Global, regional, and national burden of age-related hearing loss from 1990 to 2019

The global burden due to ARHL

Globally, the ARHL patients increased from 751.50 [95% uncertainty interval (UI): 716.45, 787.33] million in 1990 to 1456.66 (95% UI: 1395.61, 1519.30) million in 2019, and the age-standardized prevalence rate (ASPR) showed a slight upward trend from 17.33 to 17.76 per 100 during the period, with an estimated annual percentage change (EAPC) of 0.08 [95% confidence interval (CI): 0.07, 0.10]. The estimated disability-adjusted life years (DALYs) of ARHL worldwide increased from 22.01 (95% UI: 14.91, 31.34) million to 40.24 (95% UI: 27.39, 57.13) million over the past 30 years, with a stable age-standardized DALY rate (ASDR) of about 0.005. The number of males with ARHL was 742.04 (95% UI 711.33, 774.94) million in 2019, accounting for 50.9% of all ARHL patients, causing 20.17 (95% UI 13.71, 28.75) million DALYs. Although women had lower ASPR and ASDR than men, female ASPR rose faster and ASDR fell more slowly (Table 1). Most patients were between 45–74 years old, and the burden of disease in this age group was also heavier (Figure 1).

Figure 1. Prevalent cases, DALYs and the corresponding rates of age-related hearing loss by sex, age group, and severity in 2019. (A) Prevalent cases and prevalence rate; (B) DALYs and DALY rate. Abbreviation: DALY: disability-adjusted life year.

Variation in ARHL burden at the national and regional level

The global variety of ASPR and ASDR of ARHL was around 2.4 and 3.5 times in 2019, respectively, with the highest ASPR (21.35/100) and ASDR (6.77/1000) in Kenya, and the lowest ASPR (8.75/100) and ASDR (1.96/1000) in Sweden. Overall, the ASPR in 2019 was higher than 17/100 in 98 countries and territories, including China, India, South Africa, etc., (Figure 2A, Supplementary Table 1), which also showed a severe burden in ASDR (Figure 2B, Supplementary Table 1). During the period from 1990 to 2019, 24 countries and territories showed an upward trend in ASPR, including China and the USA. ASDRs in most countries and territories showed a downward trend. Only a very small number of countries and territories, such as Niger and Burkina Faso, show an upward trend both in ASPR and ASDR (Figure 2C, Figure 2D, Supplementary Table 1, Supplementary Table 2). The number of DALYs and prevalent cases was high in high-middle and middle socio-demographic index (SDI) regions and increased in all SDI regions compared with 1990. ASPR and ASDR in other SDI regions showed a downward or stable trend, excluding the rising trend of ASPR in high and high-middle SDI regions (Table 1, Supplementary Figure 1). Except for the ASPR in East Asia, which showed an upward trend, the ASPR and ASDR in other GBD regions showed a downward or stable trend (Supplementary Table 2).

Figure 2. Global distribution of ASRs and the corresponding EAPCs in age-related hearing loss. (A) ASPR in 2019; (B) ASDR in 2019; (C) EAPC of ASPR from 1990 to 2019; (D) EAPC of ASDR from 1990 to 2019. Abbreviations: ASR: age-standardized rate; EAPC: estimated annual percentage change; ASPR: age-standardized prevalence rate; ASDR: age-standardized DALY rate; DALY: disability-adjusted life year.

ARHL burden due to severity

Compared with 1990, the number of DALYs and prevalent cases of ARHL of all severity in 2019 has increased significantly. Among them, the number of DALYs and prevalent cases of mild ARHL was the highest, 10.28 (95% UI: 4.57, 19.82) million and 1083.06 (95% UI: 1029.88, 1136.29) million respectively, which almost doubled from 1990. The ASPR and ASDR of all other severity ARHL were lower than 0.1/1000 and 5/100, and both showed a downward trend, except for mild ARHL. The ASDR and ASPR of mild ARHL were 0.12/1000 and 13.09/100 in 2019, respectively, and both showed an upward trend, with EAPC of 0.18 and 0.18 respectively (Table 1). In terms of age distribution, people over 75 years old account for a larger proportion of moderate/moderately severe/severe ARHL, and people under 49 years old account for a larger proportion of profound/complete ARHL (Supplementary Figure 2). Except for the DALY rate of mild ARHL and the moderately severe ARHL of the age group over 70 years old, which showed a slight increase, the DALY rate of ARHL of each severity of all age groups showed a downward trend or remained unchanged (Supplementary Figure 3A). Except for the increase in DALY rate attributable to occupational noise for males aged 15–49 in the high/high middle SDI region, the DALY rate attributable to occupational noise of males in all age groups and SDI regions was decreasing. The DALY rate attributable to occupational noise for females aged 15–49 in all SDI regions and females over 50 in high/high-middle SDI regions was increasing (Supplementary Figure 3B).

The influential factors for EAPC

Figure 3 shows the impact of the ASDR in 1990 and SDI in 2019 on ASDR trends in each region. There was no significant correlation between ASDR of ARHL in 1990 and EAPC of ASDR from 1990 to 2019 at the national level (ρ = 0.0529, P = 0.4516), indicating that ARHL burden may not be given priority intervention in countries with high burden rates. Compared with other countries, some countries with a medium disease burden, such as Egypt, Iran, had a faster decline in disease burden (Figure 3A). Similarly, the temporal trend of ARHL related ASDR from 1990 to 2019 and SDI in 2019 still seemed to have no significant associations at the national level (ρ = −0.0287, P = 0.6391). Some countries in North Africa and Middle East had the most obvious downward trend, despite their varying SDI in 2019 (Figure 3B). In terms of ASPR, we found a positive correlation between the ASPR in 1990 and the EAPC of ASPR from 1990 to 2019 at the national level (ρ = 0.2437, P = 0.0004) (Supplementary Figure 4). Further, we investigated the correlation between SDI and ASDR by severity from 1990 to 2019 in 21 GBD regions around the world. The results showed that, except for the burden of mild ARHL, which showed a significant upward trend with the increase of SDI, the ASDR of ARHL in other severity was significantly negatively correlated with SDI in the 21 GBD regions (Supplementary Figure 5).

Figure 3. The association between ASDR in 1990, SDI in 2019, and the EAPC of ASDR from 1990 to 2019. (A) ASDR in 1990 and the EAPC of ASDR from 1990 to 2019; (B) SDI in 2019 and the EAPC of ASDR from 1990 to 2019. The blue line was an adaptive association fitted with adaptive Loess regression based on all data points. Abbreviations: EAPC: estimated annual percentage change; SDI: socio-demographic index; ASDR: age-standardized DALY rate; DALY: disability-adjusted life year.

The influential factors for ARHL burden attributable to occupational noise

Among 204 countries and territories, the ASDR attributable to occupational noise in most countries in 2019 was lower than that in 1990 and countries with increased disease burden due to occupational noise were mostly in East Asia and Latin America and Caribbean (Figure 4A). With the increase in SDI in 2019, the ASDR attributed to occupational noise in 2019 was declining at the national level (ρ = −0.7662, P < 2.2e-16), which reflected the countries with higher SDI levels may be better able to reduce the burden of ARHL (Figure 4B). Similar results were observed in the aspects of the proportion of ARHL burden attributable to occupational noise (Supplementary Figure 6).

Figure 4. The relationship between ASDR due to occupational noise in 1990, SDI in 2019 and ASDR due to occupational noise in 2019 in age-related hearing loss. (A) ASDR due to occupational noise in 1990; (B) SDI in 2019. The blue line was an adaptive association fitted with adaptive Loess regression based on all data points. Abbreviations: ASDR: age-standardized DALY rates; SDI: socio-demographic index; DALY: disability-adjusted life year.

Study data

In our study, we collected annual prevalent cases, ASPR, DALYs and ASDR of ARHL by 5-year age groups, genders, severity, regions, and risk factors from 1990 to 2019 from the GBD 2019 database via the Global Health Data Exchange (GHDx) query tool (http://ghdx.healthdata.org/gbd-results-tool). Hearing loss was defined as the average intensity of the softest sound that can be heard in one or both ears at a frequency of 0.5, 1, 2, and 4 kHz equal to or greater than 20 dB. According to the average intensity, hearing loss can be divided into mild (threshold in 20–34 dB), moderate (threshold in 35–49 dB), moderately severe (threshold in 50–64 dB), severe (threshold in 65–79 dB), profound (threshold in 80–94 dB) and complete (threshold more than 95 dB). The data used in this study excluded hearing loss caused by congenital, meningitis, and otitis media. The occupational noise exposure was defined as the proportion of the population occupationally exposed to 85+ decibels of noise based on population distributions across 17 economic activities, such as agriculture, hunting, forestry, fishing, mining and quarrying. The burden of ARHL for occupational noise was estimated using the population attributable fraction (PAF), which requires information on the relative risk of hearing loss due to occupational noise and occupational noise exposure levels. The relative risks were obtained from published meta-analyses or pooled studies and the formula of PAF was:

$$PAF=\frac{{\displaystyle {\sum }_{x=1}^{n}RR(x)P(x)-1}}{{\displaystyle {\sum }_{x=1}^{n}RR(x)P(x)}}$$

where P(x) is the proportion of population exposed to occupational noise at level x in the target population and RR(x) is the relative risk of occupational noise exposure level x [23, 24]. In the GBD database, a total of 204 countries and territories were included and were classified into 5 regions according to their SDIs, which were calculated by combining educational attainment, total fertility rate, and the lag-distributed income per capita. Further, the 204 countries and territories were divided into 21 GBD regions based on geographical proximity and epidemiological similarity, and further simplified into seven Super GBD regions.

Previous studies have introduced the method of disease burden estimation detailly. In short, DisMod MR 2.1 was used to estimate the ARHL disease burden based on published literature and cross-sectional studies. The exposure to risk factors and its attributable burden were quantified by GBD comparative risk assessment [1, 19, 20]. The age-standardized rates (ASRs) were estimated according to the world population by the GBD study. Relevant data were reported in numbers and 95% UIs, which were determined by 2.5% and 97.5% of the ordered 1000 estimates.

Statistical analysis

We applied ASPR, ASDR, and EAPC to quantify the trends of ARHL disease burden by age, gender, regions, severity, and risk factors from 1990 to 2019. Standardization is important for this study because it can avoid the difference in age compositions of different groups even of the same population in different periods. The EAPC, a widely used measure that summarized the ASR trends in a specified time interval, was calculated to describe the temporal trends of ASRs of ARHL burden. We put ASR in the regression line model “ln (ASR) = α + β × calendar year + ɛ”. The calculation formula of EAPC is 100 × (exp (β)-1). The 95% CI for EAPC is also generated from this model. If both the EAPC estimate and the lower boundary of its 95% CI are greater than 0, the ASR is considered to be on the rise. Conversely, if the upper boundary of the EAPC estimate and its 95% CI is less than 0, then ASR is considered to be a downward trend. Otherwise, ASR is considered stable over time [25–27]. In order to explore the influencing factors of EAPC, we used Spearman rank test to evaluate the relationship between EAPC and ASR (1990) and SDI (2019) respectively at the national level. ASR (1990) reflects the initial disease burden, and SDI (2019) can replace the socioeconomic level and availability of medical services in different countries/territories [25–27]. R program was used in all statistical analyses of our study (version 4.0.3; https://www.R-project.org/), and the P-value on both sides of less than 0.05 was considered statistically significant.

Data sharing statement

To download the data used in these analyses, please visit the Global Health Data Exchange GBD 2019 website: http://ghdx.healthdata.org/gbd-results-tool.

Ethical statement

Our study protocol was approved by the Institutional Review Boards of Qilu Hospital of Shandong University with approval number KYLL-202011(KS)-239.