacoustic emissions, HNO aktuell. 40 (11) (1992) 415–421. [13] M.P. Richardson, T.J. Williamson, S.W. Lenton, M.J. Tarlow, P.T. Rudd, Otoacoustic emissions as a screening test for hearing impairment in children, Arch. Dis. Child. 72 (1995) 294 – 297. [14] C.R. Kennedy, L. Kimm, D. Cafarelli Dees, P.I.P. Evans, M. Hunter, S. Lenton, R.D. Thornton, Otoacoustic emissions and auditory brainstem resposes in the newborn, Arch. Dis. Child. 66 (1991) 1124–1129. [15] E.T. Swigart, Neonatal Hearing Screening. Taylor and Francis, London, 1986.

.

89

[16] M. Hess, A. Lamprecht, R. von Kries, Detection of hearing impairment in premature infants-is a single determination of auditory thresholds sufficient?, Otorhinolaryngol Nova 1 (1991) 87 – 93. [17] M. Hess, C. Scholand, M. Gross, Parental compliance for hearing screening. Aktuelle phoniatrisch-pa¨daudiologische Aspekte, Berlin RGV 2 (1995) 46 – 49. [18] G.J. Sutton, P. Gleadle, S.J. Rowe, Tympanometry and otoacoustic emissions in a cohort of special care neonates, Br. J. Audiol. 30 (1) (1996) 9 – 17. [19] R.D. Eavey, M. Bertero, C. do, A.R. Thornton, B.S. Herrmann, J.M. Joseph, R.E. Gliklich, K.S. Krishnamoorthy, I.D. Todres, Failure to clinically predict NICU hearing loss, Clin. Pediatr. (PA) 34 (3) (1995) 138 – 145. [20] A. Johnson, P. Townshend, P. Yudkin, D. Bull, A.R. Wilkinson, Functional abilities at age 4 years of children born before 29 weeks of gestation, Br. Med. J. 306 (1993) 1715 – 1718. [21] K. Aziz, D.B. Vickar, R.S. Sauve, P.C. Etches, K.S. Pain, C.M.T. Rebertson, Province-based study of neurologic disability of children weighing 500 through 1249 grams at birth in relation to neonatal cerebral ultrasound findings, Pediatrics 95 (6) (1995) 837 – 844. [22] S. Saigal, P. Szatmari, P. Rosenbaum, P. Campbell, S. King, Cognitive abilities and school performance of extremely low birth weight children and matched term control children at age 8 years: A regional study, J. Pediatr. 118 (5) (1991) 751 – 760. [23] M. Hack, B. Weissman, N. Breslau, N. Klein, E. Borawski-Clark, A. Fanaroff, Health of very low birth weight children during their first eight years, J. Pediatr. 122 (6) (1993) 887 – 892. [24] I. Bergman, R.P. Hirsch, T.J. Fria, S.M. Shapiro, I. Holzman, M.J. Painter, Cause of hearing loss in the high-risk premature infant, J. Pediatr. 106 (95) (1985) 95 – 101. [25] The Scottish Low Birthweight Study Group. The Scottish low birthweight study: I. Survival, growth, neuromotor and sensory impairment. Arch. Dis. Child. (67) (1992) 675 – 681. [26] J.S. Gravel, I.F. Wallace, R.J. Ruben, Auditory consequences of early mild hearing loss associated with otitis media, Acta Otolaryngol. Stockh. 116 (1996) 219 – 221. [27] Effective screening in child health-give deaf children a better deal and look more broadly at other problems. Editorial, Br. Med. J. (1998) 316 (January 3). [28] J.L. Johnson, G.W. Mauk, K.M. Takegawa, P.R. Simon, C.C.J. Sia, P.M. Blackwell, Implementing a statewide system of services for infants and toddlers with hearing disabilities, Semin. Hear. 14 (1993) 105 – 118.