(34) Neonatal Risk Factors for Hearing Loss Diagnosed Before and After Discharge from a Neonatal Intensive Care Unit

Background: Hearing loss incidence is increased in infants in the Neonatal Intensive Care Unit (NICU), but few studies capture hearing loss diagnosis after NICU discharge.

Objectives: To measure hearing loss diagnosis in NICU graduates using electronic health record (EHR) data and to determine the significance of studied risk factors during the NICU stay as predictors for hearing loss.

Design/Methods: A retrospective cohort analysis of EHR data of neonates in a level IV NICU from 2008 – 2019 who met inclusion criteria (Figure 1) and had a documented hearing screen either as an inpatient or outpatient was performed. Univariate analyses were performed using chi-squared tests and t-tests for categorical and continuous variables, respectively. Multivariate analyses were performed using a traditional cox proportional hazard regression to account for variable time to hearing loss diagnosis.

Results: A total of 6,229 infants met study inclusion criteria and 1,414 were found to have a hearing loss diagnosis, with 1,356 (96%) diagnosed during follow up visits. Table 1 and Table 2 display the univariate analysis of the cohort. Those with hearing loss had significantly lower gestational age (especially < 29 weeks gestational age) and birth weight compared to those without hearing loss. Many factors related to severity of illness (ventilator days, admission length, etc.) were significantly enriched among those with hearing loss. The multivariate adjusted risk factors significantly associated with hearing loss (Figure 2) included cytomegalovirus (CMV), Hypoxemic Ischemic Encephalopathy (HIE), Trisomy 21, and Cleft lip and/or palate (p < 0.01). Other adjusted risk factors during the NICU stay included total bilirubin >15 mg/dL and blood urea nitrogen (BUN) > 20 mg/dL as well as prolonged days of exposure to dopamine and milrinone.

Conclusions: The study methodology identified conditions commonly associated with hearing loss such as congenital CMV, HIE, Trisomy 21, and cleft lip/palate. Additional clinical, laboratory and medication risk factors identified likely reflect illness severity and inflammation. Increased days of exposure to classic ototoxic medications like gentamicin and vancomycin were not significant predictors in multivariate analysis. This study demonstrates the importance of incorporating longitudinal follow-up data into neonatal research with the need to develop predictive models to identify and mitigate evolving risks.