The association between prenatal nicotine exposure and offspring’s hearing impairment

Abstract

Objective:

To evaluate whether there is an association between in-utero exposure to nicotine and subsequent hearing dysfunction.

Patients and Methods:

Secondary analysis of a multicenter randomized trial to prevent congenital cytomegalovirus (CMV) infection among gravidas with primary CMV infection. Monthly intravenous immunoglobulins hyperimmune globulin therapy did not influence the rate of congenital CMV. Dyads with missing urine, fetal or neonatal demise, infants diagnosed with a major congenital anomaly, congenital CMV infection, or with evidence of middle ear dysfunction were excluded. The primary outcome was neonatal hearing impairment ≥1 ear defined as abnormal distortion product otoacoustic emissions (DPOAE; 1 to 8 kHz) that were measured within 42 days of birth. DPOAEs were interpreted using optimized frequency-specific level criteria. Cotinine was measured via ELISA kits in maternal urine collected at enrollment and in the third trimester (mean gestational age 16.0 and 36.7 weeks, respectively). Blinded personnel ran samples in duplicates. Maternal urine cotinine >5 ng/ml at either time point defined in-utero exposure to nicotine. Multivariable logistic regression included variables associated with the primary outcome and with the exposure (P <0.05) in univariate analysis.

Results:

Of 399 enrolled patients in the original trial, 150 were included in this analysis, of whom 46 (31%) were exposed to nicotine. The primary outcome occurred in 18 (12%) newborns and was higher in nicotine-exposed infants compared with those non-exposed (15.2% vs. 10.6%, OR 1.52, 95% CI 0.55 – 4.20), but the difference was not significantly different (aOR=1.0, 95% CI 0.30–3.31). This association was similar when exposure was stratified as heavy (>100ng/ml, aOR 0.72, 95% CI 0.15–3.51) or mild (5–100ng/ml, aOR 1.28, 95% CI 0.33–4.95). There was no association between nicotine exposure and frequency-specific DPOAE amplitude.

Conclusions:

In a cohort of parturients with primary CMV infection, nicotine exposure was not associated with offspring hearing dysfunction assessed with DPOAEs.

Introduction

Over 10,000 infants are born each year in the United States with bilateral sensorineural hearing loss (SNHL), with 5000 having severe to profound SNHL^1,2 The 0.2% prevalence at birth makes SNHL more common than Down Syndrome and spina bifida³. Genetic causes account for more than 60% of congenital hearing loss, and cytomegalovirus (CMV) is the most common cause of nonhereditary SNHL in children^4,5. Recent evidence suggests an association between prenatal exposure to nicotine from tobacco smoke and hearing impairment in offspring.^6,7 Nicotine exposure affects the fetal outer hair cells (OHCs) either directly via nicotinic acetylcholine receptors, which have been identified in the hair cells, or indirectly through the vasoconstrictive effects of nicotine.^6,7 Otoacoustic emission (OAE) testing is widely utilized in newborn hearing screening programs and as a diagnostic tool. OAEs are sounds measured in the external ear canal that reflect movement of the OHCs in the cochlea.⁶

The majority of insults to the cochlea first affect the OHCs, and OAEs are exquisitely sensitive to even subtle OHC dysfunction. Hypoxic cochlear deficits, for example because of nicotine-induced vasoconstriction, are reflected by reduced OAE amplitude; therefore, OAE abnormalities provide early and compelling evidence of cochlear (OHC) dysfunction.⁸ The overarching hypothesis is that in-utero exposure to nicotine has a detrimental effect on the developing OHCs and cochlea, and leads to decreased OAE amplitude and signal to noise (SNR) responses. Prior studies suggest these effects of prenatal tobacco exposure may be apparent at birth and may be indicative of future hearing impairment^6,9–13.

Investigating the effect of nicotine exposure during pregnancy and offspring hearing and neurodevelopment is important and relevant, as up to 30% of pregnant women are exposed to tobacco smoke.⁷ In addition, while the evidence suggests an association between prenatal exposure to nicotine and hearing impairment in offspring, prior studies are limited by retrospective nature, small sample size, and potential confounding effects of post-natal tobacco exposure. Therefore, we sought to examine the association of nicotine exposure during pregnancy on the offspring’s hearing and neurodevelopment.

Materials and Methods

Study Design

This is a secondary analysis of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network multicenter randomized trial of hyperimmune globulin to prevent congenital cytomegalovirus infection, among gravidas with primary CMV infection. (ClinicalTrials.gov number, NCT01376778) In the original study, patients with evidence of primary CMV before 24 weeks gestation were randomized to monthly infusions of CMV hyperimmune globulin or an identical placebo until delivery. Diagnosis of primary maternal CMV infection was defined as either a positive CMV IgM antibody (≥ 1.00 Index) and low-avidity maternal CMV IgG antibody screen (< 50.0%) or evidence of maternal seroconversion with development of CMV IgG antibody (≥ 6.0 AU/ml) following a prior negative CMV screen (< 6.0 AU/ml). The details of the study, which was conducted at 17 centers in the U.S. between 2012 and 2018, are described elsewhere¹⁴. The primary outcome of the original trial was a composite of confirmed fetal CMV infection or congenital infection diagnosed by three weeks of life, or fetal or neonatal death, and hyperimmune globulin therapy was not found to not influence the primary outcome rate. All variables and data from these studies were collected prospectively by trained research staff following strict and specific protocols outlined in a manual of operations.

For this secondary analysis, we excluded mother-infant dyads with missing urine samples, fetal or neonatal demise, and those with infants diagnosed with a major congenital anomaly, congenital CMV infection or unknown status of CMV congenital infection, or with evidence of middle ear dysfunction. Middle ear dysfunction may result from fluid in the ear canal, common immediately after birth or in the setting of infection, and renders DPOAE testing unreliable with SNR <3dB. In the original trial, urine samples were collected from all participants at randomization and before each monthly infusion, frozen, and stored. Samples collected prior to randomization and samples from the third trimester were used in our analysis.

Study outcomes

For this analysis, the primary outcome was neonatal hearing impairment in at least one ear defined as amplitudes on distortion product otoacoustic emissions (DPOAE) test below frequency-specific cutoff¹⁵ for ≥ 50% of the frequencies tested, assessed at baseline within 42 days of birth or prior to neonatal intensive care unit discharge (Table 1). DPOAE and acoustic immittance testing was performed at the baseline neonatal assessment with audiometers calibrated according to American National Standards Institute (ANSI) standards for insert earphones (2-cc coupler). If hearing loss was suspected, air and bone conduction auditory brainstem response (ABR) testing was also performed to determine sensorineural or conductive hearing loss. For the DPOAE screen, the optimal data collection goal was to obtain DPOAEs at F2 frequencies of 1 kHz, 2 kHz, 3 kHz, 4 kHz, 5 kHz, 6 kHz, and 7 kHz for each ear. DPOAE amplitude and SNR were reported for the OAE at each frequency band, though results at 1.0 kHz were not included in this secondary analysis as they less reliably predicted hearing loss based on prior studies¹⁵. With regards to acoustic immittance, tympanometry was performed using a 1000 Hz probe tone, with the presence and absence of a peak used to define normal or abnormal middle ear function, respectively. Optimal tympanometric and acoustic reflex data collection measurements included ear canal volume, peak-compensated static admittance, middle ear peak pressure, and tympanometric width.

Table 1.

Direct Proportion Otoacoustic Emissions (DPOAE) Level¹⁵

F2 frequency	DPOAE level (dB SPL) cutoff
2 kHz	6.3
3kHz	4.4
4kHz	5.9
5kHz	0.8
6kHz	−1.7
7kHz	−4.2

Adapted from Blankenship et al (2018), Optimizing Clinical Interpretation of Distortion Product Otoacoustic Emissions in Infants Ear Hear; DPOAE levels 5–8kHz were extrapolated from available reported data

Laboratory testing

In-utero tobacco exposure was defined as positive urine cotinine (levels of >5 ng/mL) on maternal urine collected at enrollment (mean GA 16.0 +/− 3.8 weeks) and/or in the third trimester (mean GA 36.7 +/− 1.7 weeks). Cotinine is the principal metabolite of nicotine, and is considered an ideal analyte to validate active and passive cigarette smoking due to its specificity and retention time of 18–20 hours¹⁶. Cotinine levels >100ng/ml strongly suggest regular active smoking, while levels between 10–100ng/ml may result from heavy passive exposure or less frequent active smoking, and levels ≤5ng/ml represent nonsmokers^17,18.

Urine cotinine was measured via Abnova ELISA kits (KA0930, Taiwan) according to the manufacturer’s protocol. The detection limit for the assays is 1 ng/mL, and the assay sensitivity based on the minimum cotinine concentration required to produce three standards deviation from assay Ao is 1 ng/mL. Masked personnel ran all samples in duplicate and the mean values were used in analysis. As per the assay protocol, 10 μL of samples or standards was collected in each well along with 100 μL of enzyme conjugate and incubated at room temperature (20–25°C) in the dark for 60 minutes. The wells were washed 3 times with 300μL of 1X Wash Buffer before inverting wells to empty and ensuring all residual moisture was removed. We then added 100 μL substrate reagent to each well and incubated for an additional 30 minutes in the dark. Finally, we added 100 μL Stop Solution to each well. The absorbance was read at 450 nm within 15 minutes of adding the stopping solution.

Statistical Analysis

Maternal and neonatal continuous variables were compared with the use of the Wilcoxon rank-sum test, and categorical variables with the chi-square or Fisher’s exact test. Baseline characteristics were assessed as potential confounders and included maternal age, race and ethnicity, marital and employment status, primary source of medical payment for prenatal care, maternal body mass index (BMI; kg/m²) at enrollment, alcohol and drug use during pregnancy, family history of childhood hearing loss, treatment assignment, and neonate sex. Variables that were imbalanced at baseline (P<0.05) and associated with the primary outcome (P<0.1) were evaluated in a multivariable logistic model. In pre-specified secondary analyses, a continuous exposure variable and a 3-level categorical exposure variable were analyzed (heavy, mild and non-exposed). In addition, secondary outcomes analyzed included neonatal bilateral hearing impartment and a dichotomous indicator for neonates with any frequency below the DPOAE amplitude cutoffs (Table 1). Similar to the primary analysis, multivariable logistic regression was used in secondary analyses. Statistical analyses were performed using SAS statistical software (SAS Institute, Cary, NC).

Results

From 2012 to 2018, 712 individuals were found to have primary CMV infection, of whom 399 (56%) were enrolled in the parent trial. One hundred and fifty were included in this analysis (Figure 1), of whom 46 (31%) had evidence of nicotine exposure (Table 2). Baseline characteristics differed significantly between tobacco-exposed and non-exposed groups. Pregnant patients with antepartum tobacco exposure had higher BMI, were more likely to be younger, of self-reported Black race, have government insurance or self-pay, have drug use in pregnancy and self-report cigarette smoking, and were less likely to be married or living with a partner (Table 2).

Figure 1.

Flowchart

*Middle ear dysfunction may result from fluid in the ear canal, common immediately after birth or in the setting of infection, and renders DPOAE testing unreliable with SNR <3dB

Table 2.

Maternal and neonatal characteristics of patients

	Exposed (n=46)	Non exposed (n=104)	P value*
Maternal age (years)	25.1 ± 5.2	30.4 ± 5.5	<0.001
Black	16 (34.8)	4 (3.8)	<0.001
Hispanic	6 (13.0)	14 (13.5)	0.94
Married or living with partner	27 (58.7)	91 (87.5)	<0.001
Not employed	17 (37.0)	20 (19.2)	0.02
Government insurance or self-pay	31 (67.4)	16 (15.4)	<0.001
BMI at randomization (kg/m2)	28.2 ± 7.6	25.5 ± 5.2	0.03
Alcohol use in pregnancy	4 (8.7)	11 (10.6)	1.0
Drug use in pregnancy	7 (15.2)	1 (1.0)	0.001
Self-reported cigarette smoking in pregnancy	16 (34.8)	0 (0)	<0.001
Family history of childhood hearing loss	2 (4.3)	4 (3.8)	1.0
Assigned to CMV hyperimmune globulin	23 (50.0)	44 (42.3)	0.38
Infant male sex	25 (54.3)	44 (42.3)	0.17

Abbreviations: BMI, body mass index. Data are reported as mean ± standard deviation or n (%), unless otherwise noted.

^*Based on Wilcoxon rank sum test, chi-square test, or Fisher’s exact test.

Potential confounder variables that differed significantly between the exposed and non-exposed groups (i.e., BMI, maternal age, race, insurance type, and drug use) were further evaluated in univariate models with the primary outcome. Self-reported race was the only variable that was significant at the p<0.1 and was therefore included in adjusted models.

The primary outcome occurred in 18 newborns; the association of the primary outcome with tobacco exposure was non-significant in the unadjusted and adjusted model with 15.2% in tobacco-exposed infants versus 10.6% in non-exposed (OR 1.52, 95% CI 0.55 – 4.20; aOR 1.0, 95% CI 0.30 – 3.31) (Table 3). In secondary analysis this association was also non-significant when exposure was stratified as heavy (>100ng/ml, OR 1.27, 95% CI 0.32–4.97, aOR 0.72, 95% CI 0.15–3.51) or mild (5–100ng/ml, OR 1.78, 95% CI 0.51–6.19, aOR 1.28, 95% CI 0.33–4.95) or when exposure was included as a continuous variable in the model (Table 4). Similarly, there was no association between nicotine exposure and the secondary outcomes of neonatal bilateral hearing loss (OR 0.75, 95% CI 0.08–7.39) and frequency-specific amplitudes (OR 1.32, 95% CI 0.64 – 2.72 ; aOR 1.4 , 95% CI 0.63 – 3.08). (Table 3).

Table 3.

Study outcomes

	Exposed (n=46)	Non exposed (n=104)	Unadjusted OR (95% CI)	Adjusted OR (95% CI)*
Primary outcome†	7 (15.2)	11 (10.6)	1.52 (0.55–4.20)	1.0 (0.30–3.31)
Secondary outcomes
Neonatal bilateral hearing loss	1 (2.2)	3 (2.9)	0.75 (0.08–7.39)	N/A
Any amplitude below frequency∞	18 (39.1)	34 (32.7)	1.32 (0.64–2.72)	1.40 (0.63–3.08)

Data are reported as n (%).

^*Adjusted for race (Black versus other race).

^†Defined as neonatal hearing impairment in at least one ear defined as amplitudes on otoacoustic emissions (DPOAE) test below frequency-specific cutoff and a signal to noise ratio (SNR) ≥3dB for ≥50% of the frequencies tested, assessed within 42 days of birth or at hospital discharge.

^∞Defined as neonatal hearing impairment in at least one ear defined as amplitudes on otoacoustic emissions (DPOAE) test below frequency-specific cutoff and a signal to noise ratio (SNR) ≥3dB for at least 1 of the frequencies tested, assessed within 42 days of birth or at hospital discharge.

Table 4.

Stratified Outcomes

	Primary Outcome†	Adjusted OR (95% CI)*
Heavy (n=26)   Cotinine >100ng/ml	4 (15.4)	0.72 (0.15–3.51)
Mild (n=20)   Cotinine 5–100ng/ml	3 (15.0)	1.28 (0.33–4.95)
Non-exposed (n=104)   Cotinine <5ng/ml	11 (10.6)	Referent group

Data are reported as n (%).

^*Adjusted for race (Black versus other race).

^†Defined as neonatal hearing impairment in at least one ear defined as amplitudes on otoacoustic emissions (DPOAE) test below frequency-specific cutoff and a signal to noise ratio (SNR) ≥3dB for ≥50% of the frequencies tested, assessed within 42 days of birth or at hospital discharge.

Discussion

Our analysis did not demonstrate an association between in-utero nicotine exposure and congenital hearing dysfunction. To our knowledge, our study is the first to evaluate laboratory-proven biochemically-verified tobacco exposure and infant hearing dysfunction, as prior studies relied on maternal recollection or documentation of tobacco use within the medical record. Organogenesis is a critical period for the development of the auditory system and is subject to direct damage by nicotine or other chemicals present in cigarette smoke. Nicotinic acetylcholine receptors (nAChRs) are the principal binding sites of nicotine.^6,7,19 The receptors are present throughout the ascending auditory pathway and expressed on inner and outer hair cells (OHCs) where they promote synaptic maturation and stabilization during early development. Chronic exposure to nicotine, particularly during critical periods of development, activates and subsequently desensitizes nAChRs.⁶ Human and animal studies have found nicotine-induced vasoconstriction and elevated carbon monoxide levels reduced cochlear blood oxygen levels in-utero in a dose-dependent manner, resulting in cochlear hypoxia and damage to offspring’s OHCs.^20,21

Newborn hearing screening is recommended for all infants, and historically risk-based and universal approaches have been employed^22,23. Otoacoustic emission (OAE) testing is widely utilized in newborn hearing screening programs and as a diagnostic tool. OAEs are sounds measured in the external ear canal that reflect movement of the OHCs in the cochlea.⁶ OAEs are produced by the vibrational energy from OHC motility that travels outward from the cochlea through the middle ear, vibrating the tympanic membrane and propagating into the external ear canal.⁸

The impact of toxic substances from tobacco smoke on cochlear OHCs can be measured using OAEs. Durante et al examined the impact of in-utero tobacco exposure on the cochlea in neonates using OAEs and concluded that exposed neonates exhibit lower OAEs amplitude and signal-to-noise ratio (SNR) responses.¹⁰ Furthermore, Korres et al measured OAEs in newborns born to mothers who smoked cigarettes during pregnancy and revealed a significant decline in OAE response amplitude in newborns exposed to high levels of tobacco smoking compared to a control group of neonates with no exposure to tobacco smoke.¹¹

These hearing deficits may contribute to the significant negative association between prenatal tobacco smoke exposure and the infant’s cognitive developmental deficits that persists throughout life. Exposure to tobacco smoke during pregnancy can affect language aptitude and auditory performance through underlying physiologic mechanisms (e.g., OHCs in the ear), thus leading to poorer performance on language-related tasks in the offspring.⁶ Studies have shown that prenatal smoke exposure leads to poorer performance on auditory-related tasks, such as altered cognitive maturity and decreased responsiveness to sound stimuli as early as the neonatal period.^6,12 Furthermore, children of mothers who were passive smokers during their pregnancy have decreased neurodevelopmental abilities compared with children of non-exposed mothers.⁷

There is clear physiologic plausibility that nicotine may cause congenital hearing dysfunction, and data have demonstrated an association between tobacco use during pregnancy and infant and child hearing loss^6,7,10,11.

While the differences in newborn hearing outcomes were not significantly different between neonates with laboratory-proven biochemically-verified nicotine exposure in-utero, our study has several strengths. Maternal use was proven established with urine samples collected at multiple points in the pregnancy and quantified with ELISA testing. This is, to our knowledge, the first study quantifying in-utero nicotine exposure in this manner. Furthermore, inclusion criteria ensured newborn hearing testing was performed within the first 42 days of life, minimizing post-natal exposures which could confound results. We employed strict criteria in the audiology testing results to analyze only interpretable results by excluding those with middle ear dysfunction and applied frequency-specific thresholds with clinical application.

Nevertheless, our study is not without limitations. Due to the study design of a secondary analysis of a large randomized controlled trial, we were unable to expand our sample size beyond that of the original study. As a result, post-analysis power calculations suggest we may be underpowered to detect a difference in the defined primary outcome of infant hearing dysfunction as defined for the primary outcome. Furthermore, a liberal definition of exposure at >5 ng/ml was employed, with only roughly half of exposed dyads falling into a heavy exposure group representing active smokers. Interestingly, the number of individuals in this group is higher than those with self-reported tobacco (n=26 vs n=16), shedding light on the potential inaccuracy of using only self-reported data. It is possible that gestational age timing, nicotine dose, or a combination of the two may be key in which neonates demonstrate hearing loss at birth. As a result, further studies quantifying nicotine exposure earlier and throughout gestation are needed. Finally, while infants with congenital CMV were excluded from this study, inclusion in the parent CMV trial required primary maternal CMV infection during pregnancy, with potential detrimental effects to the placenta resulting in increased risk of adverse perinatal outcomes²⁴. The pathogenesis of hearing loss caused by CMV remains incompletely elucidated, and a large study including pregnant individuals with and without CMV infection may yield more generalizable results.

Conclusion

We did not demonstrate that prenatal nicotine exposure is associated with hearing dysfunction assessed with DPOAE in newborns exposed to maternal CMV infection. A follow up analysis of infant hearing testing in this cohort at age 2 years of life is planned. Nevertheless, additional large studies with long-term follow up are needed.

Key points:

• Nicotine exposure was quantified from maternal urine

• Nicotine exposure ientified in 30% of the cohort

• Neonatal hearing impairment assessed within 42 days

• Exposure was not associated with offspring hearing dysfunction