Genetic testing for congenital bilateral hearing loss in the context of targeted cytomegalovirus screening

Abstract

Objective:

To determine the prevalence of children with genetic hearing loss who are CMV positive at birth and the relative proportion of genetic and CMV etiology among children with congenital bilateral hearing loss.

Methods:

We performed a review of clinical test results for patients undergoing comprehensive genetic testing for all known hearing loss-associated genes from January 2012 to January 2019. This population was reviewed for reported CMV status and genetic causes of congenital bilateral hearing loss.

Results:

In the OtoSCOPE® database 61/4282 patients were found to have a documented CMV status and 661/4282 had documented bilateral congenital hearing loss. Two patients were identified who had both a positive CMV result and a genetic cause for their hearing loss. Forty-eight percent of patients with bilateral congenital hearing loss (320/661) were found to have a genetic etiology. In 62% (198/320), the hearing loss was associated with pathogenic variants in GJB2, STRC, SLC26A4 or an Usher syndrome-associated gene.

Conclusions:

We estimate that ~2% of CMV-positive newborns with hearing loss have a known genetic variant as a cause. The sub-cohort of CMV-positive newborns with symmetric mild-to-moderate bilateral hearing loss will have at least a 7% chance of having pathogenic gene variants associated with hearing loss. In a CMV-positive neonate who failed their newborn hearing screen bilaterally, genetic screening needs to be considered for accurate diagnosis and possible deferment of antiviral treatment.

Level of Evidence:

4

1. INTRODUCTION

Newborn hearing screening in developed countries is ubiquitous, with recent statistics showing that 98.2% of newborns in the United States undergo screening.¹ There remains however, variability in practice on subsequent diagnostic workup following confirmed congenital hearing loss.

1.1. CMV screening for congenital hearing loss

In developed countries, congenital cytomegalovirus (cCMV) is the most common intrauterine infection diagnosed at birth. Congenital CMV incidence is ~0.6% (1 in 200) of all live births and its presentation is highly variable.² Approximately 10% of newborns are symptomatic at birth.⁵ With some variation, hearing loss has not been included in the definition of a symptomatic cCMV infection and consequently asymptomatic cCMV infection with isolated hearing loss is a distinct diagnostic category. Approximately 10% of asymptomatic cCMV infection cases present with hearing loss at birth as the only symptom.^3–5 A portion of asymptomatic children without hearing loss will subsequently develop a progressive fluctuating loss. ^3–5

Several guidelines and treatment algorithms now include recommendations for cCMV testing if congenital sensorineural hearing loss (SNHL) is detected on a newborn hearing screen (NBHS).^6,7 Screening for asymptomatic cCMV infection is time sensitive for two reasons: 1) if screening does not take place within the first 14-21 days of life⁸ a positive result cannot be unequivocally distinguished from acquired CMV, and 2) in trials validating anti-viral medication all treatment began before 30 days of life.^9,10 Hearing-targeted cCMV screening is being progressively implemented.¹¹ In the United States, four states currently have legislation mandating CMV screening after failed NBHS, with a fifth directing that the test be offered. In addition, many birth hospitals have enacted their own mandatory protocols.¹²

In two large randomized controlled trials (RCT), antivirals have shown benefit in preventing progression of hearing loss and neurodevelopment in symptomatic cCMV infections.^9,10 A subsequent retrospective review showed antiviral treatment prevented the progression of hearing loss in newborns with an asymptomatic cCMV infection-related hearing loss and a prospective trial to validate these results is currently ongoing.^{13, 14} The two large RCT have also shown that at least 1/5 of patients have some degree of treatment-related neutropenia.^9,10 Consensus recommendations from 2017 advise treatment for cCMV only in those newborns with moderate to severe systemic symptoms.¹⁵ Nevertheless, based on developing evidence, treatment with valgancyclovir will likely be offered to many neonates with an asymptomatic cCMV infection.

1.2. Genetic screening for congenital hearing loss

The role of genetic testing in neonates born with hearing loss is more ambiguous. There is consensus that unless syndromic hearing loss is suspected, obtaining genetic testing in unilateral congenital hearing loss has limited utility.^6,7 The International Pediatric Otolaryngology Group (IPOG) and American College of Medical Genetics and Genomics recommend that the implementation of genetic testing for congenital hearing loss be guided by audiologic phenotype, ethnicity and clinical exam.^7,16 Some clinicians prefer initial GJB2 testing prior to comprehensive genetic testing (CGT).⁷ Directed gene testing at the gene and/or mutation level, however, is challenging secondary to the heterogeneity of SNHL and the variability in pathogenic variant frequencies according to ethnicity.¹⁷

None of the recommendations addresses the role of genetic testing in a neonate who has tested positive for cCMV. A positive cCMV test may dissuade further etiologic investigation. Discontinuing diagnostic workup based on a positive CMV test risks missing syndromic and non-syndromic genetic causes of hearing loss. If a child who is positive for cCMV has a genetic cause for hearing loss, antiviral treatment may not provide a significant improvement in hearing while introducing unnecessary risk.

2. MATERIALS AND METHODS

A retrospective review was conducted on patients referred to the Molecular Otolaryngology and Renal Research Laboratories (MORL) for CGT between January 2012 to January 2019 to identify CMV status at the time of genetic testing. Genetic screening was completed using OtoSCOPE®, a custom panel that uses targeted genomic enrichment with massively parallel sequencing to simultaneously screen all genes associated with non-syndromic SNHL and non-syndromic mimics.¹⁸ CMV status was defined based on a positive or negative result on a review of the clinical data recorded on the genetic testing requisition forms. Only saliva culture and saliva/blood/urine PCR taken within the first three weeks of life and/or dried blood spot PCR results were included. The Institutional Review Board of the University of Iowa approved this study and the described research was performed in accordance with the Declaration of Helsinki.

2.2. Library preparation and variant interpretation:

Library preparation and variant interpretation were performed as described in Sloan-Heggen (2016), with the additional inclusion of OtoSCOPE® versions 4-8.¹⁹ These platforms differ in the number of deafness-associated genes that are included, which ranged from 66 to 152 reflecting the discovery of additional relevant genes over time. Variant interpretations were completed on a patient-by-patient basis by a multidisciplinary group by correlating phenotypic and genotypic data as we have described.¹⁹

3. RESULTS

3.1. Congenital CMV

Of 4282 patients who underwent CGT 61 patients were identified with a reported congenital CMV status. In 14 cases the CMV result was positive and consistent with an asymptomatic cCMV infection and all had hearing loss. Twelve of these patients had available audiometric data and in two patients a genetic cause for hearing loss was identified. One patient had GJB2-related hearing loss and another had CDH23-related hearing loss. These results are summarized in Tables 1 and 2.

Table 1.

Patients with CMV data submitted to OtoSCOPE®

	Genetic etiology	No known genetic etiology
CMV Positive	2	12
CMV Negative	24	23

Table 2.

OtoSCOPE® patients with congenital SNHL and +CMV

Patient	Symmetry	Severity	OtoSCOPE® result
1	NO	R-severe	Negative
2	NO	R-mod to severe, L-profound	Negative
3	YES	B/L-severe to profound	Negative
4	YES	B/L-severe to profound	Negative
5	NO	R-mod to severe	Negative
6	YES	B/L-severe	Negative
7	NO	R-moderate, L-mild	Negative
8	NO	R-mild to moderate, L-profound	Negative
9	NO	R-profound, L-severe	Negative
10	YES	B/L mod to severe	Negative
11	YES	B/L severe to profound	GJB2 +
12	YES	B/L severe to profound	CDH23 +
13	unknown	Unknown	Negative
14	unknown	Unknown	Negative

3.2. Cases of positive CMV with genetic variants associated with hearing loss

Case #1:

This male child born at term from an uncomplicated pregnancy failed his NBHS on the right but passed on the left. Repeat testing showing bilateral hearing loss. cCMV was positive by saliva PCR and a diagnosis of an asymptomatic cCMV infection was made. Antiviral treatment was started with valgancyclovir at 16mg/kg/dose. An otolaryngology follow-up with sedated ABR showed bilateral symmetric severe SNHL and genetic testing via massive parallel sequencing was obtained which identified homozygosity for NM_004004:c.35delG,p.Gly12ValfsTer2 in GJB2. This was confirmed with Sanger sequencing. This pathogenic variant (commonly called the 35delG variant) has an approximately three percent carrier rate in European-American population and is a common cause of congenital severe-to-profound autosomal recessive hearing loss.²⁰ Continued antiviral therapy was deemed unlikely to provide any hearing benefit for this patient and was discontinued.

Case #2:

This female child born at term from an uncomplicated pregnancy failed her initial newborn hearing screen and underwent an ABR that showed bilateral severe-to-profound sensorineural hearing loss. Dried blood spot testing for cCMV was positive but treatment was deferred based on severity of hearing loss and a family history positive for severe-to-profound hearing loss in an older sister. CGT identified two pathogenic missense variants in CDH23 (NM_022124.5:c.4562A>G,p.Asn1521Ser and NM_022124.5:c.2591G>A,p.Gly864Asp). The first variant, NM_022124.5:c.4562A>G,p.Asn1521Ser, was manually curated as pathogenic based on data from an Iranian family. ²¹ The second variant, NM_022124.5:c.2591G>A,p.Gly864Asp, was assigned as a variant of unknown significance based on American College of Medical Genetics and Genomics criteria. Segregation analysis has not yet been completed. It was therefore rated as PM2 (absent in databases indicating extremely low frequency), PP3 (three of four in cilico algorithms predicted pathogenicity) and PP4 (multiple family members with congenital hearing loss).²² CDH23 is implicated in both DFNB12 and USH1D. Based on developmental motor milestones, which were normal in both the patient and her older sister, a diagnosis of DFNB12 hearing loss was made. Close observational follow up will be necessary to monitor for an atypical Usher presentation.

3.3. Common genetic causes of congenital bilateral SNHL

Congenital bilateral symmetric hearing loss was documented in 661 of 4282 persons who underwent CGT using the OtoSCOPE® platform. In 320 of these 661 patients (48%), a genetic etiology of was identified. Commonly implicated genes include GJB2, Usher-related genes (CDH23, USH2A, USH1C, USH1G, MYO7A, ADGRV1, GPR98, and PCDH15), STRC and SLC26A4. These results are summarized in Table 3.

Table 3.

Genetic causes of bilateral congenital hearing loss in OtoSCOPE® database

Hearing loss cause	# of patients	Percentage of pts
All genetic mutations	320/661	48%
GJB2/GJB6 mutations	75/661	11%
Usher-related mutations	63/661	10%
STRC mutations	48/661	7%
SLC26A4 mutations	17/661	3%
Other genetic mutations	122/661	18%

4. DISCUSSION

4.1. Relative prevalence of bilateral congenital hearing loss

The proportion of congenital hearing loss attributed to an asymptomatic cCMV infection is often cited between 20-25%.²³ Due to the known variability in cCMV-related hearing loss, a large portion of these patients present with unilateral or asymmetric hearing loss. One meta-analysis estimated that bilateral hearing loss from asymptomatic cCMV infections accounts for only ~2/10,000 births.³ This estimate includes those children with delayed hearing loss (i.e., hearing loss not found on the NBHS). Bilateral congenital hearing loss occurs in ~1/500 births, ²³ suggesting that asymptomatic cCMV infections account for ~10% of this population. In comparison, genetic etiologies account for at least 50% of congenital symmetric hearing loss.^24,25 As a result, screening for genetic hearing loss has a significantly higher diagnostic yield than screening for cCMV.

This estimated relative prevalence is consistent with studies that have screened for both cCMV and genetic causes of hearing loss. Furutate et al (2011) tested 46 children with bilateral hearing loss for cCMV and GJB2/SLC26A4 on dried blood spots (DBS) collected at birth and found 8% positive for cCMV and 22% with a genetic cause for hearing loss. If restricted to children with bilateral severe-to-profound SNHL these rates are 14% and 32% respectively. Of the 4 children with positive cCMV and bilateral SNHL all were late-onset hearing loss. ²⁶ Meyer et al (2017) screened 57 children for DBS cCMV with 15 positive results.²⁷ A chart review of this population found 14 patients with a genetic cause for hearing loss. The largest study by Dahl et al (2013) used DBS to test 364 patients for both cCMV and GJB2/SLC26A4/MT-RNR1, and found 28 patients with cCMV and 56 with genetic hearing loss. One patient positive for cCMV in this population also carried two SLC26A4 pathogenic variants.²⁸

The availability and urgency of treatment underlines the importance of incorporating cCMV screening early in the newborn period. In comparison, the early implementation of genetic testing offers no immediate benefits when incorporated in the newborn hearing screen. However, as noted above, genetic testing has both a higher diagnostic yield and may inform decisions to use anti-viral treatment for cCMV positive children. Added benefits of establishing a genetic diagnosis include the ability to provide prognostic information regarding hearing loss progression, the opportunity for meaningful parental genetic counselling, and the identification of associated medical concerns that may require intervention.

4.2. Concurrent cCMV and genetic hearing loss

We identified two patients illustrating the co-occurrence of cCMV and genetic hearing loss. In Case #1, genetic testing allowed the cessation of treatment for cCMV thereby avoiding potential side effects and provided an opportunity for parental genetic counselling. This is an important benefit as the recurrence chance for these parents to have another deaf child is 25%. In comparison, we could find no documented reports of multiple sequential children born with cCMV. Mothers with known seroprevalence would have a lower chance of having a child with cCMV-related hearing loss than the general population.¹⁵

For Case #2, recessive mutations in CDH23 were diagnostic of non-syndromic hearing loss at the DFNB12 locus based on a review of the child’s phenotype, which included normal developmental motor milestones. Variants in CDH23 also cause Usher syndrome type 1D.²⁹ There are no gene-specific treatments for USH1D, however gene therapy trials have started for USH2a and USH1B involving retinal injections.^30,31 Nonspecific treatment includes the use of sunglasses to reduce UV, lutein, omega-3 fatty acids, and vitamin A.³²

Based on relative prevalence data, the proportion of patients who are cCMV positive and have genetic hearing loss can be estimated. Patients with genetic hearing loss (at least 50% of all congenital bilateral hearing loss^24,25) have the same prevalence of asymptomatic cCMV infection (0.52% of all births³) as the general population. Thus 0.26% of patients with congenital genetic hearing loss will have a positive cCMV test (approximately 2.3% of all bilateral cCMV hearing loss).

Creating a cCMV audiometric profile assists in identifying patients with cCMV who have a higher likelihood of genetic hearing loss however the scarcity of reported NBHS results in cCMV-positive newborns and the variability of cCMV hearing loss are limited. Symmetry is important because genetic non-syndromic SNHL is bilaterally symmetric whereas infection-related hearing loss is typically asymmetric. In Case #1, for example, the exact symmetry of the ABR result raised suspicion for a genetic cause. In addended data to Dahl et al (2013), the highest threshold for both ears for 30 patients with bilateral hearing loss was provided. Using a difference of <20dB between ears to establish symmetry, only 19 of 30 (~60%) children had symmetric hearing loss.²⁸ Using this loose definition of symmetry doubles the proportion of genetic loss in cCMV-positive neonates to 4%. It is likely that ABR data at multiple contiguous frequencies would reveal that only a small percentage of patients with CMV hearing loss have actual symmetric hearing loss.

The likely severity of cCMV hearing loss can also be estimated. Three studies with at least five patients that included CMV status, bilaterality and severity were found. The highest thresholds for the worst hearing ear were used to create Table 4. In 78% of patients (54/69) hearing loss was moderately severe or worse, suggesting that if cCMV involves both ears, it is typically severe enough to cause more significant damage and inflammation. It is also possible however, that these data are biased as mild hearing loss can be missed on the NBHS. In an analysis of all-cause genetic testing, ~43% (306/705) of genetic loss was mild-to-moderate. Using this population we calculated that ~7% of asymptomatic cCMV newborns with bilateral symmetric mild-to-moderate hearing loss will have genetic hearing loss.¹⁹ Copy-number variants in the STRC gene are the most common genetic cause of mild-to-moderate hearing loss.¹⁹ This type of genetic change cannot be detected by Sanger sequencing and requires the incorporation of read-depth analysis in the CGS bioinformatics pipeline.

Table 4.

Degree of bilateral hearing loss in congenital CMV

	Mild	Moderate	Mod-severe	Severe	Profound	Total
Dahl 2013 26	3	4	2	6	15	30
Madden 2005 33	2	1	2	3	8	16
Royackers 2011 34	2	2	0	0	7	11
OtoSCOPE® database	0	1	1	1	9	12
Total	7	8	5	10	39	69

4.3. Study Limitations

There are several limitations to this study including variable reporting in the literature of clinical details such as laterality, audiogram data, CMV status and onset of hearing loss. There is also a likely propensity to not refer children with known cCMV for genetic testing.

5. CONCLUSION

Screening for asymptomatic cCMV infections is important as a positive test result offers an opportunity for intervention with valgancyclovir, which may mitigate the hearing loss. However, neonates who fail their NBHS and have bilateral hearing loss are more likely to have genetic as opposed to cCMV-related hearing loss. Even in the context of a positive cCMV result, symmetric SNHL can still have a genetic etiology. Offering CGT to babies with bilateral symmetric SNHL should be considered, regardless of cCMV status, in order to establish an appropriate diagnosis so that informed treatment decisions can be made. Studies assessing the characteristic of asymptomatic cCMV infection related hearing loss, such as bilaterality and symmetry are needed, as are longitudinal studies of neonates who undergo cCMV and CGT.