MANUSCRIPT CITATION

Hemmingsen D, Moster D, Engdahl BL, Klingenberg C. Sensorineural hearing impairment among preterm children: a Norwegian population-based study. Arch Dis Child Fetal Neonatal  Ed 2024;0: F1–F7. doi:10.1136/archdischild-2024-326870

REVIEWED BY

Dr Chidi Anakebe, Dr Sohaib Bin Nawaz, Dr Haji Sheeraz Khan

TYPE OF INVESTIGATION

Observational Study

QUESTION

In Neonates born between 22^nd-36^th week gestational age (GA), what is the risk of development of sensorineural hearing impairment (SNHI), how is it stratified across gestational ages, and how much of it can be attributed to perinatal morbidities and therapies compared with gestational age alone?

METHODS

• Design: Observational Study – Population-Based Cohort (using data registries)
• Allocation: All live-born infants over a period of 16 years were included in the study group (preterm i.e., 22^nd to 36^th week GA) or reference group (term i.e., ≥37 weeks GA).
• Blinding: Not applicable as an observational study.
• Follow-up period: 5 years
• Setting: Population-based cohort of all children born in Norway from 1999 to 2014 with data available on 5 nationwide registries and records linked together using 11-digit personal identification numbers.
• Patients: 60,023 infants born preterm (22^nd to 36^th GA), and 869,797 infants born at term (>37 weeks GA)
• Inclusion Criteria: All live-born infants in Norway born from the start of 1999 until the end of 2014
• Exclusion Criteria: Neonates whose birth weight (Z-score) for GA was outside 3SD, Neonates without an 11-digit personal identification number, and infants who died before the age of 2 years.

• Intervention/ Exposure: The main exposure was gestational age, further classified into:

1. 1. moderate to late preterm (MLP) infants (32–36 weeks GA)
   2. very preterm (VP) infants (28-31 weeks GA), and
   3. extremely preterm (22-27 weeks GA) infants.
   4. Other factors studied included maternal variables – daily smoking early in pregnancy, immigrant status, parental consanguinity, educational level, mode of delivery, and prolonged (> 24 hours) rupture of membranes.
   5. Neonatal variables studied include – Sex, birth weight, and length, small for GA (SGA), antibiotic therapy, intracranial hemorrhage, jaundice therapy, non-invasive respiratory support, and mechanicalventilation.

• Outcomes: The main outcome studied was Sensorineural hearing impairment (SNHI) identified during 5 years after birth and documented at least on two occasions.

• Primary outcome: Sensorineural Hearing Impairment (SNHI)
• Secondary outcome: None

• Analysis and Sample Size:

1. 1. The sample size was 60,023 infants in the preterm (study) group and 869,797 infants in the term (reference) group.
   2. The primary outcome was further analyzed against different variables, within different (predefined) subgroups of preterm infants.
   3. The author did all statistical analyses using the SPSS software (V.28.0.1.0).
   4. Comparisons of groups were performed with chi-squared (χ2) and non-parametric tests and two-tailed P values <0.05 were considered statistically significant.
   5. The authors used log-binomial regression analysis to evaluate the association between exposure and SNHI.
   6. Literature search was done to identify mediators for SNHI: antibiotic therapy, intracranial haemorrhage, jaundice therapy, non-invasive respiratory support, and mechanical ventilation; and foetal growth restriction was identified as a confounder.
   7. Directed Acyclic Graph (DAG) was drawn for the variables to identify cofounders, mediators, and covariates.
   8. The authors performed stratified univariable log-binomial regression analyses within subgroups to assess whether the association of mediators and covariates with SNHI varies with the degree of prematurity.
   9. The authors stated in their analysis, that they did not perform formal mediation analysis because they identified a significant interaction between exposures and mediators, and the effects of the mediators varied substantially between different subgroups.
   10. A log binomial regression analysis was also performed again with adjustment for small for gestational age (SGA).
   11. The authors presented the final results as crude risk ratio (RR) and adjusted risk ratio (aRR) with a 95% confidence interval (CI) mentioned.

• Patient follow-up / Percent Analyzed: 949,712 babies were born during the 1999-2014 period however, 13,505 did not meet the study criteria, and out of the remaining 936,207 babies, authors have excluded 6,351 due to unavailability of required data and 36 due to GA documented as <22 weeks. As a result, 929,820 (99.3%) babies were analyzed, with 60,023 in the study cohort and 869,797 term babies in the reference group.
• Funding: Northern Norway Regional Health Authority (grant number HN-1355-17), Gerda Meyer Nyquist Gulbrandson and Gerdt Meyer Nyquist’s Fund, and the Norwegian SIDS and Stillbirth Society.

MAIN RESULTS

60,023 (53% male) preterm infants born during the 16 years (1999-2014) and followed up until 2019 were analyzed against a reference population of 869,797 (51.2% male) term infants born during the same years in Norway.

In the preterm cohort, 3.4% (n=2065) were extremely premature (EP) infants, 10.3% (n=6192) were very premature (VP) infants, and 86.3% (n=51 766) were moderate to late premature (MLP) infants. The prevalence of SNHI in the preterm cohort was 1.4% compared to 0.7% in the reference group. The degree of prematurity had a profound effect, with the relative risk (RR) of SNHI increasing from 1.7 (95% CI: 1.6–1.9) in MLP infants, to 3.5 (95% CI:  3.0–4.1) in VP infants and 7.9 (95% CI: 6.5–9.5) in EP infants.

Most cases of SNHI were diagnosed early in smaller gestational age infants. The median age at first diagnosis of SNHI was lower in EP and VP infants at 1year (average of 0-4.5), compared to MLP and term-born infants who are diagnosed at age 4 years (average 1-6) P<0.002. Before the age of one, the percentage of infants diagnosed with SNHI varies from 38%, 44%, 21% and 24% amongst EP, VP, MLP and term-born infants respectively. These decreased by age five to 15%, 5%, 14% and 13% in EP, VP, MLP and term-born infants respectively.

In EP infants, intracranial hemorrhage was the only risk factor associated with increased risk of SNHI in univariable unadjusted analysis. However, in VP and MLP infants, delivery by caesarean section, SGA, intracranial hemorrhage, antibiotic therapy, and mechanical ventilation, were all associated with an increased risk of SNHI.

The authors then analyzed the independent impact of gestation in VP and MLP groups by comparing ‘healthy’ preterm babies who did not have antibiotic therapy or mechanical ventilation to term babies who were not admitted to the NICU. This confirmed the causal association of prematurity to SNHI, with a prevalence of 1.6% in VP infants, 1.0% in MLP infants, and 0.6% in term infants.

Prevalence, crude and adjusted risk ratio for sensorineural hearing impairment in children born in Norway 1999- 2014

Univariable analysis and RR (95% CI) for sensorineural hearing impairment in preterm and term children born in Norway 1999-2014

CONCLUSION

Prematurity is an independent risk factor for SNHI, with an increasing prevalence with decreasing gestational age (GA). Invasive therapies and comorbidities increase this risk, with a more pronounced effect on babies with GA ≥28 weeks. GA of 22 to 27 weeks has a higher risk of SNHI, which is further increased by intracranial haemorrhage but not by other comorbidities or invasive therapies.

COMMENTARY

The incidence of moderate to severe bilateral hearing deficit (>50 dB) is estimated to occur in 1 to 3 per 1,000 live births in healthy infants and 2 to 4 per 100 infants admitted to the NICU.¹ SNHI can lead to developmental delays as well as psychological and mental health issues, which adds to the existing challenges of being born prematurely.²

 

Risk factors associated with SNHI include family history of hearing loss, craniofacial anomalies, complex congenital abnormalities, congenital infections (such as TORCH), low birth weight, prematurity, hyperbilirubinemia requiring exchange transfusion, use of ototoxic medications, bacterial meningitis, a low Apgar score (<7 at 5 minutes), mechanical ventilation for at least five days, and NICU care lasting > 7 days.^2,3

 

This extensive study included over 60,000 preterm infants over a 16-year period, focusing on the risk of SNHI in preterm infants born between 22 and 36 weeks. It allowed a follow-up period of five years to capture cases of late-onset hearing impairment. The findings align with previous studies, showing a higher prevalence of SNHI (1.4%) in preterm infants compared to the reference group (0.7%), with the highest risk observed (5.2%) in extreme preterm infants (gestational age 22–27 weeks). The increased incidence of SNHI in younger gestational age is related to the development of the auditory system, as the structural auditory system develops in the first 20 weeks of life, while the neurosensory system only becomes functional around 25 weeks.⁴ Insults to this developmental process can result in varying degrees of hearing impairment.

 

Routine hearing screenings for infants involve non-invasive methods such as otoacoustic emissions (OAE) and confirmatory Auditory Brainstem Response (cABR) testing.¹ Infants with hearing loss due to neural conduction disorders or auditory neuropathy may require ABR testing, as OAE alone may not detect their SNHI.³ This study made use of either diagnostic methods which may not have captured all infants with hearing impairment (HI) due to neurodysfunction. It also did not grade the severity of HI nor disclose cases of isolated conductive HI, and other forms of HI were broadly classed as “unspecified”.

 

Other limitations of this study include:

• No grade of the severity of intracranial haemorrhage or periventricular leukomalacia, despite considering them as risk factors for SNHI.
• No clear definition of non-invasive ventilation, making it unclear what types of respiratory support were included.
• Did not specify the treatment methods used for jaundice in infants.
• No data analysis of uni/bi lateral hearing impairment.
• No information on infants who received hearing aids or cochlear implants.
• Data were collected from five mandatory health and social registries in Norway, though Norway has 17 national health registries,⁵ and no explanation was given on registry selection.
• The study did not mention other potential congenital infectious variables, such as congenital cytomegalovirus infection.⁶

 

Overall, while this study provides valuable insights into the prevalence and risk factors of SNHI in preterm infants, addressing these limitations could further enhance understanding in this area.

REFERENCES

1. Ohl C, Dornier L, Czajka C, Chobaut JC, Tavernier L. Newborn hearing screening on infants at risk. Int J Pediatr Otorhinolaryngol. 2009 Dec;73(12):1691-5. doi: 10.1016/j.ijporl.2009.08.027. Epub 2009 Sep 30. PMID: 19796829.

2. Wroblewska-Seniuk K, Greczka G, Dabrowski P, Szyfter-Harris J, Mazela J. Hearing impairment in premature newborns-Analysis based on the national hearing screening database in Poland. PLoS One. 2017 Sep 14;12(9):e0184359. doi: 10.1371/journal.pone.0184359. PMID: 28910311; PMCID: PMC5598959.

3. American Academy of Pediatrics, Joint Committee on Infant Hearing. Year 2007 position statement: Principles and guidelines for early hearing detection and intervention programs. Pediatrics. 2007 Oct;120(4):898-921. doi: 10.1542/peds.2007-2333. PMID: 17908777.

4. Graven, S. N., & Browne, J. V. (2008). Auditory development in the fetus and infant. Newborn and Infant Nursing Reviews, 8(4), 187- 193.https://doi.org/10.1053/j.nainr.2008.10.010

5. Health data | Biobank Norway. (n.d.). https://bbmri.no/health-data (Accessed August 2024)

6. De Cuyper, E., Acke, F., Keymeulen, A., De Leenheer, E. M. R., Van Hoecke, H., Padalko, E., et al. (2023). Risk factors for hearing loss at birth in newborns with congenital cytomegalovirus infection. JAMA Otolaryngology–Head & Neck Surgery, 149(2), 122. https://doi.org/10.1001/jamaoto.2022.4109

ACKNOWLEDGEMENT

BRIDGE Paediatric Research Network, Yorkshire and Humber, for article selection and editing.

FUNDING

None

CONFLICTS OF INTEREST

Dr Khan has received a speaker honorarium from Vertex Pharmaceuticals (Europe) Limited for an educational non-promotional talk.