The association of sex steroid hormone concentrations with hearing loss: a cross-sectional study

Abstract Background Hearing loss is the most prevalent sensory disorder worldwide. Several studies have indicated that sex steroid hormone levels may be vital to hearing. Objective We aimed to explore the associations between speech-frequency hearing loss and sex steroid hormones. Methods We conducted a secondary analysis based on 3558 adult participants’ data from the National Health and Nutrition Examination Survey (NHANES) from 2015 to 2016. We defined hearing loss as a pure-tone average (PTA) at 0.5, 1, 2, and 4 kHz ≥20 dB in the better ear. Multivariate logistic regression analysis was used to evaluate the association between sex steroid hormones and hearing loss risk. A nomogram model for the risk of hearing loss was constructed. Results There were 560 (15.7%) cases who had hearing loss among the participants enrolled in this study. Participants with hearing loss had a higher total testosterone level and a lower estradiol level. Individuals with estradiol levels in the highest tertile still had lower hearing loss risks than those in the lowest tertile. Nevertheless, the total testosterone level had no influence on the risk of hearing loss. Conclusion Our research indicated that low estradiol concentrations were significantly associated with hearing loss, especially in menopausal women.


Introduction
Hearing loss is the most prevalent sensory disorder according to the WHO 2021 World Report on Hearing [1]. Approximately 1.6 billion people worldwide live with some degree of hearing loss. This number is expected to reach nearly 2.5 billion by 2050. The Global Burden of Disease Study reported that hearing loss was the fourth leading cause of disability worldwide [2]. Hearing loss in adults primarily influences communication, which can create difficulties in interpersonal relationships. Hearing loss-related costs and unemployment have also increased year by year. This undoubtedly places a huge burden on society. Studying risk factors and preventing and intervening in hearing loss is a preferential and essential strategy.
Sex steroid hormones include estrogens and androgens, whose concentrations and functions are different in men and women. Traditionally, estrogens and androgens were considered to be female and male hormones, respectively. They were also shown to be multitasking hormones that played prominent roles in libido and reproduction and contributed to physiological and growth mechanisms [3,4]. It has been reported that estradiol protected against noise-induced hearing loss in female mice [5]. Polycystic ovary syndrome patients with increased testosterone levels had a higher rate of hearing loss [6]. Those studies indicated that sex steroid hormone levels may be vital to hearing. Therefore, we extracted hearing condition data from the National Health and Nutrition Examination Survey (NHANES). This study aimed to explore the associations between sex steroid hormones and speech-frequency hearing loss.

Research population
Research data came from the NHANES survey. NHANES was a stratified, multistage probability survey for the noninstitutionalized US population. The unique feature of this survey was that it combined interviews and physical examinations. Health interviews were conducted at the interviewee's home, and physical examinations were carried out in a specially designed and equipped mobile center, located all over the country (http://www.cdc.gov/nchs). The research team consisted of a doctor, a dentist, medical and health technicians, and diet and health interviewers.
The participants in this research were enrolled from the NHANES cycle 2015-2016, which contains both sex steroid hormone and audiometry examination data in adults aged 20-69 years old. Figure 1 shows a flow chart for participant selection in this study. A total of 9971 potential participants were exacted from the dataset. Participants missing audiometry data (n = 5658) or missing total testosterone or estradiol (n = 269) were excluded. In addition, 486 participants missing body mass index (BMI), race, hypertension or diabetes were also excluded. Eventually, a total of 3558 adults were enrolled in this study for further analysis. All data could be obtained on the internet and were exempt from ethical review.

Audiometric measurement
The detailed procedures and protocols of the audiometric examinations were described in the NHANES online manual (https://wwwn.cdc.gov/Nchs/Nhanes/2015-2016/TST_I.htm). Briefly, all audiometry examinations were performed by trained examiners on participants in a dedicated soundisolation room in the mobile examination center. Hearing threshold testing was conducted on both ears of participants at seven frequencies (0.5, 1, 2, 3, 4, 6, and 8 kHz). Hearing loss was defined as pure-tone average (PTA) at 0.5, 1, 2, and 4 kHz ≥20 dB in the better ear according to The World Report on Hearing published by the World Health Organization in 2021 [1]. Hearing loss is categorized as mild (20 to <35 dB), moderate (35 to <50 dB), moderately severe (50 to <65 dB), severe (65 to <80 dB), profound (80 to <95 dB) and complete (95 dB or greater).

Sex steroid hormones and SHBG measurements
Serum samples were processed, stored, and shipped to the Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA for analysis (https://wwwn.cdc.gov/ Nchs/Nhanes/2015-2016/AUX_I.htm). Total testosterone and estradiol in serum were measured. Sex hormone binding globulin (SHBG) was measured with a chemiluminescence measurement based on the reaction of SHBG with immune antibodies.

Covariates
We included the covariates on demographic characteristics (age, gender, race, and BMI), personal medical history (hypertension and diabetes), serum lipid level (total cholesterol and triglycerides), lifestyle habits (smoke and alcohol status), and noise exposure (firearm, occupational and off-work noise).

Statistical analysis
Normally distributed variables are presented as the mean ± standard deviation (SD) and were analysed with independent sample Student's t test. Nonnormally distributed variables are presented as medians with interquartile ranges (IQRs) and were compared by nonparametric Mann-Whitney U tests. Categorical variables are shown as numbers and percentages and were assessed by the chi-square test. We distributed the lg-transformed sex steroid hormone concentrations into tertiles to facilitate interpretation.
The associations between sex steroid hormone concentrations and hearing loss were assessed by univariate and multivariate logistic regression analyses. Model 1 was adjusted for age, gender, BMI and race. Model 2 was adjusted for age, gender, BMI, race, hypertension, diabetes, triglycerides, smoking status, and noise exposure (firearm, occupational and off-work noise). The odds ratio (OR) with 95% confidence intervals (CI) was calculated. Any nonlinear relationship between estradiol concentrations and hearing loss was explored by a restricted cubic spline regression model with 4 knots at the 5 th , 35 th , 65 th , and 95 th percentiles of estradiol levels. The variables in Model 2 were used to construct the nomogram. The receiver operating characteristic (ROC) curve and calibration curve were performed to assess the nomogram. An interactive web-based dynamic nomogram application was built with Shiny. Subgroup analyses stratified by age, gender, BMI, hypertension, diabetes, serum triglyceride level and smoking status were conducted to determine the associations between sex steroid hormones and hearing loss.
All computations were conducted in the R environment (version 4.2.1). A P value < 0.05 was considered statistically significant.

Descriptive analyses
Among the 3,558 participants in this study, 560 (15.7%) had hearing loss. The characteristics of all participants are shown in Table 1 according to their hearing status. There were significant differences between the participants with and without hearing loss and non-hearing in age (p < 0.001), gender (p < 0.001), race (p < 0.001), BMI (p < 0.001), hypertension Figure 1. Flow chart of participant selection. *the participants outside the age limits of 20-69 years did not detect audiometry data. the participants using hearing aids who were not able to remove them for testing and participants who had sufficient ear pain at the time of the exam that they could not tolerate headphones were also excluded from the audiometry component.
(p < 0.001), diabetes (p < 0.001), serum triglycerides level (p < 0.001), smoking status (p < 0.001), firearm noise exposure (p = 0.016), occupational noise exposure (p < 0.001), and off-work noise exposure (p < 0.001). Regarding sex steroid hormone concentrations, the hearing loss group had a higher level of total testosterone (p = 0.001) and a lower level of estradiol (p < 0.001). Higher severity degree of hearing loss was significantly associated with elevated total testosterone levels (p = 0.008) and decreased estradiol levels (p < 0.001) ( Figure S1). However, there were no significant differences between the two groups in SHBG concentrations.

Association between sex steroid hormones and hearing loss
In our logistic regression analysis, we converted lg-transformed total testosterone and estradiol concentrations from a continuous variable to a categorical variable (tertiles). We also used continuous variables of lg-transformed total testosterone and estradiol concentrations to calculate the linear trend. In the crude model, when compared with the reference of the lowest tertile, the ORs with 95% CIs for hearing loss across the highest tertile were 1.487 (1.192-1.857) for total testosterone and 0.275 (0.212-0.355) for estradiol ( Table 2). The P value for trend showed that total testosterone levels were positively associated with hearing loss, while estradiol levels were negatively related to hearing loss. In the partially adjusted model (Model 1, adjusted for age, gender, BMI and race) and fully adjusted model (Model 2), individuals with estradiol levels in the highest tertile still had lower risks of hearing loss than those in the lowest tertile. However, no significant P value for trend was observed among tertiles of total testosterone levels and hearing loss in both adjusted models (all p > 0.05). A restricted cubic spline regression model indicated that the association between estradiol levels and hearing loss risk was not nonlinear (p > 0.05).  The risk of hearing loss was increased at estradiol concentrations below 24.5 pg/mL and decreased at estradiol concentrations above 24.5 pg/mL ( Figure 2).

Subgroup analysis
As shown in Table 3, we stratified by age, gender, BMI, hypertension, diabetes, triglycerides and smoking status to determine the effect size of these factors on hearing loss. At the highest tertile, female sex, age from 40 to 60 years old, higher BMI and smoking status, and no diabetes had modified effects on the association between hearing loss and estradiol concentrations when compared with the lowest tertile. Interestingly, participants without hypertension were more likely to have a hearing loss risk in the second tertile of estradiol. Considering the differences in sex steroid hormone levels between the genders, we also stratified by gender to determine the effect size of this factor on hearing loss. However, in adjusted models, no significant P value was observed among tertiles of total testosterone levels and hearing loss in the gender-specific participants (Table S1). The scattergrams with total testosterone/estradiol levels and hearing threshold in men and women separately are shown in Figure S2.

Nomogram model
The nomogram model was drawn to predict the risk of hearing loss according to the multivariate logistic regression Figure 2. the restricted cubic spline of association between the estradiol concentration and the risk of hearing loss. the model was adjusted for age, gender, race, BMi, hypertension, diabetes, triglycerides, noise exposure (firearm, occupational, recreational, and off-work noise) and smoking. the solid red lines and shaded areas represent the lg-transformed odds ratios and corresponding 95% confidence intervals. analysis model (Model 2) (Figure 3(A)). The following equation was obtained based on these results, i.e. Risk of hearing loss = −6.5577 + 0.0968*Age +0.8483* (Gender = Male) +0.2403*BMI −0.1484*Race (1= Mexican American; 2 = Other Hispanic; 3 = Non-Hispanic White; 4 = Non-Hispanic Black; 5 = Other Race) −0.0879* (Hypertension = yes) +0.4088* (Diabetes = yes) +0.0086*Log 2 (Triglycerides) −0.1672* (Firearm noise = yes) +0.1273* (Job noise = yes) +0.5374* (Off-work noise = yes) +0.2631* (Smoke = yes) −0.3204*Lg (Estradiol). An available dynamic online nomogram application was also constructed, as shown in Figure  3(B). The ROC curve of the nomogram is presented in Figure 4(A), and its area under the curve (AUC) was 0.845. The corresponding calibration plot was also drawn, which demonstrated a good correlation between the observed and predicted hearing loss risk with a mean absolute error of 0.003.

Discussion
This cross-sectional study explored the relationships between sex steroid hormone concentrations and speech-frequency hearing loss in US adults using nationally representative data. The results obtained in this study support the findings of recently published studies [4,5]. There are several main findings of this research. First, estradiol was found to be negatively associated with the risk of speech-frequency hearing loss in a large sample aged 20-69 years old. Second, this effect was especially significant in menopausal women. Third, although total testosterone levels were found to be positively related to hearing loss, there was no significant influence on the risk of hearing loss after adjusting for confounders.
Hearing loss is a major public health problem that contributes to decreased quality of life. In the matter of acquired hearing loss, there are several risk factors for hearing loss, including age, BMI, drugs, noise exposure, infections, nutrition, trauma and so on [7]. Hypertension, diabetes and hyperlipemia were also shown to increase the risk of hearing loss [8][9][10]. Our study also demonstrated that hearing loss was significantly related to age, BMI, hypertension, diabetes, hyperlipemia and noise exposure. This means the participants enrolled in this study are representative. Additionally, the prevalence rate of hearing loss between males and females was shown to be very different [11]. We thought that sex steroid hormones may be involved in the pathogenesis of hearing loss. Therefore, we explored the relationship between sex steroid hormones and gender.
In a previous study, Kim et al. reported that postmenopausal women with lower levels of serum estradiol had a higher hearing loss rate [12]. Similarly, postmenopausal women treated with estradiol had consistently lower air conduction thresholds than those who did not receive treatment [13]. The results from Béatrice et al. identified biological sex as a critical factor that influenced the severity of hearing loss [14]. However, male and female mice were also shown to be vulnerable to noise exposure when they lacked endogenous gonadal hormones. Estradiol replacement therapy ameliorated threshold shifts in gonadectomized female mice [5]. These effects were realized by reducing the loss of outer hair cells and pairing synapses. In our subgroup analysis, the protective effects of estrogen seemed to be more significant in women aged 40 to 60 years. This was consistent with the results of previous studies.
On the other hand, patients with polycystic ovary syndrome showed high levels of testosterone and had significantly higher high-frequency hearing loss than control women [6]. Excessed androgen was shown to exert proatherogenic effects on macrophage function by facilitating the uptake of modified lipoproteins [15]. Atherosclerotic occlusion in the cochlear arteries or arterioles was shown to be an important cause of hearing loss [16]. Meanwhile, antiandrogen treatment protected against kanamycin-induced hearing loss in rats by downregulating megalin expression and reducing the inflammatory response and apoptosis level [17]. The total testosterone level was positively associated with hearing loss in this study. It was also shown to increase the risk of hearing loss. However, this association was no longer statistically significant after adjusting for confounders. These results indicate that the influence of testosterone on hearing may have been regulated by multiple factors. Further studies of the mechanism are needed.
Interestingly, our result indicates that the morbidity of hearing loss is lower in blacks than in whites. This is consistent with the results from the research of Lin et al. [18]. They also demonstrated an association between certain Fitzpatrick skin phototypes and lower hearing loss rates among Hispanics, which they suggested might be related to skin pigmentation. On the other hand, estradiol and progesterone were reported to reciprocally regulate skin pigmentation through nonclassical membrane-bound receptors [19]. Compared with Caucasian women, African-America women had higher estradiol levels, due to increased ovarian aromatase activity and expression [20]. This may be one reason why blacks have less hearing loss than whites. There were some limitations in our study. First, this research was designed as a cross-sectional study based on data from the NHANES. Although the number of participants enrolled in that database was large and nationally representative, the results could only infer correlation, not causality. Longitudinal and well-designed follow-up studies are needed. Second, all the participants were from the United States, and more surveys from different countries are needed to avoid the effects of geographical drift. Third, although we adjusted the confounders related to hearing as much as possible, unmeasured or unknown covariates may have affected the validity of the conclusions. Fourth, the hearing measurement methods used in the NHANES included pure tone audiometry, which is a standard approach but not the best. Finally, a mechanistic study of the effects of estradiol deficiency on hearing loss is necessary.

Conclusion
In the current study, we observed a negative relationship between estradiol levels and speech-frequency hearing loss in adults aged 20-69 years old, especially in menopausal women. However, there was no statistically significant association between total testosterone levels and the odds of hearing loss. Further studies are required to confirm these findings.