Impact of the ACEMg Biomedicine on Sensorineural Hearing Loss and Auditory Function: Analysis of Real-World Clinical Data

Abstract

Background

Hearing loss is a chronic disability that affects the lives and livelihoods for over 1.5 billion people worldwide. Hearing aids are the predominant treatment; however, their impact is modest, and they are less available in low- and middle-income countries where the need is greatest. Moreover, hearing loss is the #1 mid-life risk factor for dementia. Sensorineural hearing loss (SNHL), the most common form of hearing loss, is a metabolic stress disorder.

Objective

The micronutrient formula ACEMg was developed to block SNHL and successfully tested on animal models. We report real-world data (RWD) from a real-world evidence (RWE) study (N = 190) conducted to assess whether ACEMg impacts SNHL in humans.

Methods

This was a retrospective observational cohort study design with a historical comparison group. RWD from distortion product otoacoustic emissions (OAE) examinations were collected from patients previously diagnosed with SNHL. OAE testing is a recognized marker of cochlear health, producing objective measures of the auditory function of outer hair cells (OHC) in the cochlea.

Results

Analyses of RWD from OAE examinations in the treatment group (N = 93), who used ACEMg as softgel capsules, compared with OAE RWD from the untreated group (N = 97) suggest that ACEMg is significantly associated with the preservation and improvement of auditory function. Overall, OAE scores remained unchanged or improved for 75.3% of those in the ACEMg treatment group, as contrasted with 26.8% in the no ACEMg group. OAE scores over 2 years remained unchanged for 37.6% and improved for 37.7%. OAE scores decreased for 73.2% of those in the untreated group (X ² _2,190 = 55.94, P < .001).

Conclusion

This study yielded statistically reliable, objective, real-world clinical data suggesting a potential association between ACEMg supplementation and higher OAE scores over time, although alternative explanations cannot be excluded. These findings warrant additional investigation, given the widespread incidence of SNHL.

Introduction

The World Health Organization (W.H.O.) describes hearing loss as a major global health issue, an unmet health and medical need that decreases the quality of life at any age, estimating current annual economic costs to be nearly 1 trillion dollars. Hearing loss is associated with social isolation, cognitive impairments, educational challenges, underemployment, and higher rates of unemployment. ¹ Moreover, hearing loss in midlife is the number one risk factor for dementia (Alzheimer’s disease). ² By 2050, approximately 10% of the world population, or 2.5 billion people, are projected to have a degree of hearing loss.^3,4

The hearing loss problem is most acute in the younger population. The W.H.O. estimates “Over 1 billion people aged 12 to 35 years risk losing their hearing due to prolonged and excessive exposure to loud music and other recreational sounds. This can have devastating consequences for their physical and mental health, education, and employment prospects.” ⁵

SNHL is the most common form of hearing loss arising from a variety of etiologies including age (presbycusis), ototoxic pharmaceuticals, certain genetic mutations, viral or bacterial infection, but principally from noise exposure. ⁶

Measurably reducing the economic and social burdens associated with SNHL was the overarching motivation for the basic research on SNHL beginning in the late 1980s at the Kresge Hearing Research Institute at the University of Michigan Medical School under the direction of its director, Josef M. Miller, Ph.D., supported by grants from the National Institute on Deafness and Other Communication Disorders (NIDCD) at the National Institutes of Health (NIH).

The evidence on the key role of free radicals in cell dysfunction provided theoretical and empirical support for the hypothesis that it plays a role in hearing impairment. Indeed, evidence for the efficacy of antioxidant treatment to prevent free radical-induced pathology in the eye and ear, supported the idea that antioxidant interventions might be beneficial in preventing or ameliorating SNHL.^7,8

Thus, SNHL can be described as a metabolic stress disorder caused by excess inner ear free radicals, reactive oxygen species (ROS) formed as byproducts of mitochondrial metabolism. The resulting oxidative stress disrupts normal biochemical processes, damages mitochondria and other organelles, upregulates genes involved in cell death pathways, leading to auditory path dysfunction experienced as hearing loss. ⁹

It is well established that mammalian cochlear hair cell loss is permanent. Unlike birds and fish, adult humans cannot regenerate destroyed hair cells. Current research into hair cell regeneration through gene therapy and stem cell approaches remains in early experimental stages with no clinical applications. However, the distinction between permanent structural loss and reversible functional impairment is clinically relevant. Hair cells under chronic oxidative stress may become metabolically compromised and cease producing detectable otoacoustic emissions while remaining structurally intact. If antioxidant intervention can rescue such stressed-but-surviving cells before irreversible damage occurs, functional improvement might be achievable without requiring biological regeneration. The present study examines whether ACEMg supplementation is associated with OAE outcomes, with findings interpreted in the context of potential metabolic rescue rather than structural regeneration.

The high-potency ACEMg supplement formulation of antioxidants, Vitamins A, C, and E, and the vasodilator Magnesium, developed in the Miller laboratory, was demonstrated to protect against noise-induced hearing loss in animal models. ¹⁰ Additional Phase I, II and III translational studies under FDA IND were funded by the NIDCD and a BfArM (Germany) IMPD RCT study was funded by a European Commission medical innovation grant.^11-15

The ACEMg formulation was licensed to a public benefit corporation that produces a softgel capsule oral dose format of the ACEMg supplement. Availability enabled conducting an initial post-marketing RWD clinical study to assess whether ongoing, daily oral self-administration of ACEMg would inhibit further hearing loss in those who had been previously diagnosed with SNHL. The present study evaluates ACEMg as a disease-modifying intervention in individuals with established SNHL, assessing whether daily supplementation is associated with stabilization or improvement of objective auditory function. This positions ACEMg as a secondary prevention strategy aimed at attenuating progression in those already affected, rather than as primary prevention in unaffected individuals or as a treatment to restore lost hearing.

It should be noted that SNHL encompasses damage not only to OHCs, but also to inner hair cells (IHCs), ribbon synapses connecting IHCs to auditory nerve fibers (cochlear synaptopathy or “hidden hearing loss”), and the auditory nerve itself. The present study specifically evaluates OHC function via OAE and does not directly assess synaptopathy or neural components of auditory processing, which may contribute to functional hearing difficulties independent of OHC status.

Method

Study Design

This study employed a retrospective observational cohort design with historical comparison groups to evaluate the impact of ACEMg supplementation on auditory function in individuals with sensorineural hearing loss (SNHL). The study analyzed RWD from OAE examinations collected as part of routine clinical care. Participants were not randomly assigned to treatment conditions; rather, the study compared outcomes between patients who elected to use ACEMg supplementation (the ACEMg treatment group) and historical controls who received standard audiological care without ACEMg (no-treatment group).

The comparison groups were defined by temporal availability of the ACEMg intervention: the treatment group consisted of patients examined between 2020-2023 who began ACEMg supplementation after 3 years of baseline OAE monitoring, while the no-treatment group consisted of patients examined between 2015-2020, before ACEMg softgel capsules became available. The RWD study design assessed the potential effectiveness of ACEMg under conditions of routine clinical practice rather than the controlled conditions of a randomized trial.

The study analyzed deidentified secondary data and was verified as exempt by IRB review (Solutions IRB protocol #0439, February 9, 2024) according to 45CFR46.104(d)(4). The study is exempt from FDA Investigational New Drug (IND) requirements as it assesses the impact of a dietary supplement, regulated under the Dietary Supplement Health and Education Act (DSHEA), on normal physiological function rather than investigating a drug for disease treatment. As a retrospective analysis of deidentified clinical records, individual patient consent was waived by the IRB under 45 CFR 46.116(d). All data handling procedures complied with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule.

Study Setting and Participants

Setting

Patient data were obtained from Lake Forest Hearing Professionals, a private audiology clinic in Lake Forest, Illinois, United States. The clinic serves a suburban population in the greater Chicago metropolitan area and provides comprehensive audiological services including diagnostic hearing evaluations, hearing aid dispensing, and auditory rehabilitation. The clinic operates as a specialty audiology practice accepting both self-referred patients and physician referrals from primary care providers, otolaryngologists, and other healthcare professionals. The patient population represents a mixed socioeconomic demographic typical of suburban northeastern Illinois communities.

Participant Selection and Eligibility Criteria

Medical records from January 2015 through December 2023 were systematically screened to identify eligible patients from historical controls (2015-2020) and treatment group (2020-2023), to identify all patients meeting the following criteria.

Inclusion Criteria

• 1. Previous clinical diagnosis of sensorineural hearing loss (SNHL) documented in the medical record.

• 2. Age ≥18 years at the time of first OAE examination in the study period.

• 3. Availability of OAE examination data for 5 consecutive annual time points

• 4. For the ACEMg treatment group, documented commencement of ACEMg supplementation immediately following the third annual OAE examination, with additional OAE examinations at 3-, 6-, and 9-months post-initiation.

• 5. For the no-treatment group, OAE examinations conducted between 2015-2020, before ACEMg softgel capsules became available.

Exclusion Criteria

• 1. Conductive hearing loss or mixed hearing loss (conductive and sensorineural components).

• 2. Fluctuating or idiopathic hearing loss conditions (eg, Meniere’s disease).

• 3. Active middle ear pathology at any examination timepoint.

• 4. Cochlear implant placement during the study period.

• 5. Incomplete OAE data (missing any of the required periodic examinations).

• 6. For the ACEMg treatment group, reported ACEMg supplementation prior to the initiation timepoint.

• 7. For the ACEMg treatment group, documented non-adherence to ACEMg supplementation protocol during the treatment period.

Record Selection Process

A total of 847 patient records were initially screened from the clinic’s audiological database for the combined time periods (2015-2023). Of these, 421 patients had documented diagnoses of SNHL. Among SNHL patients, 190 met all inclusion criteria and had no exclusions: 93 in the ACEMg treatment group (examined 2020-2023) and 97 in the no-treatment group (examined 2015-2020). The primary reasons for exclusion from the no-treatment group were incomplete 5-year OAE data (N = 187), conductive or mixed hearing loss components (N = 31), and cochlear implant placement during follow-up (N = 13).

The requirement for 5 consecutive years of OAE data was implemented to ensure adequate baseline characterization (3 pre-intervention years) and follow-up duration (2 post-intervention years) to detect meaningful changes in auditory function. We acknowledge this criterion introduces survivorship bias, potentially excluding patients with more rapid deterioration who may have discontinued clinic attendance, or those with unstable healthcare access. This limitation may result in underestimation of SNHL progression rates in both groups and potentially biases the sample toward patients with better overall health engagement.

Concurrent controls were not feasible within this real-world evidence framework because ACEMg supplementation was offered to all eligible patients once available (2020 onward), and withholding a potentially beneficial intervention for research purposes was considered ethically inappropriate in routine clinical practice. The historical control design, while methodologically limited, reflects the pragmatic constraints of observational research in clinical settings. See Figure 1, Participant Flow Diagram.

Figure 1.

The participant flow diagram illustrates record screening, eligibility assessment, group assignment, and follow-up for the retrospective observational cohort study of ACEMg supplementation effects on auditory function in patients with sensorineural hearing loss. Otoacoustic emissions examinations are abbreviated OAE. The no-treatment group consisted of historical controls examined before ACEMg softgel capsules became available (2015 2020). The ACEMg treatment group consisted of patients who elected to use ACEMg supplementation starting immediately after their Year 3 examination (2020-2023). All participants had complete data. No data loss follow-up was necessary as this was an analysis of existing clinical records from patients who completed all required examinations

All data were extracted from electronic medical records by licensed audiologist L.A. Halvorson and deidentified prior to analysis. Patient identifiers were replaced with unique study codes, and all protected health information was removed in accordance with HIPAA regulations.

ACEMg Supplementation Intervention

The ACEMg supplement contains high-potency antioxidant vitamins and a vasodilatory mineral developed specifically for auditory health. Each ACEMg softgel capsule contains beta-carotene (provitamin A) 9000 IU (as mixed carotenoids; Vitamin C (ascorbic acid) 500 mg; Vitamin E (d-alpha tocopherol), 200 IU (as natural source d-alpha tocopheryl acetate); and Magnesium (as magnesium oxide): 167.5 mg elemental magnesium. The daily dose is 2 softgel capsules (See Supplemental Materials for additional information).

Administration Protocol

Patients in the treatment group were instructed to self-administer 2 ACEMg softgel capsules once daily by mouth. Specific timing during the day was left to participant preference to reflect real-world usage patterns. Participants received a 3-month supply at each clinic visit.

Adherence Monitoring

Supplementation adherence was assessed by and certified through structured patient interviews by the audiologist at each OAE examination. Patients were asked to report their typical daily consumption pattern and any missed doses. No pill counts, pharmacy records, or biomarker verification (eg, serum vitamin levels) were used. Patients who reported taking ACEMg fewer than 5 days per week on average, or who reported extended gaps in use (>2 consecutive weeks), were excluded from the treatment group. We acknowledge that self-reported adherence may overestimate actual consumption, and objective verification was not possible within this retrospective framework.

Concomitant Treatments

Both groups received standard audiological care as determined by their treating audiologist, including hearing aid recommendations when clinically indicated. Use of hearing aids was permitted. Participants were not restricted from using other dietary supplements or medications during the study period, though use of medications with known ototoxic side effects such as aminoglycoside antibiotics and platinum-based chemotherapy were documented as exclusion criteria. Use of other dietary supplements was not systematically documented and represents a potential confounder. Patients taking ACEMg may also use other antioxidant supplements (eg, multivitamins, omega-3 fatty acids, CoQ10), which could independently or synergistically affect auditory outcomes. We cannot determine whether observed associations are specific to the ACEMg formulation or might be partially attributable to concurrent supplement use. No other formal hearing preservation interventions were provided to either group during the study period.

Hearing Measures

Outcome Measures

The most common hearing test is the pure tone audiometry (PTA) examination in which tones at varying frequencies and intensities are played through headphones, and the individual indicates whether they can hear the tone. As such, the PTA examination requires a subjective response.

For the clinical research purpose of assessing the potential impact of ACEMg on OHC biological function, however, an objective technique for testing hearing is required. The OAE examination assesses OHC auditory amplification function by recording responses generated by mechano-electrical transduction of OHC in the Organ of Corti within the cochlea. A tiny probe containing a microphone and a speaker is inserted into the ear canal, which is sealed with an earplug. Paired tones at various frequencies are emitted by the speaker, producing distortion frequencies. In response, OHC function at or above the lower detectable limit emits an acoustic wave (an otoacoustic emission) which can be detected by the microphone and measured by the OAE system. If OHC do not respond at a given frequency at a lower limit of sensitivity, cochlear OHC function at that frequency is classified as below the threshold of detection. This has been considered a final determination, as absent OAE at a given frequency indicates severely compromised OHC function at that frequency, though it does not provide information about inner hair cell or auditory nerve function. OAE reflects outer hair cell integrity specifically, not the complete auditory pathway. Thus, OAE is an objective measurement of cochlear function not subject to expectation or attention effects.^16,17

It should be noted that OAE generation depends not only on OHC function but also on the endocochlear potential (EP), the positive voltage maintained by the stria vascularis that provides the electrochemical driving force for hair cell transduction. Therefore, improvements in OAE measures may reflect beneficial effects on OHC function, EP maintenance, or both. This mechanistic consideration is relevant when interpreting ACEMg effects, as the antioxidant and vasodilatory components of the formulation could theoretically influence both OHC metabolic function and strial vascular health.

Primary Outcome

The primary outcome measure was the percentage of frequencies with detectable otoacoustic emissions, calculated as: (number of frequencies with detectable OAE/22 total possible frequencies) × 100%. This metric ranges from 0% (complete absence of detectable cochlear function bilaterally across all tested frequencies) to 100% (normal cochlear function bilaterally across all frequencies). A higher percentage indicates better preserved auditory function.

The primary endpoint for the ACEMg treatment group compared the change in OAE percentage from Year 3 baseline, immediately before ACEMg initiation, to Year 5, 2 years post-baseline, to the fifth annual exam for no-treatment group. This 2-year post-intervention period was selected to assess sustained effects of daily ACEMg supplementation on auditory function.

Secondary Outcomes

• 1. For the no-treatment group, trajectory of OAE percentage across 5 annual timepoints.

• 2. For the ACEMg treatment group, changes in OAE percentage at 3, 6, and 9 months post-ACEMg initiation.

• 3. Categorical classification of individual patient outcomes as: Improved (increase in OAE percentage >0%), Stable (no change in OAE percentage), or Declined (decrease in OAE percentage).

• 4. Frequency-specific analyses examining OAE response patterns across the tested range (1000-10000 Hz).

OAE Examination Procedures

OAE measurements were made by a licensed, trained professional in private clinical examination rooms using an Interacoustics Titan OAE system. Measurements were performed in both ears at 11 different frequencies (1000, 1500, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, and 10000 Hz). The measurements are binary: either a detectable otoacoustic emission is emitted at each of the frequencies (indicating OHC function at that frequency) or it is not. With 11 otoacoustic measures for each ear, a total of 22 binary measurements were available for analysis.

An individual with hearing within normal parameters would have detectable otoacoustic emissions at all frequencies in both ears; as SNHL progresses, the percentage of frequencies at which otoacoustic emissions are detectable decreases. Therefore, in this study analyses were conducted on the percentage of the potential total of 22 frequencies present in each exam, with a range of 0% (no detectable cochlear function in either ear at any frequency) to 100% (cochlear function within normal parameters in both ears at all frequencies).

Outcome Assessment Blinding

The audiologist performing or supervising OAE examinations, L.A. Halvorson, was not blinded to patient treatment status, as this was a retrospective analysis of clinical care data. Note that OAE examinations minimize performance bias. In contrast to PTA examinations, which are inherently subjective, OAE examinations are inherently objective physiological recordings that assess neuronal response and do not depend on clinician or patient judgment.

The statistician conducting primary data analyses, R.A. Detweiler, performed initial analyses blinded to which group represented the treatment vs no-treatment condition until the primary analysis was complete, at which point treatment assignment was revealed for interpretation and secondary analyses.

Exploratory Analyses

The following analyses were conducted as exploratory, hypothesis-generating investigations and should be interpreted with appropriate caution. There is the increased risk of Type I error from multiple comparisons in the secondary analyses: (1) frequency-specific analyses examining OAE response patterns across low, mid, and high frequency ranges; and (2) time course analyses examining OAE changes at 3-, 6-, and 9-months post-ACEMg initiation.

Statistical Analysis

Baseline Comparability Assessment

To evaluate whether the treatment and no-treatment groups were comparable at baseline (prior to ACEMg initiation), we compared:

• 1. Mean OAE scores across the 3 pre-intervention years using two-way repeated measures ANCOVA (Group × Year)

• 2. Demographic characteristics (age, sex) using one-way ANCOVA (continuous variables) and chi-square tests (categorical variables)

• 3. Variance homogeneity in age distribution using Levene’s test

Groups were considered comparable if: (a) mean OAE scores did not differ significantly between groups (no main effect of Group); (b) both groups suggested similar rates of decline over Years 1-3 (no Group × Year interaction); and (c) demographic distributions were similar.

Primary Analysis

The primary analysis tested the hypothesis that ACEMg supplementation affects the trajectory of auditory function over time. A 2 × 5 mixed-design ANCOVA was conducted with (a) Between-subjects factor: Group (ACEMg treatment vs no-treatment); (b) Within-subjects factor: Year (5 annual OAE measurements); and (c) Covariate: Age.

The critical test was the Group × Year interaction, indicating whether the 2 groups indicated different patterns of change over the 5-year period. A significant interaction would suggest that ACEMg altered the trajectory of hearing decline observed in the no-treatment group.

Age was included as a covariate because: (1) the groups differed significantly in mean age, and (2) age is an established predictor of hearing function. The Group × Age interaction was tested to determine whether ACEMg effects varied by patient age. Posthoc pairwise comparisons of mean OAE scores between groups at each year were conducted using Tukey’s HSD test to identify specific time points where groups differed significantly. The Tukey procedure adjusts for multiple comparisons while maintaining family-wise error rate at α = 0.05.

Secondary Analyses

• 1. Time course of ACEMg effect: For the ACEMg treatment group only, OAE scores were examined at baseline (Year 3) and at 3-, 6-, and 9-months post-initiation using one-way repeated measures ANCOVA to identify when significant changes first emerged.

• 2. Categorical outcome analysis: Each participant’s outcome from Year 3 to Year 5 was classified as Improved, Stable, or Declined based on change in OAE percentage. The distribution of these categories was compared between groups using chi-square test (2 groups × 3 outcome categories).

• 3. Frequency-specific analyses: Exploratory analyses examined whether ACEMg effects varied across the frequency spectrum (low: 1000-3000 Hz; mid: 4000-6000 Hz; high: 7000-10000 Hz) using three-way mixed ANCOVA (Group × Year × Frequency Range).

Missing Data

All participants included in the analysis had complete OAE data for all required time points (5 annual exams for both groups, plus 3-, 6-, and 9-month exams for the ACEMg treatment group). Patients with any missing OAE examinations were excluded during the record selection phase. Therefore, no imputation of missing outcome data was required. Baseline demographic data (age, sex) were complete for all participants.

Statistical Assumptions and Diagnostics

Prior to conducting ANCOVA, outcome data were screened for violations of parametric assumptions using Q-Q plots, Shapiro-Wilk tests, and Mauchly’s test sphericity. All analyses were conducted using JASP 18.2 (JASP Team, 2024, University of Amsterdam), which implements R statistical routines. Statistical significance was set at α = 0.05 (2-tailed) for all tests.

Results

Study Sample

Two groups of patients diagnosed with SNHL (N = 190) were compared: the ACEMg treatment group (N = 93) and the no-treatment group (N = 97). As a real-world evidence study, subjects were not randomly assigned to condition but were defined by temporal availability of ACEMg (see the Methods section for detailed eligibility criteria).

The no-treatment group (N = 97) consisted of all individuals diagnosed with SNHL who had annual OAE test data for 5 consecutive years; these were individuals for whom this data was available from the years before ACEMg softgel capsules were available. The purpose of the no treatment group is to provide a baseline of the typical rate of change in auditory function for those with SNHL over 5 years.

The analysis included all available patient records meeting eligibility criteria (N = 190). No a priori sample size calculation was performed, as this was a retrospective observational study. To provide context regarding the sensitivity of the analysis, we note that the available sample size (N = 190) allows detection of medium effect sizes (Cohen’s f ≈ 0.25 or larger, corresponding to partial η² ≈ 0.06) for the main effects and the interaction in a 2-way mixed-effects analysis of variance at a 2-sided α = 0.05. This information is provided for descriptive purposes only and is not intended as justification for the observed findings.

To serve as an appropriate comparison, it is important that the degree of SNHL in the 2 groups – ACEMg treatment and no-treatment – in the 3 years before the administration of ACEMg are similar.

First, the mean OAE scores of the 2 groups are not significantly different (F_1,188 = 0.50, n.s.). Second, both groups show a reduction in OAE scores over the first 3 years (F_2,376 = 36.81, P < .001) with a mean OAE score of 39.81% in the first year and 32.21% in the third year. This is consistent with other longitudinal research which indicates that hearing loss appears to be continuous and gradual.^18,19 Finally, and importantly, there is no difference in the amount of reduction in OAE scores for the 2 groups as indicated by the lack of an interaction between Group and Year (F_2,376 = 1.01, n.s.).

While a smaller proportion of the females received ACEMg, this difference is not significant (43.40% females and 55.95% males; X ² _1,190 = 2.96, n.s.). There is a statistically significant difference in average age (62.3 in the ACEMg group and 75.6 in the no-treatment group; (F_1,188 = 28.57, P < .001). With a range of ages of 19 to 88 in the ACEMg treatment group and 22 to 100 in the no-treatment group, the distribution of ages is statistically equal as indicated by the equality of variances (F_96,92 = 1.34, n.s.).

The significant age difference between groups (ACEMg treatment: M = 62.3 years; no-treatment: M = 75.6 years; F₁,₁₈₈ = 28.57, P < .001) likely reflects temporal changes in the clinic’s patient population and/or differential willingness to try supplementation by age. Importantly, age was included as a covariate in the primary analysis, and the non-significant Year × Age interaction (F₃.₁₅,₅₈₈.₉₇ = 1.74, n.s.) indicates that ACEMg-associated effects did not vary as a function of patient age. Residual confounding by age-related factors cannot be excluded; subgroup analyses stratified by age were not conducted due to sample size limitations. Further analysis of the possible confounding effects of age based on matched group analyses are reported below.

Statistical Analysis of ACEMg Effects

The primary analysis of the effect of ACEMg is a comparison of the change in hearing over 5 years as measured by OAE scores for those in the ACEMg treatment group with those in the no-treatment group. Figure 2 presents this comparison.

Figure 2.

Mean OAE scores (percentage of 22 tested frequencies with detectable otoacoustic emissions) over 5 annual examinations for the ACEMg treatment group (N = 93, solid line) and no treatment historical control group (N = 97, dashed line). Higher percentages indicate better preserved outer hair cell function. Error bars represent standard error of the mean. The ACEMg treatment group began supplementation immediately after the Year 3 examination (arrow). Asterisks indicate significant between-group differences at P < .001 (Tukey HSD)

As shown in Table 1, Year and Group are the first significant effects to consider. However, the significant interaction of Year with Group makes interpreting these effects not meaningful. That is, whether hearing on average changes over 5 years or the ACEMg treatment group on average hears better is not relevant. Whether hearing changes in the years of the study after 1 group begins taking ACEMg is important. This effect is tested statistically by the interaction of Year by Group, which is highly significant (F₃.₁₅,₅₈₈.₉₇ = 25.24, P < .0001), indicating divergent longitudinal trajectories of auditory function between groups. The interaction effect size was partial η² = 0.12, representing a medium-to-large effect at the model level. There is also a significant effect of Age on OAE scores; this is to be expected since subjects varied in age from 19 to 100. The fact that the Year-by-Age interaction is not significant indicates that the age of subjects did not affect the impact of ACEMg.

Table 1.

Effects of ACEMg Group Over 5 Years

Source	Sum of squares	df	Mean square	F	Significance
Year	960.65	3.15	305.01	2.39	n.s.
Year by group	10153.39	3.15	3223.74	25.24	<.0001
Year by age	700.96	3.15	222.56	1.74	n.s.
Residual	75234.83	588.97	127.74
Group	20246.96	1	20246.96	7.08	<.01
Age	230516.37	1	230516.37	80.63	<.0001
Residual	534651.07	187	2859.1

Note. Fractional df are based on conservative Greenhouse-Geisser adjustments due to a violation of sphericity.

Post hoc pairwise comparisons were conducted using Tukey’s HSD, which controls for the probability of making Type 1 errors when making multiple pairwise comparisons, to identify specific time points at which group means differed. Statistical inference was based on the omnibus mixed-effects ANCOVA. With 5 pairwise comparisons (Year 1 ACEMg treatment vs no-treatment, Year 2 ACEMg treatment vs no-treatment, etc.), a difference in means of 4.53 was required for significance at the P < .05 level. This threshold was exceeded only in Year 4 (mean difference = 8.85) and Year 5 (mean difference = 11.25), both of which were significant at the P < .001 level (critical difference = 7.03). For interpretability, standardized mean differences (Cohen’s d) were calculated using pooled standard deviations of group means at each time point and are reported for descriptive purposes only. The resulting effect sizes were d ≈ 0.31 at Year 4 and d ≈ 0.39 at Year 5, indicating small-to-moderate between-group differences. These results indicate that ACEMg supplementation is statistically associated with a differential trajectory of auditory function over time relative to no treatment.

Consistent with these findings, categorical outcome analysis demonstrated a strong association between treatment group and individual hearing trajectory (χ²₂,₁₉₀ = 55.94, P < .001); Cramér’s V = 0.54), indicating a large effect for the relationship between ACEMg use and the likelihood of improvement or stabilization vs decline. As shown in Table 2, for patients who received ACEMg, 75.3% had no change or improved OAE scores (37.6% had no change, 37.7% improved). In contrast, for patients who did not receive ACEMg, 73.2% had a decline in their OAE scores over the 5 years of the study; only 26.8% improved or had no change in OAE scores (2.1% improved and 24.7% had no change). This difference is highly significant (X²_2,190 = 55.94, P < .001).

Table 2.

Distribution of Individual Patient Outcomes by Treatment Group

Percent of patients with OAE change year 1 to year 5
Group	Decrease (%)	No change (%)	Increase (%)
ACEMg	24.7	37.6	37.7
No ACEMg	73.2	24.7	2.1

Categorical classification of auditory function trajectory from baseline (Year 1) to final assessment (Year 5). Decrease = reduction in OAE percentage; No Change = stable OAE percentage; Increase = improvement in OAE percentage. Chi-square test: χ² (2, N = 190) = 55.94, P < .001.

Sensitivity Analysis: Propensity Score Matching

To address possible confounding by age, we conducted propensity score matching (PSM) using logistic regression with age, sex, and baseline OAE scores. Using 1:1 nearest-neighbor matching (caliper = 0.2 × SD), 51 of 93 treatment patients (54.8%) were successfully matched to controls with similar propensity scores. In the matched sample (N = 102), covariate balance was achieved for age (Treatment: M = 67.4 y vs Control: M = 66.3 y; SMD = 0.065). The treatment association remained statistically significant in the matched sample (paired t = 4.00, P = .0002; Cohen’s d = 0.57), with 58.8% of treatment patients showing improvement compared to 5.9% of controls (χ² = 36.45, P < .0001). However, 42 treatment patients, predominantly younger adults, had no comparable matches in the historical control group, confirming partial non-comparability between groups. While PSM strengthens the observed association by controlling for measured confounders, unmeasured confounders, including hearing protection behaviors, remain unaddressed.

To assess the time course of treatment response, the ACEMg treatment group underwent additional OAE testing at 3-, 6-, and 9-months post-initiation (Figure 3). Measurable improvement in auditory function emerged within 6 months of beginning daily supplementation, with sustained effects observed through 24 months.

Figure 3.

Time course of auditory function changes following ACEMg initiation in the treatment group (N = 93). Mean OAE scores (percentage of 22 tested frequencies with detectable otoacoustic emissions) are shown at baseline (Year 3, immediately prior to ACEMg supplementation) and at 3-, 6-, and 9-months post-initiation. Higher percentages indicate better preserved OHC function. Red data points with numerical labels indicate mean OAE percentage at each time point. Measurable improvement in auditory function emerged by 6 months post-initiation, with continued improvement through 9 months. These short-term findings are consistent with the sustained improvement observed at 24-month follow-up (Year 5; see Figure 2). This analysis was exploratory and should be interpreted with appropriate caution

Discussion

Pharmaceuticals to treat SNHL are unavailable today. Hearing regeneration pharmaceuticals are years away, and general availability is unlikely as is access to cochlear implant devices.^20,21 Earplugs, earmuffs, noise-canceling headphones, and earbuds mitigate noise exposure, but they are only modestly effective, and unlikely to be continuously worn. Hearing aids employ audio engineering technologies to remediate hearing deficits but are either unaffordable or unused by the majority of those with hearing deficits. “For LMIC regions that make up 85% of the world population, [hearing aid] coverage ranges from 1.5% to 12%”. ²²

This study tested the potential of ACEMg to impact hearing preservation, relying on objective auditory function data from annual OAE examinations among audiology patients previously diagnosed with SNHL. The no-treatment group (N = 97) represented the current standard of care for SNHL. Analysis of real-world, no-treatment data validated the common understanding that hearing loss progresses over time.

OAE annual examination data from the 2 most recent years was available for those in the ACEMg treatment group (N = 93). This group started taking ACEMg immediately after the third-year OAE examination.

Annual OAE examination data among adults with SNHL is unusual. Conventional wisdom is that hearing loss is permanent; therefore, repeated OAE examinations are considered unwarranted if an OAE examination cannot detect OHC function, especially in speech range frequencies. The study hypothesized that ACEMg would be shown to preserve auditory function if OAE scores in the ACEMg treatment group declined less than OAE scores in the no-treatment group. OAE scores in the ACEMg treatment group markedly improved within 6 months of taking ACEMg daily, suggesting that ACEMg supplementation was associated with attenuation of SNHL progression. Moreover, ACEMg improved auditory function.

To our knowledge, based on available published literature, while these observational findings suggest a potential association between ACEMg supplementation and OAE outcomes, methodological limitations including non-randomized controls preclude confident conclusions about efficacy. The apparent association of ACEMg with higher OAE scores may reflect potential stabilization or recovery from temporary fluctuations in a self-selected population rather than therapeutic reversal of chronic pathology. These findings contrast with the natural history of SNHL progression reported in previous longitudinal studies that documented progressive hearing decline over time in observational cohorts, consistent with our no-treatment group.^18,19 To our knowledge, no previous clinical study has demonstrated improvement in objective OAE measures of OHC cochlear function in adults with established SNHL.

The observed increase in OAE scores in the treatment group should not be interpreted as evidence of hair cell regeneration, which remains biologically implausible in adult humans. Rather, if the association reflects a genuine treatment effect, which cannot be established from this observational design, the mechanism would likely involve metabolic rescue of hair cells that were functionally compromised but structurally intact. Such cells may have been under oxidative stress sufficient to impair their mechanoelectrical transduction and OAE production, yet not so severe as to trigger apoptosis. Antioxidant support could theoretically restore function in this subpopulation of stressed cells. Alternative explanations, including regression to the mean following temporary threshold shifts and concurrent adoption of hearing-protective behaviors, are equally consistent with the observed data.

While cochlear regeneration research shows promise in animal models, translation to humans remains distant. The observed association with higher OAE scores in the ACEMg treatment group, therefore, represent a potentially novel finding warranting replication in further RWE/RWD studies. If causal effects are confirmed in prospective trials, ACEMg could theoretically have implications for global hearing health given its oral formulation, relative affordability, and established safety profile. However, such extrapolation is premature based on the present observational findings. Similarly, while hearing loss is an established modifiable risk factor for dementia, we cannot extrapolate from the present OAE findings to cognitive outcomes without direct investigation.

The free radical theory of cell aging was first proposed as a major risk factor for disease, cell dysfunction, and the aging process in 1954. In theory, excess reactive oxygen species (ROS), byproducts of metabolism in cell mitochondria not removed by antioxidant defenses, could be expected to damage cell function. ²³ ACEMg applies the free radical theory of cell aging to inner ear pathology. SNHL pathogenesis is a consequence of increased demand for cochlear energy production, generating a vast influx of cochlear ROS that overwhelms the endogenous antioxidant system. Residual superoxides and singlet oxygen ROS interfere with redox reactions, causing lipid peroxidation that weakens membranes, alters cell proteins, and hastens cell aging. ROS penetrate the outer mitochondrial membrane, interfering with the function of other organelles. Finally, vasospasm induces mtDNA expression of abnormal proteins leading to cell dysfunction or cell death at worst, experienced as hearing loss. ²⁴ A consensus has emerged from ongoing research to treat SNHL with mitochondrial-targeted antioxidants, as SNHL pathogenesis and pathophysiology are consistent with pathologies in other organs and the brain, peripheral and central nervous system disorders, aging disorders, and environmental stress.^25,26

Each component in the ACEMg formula contributes a necessary mechanism of action resulting in a prophylactic composition that blocks the initiating events triggering SNHL. β-carotene (provitamin A) scavenges singlet oxygen. Singlet oxygen reacts with lipids to form lipid hydroperoxides; removing singlet oxygen prevents lipid peroxidation. ²⁷ Donor antioxidant vitamin E removes free radicals from the lipid compartments, also reacting with and reducing peroxyl radicals, inhibiting the spread of lipid peroxidation. Vitamin C detoxifies by scavenging oxygen radicals in the aqueous phase and blocks lipid peroxidation by free radicals that “escape” neutralization by vitamins A and E.^28-30 Supplemental magnesium reduces vasoconstriction and reperfusion injury, which attenuates SNHL. ³¹

In laboratory research, ACEMg demonstrated an unexpected beneficial synergistic effect, measurably reducing NIHL at 4, 8, and 16 kHz far better than treatment with antioxidants A, C, and E together, or treatment with magnesium Mg alone, or saline, the control. ACEMg increased the noise tolerance of cochlear cells by an average of 31 dB over the control, maintaining normal auditory function when sound pressure level increased by as much as 10 times, reducing hearing loss from noise by 75%. ¹⁰

Limitations

This study has several limitations that must be considered when interpreting the findings. Future prospective studies with contemporaneous control groups are needed to confirm these findings.

Observational Design and Causality

• 1. Selection bias: Patients who chose to use ACEMg may differ systematically from those who did not in ways that affect hearing outcomes (eg, higher health consciousness, greater resources, more motivated to preserve hearing). These unmeasured differences could confound the observed association between ACEMg use and hearing preservation.

• 2. Age Disparity and Control Validity: The significant age difference between groups (Treatment: M = 62.3 years vs Control: M = 75.6 years) represents an important limitation. While age was included as a covariate, ANCOVA assumes a linear relationship between age and outcome. However, non-linear presbycusis typically accelerates in the 8th decade compared to the seventh.^18,19 Therefore, the observed stabilization in the younger treatment group may partially reflect the naturally slower rate of age-related decline expected in a seventh-decade cohort compared to an eighth-decade cohort. The Year × Age interaction was non-significant, and results of the age-matched analysis confirmed age was not the determining factor. Nevertheless, given the study design limitations, our analyses should be interpreted with caution.

• 3. Healthy user bias: Healthy-user bias represents a significant concern in this observational study. Patients who elected to use ACEMg supplementation may have differed systematically from those who did not in multiple ways: greater health consciousness, higher socioeconomic status enabling supplement purchase, stronger motivation to preserve hearing, better overall health behaviors (diet, exercise, noise avoidance), and higher adherence to healthcare recommendations generally. These unmeasured characteristics could independently affect hearing outcomes through pathways unrelated to ACEMg itself. The observed association between ACEMg use and hearing preservation may therefore reflect, in part, the characteristics of individuals who choose to use supplements rather than a direct effect of supplementation.

• 4. Historical control bias (Temporal confounding): The no-treatment group was examined in an earlier time period (2015-2020) than the ACEMg treatment group (2020-2023). The use of historical rather than concurrent controls represent a significant methodological limitation. Systematic differences between these time periods, including changes in clinical protocols, audiological equipment, or patient demographics, could contribute to observed differences independent of ACEMg effects.

• 5. Regression to the mean: To evaluate whether observed Year 5 group differences could be attributed to regression to the mean, ANCOVA was conducted with Year 5 outcome scores as the dependent variable, treatment condition as the fixed factor, and Year 3 baseline scores entered as a continuous covariate. Because regression to the mean arises from imperfect test–retest correlation and baseline extremity, adjusting for baseline scores provides a direct test of whether post-treatment differences persist after accounting for initial standing. Baseline scores strongly predicted Year 5 outcomes (F_1,83 = 261.03, P < .001). After controlling for baseline, the treatment effect remained significant, (F_1,83) = 9.25, P = .003, partial η² ≈ 0.10). These findings indicate that group differences at Year 5 cannot be attributed to regression to the mean.

• 6. Equipment consistency: OAE measurements in both time periods were conducted using the same Interacoustics Titan system, which was calibrated annually according to manufacturer specifications. No equipment upgrades occurred during the study period.

• 7. Clinical practice changes: The same licensed audiologist (L.A. Halvorson) performed or supervised all OAE examinations throughout both study periods, minimizing inter-examiner variability.

• 8. Secular trends: We cannot exclude the possibility that broader environmental factors (eg, changes in noise exposure patterns, hearing protection awareness) differed between time periods.

The associations observed are consistent with a causal effect, but due to these limitations we cannot establish with total certainty that ACEMg caused the observed association with higher OAE scores in auditory function; alternative explanations cannot be completely ruled out.

Baseline Comparability

We note future prospective studies should systematically collect these variables at baseline.

• 1. PTA data: Pure tone audiometry (PTA) data were available for most participants; however, we selected OAE as the primary outcome measure because it provides objective physiological assessment of cochlear function independent of patient response behavior. PTA data showed similar patterns but are not reported here to maintain focus on the objective outcome measure.

• 2. Comorbidities: Detailed comorbidity data were not systematically documented in clinical records and could not be analyzed. This represents an unmeasured potential confounder.

• 3. Hearing aid use: Hearing aid use was permitted in both groups and was not an exclusion criterion. Rates of hearing aid use were not systematically documented and could not be compared between groups.

• 4. Noise exposure: Occupational and recreational noise exposure history was not consistently documented and represents an unmeasured potential confounder.

Generalizability and External Validity

The study was conducted at a single private audiology clinic in suburban Illinois, limiting generalizability in several ways:

• 1. Geographic and demographic constraints: Findings may not generalize to other geographic regions, healthcare settings (eg, academic medical centers, community health clinics), or populations with different demographic characteristics (eg, racial/ethnic diversity, socioeconomic status).

• 2. Clinic-specific factors: Patients at this clinic may have better access to audiological care, higher health literacy, or greater adherence to recommendations than the general population with SNHL.

• 3. Age distribution: Although the study included participants aged 19-100, the mean age was higher in the no-treatment group (75.6 years) than the treatment group (62.3 years). While age was statistically controlled as a covariate, residual confounding by age-related factors cannot be excluded.

• 4. Inclusion criteria: Requiring 5 years of complete OAE data selected patients with stable clinic relationships and good follow-up adherence. SNHL patients with less consistent healthcare engagement were excluded, potentially limiting applicability of findings to less adherent populations.

Measurement and Outcome Limitations

• 1. OAE as surrogate endpoint: OAE is an objective measure of OHC function. It does not directly measure hearing sensitivity or speech understanding – consequences of synaptopathy, hidden hearing loss involving the inner hair cell-to-auditory-nerve pathway to the brain, which also impact common consequences of OHC disfunction including tinnitus (ringing in the ears) and hyperacusis (intolerance to normal sound pressure levels). Further investigation of ACEMg and synaptopathy-related outcome measures appears to be warranted.

• 2. Lack of blinding: The audiologist performing OAE examinations was not blinded to treatment status. Although OAE is an objective physiological measurement, knowledge of treatment could theoretically influence technical factors (eg, probe placement, exam duration). However, this risk is mitigated by the standardized, automated nature of OAE testing.

Intervention and Adherence Limitations

• 1. Self-reported adherence: Actual consumption of the capsules or the exact timing of consumption cannot be independently verified.

• 2. Concomitant interventions: Interventions or lifestyle factors that could affect hearing were not controlled (eg, hearing aid use, noise exposure changes, dietary modifications, other supplement use).

• 3. Dose-response: The study examined only a single dose of ACEMg (2 capsules daily). The potential impact of alternative dosing levels was not studied.

Statistical and Analytical Limitations

• 1. Multiple comparisons: We corrected for multiple comparisons in post-hoc tests (Tukey HSD).

• 2. Sample size: Although adequate for detecting the observed effects, sample size (N = 190) limited our ability to conduct certain subgroup analyses (eg, by frequency range, by severity of baseline hearing loss) or to detect smaller effect sizes that might still be clinically meaningful.

• 3. Missing data handling: Our complete case analysis excluded patients with incomplete OAE series. If missing data was related to hearing outcomes (eg, patients with worse hearing were less likely to return for follow-up), this could bias results.

Data Quality and Documentation

• 1. Retrospective data extraction: Data were extracted from clinical records not originally designed for research purposes. Certain variables of potential interest (eg, detailed noise exposure history, comorbidities, medication use) were not systematically documented and could not be analyzed.

• 2. Historical controls: The use of historical rather than concurrent controls raises concerns about equivalence of care, measurement procedures, and population characteristics between time periods.

Unmeasured Confounding

Even with statistical adjustment for age and demonstration of baseline comparability in OAE scores, numerous unmeasured variables could confound the observed associations. These include overall health status, socioeconomic factors, educational level, occupational noise exposure, cardiovascular health, diabetes, smoking status, dietary patterns, stress levels, and genetic susceptibility to hearing loss. The absence of randomization means these potential confounders may be unevenly distributed between groups. Baseline demographic and clinical characteristics are summarized in Table 3.

Table 3.

Baseline Demographic and Clinical Characteristics of Study Participants

Characteristic	ACEMg group (N = 93)	No-treatment group (N = 97)	Statistical test	P-value
Demographics
Age (years)
Mean (SD)	62.3 (15.84)	75.6 (18.32)	F1,188 = 28.15	<.001
Range	19-88	22-100	F96,92 = 1.34 a	n.s.
Sex, N (SD)
Female	40 (43.0%)	54 (55.7%)	X21,190 = 2.96	n.s.
Male	53 (57.0%)	43 (44.3%)
Baseline auditory function (pre-intervention years)
Mean (SD)
Year 1 OAE score (SD)	38.27 b (28.81)	41.281 b (33.38)	F1,188 = 0.44	n.s.
Year 2 OAE score (SD)	32.26 b (28.33)	38.64 b (32.81)	F1,188 = 2.04	n.s.
Year 3 OAE score (SD)	31.28 b (29.17)	33.08 b (30.85)	F1,188 = 0.68	n.s.
Rate of decline (Years 1-3)
Change in OAE (%)	−6.99	−8.20	F1,188 = 0.58	n.s.

OAE Score = Percentage of 22 tested frequencies (11 frequencies × 2 ears) with detectable otoacoustic emissions. Range: 0-100%, with higher percentages indicating better auditory function.

Study Design: The treatment group began ACEMg supplementation immediately after Year 3 examination. The no-treatment group was examined during an earlier time period (2015-2020) before ACEMg became available.

^aLevene’s test for equality of variances - indicates age distribution is statistically equivalent between groups despite different means.

^bCombined mean - The manuscript reports overall means for both groups combined at Years 1 and 3, with F₁,₁₈₈ = 0.50 (n.s.) indicating no significant difference between groups at baseline.

The clinical significance of OAE percentage changes warrants consideration. While no validated minimum clinically important difference (MCID) exists for OAE percentage outcomes, we note that our primary comparison grouped the outcomes categorically (improved vs stable vs declined) rather than relying on continuous percentage changes. The categorical approach avoids assumptions about clinical meaningfulness of specific percentage thresholds. Furthermore, OAE provides information about cochlear OHC function at specific frequencies; clinical interpretation depends on which frequencies are affected and their relevance to speech communication. Future research should establish clinically meaningful thresholds and correlate OAE changes with patient-reported outcomes and speech understanding measures.

Despite these limitations, this study provides the first real-world evidence that daily ACEMg supplementation is associated with preserved and improved auditory function in patients with SNHL. The consistency of effects across multiple time points, the dose-dependent time course (effects emerging at 6 months), and the biological plausibility based on mechanistic research strengthen confidence in the findings.

Future Directions

Well-designed longitudinal RWE studies may produce more valid and generalizable evidence than traditional RCTs for SNHL preventive care interventions such as ACEMg, due to SNHL’s heterogeneous etiologies and variable progression, the gap between efficacy and effectiveness, and the current absence of approved pharmacological treatments.

SNHL encompasses a wide range of etiologies, from noise and aging to ototoxic, genetic, infectious, and idiopathic. RCT participants are typically healthier, more adherent, and more motivated than the general SNHL population. Strict inclusion/exclusion criteria create artificial patient populations whose results may not generalize to the heterogeneous patients seen in clinical practice. An RCT that restricts enrollment to 1 etiology would limit generalizability. An RCT that attempts to stratify across etiologies would require enormous, impractical sample sizes. Neither approach is scientifically satisfying or economically feasible.

RCT measure efficacy (is it better than the existing treatment under controlled conditions?) RWE measure effectiveness (does it work in nonrandomized populations?) Historically, there is typically a large gap between RCT efficacy and RWE effectiveness; effectiveness is what actually matters for preventive care generally, and ACEMg specifically. RWE has several advantages over RCT for SNHL research, including lower cost, larger sample sizes, longer study duration, population diversity, and inclusion: no patient is denied a potentially beneficial intervention. Notably, the FDA increasingly accepts RWE findings for label claims subsequent to passage of the 21st Century Cures Act.

Given SNHL characteristics and RCT limitations above, prospective longitudinal RWE study designs for ACEMg, with pre-registered hypotheses and protocols may provide more generalizable evidence than traditional RCTs for a chronic progressive condition like SNHL. We emphasize that the appropriate comparator for ACEMg is standard audiological care, not placebo, as no approved pharmacological treatments for SNHL currently exist. The present study’s historical control group, which received standard care without supplementation, represents this clinically relevant comparison. The ACEMg Otology Intervention Study (OTIS) was designed as a pilot to address several of these priorities, assessing tinnitus and hyperacusis outcomes in a prospective framework.

Conclusion

This retrospective observational study provides preliminary real-world evidence suggesting an association between daily ACEMg supplementation and preservation or improvement of objective auditory function in adults with sensorineural hearing loss. While these findings are consistent with the hypothesized mechanism of action, the observational design precludes causal inference. The results should be interpreted as hypothesis-generating, providing justification for prospective controlled studies to confirm whether ACEMg supplementation has a causal effect on auditory function. Given the study limitations, including historical control bias, self-selection, and unmeasured confounding, replication in independent samples with contemporaneous controls is warranted before drawing clinical practice or policy implications.

ORCID iDs

Barry S. Seifer https://orcid.org/0009-0000-2620-6785

Louise A. Minor https://orcid.org/0009-0003-3144-0412

Data Availability Statement

The dataset generated during and/or analyzed during the study are publicly available on the Open Science Framework at https://osf.io/9xm4j/.