Monitoring neonates for ototoxicity

Abstract

Objectives: Neonates admitted to the neonatal intensive care unit (NICU) are at greater risk of permanent hearing loss compared to infants in well mother and baby units. Several factors have been associated with this increased prevalence of hearing loss, including congenital infections (e.g. cytomegalovirus or syphilis), ototoxic drugs (such as aminoglycoside or glycopeptide antibiotics), low birth weight, hypoxia and length of stay. The aetiology of this increased prevalence of hearing loss remains poorly understood. Design: Here we review current practice and discuss the feasibility of designing improved ototoxicity screening and monitoring protocols to better identify acquired, drug-induced hearing loss in NICU neonates. Study sample: A review of published literature. Conclusions: We conclude that current audiological screening or monitoring protocols for neonates are not designed to adequately detect early onset of ototoxicity. This paper offers a detailed review of evidence-based research, and offers recommendations for developing and implementing an ototoxicity monitoring protocol for young infants, before and after discharge from the hospital.

Key Words:

• Newborn hearing screening
• conditions/pathology/disorders
• paediatric
• anatomy and physiology

Introduction

In the United States (U.S.), 12% of ∼4 million births each year are admitted to the neonatal intensive care unit (NICU) (Osterman et al, 2011). The majority of these NICU admissions (∼80%, or ∼384,000 annual births) receive prophylactic or empiric treatment with the aminoglycoside antibiotic gentamicin and ampicillin (a β-lactam), until serious bacterial infection (or early-onset sepsis) is ruled out during the first 48–72 h after admission (Escobar, 1999; Osterman et al, 2011; Machado et al, 2014). During this period, there is a high risk of mortality if systemic infection is untreated (Escobar, 2005). Between 1% and 8% of NICU infants treated for sepsis with antibiotics have confirmed sepsis, indicating that the majority of NICU infants receive antibiotics prophylactically (Simiyu, 2003; Jordan et al., 2006; Kros & Desmonds, 2015; Medhat et al, 2016). Infants who continue to show health declines in the NICU with confirmed, or continuing suspicion of, sepsis can remain on aminoglycosides for 7–10 days or more (Camacho-Gonzalez et al, 2013).

Aminoglycoside antibiotics are widely-used in the U.S. because of their extreme effectiveness and broad-spectrum specificity toward organisms common in neonatal sepsis. Yet, there is a risk of acute nephrotoxicity and permanent sensorineural hearing loss (SNHL) in as many as 20% of adult patients receiving aminoglycosides for extended periods of time (Forge & Schacht, 2000). Hearing loss in neonates and infants negatively impacts the acquisition of listening and spoken language skills, academic success, and psychosocial integration (Yoshinaga-Itano et al, 1998). These delays are associated with an estimated socioeconomic cost of >$1,393,000 (in 2015 U.S. dollars after inflation) over the lifetime of each child with severe pre-lingual hearing loss (Mohr et al, 2000). Thus, the personal, medical, and socioeconomic impacts of hearing loss for infants discharged from the NICU are significant.

More importantly, early intervention following identification of hearing loss improves academic, language, and socio-emotional outcomes in infants and children. Widespread efforts have been made to decrease the age at which hearing loss is identified in the paediatric population via universal newborn hearing screening (NBHS) programmes. Toward that end, approximately 98% of births in the U.S. were screened for hearing loss in 2009–2010, an increase from 95% in 2005–2006 (Gaffney et al, 2014). Furthermore, the percentage of infants who did not pass the hearing screen and were subsequently diagnosed with a permanent hearing loss increased from 5% to 9% in that same time period (Gaffney et al, 2014). The introduction of NBHS has resulted in improved diagnosis of hearing loss before nine months of age (Kennedy et al, 2006). For infants with confirmed, permanent SNHL, 62% received early intervention in 2012 compared with 55% in 2006 (Williams et al, 2015). When hearing losses are identified early and prompt intervention provided, infants and children can demonstrate age-appropriate development in many areas (Stika et al, 2015). Current NBHS protocols are designed to identify hearing loss within the 0.5–4 kHz frequency range important for speech understanding; however, drug-induced ototoxicity typically starts at frequencies not tested in typical NBHS protocols (>4 kHz). Infants tend to listen with broad filters, meaning that they provide equal attention to all frequencies simultaneously, whereas adults are able to filter out frequencies not associated with important speech or environmental cues (Werner & Boike, 2001). It is critical for infants to have high frequency (>4 kHz) audibility to refine their auditory discrimination skills for communication (listening and spoken language) skills as well identifying critical environmental cues (Werner & Boike, 2001; Stelmachowicz et al, 2004).

Here, we discuss the need to monitor for hearing loss in neonates and infants during and after treatment with clinically-essential medications that are potentially ototoxic, e.g. aminoglycosides. We also discuss the associated risk factors that can potentiate drug-induced hearing loss, and the challenges of implementing effective monitoring protocols for ototoxicity in neonates.

Incidence of dosing with ototoxic medications in neonates

The incidence of hearing loss in infants discharged from the NICU (2–4%) is 10-fold greater relative to neonates in the well mother and baby units (0.1–0.3%; Erenberg et al, 1999). The aetiology of these hearing losses is largely unknown, and one potential risk factor is dosing with aminoglycosides for infection control. Of NICU admissions with confirmed infections, 80% are bacterial (sepsis), a major cause of morbidity and mortality (Edwards & Baker, 2004; Remington, 2011). The majority of the remaining 20% of confirmed infections are viral in origin, for which the bactericidal aminoglycoside antibiotics are ineffective and discontinued in favour of anti-viral medications. The incidence of confirmed or suspected sepsis increases with decreasing gestational age at birth (Camacho-Gonzalez et al, 2013).

Loop diuretics are also associated with hearing loss. Many NICU admissions (∼30%, or ∼120,000 neonates annually) receive loop diuretics, like furosemide and bumetanide, for fluid overload in both acute and chronic disease states, including sepsis, bronchopulmonary dysplasia, persistent pulmonary hypertension, meconium aspiration syndrome, cardiac defects and congenital diaphragmatic hernia (Pacifici, 2012; Laughon et al, 2015). The most commonly-reported side effect of loop diuretic administration in neonates is thrombocytopenia (Laughon et al, 2015); however, lower doses of loop diuretics can induce transient hearing loss (most frequently), or lead to permanent hearing losses in paediatric admissions (Rybak, 1982; Saha et al, 2013). Although the culprit for SNHL in clinical settings is most likely the prolonged duration of aminoglycoside treatment (Fligor et al, 2005; Dennett et al, 2014), it is unclear how other co-administered treatments, such as loop diuretics, glycopeptide antibiotics, neuromuscular blockers, may contribute to the development of drug-induced SNHL (Cheung et al, 1999; Rubin et al, 2002; Robertson et al, 2006; Pacifici, 2012; Laughon et al, 2015), as discussed further below.

Paediatric oncology patients, including neonates, can also be exposed to platinum-based drugs with potent anti-cancer activity, and often substantial ototoxic effects (Knight et al, 2005, 2007; Veal et al, 2015). However, this is a relatively smaller population of ∼10,500 new cancer patients (<14 years old) each year in the U.S., not all of whom receive ototoxic platinum-based drugs (Ward et al, 2014). For those who do, with each additional course of treatment, the risk of ototoxicity continues to increase (Knight et al, 2007; see also Ototoxicity monitoring in adult patients: Service delivery gaps, barriers and solutions in select civilian and government sector programs by Konrad-Martin and colleagues, this issue).

Factors potentiating drug-induced ototoxicity

The side effects of aminoglycosides can be potentiated in specific clinical situations. In preclinical models, the glycopeptide antibiotic vancomycin readily augments gentamicin-induced ototoxicity (Brummett et al, 1990), and is commonly-prescribed in the NICU (Rubin et al, 2002). Administration of vancomycin alone can lead to acute nephrotoxicity in ∼1–9% of neonates receiving this drug, even at currently recommended (yet apparently) sub-therapeutic levels (Lestner et al, 2016). However, there is conflicting evidence for vancomycin-induced ototoxicity in preclinical models or clinical studies in adults, and disease-specific factors and other confounders need to be considered (Lestner et al, 2016). Initial concerns about the ototoxicity of vancomycin were related to impure (initial) formulations and/or concomitant dosing with aminoglycosides.

Neuromuscular blocking agents, such as pancuronium bromide and vecuronium bromide, are frequently used in sick neonates requiring respiratory assistance (via intubation and ventilation) or surgical procedures. Pancuronium bromide is specifically associated with an increased risk of SNHL and both agents can potentiate the risk of aminoglycoside-induced ototoxicity (Cheung et al, 1999; Masumoto et al, 2007).

The ototoxicity of the above-mentioned drugs can be potentiated by inflammatory status, co-administration with other potential ototoxic compounds, and ambient sound levels. The ototoxicity of aminoglycosides and cisplatin (a platinum-based anti-neoplastic [anti-cancer] drug), is potentiated by a host-induced inflammatory response to infection or bacterial immunogens (Koo et al, 2011; Oh et al, 2011; Hirose et al, 2014; Koo et al, 2015). For aminoglycosides, this is a particular concern as these bactericidal antibiotics lyse bacteria releasing more immunogens that then enhance the inflammatory response, potentiating the risk of ototoxicity (Koo et al, 2015). For platinum-based drugs, systemic inflammation can occur when abdominal tumours are irradiated, killing commensal bacteria that then increase vascular uptake of bacterial proteins to trigger an inflammatory response (Paulos et al, 2007).

Other clinical factors that increase the ototoxic potential of drugs have been recognised, particularly for aminoglycosides, including fever (higher-than-normal body temperature), hypoxia, poor renal function, and poor nutritional or low antioxidant status (Hoffman et al, 1988; Manian et al, 1990; Lautermann et al, 1995; Lin et al, 2011). The co-administration of other (less) ototoxic drugs, like vancomycin, loop diuretics, and neuromuscular blockers also significantly potentiate the ototoxicity of aminoglycoside antibiotics (Rybak, 1982, 1988; Bates et al, 2002; Robertson et al, 2006).

In addition to prematurity and ototoxic medications, genetic factors as well as common clinical conditions, such as hyperbilirubinemia or hypoglycaemia, can also increase the risk of hearing loss in NICU admissions (ACMG, 2002; Declau et al, 2008; Smit et al, 2013; Zimmerman & Lahav, 2013; Vedovato et al, 2015; Olds & Oghalai, 2016). Crucially, polymorphisms in the mitochondrial gene MTRNR1 lead to an “idiosyncratic” adverse event of hearing loss with aminoglycoside dosing that is distinct from the ototoxicity associated with supra-therapeutic levels of these drugs. Aminoglycosides bind more avidly to MTRNR1 polymorphism in the ribosomal subunit, impairing mitochondrial protein synthesis, leading to reduced mitochondrial activity that is insufficient for optimal hair cell function (Guan, 2011; Gopel et al, 2014). The American College of Medical Genetics (ACMG) recommends testing for MTRNR1 mutations in patients exposed to aminoglycosides who already have hearing loss (ACMG, 2002). Although this panel notes that avoiding aminoglycosides in polymorphism-positive individuals reduces the risk of developing hearing loss, clinical testing for this maternally-inherited trait in pregnant women is not yet performed in an anticipatory manner (Linden Phillips et al, 2013).

Although less studied, environmental aspects can also potentiate or independently contribute to acquired hearing loss in infants. NICUs typically have ambient sound levels substantially (∼7.5×) greater than those recommended by the American Academy of Paediatrics (AAP) (<45 dB(A) L_eq) and the Journal of Perinatology (<50 dB(A) L_eq) over a 24-h period (American Academy of Pediatrics C.o.E.H, 1997; Graven, 2000; Kramer et al, 2016). Levels of ambient sounds higher than these guidelines cause significant changes in vital signs of patients in the NICU, including drops in oxygen saturation, altered heart rate, blood pressure, and disturbed sleep, as well as reduced healing and growth rates (Zahr & Balian, 1995; Wachman & Lahav, 2011). Sustained levels of ambient sound that can induce temporary threshold shifts in preclinical models can also enhance the cochlear uptake of aminoglycosides and, presumptively, permanent hearing loss (Li et al., 2011, 2015). Researchers are now focussing on demonstrating whether ambient sound levels in the NICU can potentiate the ototoxicity of aminoglycosides (Zimmerman & Lahav, 2013). The ototoxicity of cisplatin can also be potentiated by higher levels of prior sound exposure (Gratton & Kamen, 1990; Gratton et al, 1990).

Approaches to monitoring neonates for ototoxicity

In 1973, the AAP and the Joint Committee on Infant Hearing (JCIH) recommended screening for hearing loss in infants with significant medical or family risk factors for hearing loss, including those in the NICU (JCIH, 1973). In 1990, the JCIH refined their recommendations to specify that screening should occur prior to discharge from the NICU (JCIH, 1990). The most recent JCIH position statement reported that all 50 states and the District of Columbia have set protocols for NBHS in the well baby unit and NICU prior to hospital discharge as part of a mandated law or voluntary state programmes. This position statement was developed to promote earlier detection and intervention for infants with hearing loss and stressed the necessity of separate monitoring protocols for infants with high-risk indicators for hearing loss, such as NICU graduates.

To date, specific hearing screening and monitoring protocols for ototoxicity in neonates and infants have not been developed. The American Academy of Audiology (AAA, 2009) and the American Speech-Language-Hearing Association (ASHA, 1994) each recommend audiologic management of children undergoing ototoxic treatment. However, because of the complex interactions between drugs, dosages, length of treatment and unique disease processes, both stop short of recommending specific protocols.

The criteria for defining ototoxic hearing loss in children have typically relied on behavioural auditory thresholds (ASHA, 1994; Brock et al, 2012), which are not used for NBHS programmes. The main aims of monitoring the hearing of infants (and others) receiving ototoxic drugs are to reveal early indicators of hearing dysfunction for the purpose of adjusting treatments and avoiding further hearing loss. Monitoring the hearing levels of infants at risk also allow clinicians to determine the need for diagnostic testing and, ultimately, audiologic interventions such as hearing aids and/or cochlear implants for habilitative purposes.

Newborn hearing screening protocols typically consist of two physiologic screening tests: an automated click auditory brainstem response (aABR) screen that provides information about neural transmission of acoustic stimuli from the cochlea to upper brainstem, and otoacoustic emissions (OAEs) that are sounds originating from healthy cochlear outer hair cells in response to externally introduced sounds. Both screening procedures are clinically useful for identifying auditory dysfunction when middle ear (ME) integrity is normal. These procedures also provide an objective, non-invasive measure of auditory function without requiring a conscious response from the infant. For screening purposes, the aABR uses computer algorithm analyses of waveform responses to low-intensity broad-band clicks and is recommended in NICUs to screen for some auditory-neural deficits in neonates that are missed by distortion-product (DP)OAE screens of sensory hair cell function (Hall et al, 2004; Shapiro & Popelka, 2011). OAE screening has been shown to provide early indications of ototoxicity (Konrad-Martin et al, 2005). A two-stage NBHS protocol involving both techniques might capture a greater percentage of infants with hearing loss (Johnson et al, 2005) given that each approach (OAE and aABR) targets different types of auditory dysfunction.

All parents of infants who have spent time in the NICU, regardless of their hearing screening result, should be strongly encouraged to follow-up with a diagnostic audiometric evaluation. This follow-up is essential to identify infants with late-onset, progressive, or temporary (e.g. secondary to ME effusion) hearing loss that might not be identified via NBHS. At least one study found that as many as 38% of extremely premature babies had delayed onset or progressive hearing losses (Robertson et al, 2009).

Challenges in monitoring neonates for ototoxicity

Several challenges exist with monitoring the hearing of neonates and young infants for ototoxicity, especially for those graduating from the NICU. First, both the AAA and ASHA protocols recommend obtaining baseline hearing test results prior to the initiation of potentially ototoxic medication. These baseline thresholds can then be compared to subsequent evaluations during and after treatment. Given that approximately 80% of neonates receive prophylactic or empiric treatment with the aminoglycoside antibiotic gentamicin and ampicillin (a β-lactam) upon admission to the NICU, it is virtually impossible to obtain baseline hearing data on these neonates. Even if a prior screening does occur, current NBHS protocols do not typically test for auditory function above 4 kHz, characteristic for initial onset of ototoxicity, and are crucial for the development of speech and language understanding (Knight et al, 2005). The goal of NBHS is to identify newborns that have permanent hearing loss that is mild to moderate in degree or greater in one or both ears in the mid-frequency range (0.5–4 kHz; JCIH, 2007). It is important to note that passing a NBHS with current screening equipment and protocols does not ensure normal hearing across the frequency range (ASHA, 2013). This is especially true of infants with minimal to mild hearing losses or those with isolated frequency ranges of normal to near-normal hearing. Therefore, without a diagnostic auditory brainstem response (ABR) prior to the initiation of ototoxic therapy, it is rarely possible to determine if any identified hearing loss is definitively the result of the therapy or pre-existing secondary to another disease or genetic process. It is critical to differentiate the causes of hearing loss, as ototoxicity may be preventable if an alternative dosing or treatment regimen becomes available.

As noted previously, broad-band clicks utilised by aABR screening are not sensitive to the initial high-frequency (>8 kHz) losses typical of ototoxicity. Because test-retest reliability of OAE levels is generally good, changes resulting from ototoxicity might be differentiated from natural variations (Abdala et al., 2017). DPOAEs and click-evoked (CE) OAEs are often reduced or absent in patients receiving aminoglycosides antibiotics (Hotz et al, 1994; Mulheran & Degg, 1997). Furthermore, DPOAEs might reflect ototoxic effects earlier, probably because they are more effectively measured at high frequencies than CEOAEs (Lonsbury-Martin & Martin, 2001) and, thus, have greater potential for monitoring drug-induced hearing loss. To date, measurement of DPOAEs at ultra-high frequencies (>10 kHz) is not available on standard clinical equipment (Abdala et al, 2017).

Another challenge to ototoxic monitoring using serial OAE or ABR testing is that they are labour-intensive and difficult to interpret in premature infants (<34 weeks gestational age) because of their immature cochleae and auditory pathways (Norton et al, 2000). Given the widespread use of ototoxic medications in the NICU setting, equipment specially designed for screening high frequency regions (>4 kHz) will be required for efficient clinical implementation of ototoxicity monitoring protocols specifically designed for neonates. This would ideally include assessment of ME function tests, as deficits in ME status (e.g. fluid, negative pressure) degrade both the stimulation and auditory response which may trigger false positive for hearing loss (Hunter et al, 2016). Identifying and treating the ME issue before testing is critical to determine if a permanent hearing loss exists, or if further testing is required.

Finally, medically necessary interventions, such as mechanical ventilation, can introduce acoustic or electrical interference that can disrupt aABR screening or diagnostic ABR testing. Although current JCIH recommendations indicate that infants should be screened prior to one month of age, extremely premature infants can reach one month of age before their neural responses have matured to facilitate accurate testing - typically at about 34 weeks corrected gestational age (JCIH, 2007).

Additional challenges to long-term monitoring of ototoxicity in infants include: (i) shortage of audiologists or professionals with the appropriate resources to perform the follow-up tests; (ii) lack of a timely referral for diagnostic follow-up due to the high-risk status of these infants; (iii) lack of well-coordinated follow-up health visits; and (iv) lack of educational and information resources provided to parents and professionals regarding the importance of follow-up, particularly for infants at higher risk of progressive hearing loss due to cytomegaloviral infections, congenital diaphragmatic hernia (Robertson et al, 1998; Dahle et al, 2000) or late-onset hearing loss following completion of treatment with some ototoxic drugs (Kolinsky et al, 2010).

Clinical issues specific to neonates

As with all audiological monitoring, it is important to distinguish between conductive and sensorineural hearing loss (Chang, 2011). Audiological screening methods do not differentiate between different types of hearing losses, therefore we rely on diagnostic evaluations to determine the nature of the hearing loss. For neonates, it is critical these diagnostic evaluations include physiologic assessments (e.g., wideband reflectance, tympanometry, bone conduction via ABRs) to rule out middle ear pathologies.

Some preterm or critically ill neonates spend several months in the NICU and often screening is delayed until prior to discharge. This may increase the risk for poor neurodevelopmental outcome for these highly vulnerable infants because intervention before six months of age are thought to be associated with improved language and communication skills by age two to five years (Yoshinaga-Itano et al, 1998; Sininger et al, 1999; Thompson et al, 2001). Another specific issue for screening neonates in the NICU is the comparatively higher levels of ambient sound levels that will increase the noise floor during DPOAE or aABR screening (Garinis et al, 2017). This could result in inconclusive screens or false referrals requiring re-screening or diagnostic referrals that consume limited resources and unduly stress the parents.

Despite their high exposure rate, some studies report that neonates have reduced rates of aminoglycoside ototoxicity compared to older children or adults (Aust, 2001; Vella-Brincat et al, 2011). These data challenge the current recommendation for follow-up hearing screening at 24–36 months of age for NICU graduates exposed to aminoglycosides if the course is limited to <14 days and serum levels are monitored to be within therapeutic levels (Hess et al, 1998; So, 2009; Fuchs et al, 2016). However, these data are diluted by the large numbers of (asymptomatic) infants who receive aminoglycosides during rule-out sepsis protocols (<3 days) and thus have little risk of ototoxicity, unless carrying genetic polymorphisms that predispose individuals to aminoglycoside-induced ototoxicity.

Individuals with a mitochondrial genetic variant (m.1555A > G in MTRNR1) are more susceptible to rapidly progressive, permanent hearing loss even when aminoglycoside levels are maintained within the therapeutic range (Guan, 2011; Gopel et al, 2014). Special consideration should be given regarding prenatal or postnatal screening to avoid aminoglycoside exposure in patient populations (e.g. Northeast Asia) with a higher prevalence of these polymorphisms.

In Africa, Asia and South America, aminoglycosides are clinically relevant due to their ready availability, cost-effectiveness, and chemical stability at ambient temperatures, especially in rural areas (Forge & Schacht, 2000). NICUs represent the best level of medical care for neonates in any country, and are found mostly in tertiary hospitals, typically in urban areas (Eidelman, 2002). It will be imperative for ototoxicity monitoring to be accomplished in these countries, as established for adults with tuberculosis (Harris et al, 2012a, b).

Increasing microbial resistance to antibiotics, including aminoglycoside antibiotics due to catalytic drug-modifying enzymes, is an urgent concern for the healthcare field (Davies & Wright, 1997; Bryant et al, 2016). Gentamicin has been the treatment of choice for presumed or confirmed early onset sepsis and/or meningitis in preterm and term neonates. A recent study found little microbial resistance to aminoglycosides used for early-onset meningitis (induced by Escherichia coli), while resistance to potentially less nephrotoxic third-generation cephalosporins is emerging (Weissman et al, 2016).

Another special consideration in neonates includes the fact that the majority of drugs currently used in the NICU are “off-label,” without adequate understanding of appropriate dose, safety, or efficacy for this population. The common practice of extrapolating data from studies conducted in adults and older children to neonates is not adequate given the special physiology and rapid developmental changes especially in preterm infants (Wiles et al, 2013). Despite legislative efforts to include children in drug development, only 3.4% of all paediatric studies registered involve neonatal pharmacologic therapeutic trials (US National Institutes of Health, Clinical Trials Registry, www.clinicaltrials.gov, March 12, 2012.). Therefore, it is challenging to develop new protocols with potentially newer and less nephrotoxic drugs given their unknown pharmacokinetics and safety profiles. However, at least one ongoing large clinical trial labelled “Antibiotic Safety in Infants with Complicated Intra-Abdominal Infections” (www.clinicaltrials.gov) evaluates the pharmacokinetics, safety, and efficacy of commonly used drug combinations to establish safe alternatives to potentially prolonged and ototoxic treatments under the Best Pharmaceuticals for Children Act.

Recommendations

Currently, there are no widely-accepted existing protocols for monitoring ototoxic-induced hearing loss in neonates and infants outside of traditional NBHS programmes. Furthermore, the current audiologic screening equipment used for NBHS in nurseries is not suitable for early detection of drug-induced hearing loss. Because of the widespread use of ototoxic drugs in NICUs, it is imperative to establish efficacious OAE and ABR protocols using higher frequency (4–16 kHz) stimulation that better detect onset of ototoxicity in neonates and infants. Higher frequency information is essential for acquisition of listening and spoken language skills, and discrimination of critical environmental cues. Currently, it is rare for infants with hearing loss who do not follow-up for a diagnostic evaluation, or have progressive hearing loss, to be identified until elementary school age. Therefore, it is essential for audiologists to refine current monitoring protocols to encourage diagnostic follow-up. These activities will educate parents of the importance of audiological monitoring of their child’s hearing, as well as ensuring that appropriate clinical equipment is used for broad frequency testing for the thorough assessment of both middle-ear and cochlear function.

The combination of co-drug administration, genetic risk factors, neonatal physiology (e.g. hyperbilirubinemia, inflammation), and/or higher levels of ambient sound in the NICU may synergistically potentiate the ototoxicity of aminoglycosides, even in “stable” preterm and term infants in the NICU. More research is needed to enhance efforts to understand and prevent drug-induced hearing loss in this population. In addition, more funding and a mandate for studies in neonatal drug design will allow development of safe and effective alternatives for ototoxic medications in these vulnerable patients. In the meantime, several steps can be taken to reduce the risk of ototoxicity in the NICU population: (1) minimising the use of (aminoglycoside) antibiotics and other potentially ototoxic drugs through antimicrobial stewardship programmes and protocols restricting use to evidence-based clinical scenarios; (2) monitoring the level of ambient sound in the NICU and quality improvement projects aimed at reduction of ambient sound (and other noxious stimuli); (3) prenatal testing of genetic risk factors for hearing loss; and (4) education of parents and NICU staff on the risk of progressive or late-onset hearing loss, the importance of continued screening, and hearing preservation in neonates.

Declaration of interest

No potential conflict of interest was reported by the authors.

Abbreviations
AAA	=	American Academy of Audiology
aABR	=	automated auditory brainstem response
AAP	=	American Academy of Paediatrics
ABR	=	auditory brainstem response
ACMG	=	American College of Medical Genetics
CF	=	cystic fibrosis
CEOAEs	=	click-evoked otoacoustic emissions
dB(A)	=	A-weighted decibels, accounting for relative loudness to the human ear
DPOAEs	=	distortion-product otoacoustic emissions
JCIH	=	Joint Committee on Infant Hearing
Leq	=	equivalent continuous sound level
ME	=	middle ear
MTRNR1	=	gene for mitochondrially-encoded 12S RNA
NBHS	=	newborn hearing screening
NICU	=	neonatal intensive care unit
OAE	=	otoacoustic emission
SNHL	=	sensorineural hearing loss
US	=	United States of America

Acknowledgements

This research was funded by National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1TR000128 to the Oregon Clinical and Translational Research Institute (PSS, ACG); by National Institute on Deafness and other Communication Disorders R01s DC004555, DC012588 (PSS); the U.S. Department of Education (H325K120305; for AMT), and the U.S. Maternal Child Health Bureau (T73MC00050; for AMT). The content is solely the responsibility of the authors and does not necessarily represent the official views of the above-listed federal agencies.