Abstract
Background:
Few studies have evaluated ototoxic effects of in utero antiretroviral (ARV) exposure on hearing in children HIV-exposed uninfected (CHEU). Concerns have been identified for some ARVs: tenofovir disoproxil fumarate (TDF), emtricitabine (FTC), lamivudine (3TC), zidovudine (ZDV), and atazanavir (ATV). The aims are to describe hearing outcomes in 5-year-old CHEU in the Surveillance Monitoring for ART Toxicities study and evaluate their association with in utero ARV exposures.
Methods:
Hearing evaluations including pure tone thresholds and distortion product otoacoustic emissions (DPOAEs) were completed in CHEU at 5 years. Log-binomial regression models were fit to assess associations of maternal ARV exposures with risk of sensorineural hearing loss (SNHL) and incomplete DPOAE responses.
Results:
Among 1078 CHEU, 13% had SNHL. Among those exposed <13 weeks gestation, there was 28% higher risk (95% CI: −37%, 158%) of SNHL with TDF/FTC with ATV and 38% lower risk (95% CI: −68%, 19%) with ZDV/3TC without ATV than for regimens containing TDF/FTC without ATV. There was a higher risk of incomplete DPOAE responses for ZDV/3TC without ATV and lower risk for TDF/FTC with ATV. For CHEU with first ARV exposure at 13-26 weeks gestation, those exposed to TDF/FTC with ATV or ZDV/3TC without ATV had lower risk for SNHL compared to TDF/FTC without ATV. Despite nontrivial relative risks, all confidence intervals were wide.
Conclusion:
While SNHL was relatively common, there were no consistent associations between in utero ARV exposure and SNHL or incomplete DPOAEs. Further research is needed on ARV exposures and other hearing measures in CHEU.
Keywords: HIV, children, hearing, in utero, antiretroviral
INTRODUCTION
Few studies have examined potential ototoxic effects of antiretroviral (ARV) medications in people living with HIV. Most prior work found no association between specific ARV classes and hearing loss. In adults, case reports suggest ototoxicity from nucleoside reverse transcriptase inhibitors (NRTIs)1,2, but a larger study observed no link between years on NRTIs, non-NRTIs (NNRTIs), or protease inhibitors (PIs) and poorer hearing once demographics and noise exposure were considered3. Among children living with HIV, combination antiretroviral therapy (cART) has not been associated with hearing loss4,5.
Animal models, however, indicate the possibility of ARV-related ototoxicity, specifically affecting auditory cells6 and function7. Fourteen ARVs from multiple classes have demonstrated dose-dependent cytotoxicity in mouse auditory cells6. Five commonly-used ARV combinations, abacavir/lamivudine (ABC/3TC), abacavir/lamivudine/zidovudine (ABC/3TC/ZDV), efavirenz/lamivudine/tenofovir disoproxil fumarate (EFV/3TC/TDF), zidovudine/lamivudine (ZDV/3TC), and tenofovir disoproxil fumarate/emtricitabine (TDF/FTC) were cytotoxic to an auditory cell line (HEI-OC1)6, likely due to exposure to ABC, 3TC, EFV, ZDV, or TDF6. Although not cochlear cells, these auditory cells express prestin8, the motor protein of cochlear outer hair cells9. Mouse offspring exposed in utero to ZDV and/or EFV had higher (poorer) auditory brainstem response (ABR) thresholds, while distortion product otoacoustic emission (DPOAE, a measure of outer cell function10) thresholds were unchanged7, suggesting possible damage to inner hair cells or auditory nerve pathways.
Data on in utero ARV exposure and hearing among children who are perinatally HIV-exposed but uninfected (CHEU) are limited. In the Surveillance Monitoring for ART Toxicities (SMARTT) study conducted by the Pediatric HIV/AIDS Cohort Study (PHACS) network, 3.1% of 1,435 infants did not pass newborn hearing screening11. Exposure to atazanavir (ATV), especially late in the third trimester, was associated with higher odds of referral, suggesting that timing of maternal ART use may be important. No prior studies have assessed potential ototoxic effects of in utero ARV exposure in school-aged CHEU.
The aims of the study were: (1) to describe pure-tone thresholds and underlying outer hair cell function via DPOAEs in 5-year-old CHEU from the PHACS SMARTT study; and (2) to evaluate associations between in utero ARV exposure and both pure-tone thresholds and DPOAEs, considering the timing of earliest ARV exposure relative to fetal auditory development.
METHODS
Study Participants
Hearing data were collected around age 5 years from CHEU in the PHACS SMARTT study, which enrolled pregnant mother/infant pairs between 2007 and 2025 across 22 U.S. sites including Puerto Rico12,13. Pregnant persons with HIV were enrolled between 13 weeks’ gestation and one week postpartum. Children were followed annually to age 5, then every other year to age 17. The protocol was approved by site Institutional Review Boards and the Harvard T.H. Chan School of Public Health; pregnant participants provided written consent for themselves and their infants.
Hearing Assessments
Beginning August 2010, the 5-year study visit included otoscopy, tympanometry, and air-conduction pure-tone thresholds (at octave frequencies from 0.25 to 8 kHz)14. Bone-conduction thresholds (0.5, 1, 2, 4 kHz) were obtained if air-conduction thresholds were elevated. DPOAEs were added in July 201514. DPOAEs rely on the middle ear system to be functioning (e.g., Type A tympanogram) for accurate measurement. DPOAEs were collected using two frequencies f1 and f2 (f2>f1) with f2/f1=1.2, 65 decibels of sound pressure level (dB SPL) (f1) and 55 dB SPL (f2), and f2 at 1, 2, 3, and 4 kHz. Certified audiologists tested children seated in sound-treated rooms.
Hearing Outcome Measures
Pure-tone averages (PTA) at 0.5, 1, 2, and 4 kHz were calculated for each ear. Normal hearing was defined as PTA ≤15 dB and hearing loss was defined as worse-ear PTA >15 dB. Sensorineural hearing loss (SNHL) was defined as air- and bone-conduction thresholds >15 dB; conductive hearing loss was defined as air-conduction thresholds >15 dB with bone-conduction ≤15 dB and an air–bone gap >15 dB. DPOAE signal-to-noise ratios (SNRs) were categorized as no response (<3 dB), intermediate response (3 dB to <6 dB), or complete (≥6 dB) response.
In Utero ART Regimens
Three regimens—selected for prevalence and potential ototoxicity6,7—were evaluated. Due to small numbers, EFV-containing regimens were excluded. Regimens considered were:
ZDV/3TC/(-ATV): with or without ABC, no ATV; may include other PIs or an NNRTI;
TDF/FTC/(-ATV): no ATV; may include other PI, integrase strand transfer inhibitor (INSTI), or NNRTI;
TDF/FTC/ATV: includes ATV.
Analyses were stratified by gestational age (GA) at earliest in utero ART exposure: (1) at conception (GA=0); (2) embryonic period (GA <13 weeks); (3) second trimester (13≤GA<26 weeks); or (4) third trimester (≥26 weeks). Between 0-13 weeks GA, the development of the cochlea, the cochlear nerve, and afferent pathways occurs while between 13-26 weeks, there is rapid maturation of the cochlea and the cochlear nerve15.
Otitis Media and Other Ototoxic Medications
Because otitis media can increase conductive hearing loss risk, children with ≥4 episodes >60 days apart or at least one diagnosis of otitis media with effusion were defined as having a significant history of otitis media. This variable was not included in regression models because it may lie on the causal pathway between in utero ARV exposure and hearing outcomes. Exposures before age 5 to potentially ototoxic medications (aminoglycosides—gentamicin, amikacin, tobramycin; macrolides—clarithromycin, azithromycin; and vancomycin) were summarized.
Statistical Analyses
Descriptive statistics were used to summarize PTAs and DPOAE SNRs by timing of ARV initiation and regimen. The proportion with worse-ear PTA >15 dB was calculated by hearing loss type. Log-binomial regression models implemented with generalized estimating equations (GEE) were used to assess associations between regimens and SNHL among children with Type A tympanograms, both unadjusted and adjusted for child characteristics (sex at birth, number of ototoxic medication exposures) and maternal characteristics (age at delivery, education, household income, first-trimester substance use, and first available prenatal HIV viral load and CD4 before ART). Analyses accounted for clustering by site and were stratified by timing of exposure; with GA=0 and 0 to <13 combined in primary analyses and with sensitivity analyses restricted to GA=0. CHEU whose regimen changed later in gestation were considered to have inconsistent ART exposures and were excluded. For DPOAEs, log-binomial GEE models were fit to evaluate risk of less-than-complete response (<6 dB) across four frequencies and both ears, unadjusted and adjusted as above, with interaction terms permitting frequency-specific regimen effects.
RESULTS
Study Population
Among 1078 CHEU with pure-tone data, mean age at their hearing assessment was 5.2 years (SD=0.3); 49% were female and 64% Black (Table 1). Overall, 6% had a significant otitis media history (although 56% lacked this history information); 28% were exposed to at least one potentially ototoxic medication before testing.
Table 1.
Demographic and maternal characteristics for 1078 SMARTT CHEU with valid pure-tone data.
| Child Characteristics | Missing (n=28) |
Timing of earliest ART exposure in gestation | Total (n=1078) |
|||
|---|---|---|---|---|---|---|
| Preconception (GA=0) (n=440) |
0<GA<13 (n=154) |
13≤GA<26 (n=381) |
GA≥26 (n=75) |
|||
| Gestational age (weeks) at birth | ||||||
| <34 | 2 (7%) | 33 (7%) | 9 (6%) | 14 (4%) | 1 (1%) | 59 (5%) |
| ≥34 - <37 | 4 (14%) | 76 (17%) | 30 (19%) | 56 (15%) | 3 (4%) | 169 (16%) |
| ≥37 | 22 (79%) | 329 (75%) | 115 (75%) | 310 (81%) | 70 (93%) | 846 (78%) |
| Age at audiometric exam (yrs), mean (SD) | 5.2 (0.3) | 5.2 (0.3) | 5.2 (0.2) | 5.2 (0.2) | 5.2 (0.3) | 5.2 (0.3) |
| Female sex at birth, n (%) | 15 (54%) | 211 (48%) | 70 (45%) | 196 (51%) | 38 (51%) | 530 (49%) |
| Black race, n (%) | 22 (79%) | 269 (61%) | 100 (65%) | 250 (66%) | 53 (71%) | 694 (64%) |
| Hispanic ethnicity, n (%) | 7 (25%) | 168 (38%) | 50 (32%) | 131 (34%) | 23 (31%) | 379 (35%) |
| Considerable history of otitis media | ||||||
| Yes, n (%) | 1 (4%) | 35 (8%) | 11 (7%) | 20 (5%) | 3 (4%) | 70 (6%) |
| No record or onset date, n (%) | 16 (57%) | 239 (54%) | 90 (58%) | 219 (57%) | 35 (47%) | 599 (56%) |
| Exposures to ototoxic medications, n (%) | ||||||
| 0 | 23 (82%) | 308 (70%) | 106 (69%) | 284 (75%) | 56 (75%) | 777 (72%) |
| 1 | 5 (18%) | 77 (18%) | 39 (25%) | 71 (19%) | 15 (20%) | 207 (19%) |
| >1 | 0 (0%) | 55 (13%) | 9 (6%) | 26 (7%) | 4 (5%) | 94 (9%) |
| Maternal Characteristics | ||||||
| Age at delivery (yrs), mean (SD) | 27.7 (6.3) | 29.9 (6.1) | 29.1 (5.9) | 28.5 (6.0) | 27.3 (6.2) | 29.1 (6.1) |
| Education completed, n (%) | ||||||
| High school or less | 21 (75%) | 261 (59%) | 102 (65%) | 231 (60%) | 90 (68%) | 666 (62%) |
| GED or more | 7 (25%) | 178 (40%) | 51 (33%) | 147 (39%) | 23 (30%) | 405 (37%) |
| Family income <20K/yr, n (%) | 20 (71%) | 304 (69%) | 116 (75%) | 252 (66%) | 57 (76%) | 749 (69%) |
| First trimester substance use, n (%) | ||||||
| Tobacco | 5 (18%) | 54 (12%) | 30 (19%) | 70 (18%) | 11 (15%) | 170 (16%) |
| Alcohol | 2 (7%) | 24 (5%) | 15 (10%) | 22 (6%) | 5 (7%) | 68 (6%) |
| Marijuana | 2 (7%) | 23 (5%) | 14 (9%) | 28 (7%) | 3 (4%) | 70 (6%) |
| Illicit drugs | 2 (7%) | 24 (5%) | 17 (11%) | 30 (8%) | 5 (7%) | 78 (7%) |
| First CD4 count during pregnancy*, n (%) | ||||||
| <200 cells/mm3 | 2 (7%) | 48 (11%) | 19 (12%) | 43 (11%) | 7 (9%) | 119 (11%) |
| 200-350 cells/mm3 | 6 (21%) | 55 (13%) | 21 (14%) | 89 (23%) | 17 (23%) | 188 (17%) |
| >350 cells/mm3 | 15 (54%) | 326 (74%) | 53 (34%) | 211 (55%) | 44 (59%) | 649 (60%) |
| First HIV RNA level during pregnancy*, n (%) | ||||||
| <400 copies/mL | 9 (32%) | 324 (74%) | 26 (17%) | 53 (14%) | 19 (25%) | 431 (40%) |
| ≥400 copies/mL | 15 (54%) | 111 (25%) | 73 (47%) | 288 (76%) | 49 (65%) | 536 (50%) |
First available CD4 and viral load for all mothers in the study are summarized. In regression analyses, if the data were collected after ART started, they were considered missing.
ART Exposures and Pure-Tone Outcomes
Hearing sensitivity is summarized in Table 2. Better- and worse-ear PTAs were similar by timing of earliest ART exposure. Overall, 15% had worse-ear PTA >15 dB and ranged between 10% and 17% across different GA at ART initiation groups. Much of the worse ear hearing loss type was sensorineural (13%; 139 of 1075), which is permanent, and 2% was conductive, which is treatable (Table 2). CHEU with significant otitis media history had more hearing loss (21.4%, 15/70) than those without such history (16.6%, 68/409).
Table 2.
Hearing sensitivity characteristics for 1078 CHEU in SMARTT with valid pure-tone threshold data for at least one ear.
| Hearing Characteristics | Missing (n=28) |
Timing of earliest ART exposure in gestation | Total (n=1078) |
|||
|---|---|---|---|---|---|---|
| Preconception (GA=0) (n=440) |
0<GA<13 (n=154) |
13≤GA<26 (n=381) |
GA≥26 (n=75) |
|||
| Better ear PTA (dB), median (IQR) | 9.4 (6.3, 11.9) | 10.0 (6.3, 12.5) | 8.8 (6.3, 12.5) | 10.0 (6.3, 12.5) | 8.8 (6.3, 12.5) | 10.0 (6.3, 12.5) |
| Min, Max | 1.3, 15.0 | −2.5, 38.8 | 0.0, 21.3 | −8.8, 21.3 | 0.0, 20.0 | −8.8, 38.8 |
| Better ear PTA >15 dB: n (%) | 0 (0%) | 27 (6%) | 9 (6%) | 34 (9%) | 5 (7%) | 75 (7%) |
| Worse ear PTA (dB), median (IQR) | 10.6 (8.8, 13.8) | 11.3 (8.8, 14.4) | 11.3 (8.8, 13.8) | 11.3 (8.8, 15.0) | 11.3 (8.8, 13.8) | 11.3 (8.8, 15.0) |
| Min, Max | 5.0, 16.3 | 0.0, 40.0 | 1.3, 65.0 | −1.3, 27.5 | 0.0, 23.8 | −1.3, 65.0 |
| Worse ear PTA >15 dB, n (%) | 3 (11%) | 68 (15%) | 16 (10%) | 58 (15%) | 13 (17%) | 158 (15%) |
| Worse ear type of hearing loss, n (%) | ||||||
| Normal | 25 (89%) | 371 (84%) | 137 (89%) | 322 (85%) | 63 (83%) | 917 (85%) |
| Sensorineural | 2 (7%) | 59 (13%) | 13 (8%) | 53 (14%) | 12 (16%) | 139 (13%) |
| Conductive | 1 (4%) | 9 (2%) | 3 (2%) | 5 (1%) | 1 (1%) | 19 (2%) |
| Could not determine | 0 (0%) | 1 (1%) | 1 (1%) | 1 (0%) | 0 (0%) | 3 (0%) |
PTA=pure-tone average, GA=gestational age, ART=antiretroviral therapy, dB=decibel, IQR=interquartile range
There were 763 CHEU with consistent exposure to one of the three regimens of interest and known maternal ART initiation timing. Demographic and maternal characteristics of this subset were very similar to the full cohort (Supplemental Table 1). Eleven CHEU exposed to TDF/FTC/(-ATV) and six exposed to TDF/FTC/ATV also received EFV and were excluded from analyses. Hearing loss prevalence was similar between CHEU exposed to the three regimens (103/746, 13.8%) and those on other regimens (32/271, 11.8%). Too few infants initiated ART at GA ≥26 weeks (n=68) for fully adjusted models. Given fetal developmental differences between second and third trimester, the ≥26-weeks stratum was excluded from regressions.
In log-binomial models (Table 3), among CHEU with earliest exposure <13 weeks, CHEU exposed to TDF/FTC/ATV had a 28% higher risk of hearing loss (adjusted RR=1.28, 95% CI: 0.63, 2.58) and 38% lower risk for those exposed to ZDV/3TC/(-ATV) (adjusted RR=0.62, 95% CI: 0.32, 1.19) compared to those exposed to TDF/FTC/(-ATV), although both had wide 95% confidence intervals (CIs). Results were similar in unadjusted and adjusted models. For exposures initiated at 13≤GA<26 weeks, both TDF/FTC/ATV and ZDV/3TC/(-ATV) had lower risk than TDF/FTC/(-ATV), again with wide CIs.
Table 3.
Unadjusted and adjusted associations between in utero exposure to ART regimens and sensorineural hearing loss, by exposure timing group.
|
ART exposure initiated before 13 weeks
gestation age |
UNADJUSTED1 | ADJUSTED2 | ||||
|---|---|---|---|---|---|---|
| In utero ART Exposure | Relative Risk |
95% Confidence Interval |
p-value | Relative Risk |
95% Confidence Interval |
p-value |
| TDF/FTC/ATV | 1.25 | 0.56 to 2.79 | 0.58 | 1.28 | 0.63 to 2.58 | 0.49 |
| ZDV/3TC/(-ATV) | 0.60 | 0.30 to 1.22 | 0.16 | 0.62 | 0.32 to 1.19 | 0.15 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
|
ART exposure initiated between 13 to 26 weeks
gestation age |
UNADJUSTED3 | ADJUSTED4 | ||||
| In utero ART Exposure | Relative Risk |
95% Confidence Interval |
p-value | Relative Risk |
95% Confidence Interval |
p-value |
| TDF/FTC/ATV | 0.74 | 0.39 to 1.43 | 0.38 | 0.70 | 0.32 to 1.52 | 0.37 |
| ZDV/3TC/(-ATV) | 0.92 | 0.42 to 2.03 | 0.84 | 0.86 | 0.39 to 1.89 | 0.70 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
TDF - tenofovir disoproxil fumarate, FTC - emtricitabine, ATV - atazanavir, ZDV - zidovudine, 3TC - lamivudine -ATV denotes a regimen without atazanavir (ATV)
Model fit among 365 children
Model fit among 355 children, adjusted for sex, ototoxic medication exposure, maternal education, household income, substance use, and mother's age at delivery
Model fit among 313 children
Model fit among 299 children, adjusted for sex, ototoxic medication exposure, maternal education, household income, substance use, and mother's age at delivery
DPOAEs
Of 493 CHEU with DPOAE and tympanogram data, demographic and maternal characteristics were again similar to the full cohort (Supplemental Table 2). Most (454; 92%) had bilateral Type A tympanograms; 14 (2.9%) had Type A in one ear and Type C in the other ear. All 25 CHEU with single-ear tympanograms had Type A tympanograms.
Across GAs, >90% had DPOAE SNRs ≥6 dB at 2, 3, and 4 kHz in both ears (Table 4). One site did not have a DPOAE system that obtained 1 kHz data; among remaining sites, ~90% had ≥6 dB SNRs bilaterally at 1 kHz.
Table 4.
Summary of DPOAE data for left ear (n=475) and right ear (n=472) by gestational age at ARV initiation.
| DPOAE Characteristics, n (%) | Missing (n=13) |
Timing of earliest ART exposure in gestation | Total (n=475) |
|||
|---|---|---|---|---|---|---|
| Preconception (GA=0) (n=224) |
0<GA<13 (n=78) |
13≤GA<26 (n=124) |
GA≥26 (n=36) |
|||
| Left ear SNR at 1 kHz | ||||||
| <3 dB | 1 (8%) | 9 (4%) | 4 (5%) | 11 (9%) | 1 (3%) | 26 (5%) |
| ≥3 dB - <6 dB | 0 (0%) | 9 (4%) | 1 (1%) | 2 (2%) | 1 (3%) | 13 (3%) |
| ≥6 dB | 11 (85%) | 165 (74%) | 61 (78%) | 91 (73%) | 28 (78%) | 356 (75%) |
| Missing* | 1 (8%) | 41 (18%) | 12 (15%) | 20 (16%) | 6 (17%) | 80 (17%) |
| Left ear SNR at 2 kHz | ||||||
| <3 dB | 0 (0%) | 4 (2%) | 6 (8%) | 6 (5%) | 1 (3%) | 17 (4%) |
| ≥3 dB - <6 dB | 1 (8%) | 1 (0%) | 0 (0%) | 3 (2%) | 0 (0%) | 5 (1%) |
| ≥6 dB | 12 (92%) | 219 (98%) | 72 (92%) | 114 (92%) | 35 (97%) | 452 (95%) |
| Missing | 0 (0%) | 0 (0%) | 0 (0%) | 1 (1%) | 0 (0%) | 1 (0%) |
| Left ear SNR at 3 kHz | ||||||
| <3 dB | 0 (0%) | 6 (3%) | 3 (4%) | 7 (6%) | 1 (3%) | 17 (4%) |
| ≥3 dB - <6 dB | 0 (0%) | 4 (2%) | 1 (1%) | 1 (1%) | 0 (0%) | 6 (1%) |
| ≥6 dB | 13 (100%) | 214 (96%) | 74 (95%) | 116 (94%) | 35 (97%) | 452 (95%) |
| Left ear SNR at 4 kHz | ||||||
| <3 dB | 0 (0%) | 9 (4%) | 3 (4%) | 11 (9%) | 2 (6%) | 25 (5%) |
| ≥3 dB - <6 dB | 0 (0%) | 5 (2%) | 0 (0%) | 1 (1%) | 1 (3%) | 7 (1%) |
| ≥6 dB | 13 (100%) | 209 (93%) | 74 (95%) | 111 (90%) | 33 (92%) | 440 (93%) |
| Missing | 0 (0%) | 1 (0%) | 1 (1%) | 1 (1%) | 0 (0%) | 3 (1%) |
| Missing (n=13) |
Preconception (GA=0) (n=221) |
0<GA<13 (n=78) |
13≤GA<26 (n=126) |
GA≥26 (n=34) |
Total (n=472) |
|
| Right ear SNR at 1 kHz | ||||||
| <3 dB | 0 (0%) | 14 (6%) | 2 (3%) | 13 (10%) | 1 (3%) | 30 (6%) |
| ≥3 dB - <6 dB | 0 (0%) | 5 (2%) | 3 (4%) | 1 (1%) | 2 (6%) | 11 (2%) |
| ≥6 dB | 12 (92%) | 161 (73%) | 60 (77%) | 91 (72%) | 26 (76%) | 350 (74%) |
| Missing* | 1 (8%) | 41 (19%) | 13 (17%) | 21 (17%) | 5 (15%) | 81 (17%) |
| Right ear SNR at 2 kHz | ||||||
| <3 dB | 1 (8%) | 7 (3%) | 5 (6%) | 5 (4%) | 1 (3%) | 19 (4%) |
| ≥3 dB - <6 dB | 0 (0%) | 7 (3%) | 3 (4%) | 3 (2%) | 0 (0%) | 13 (3%) |
| ≥6 dB | 12 (92%) | 207 (94%) | 70 (90%) | 118 (94%) | 33 (97%) | 440 (93%) |
| Right ear SNR at 3 kHz | ||||||
| <3 dB | 0 (0%) | 4 (2%) | 5 (6%) | 9 (7%) | 1 (3%) | 19 (4%) |
| ≥3 dB - <6 dB | 0 (0%) | 1 (0%) | 1 (1%) | 2 (2%) | 0 (0%) | 4 (1%) |
| ≥6 dB | 13 (100%) | 216 (98%) | 72 (92%) | 115 (91%) | 33 (97%) | 449 (95%) |
| Right ear SNR at 4 kHz | ||||||
| <3 dB | 0 (0%) | 10 (5%) | 5 (6%) | 9 (7%) | 1 (3%) | 25 (5%) |
| ≥3 dB - <6 dB | 0 (0%) | 6 (3%) | 1 (1%) | 4 (3%) | 1 (3%) | 12 (3%) |
| ≥6 dB | 13 (100%) | 204 (92%) | 71 (91%) | 112 (89%) | 32 (94%) | 432 (92%) |
| Missing | 0 (0%) | 1 (0%) | 1 (1%) | 1 (1%) | 0 (0%) | 3 (1%) |
One clinical site did not have a DPOAE system that collected data at 1 kHz.
In regression models (Table 5), among 228 CHEU whose mothers initiated ART at GA <13 weeks, there was a higher risk of incomplete (<6 dB) DPOAE responses in CHEU exposed to ZDV/3TC/(-ATV) and a lower risk for CHEU exposed to TDF/FTC/ATV compared to CHEU exposed to TDF/FTC/(-ATV) across all frequencies, although all estimates had wide 95% CIs. For 113 CHEU with maternal ART initiation at 13≤GA<26 weeks, both TDF/FTC/ATV and ZDV/3TC/(-ATV) had lower risk than TDF/FTC/(-ATV) across frequencies, again with wide CIs.
Table 5.
Unadjusted and adjusted associations between in utero exposure to ART regimens and DPOAE frequencies, by exposure timing group.
| ART exposure initiated before 13 weeks gestation age | UNADJUSTED1 | ADJUSTED2 | ||||
|---|---|---|---|---|---|---|
| In utero ART Exposure | Relative Risk |
95% Confidence Interval |
p-value | Relative Risk |
95% Confidence Interval |
p-value |
| 1 kHz: | ||||||
| TDF/FTC/ATV | 0.78 | 0.29 to 2.14 | 0.63 | 0.86 | 0.38 to 1.93 | 0.71 |
| ZDV/3TC/(-ATV) | 1.22 | 0.56 to 2.64 | 0.61 | 1.17 | 0.56 to 2.44 | 0.67 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
| 2 kHz: | ||||||
| TDF/FTC/ATV | 0.78 | 0.26 to 2.35 | 0.66 | 0.85 | 0.28 to 2.57 | 0.78 |
| ZDV/3TC/(-ATV) | 1.57 | 0.59 to 4.16 | 0.37 | 1.66 | 0.61 to 4.51 | 0.32 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
| 3 kHz: | ||||||
| TDF/FTC/ATV | 0.43 | 0.10 to 1.90 | 0.27 | 0.43 | 0.12 to 1.62 | 0.21 |
| ZDV/3TC/(-ATV) | 1.54 | 0.49 to 4.87 | 0.46 | 1.40 | 0.45 to 4.33 | 0.56 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
| 4 kHz: | ||||||
| TDF/FTC/ATV | 0.93 | 0.32 to 2.76 | 0.90 | 0.87 | 0.29 to 2.64 | 0.81 |
| ZDV/3TC/(-ATV) | 1.28 | 0.47 to 3.47 | 0.63 | 1.20 | 0.45 to 3.22 | 0.72 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
|
ART exposure initiated between 13 to 26 weeks
gestation age |
UNADJUSTED3 | ADJUSTED4 | ||||
| In utero ART Exposure | Relative Risk |
95% Confidence Interval |
p-value | Relative Risk |
95% Confidence Interval |
p-value |
| 1 kHz: | ||||||
| TDF/FTC/ATV | 0.56 | 0.18 to 1.78 | 0.33 | 0.66 | 0.20 to 2.14 | 0.49 |
| ZDV/3TC/(-ATV) | 0.86 | 0.33 to 2.26 | 0.76 | 0.64 | 0.22 to 1.82 | 0.40 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
| 2 kHz: | ||||||
| TDF/FTC/ATV | 0.30 | 0.05 to 1.67 | 0.17 | 0.34 | 0.06 to 2.00 | 0.23 |
| ZDV/3TC/(-ATV) | 0.43 | 0.12 to 1.53 | 0.19 | 0.30 | 0.07 to 1.26 | 0.10 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
| 3 kHz: | ||||||
| TDF/FTC/ATV | 0.82 | 0.24 to 2.81 | 0.75 | 0.90 | 0.27 to 2.97 | 0.86 |
| ZDV/3TC/(-ATV) | 0.51 | 0.14 to 1.87 | 0.31 | 0.29 | 0.07 to 1.21 | 0.09 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
| 4 kHz: | ||||||
| TDF/FTC/ATV | 0.52 | 0.14 to 1.97 | 0.34 | 0.48 | 0.14 to 1.61 | 0.23 |
| ZDV/3TC/(-ATV) | 0.67 | 0.22 to 1.99 | 0.47 | 0.38 | 0.13 to 1.09 | 0.07 |
| TDF/FTC/(-ATV) | 1.00 (ref) | --- | --- | 1.00 (ref) | --- | --- |
Model fit among 1637 observations, including four frequencies and two ears
Model fit among 1621 observations, adjusted for sex, ototoxic medication exposure, maternal education, household income, substance use, and mother's age at delivery
Model fit among 827 observations
Model fit among 795 observations, adjusted for sex, ototoxic medication exposure, maternal education, household income, substance use, and mother's age at delivery
Relationship Between Pure-Tone and DPOAE Data
Among 440 CHEU with both hearing measures, those with DPOAE SNRs <6 dB had substantially higher rates of hearing loss defined by PTA >15 dB (27%–43%) than those with complete responses (6%–8%) for all frequencies and ears. Median PTA was 2.5–3.75 dB higher for those with <6 dB DPOAEs at all frequencies except left-ear 3 kHz.
DISCUSSION
In this cohort, 15% of 5-year-old CHEU had hearing loss, ranging from 10% to 17% across exposure-timing strata. This prevalence exceeds that reported for older (6–11-year-old) U.S. children: 7.6% using PTA at 0.5, 1, and 2 kHz and 12.2% using PTA at 3, 4, and 6 kHz16. Early identification is therefore critical to enable interventions that mitigate impacts on speech and language development17, academic performance, and social interactions18. DPOAEs were largely robust (92%–95% with ≥6 dB SNR at 2–4 kHz), consistent with norms among children with Type A tympanograms19.
There was no consistent association between earliest in utero ARV exposure and either hearing loss or DPOAE outcomes. Among CHEU with exposure initiated <13 weeks, TDF/FTC/ATV showed a nonsignificant higher risk of hearing loss compared to TDF/FTC/(-ATV), whereas ZDV/3TC/(-ATV) had a lower risk. For DPOAEs in CHEU exposed at GA <13 weeks, ZDV/3TC/(-ATV) was associated with higher risk of <6 dB DPOAEs and TDF/FTC/ATV with lower risk of <6 dB DPOAEs. The timing of exposure relative to auditory system development likely matters. Because the development of the cochlea and cochlear nerve occurs during GA <13 weeks15, this may explain the higher risk of <6 dB DPOAEs and hearing loss in CHEU with in utero exposure to ZDV/3TC/(-ATV) and TDF/FTC/ATV, respectively, for children whose mothers first initiated ARV at GA <13 weeks. Conversely, among children whose mothers first initiated ARV at 13≤GA<26 weeks, another critical period of auditory development15, there was a lower risk of hearing loss and <6 dB DPOAEs with in utero TDF/FTC/(ATV) or ZDV/3TC/(-ATV).
There are no prior studies specifically linking in utero ARV exposure to hearing outcomes in school-aged CHEU. However, multiple SMARTT analyses have examined neurodevelopment20-22, speech and language23, and cognition/academics24. For example, ATV-containing regimens in the second/third trimester have been linked with adverse language and social-emotional outcomes at age 1-year-old CHEU21. In utero TDF exposure has been associated with higher risk of speech impairment at age 3 years23, but among older CHEU (5–13 years), no significant associations with specific ARVs were observed for cognitive/academic outcomes24. More recently, 5-year-old CHEU with in utero TDF/FTC/ATV had higher adjusted cumulative odds of one or more neurodevelopmental signals compared to ZDV/3TC plus other PI or NNRTI22. This aligns with our observation of a higher risk of hearing loss with early TDF/FTC/ATV exposure.
Limitations include the absence of an HIV-unexposed, uninfected control group; small numbers for certain ARVs (e.g., EFV) and some exposure-timing strata; and exclusion of the ≥26-weeks initiation group from regressions due to both size and developmental differences. Maternal ARVs evaluated in this study are not currently the most widely used. WHO and U.S. guidelines recommend the use of integrase strand transfer inhibitors (INSTIs) for treatment of HIV infection during pregnancy25,26. Second-generation INSTIs generally offer improved efficacy and higher resistance barriers than other drug classes27, and some studies suggest lower risk of adverse perinatal outcomes with maternal use of INSTI-based regimens28,29. Some hearing measures in younger children can be less reliable, potentially contributing to inconsistent associations. Additionally, pure tones and DPOAEs assess peripheral hearing and in utero ARV exposure could also affect central auditory processing. Central measures (e.g., dichotic digits) are typically feasible at ≥7 years30, while competing sentences and frequency patterns testing are commonly performed at ≥8 years31,32.
In summary, approximately one in eight 5-year-old CHEU had permanent SNHL, underscoring the need for early screening, timely intervention, and awareness of hearing testing among parents and caregivers of CHEU. While there were no consistent regimen-specific associations, exploratory patterns suggest exposure timing, particularly GA <13 weeks, may influence risk. These results support continued hearing testing in CHEU and motivate future studies incorporating central auditory processing measures as children get older. Further work should examine newer ARVs and regimen initiation timing to clarify potential auditory effects. Clinicians caring for CHEU should ensure adherence to hearing screening as part of standard care as hearing is a fundamental component to communication and quality of life, and undiagnosed hearing loss can have lasting adverse effects.
Supplementary Material
ACKNOWLEDGEMENTS:
We thank the participants and families for their participation in PHACS, and the individuals and institutions involved in the conduct of PHACS. The Pediatric HIV/AIDS Cohort Study (PHACS) network was supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD), Office of The Director, National Institutes of Health (OD), National Institute of Dental & Craniofacial Research (NIDCR), National Institute of Allergy and Infectious Diseases (NIAID), National Institute of Neurological Disorders and Stroke (NINDS), National Institute on Deafness and Other Communication Disorders (NIDCD), National Institute of Mental Health (NIMH), National Institute on Drug Abuse (NIDA), National Cancer Institute (NCI), National Institute on Alcohol Abuse and Alcoholism (NIAAA), and the National Heart, Lung, and Blood Institute (NHLBI) through grants to the Harvard T.H. Chan School of Public Health (P01HD103133, Principal Investigators: Ellen Chadwick, Sonia Hernandez-Diaz, Jennifer Jao, Paige Williams; Program Director: Liz Salomon and HD052102: Principal Investigator: George R Seage III; Program Director: Liz Salomon) and with Tulane University School of Medicine (HD052104) (Principal Investigator: Russell Van Dyke; Co-Principal Investigator: Ellen Chadwick; Project Director: Patrick Davis). Data management services were provided by Frontier Science (Data Management Center Director: Suzanne Siminski), and regulatory services and logistical support were provided by Westat, Inc (Project Director: Tracy Wolbach).
The following institutions, clinical site investigators and staff participated in conducting PHACS SMARTT in 2022, in alphabetical order: Ann & Robert H. Lurie Children’s Hospital of Chicago: Ellen Chadwick, Margaret Ann Sanders, Kathleen Malee; Baylor College of Medicine: Mary Paul, Ruth Eser-Jose, Chivon McMullen-Jackson, Lynnette Harris; BronxCare Health System: Murli Purswani, Mahoobullah Mirza Baig, Alma Villegas, Marvin Alvarado; Children’s Diagnostic & Treatment Center: Lisa-Gaye Robinson, Jawara Dia Cooley, James Blood, Patricia Garvie; New York University School of Medicine: William Borkowsky, Nagamah Sandra Deygoo, Jennifer Lewis; Rutgers - New Jersey Medical School: Arry Dieudonne, Linda Bettica, Juliette Johnson, Karen Surowiec; St. Jude Children’s Research Hospital: Katherine Knapp, Jamie Russell-Bell, Megan Wilkins, Stephanie Love; San Juan Hospital Research Unit/Department of Pediatrics, San Juan Puerto Rico: Nicolas Rosario, Lourdes Angeli-Nieves, Vivian Olivera; SUNY Downstate Medical Center: Stephan Kohlhoff, Ava Dennie, Jean Kaye, Jenny Wallier; Tulane University School of Medicine: Karen Craig, Margarita Silio, Patricia Sirois; University of Alabama, Birmingham: Cecelia Hutto, Paige Hickman, Julie Huldtquist, Dan Marullo; University of California, San Diego: Stephen A. Spector, Veronica Figueroa, Megan Loughran, Sharon Nichols; University of Colorado, Denver: Elizabeth McFarland, Christine Kwon, Carrie Glenny, Jennifer Englund; University of Florida, Center for HIV/AIDS Research, Education and Service: Mobeen Rathore, Saniyyah Mahmoudi, Sarah El-Hassan, Jamilah Tejan; University of Illinois, Chicago: Karen Hayani, Lourdes Richardson, Renee Smith, Alina Miller; University of Miami: Gwendolyn Scott, Gustavo Gil Garcia, Gabriel Fernandez, Anai Cuadra; Keck Medicine of the University of Southern California: Toni Frederick, Mariam Davtyan, Guadalupe Morales-Avendano; University of Puerto Rico School of Medicine, Medical Science Campus: Zoe M. Rodriguez, Lizmarie Torres, Nydia Scalley
Funding for this project was supported by NIH P01HD103133
Footnotes
Note: The conclusions and opinions expressed in this article are those of the authors and do not necessarily reflect those of the National Institutes of Health or U.S. Department of Health and Human Services.
REFERENCES
- 1.Simdon J, Watters D, Bartlett S, et al. Ototoxicity associated with use of nucleoside analog reverse transcriptase inhibitors: a report of 3 possible cases and review of the literature. Clin Infect Dis. 2001;32:1623–1627. [DOI] [PubMed] [Google Scholar]
- 2.Christensen LA, Morehouse CR, Powell TW, et al. Antiviral therapy in a child with pediatric human immunodeficiency virus (HIV): case study of audiologic findings. J Am Acad Audiol. 1998;9:292–298. [PubMed] [Google Scholar]
- 3.Torre III P, Hoffman HJ, Springer G, et al. Hearing loss among HIV-seropositive and HIV-seronegative men and women. JAMA Otolaryngol Head Neck Surg. 2015;141:202–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chao CK, Czechowicz JA, Messner AH, et al. High prevalence of hearing impairment in HIV-infected Peruvian children. Otolaryngol Head Neck Surg 2012;146:259–265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Torre III P, Zeldow B, Hoffman HJ, et al. Hearing loss in perinatally HIV-infected and HIV-exposed but uninfected children and adolescents. Pediatr Infect Dis J. 2012;31:835–841. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Thein P, Kalinec GM, Park C, et al. In vitro assessment of antiretroviral drugs demonstrates potential for ototoxicity. Hear Res. 2014;310:27–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.DeBacker JR, Langenek B, Bielefeld EC. Hearing loss in offspring exposed to antiretrovirals during pregnancy and breastfeeding. Front Mo Neurosc. 2022;15:920528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Park C, Thein P, Kalinec G, et al. HEI-OC1 cells as a model for investigating prestin function. Hear Res. 2016;335:9–17. [DOI] [PubMed] [Google Scholar]
- 9.Zheng J, Shen W, He DZ, et al. Prestin is the motor protein of cochlear outer hair cells. Nat. 2000; 405:149–155. [DOI] [PubMed] [Google Scholar]
- 10.Kemp D Evidence of mechanical nonlinearity and frequency selective wave amplification in the cochlea. Arch Otorhinolaryngol-Head Neck Surg, 1979;224:37–45. [DOI] [PubMed] [Google Scholar]
- 11.Torre III P, Zeldow B, Yao T, et al. Newborn hearing screenings in human immunodeficiency virus-exposed uninfected infants. J AIDS Immunol Res. 2016;1:102–106. [PMC free article] [PubMed] [Google Scholar]
- 12.Williams PL, Hazra R, Van Dyke RB, et al. Antiretroviral exposure during pregnancy and adverse outcomes in HIV-exposed uninfected infants and children using a trigger-based design. AIDS. 2016;30:133–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Van Dyke RB, Chadwick EG, Hazra R, et al. The PHACS SMARTT Study: Assessment of the safety of in utero exposure to antiretroviral drugs. Front Immunol. 2016;7:199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.American Speech-Language-Hearing Association. (2004). Guidelines for the Audiologic Assessment of Children From Birth to 5 Years of Age [Guidelines]. Available from www.asha.org/policy. [PubMed] [Google Scholar]
- 15.Moore JK, Linthicum FH Jr. The human auditory system: a timeline of development. Int J Audiol. 2007;46:460–478. [DOI] [PubMed] [Google Scholar]
- 16.Niskar A, Kiesak S, Holmes A, et al. Prevalence of hearing loss among children 6 to 19 years of age: The Third National Health and Nutrition Examination Survey. JAMA. 1998;279:1071–75. [DOI] [PubMed] [Google Scholar]
- 17.Yoshinaga-Itano C From screening to early identification and intervention: Discovering predictors to successful outcomes for children with significant hearing loss. J Deaf Stud Deaf Educ. 2003;8:11–30. [DOI] [PubMed] [Google Scholar]
- 18.Bess F, Dodd-Murphy J, Parker R. Children with minimal sensorineural hearing loss: prevalence, educational performance, and functional status. Ear Hear. 1998;19:339–54. [DOI] [PubMed] [Google Scholar]
- 19.Lyon A, Kei J, Driscoll C. Distortion product otoacoustic emissions in children at school entry: A comparison with pure-tone screening and tympanometry results. J Am Acad Audiol. 2004;15:702–715. [DOI] [PubMed] [Google Scholar]
- 20.Sirois PA, Huo Y, Williams PL, et al. Safety of perinatal exposure to antiretroviral medications: Developmental outcomes in infants. Pediatr Infect Dis J. 2013;32:648–655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Caniglia EC, Patel K, Huo Y, et al. Atazanavir exposure in utero and neurodevelopment in infants: A comparative safety study. AIDS. 2016;30:1267–1278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Yao TJ, Malee K, Zhang J, et al. In utero antiretroviral exposure and risk of neurodevelopment problems in HIV-exposed uninfected 5-year-old children. AIDS Pat Care STDs. 2023;37:119–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rice ML, Russell JS, Frederick T, et al. Risk for speech and language impairments in preschool age HIV-exposed uninfected children with in utero combination antiretroviral exposure. Pediatr Infect Dis J. 2018;37:678–685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Nozyce ML, Huo Y, Williams PL, et al. Safety of in utero and neonatal antiretroviral exposure: Cognitive and academic outcomes in HIV-exposed, uninfected children 5-13 years of age. Pediatr Infect Dis J. 2014;33:1128–1133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.World Health Organization (2021). Consolidated Guidelines on HIV Prevention, Testing, Treatment, Service Delivery and Monitoring: Recommendations for a Public Health Approach. Geneva. [PubMed] [Google Scholar]
- 26.Panel on Treatment of HIV During Pregnancy and Prevention of Perinatal Transmission. Recommendations for the Use of Antiretroviral Drugs During Pregnancy and Interventions to Reduce Perinatal HIV Transmission in the United States. Department of Health and Human Services. Available at https://clinicalinfo.hiv.gov/en/guidelines/perinatal. Accessed 10/25/2025. [Google Scholar]
- 27.Mbhele N, Chimukangara B, Gordon M. HIV-1 integrase strand transfer inhibitors: a review of current drugs, recent advances and drug resistance. Int J Antimicrob Agents. 2021; 57:106343. [DOI] [PubMed] [Google Scholar]
- 28.Schneidman J, Lee T, Sauve L, et al. Type and timing of antiretroviral therapy during pregnancy: Impact on risk of preterm delivery and small-for-gestational age births in Canada, a retrospective cohort study. Int J Gynaecol Obstet. 2023;162:614–622. [DOI] [PubMed] [Google Scholar]
- 29.Beck K, Cowdell I, Portwood C, et al. Comparative risk of adverse perinatal outcomes associated with classes of antiretroviral therapy in pregnant women living with HIV: systematic review and meta-analysis. Front Med. 2024; 11:1323813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Musiek F. Assessment of central auditory dysfunction: The dichotic digit test revisited. Ear Hear. 1983;4:79–83. [DOI] [PubMed] [Google Scholar]
- 31.Musiek F, Pinheiro M. Dichotic speech tests in the detection of central auditory dysfunction. In Pinheiro M, Musiek F eds. Assessment of central auditory dysfunction: Foundations and clinical correlates. Baltimore: Williams & Wilkins; 1985;201–218. [Google Scholar]
- 32.Musiek F, Pinheiro M. Frequency patterns in cochlear, brainstem and cerebral lesions. Audiol. 1987;26:79–88. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.