Global burden for blindness and vision loss among women of childbearing age and perimenopause: age-period-cohort analysis 2021

Abstract

Blinding eye disease is a major threat to global public health, and adult women represent a significant portion of this population. This study aimed to characterize the prevalence and trends of blinding eye diseases and vision loss in women of childbearing age (WCBA) and Perimenopause (PMW), to provide important information for the future development of more effective preventive and interventional measures for eye diseases in women. In 2021, the global prevalence of BVL disease in WCBA and PMW was 333 million. Forty countries and regions have more than one million people with BVL disease in both WCBA and PMW. India, China, Brazil, and Indonesia are the top 4 countries with more than ten million people. The BAPC model predicts that there will be 800 million cases of BVL in WCBA and PMW worldwide by 2050. This study comprehensively analyzed global trends in the burden of BVL disease in WCBA and PMW between 1992 and 2021. The results show a significant increase in the prevalence and burden of BVL, particularly concentrated in older female populations.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-41588-y.

Introduction

The increasing prevalence of blinding eye diseases globally has become a significant public health concern. According to the World Health Organization (WHO), by 2020, an estimated 43.3 million people were expected to be blind, and an additional 295 million were projected to experience varying degrees of visual impairment¹. Notably, existing studies have shown that the global incidence of blindness and vision loss (BVL) in women is higher than in men, highlighting the necessity of focusing on women’s eye health^2–4. Women of childbearing age (WCBA) and perimenopausal women (PMW) represent an important subset of the female population, as they are in their prime working and career years. Vision problems in this age group can severely impact mental health, social participation, and family functioning, subsequently affecting societal economic productivity^5,6.

The causes of visual impairment in this group are multifactorial, mainly including uncorrected refractive errors, cataracts, glaucoma, age-related macular degeneration (AMD). Although numerous vision problems are preventable or treatable, many women fail to receive timely and effective interventions due to limited medical resources and healthcare services, particularly in low and middle SDI countries⁷. Cultural and socio-economic barriers further hinder their access to eye care. Additionally, fluctuating and disrupted hormone levels during the reproductive years, particularly during perimenopause, render women more susceptible to various eye diseases, including cataracts, glaucoma, and dry eye^8–11.

To date, no studies have systematically assessed the prevalence and trends of vision loss in WCBA and PMW. Furthermore, systematic analyses of the prevalence trends of vision loss during the reproductive and perimenopausal periods, and their associations with age, menstrual cycles, and birth cohort effects, are lacking across countries of all SDI levels. This study aims to analyze the prevalence and trends of vision loss in women of reproductive and perimenopausal age globally, utilizing age-period-cohort (APC) models with epidemiological data from the latest Global Burden of Disease (GBD) 2021 database. Through this study, we aim to provide health policymakers and public health experts with reliable data to inform the development of more effective preventive and intervention strategies for women’s eye diseases, reduce the incidence of visual impairment, and enhance the quality of life and social participation of women of reproductive and perimenopausal age.

Methods

Data sources and disease definition

The Institute for Health Metrics and Evaluation (IHME) has released the GBD 2021 study, offering a comprehensive and current epidemiological analysis of the disease burden associated with 371 diseases and injuries across 204 countries and territories. The data, which can be accessed freely at GBD 2021 (https://ghdx.healthdata.org/gbd-2021/sources), includes detailed information on data sources, statistical modeling, and methodologies from previous reports.

The GBD data sources were compiled with meticulous care through a rigorous process of identification and extraction. This systematic integration drew from a wide range of authoritative sources, including government and international websites, peer-reviewed reports, and significant data contributors. The research covered a broad spectrum of health data, such as censuses, household surveys, vital records, disease registries, health service usage, environmental air quality data, satellite imagery, and disease reporting. Each data source was uniquely identified and cataloged within the Global Health Data Exchange (GHDx), establishing a precise and accountable framework for global health data analysis. To address study heterogeneity—such as variations in sample size and selection, examination rate, and diagnostic criteria—the GBD’s DisMod estimation tool quantified between-study heterogeneity, which is the variance not attributed to fixed effects or geographical random effects. This tool then incorporated uncertainty based on these findings to ensure robust and reliable estimates¹².

In the GBD study, the definition of disease could be ascertained by the International Classification of Diseases Tenth Revision (ICD-10)¹³. Blindness and vision loss belongs to ICD-10 (H25-H28.8, H31-H36.8, H40-H40.9, H42-H42.8, H46-H54.9); glaucoma (H40-H40.9, H42-H42.8); cataract (H25-H26.9, H28-H28.8); age-related macular degeneration (H35.3-H35.389); refraction disorders (H52-H52.7) and other vision loss (H27-H27.9, H31-H35.23, H35.4-H36.8, H46-H51.9, H53-H54.9).

In the GBD study, the category of blindness and vision loss includes impairments due to refractive errors, near-vision loss, and other eye diseases in GBD categories. We adopted the GBD 2021 definition of vision loss, which is visual acuity less than 6/18 as measured by the Snellen or LogMAR chart. The severity levels are categorized as follows: (1) blindness, with visual acuity less than 3/60 or a visual field of less than 10% around central fixation; (2) severe vision loss, with visual acuity between 3/60 and 6/60; (3) moderate vision loss, with visual acuity between 6/60 and 6/18; and (4) near-vision loss, defined by near visual acuity equivalent to less than 6/12 at distance¹².

The socio-demographic index (SDI) serves as a comprehensive measure of development status, showing a strong correlation with health outcomes. It is calculated using the geometric mean of three 0–1 indices: fertility rates among females under 25, average years of education for individuals aged 15 and above, and lag-distributed income per capita. For the GBD 2021 study, SDI values were scaled by multiplying by 100 for reporting purposes. An SDI of 0 indicates the theoretical minimum development level relevant to health, whereas an SDI of 100 signifies the theoretical maximum. The GBD 2021 capstone paper detailed the construction of SDI and categorized the 204 countries into five quintiles (low, low-middle, middle, high-middle, and high) based on their 2021 country-level SDI estimates¹².

This study was conducted in accordance with the Declaration of Helsinki. As the GBD database is publicly available and our analyses did not involve any personally identifiable information, the need for ethical approval and informed consent was waived¹⁴.

Study population

WCBA and PMW with BVL were the subjects of the study. The rate and number data of BVL in WCBA and PMW from 1992 to 2021 were obtained from GBD 2021. WCBA and PMW are defined as females aged between 15 and 55^15,16. However, due to the lack of data for the age group of 15 to 19 in the database and the extremely low number of cases within this age range, we only included individuals aged between 20 and 55 in our study. The current investigation is conducted by the Guidelines for Accurate and Transparent Health Estimates Reporting, ensuring the integrity and transparency of the health estimates presented¹⁴.

Statistical analysis

We calculated the age-standardised rates of prevalence and YLDs using the GBD World Standard Population using the following formula: , where a_i is the ASR for the each age group and w_i is the corresponding number of people (or the weight) in the same age group among the GBD standard population. N is the total number of age groups^17,18.

The APC model is an advanced analytical tool that outperforms traditional methods in health and socio-economic studies. It allows for the calculation of net drift and local drift, which represent overall and specific time trends, respectively, while also assessing the effects of age, period, and birth cohort. This study employed the APC model to examine trends in the prevalence and YLDs of BVL across various age groups, time periods, and birth cohorts.

In APC analysis, the age interval typically matches the period interval, often using a 5-year age group with a corresponding 5-year period. Consequently, this study utilized the 2021 GBD database, which includes incidence data for BVL over the past 30 years (1992–2021) for the WCBA and PMW population, along with population data by location. The age range of 15–49 years was divided into seven age groups: 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, and 50–54 years. The study period from 1992 to 2021 was split into six 5-year periods: 1992–1996, 1997–2001, 2002–2006, 2007–2011, 2012–2016, and 2017–2021. This allowed for the analysis of 12 partially overlapping 10-year birth cohorts, ranging from 1942 to 1951 to 1997–2006. The APC model estimated both the overall time trend and the incidence trend within each age group¹⁹. The overall trend is expressed as the annual percentage change in incidence, indicating net drift (% per year), influenced by calendar time and continuous birth cohorts. The local drift (% per year) reflects the percentage change in annual incidence by age. Even slight changes in drift values can result in significant differences in the fitting rate over three decades. The significance of the annual percentage change trend was assessed using the Wald χ2 test. In the APC model, the age effect is shown through age-specific rates aligned with birth cohorts, while the period/cohort effect is represented as the relative risk of prevalence compared to a reference period/cohort, with the choice of reference being arbitrary and not affecting the interpretation of results.

To forecast future trends in the global burden of BVL, we utilized a Bayesian age-period-cohort model. This model employs the Bayesian formula to calculate hypothetical probability distributions based on three factors: age, period, and cohort. It integrates prior information with sample data to generate posterior information²⁰.

We conducted Pearson’s correlation analysis to explore the relationship between ASR and the SDI across 21 GBD regions and 204 countries and territories²¹. All analyses were performed using R software (version 4.3.2) and joinpoint software (version 5.0.2), with statistical significance defined as p < 0.05.

Results

Global BVL burden in WCBA and PMW, 1992–2021

The global prevalence of BVL burden in WCBA and PMW increased significantly from 1992 to 2021, reaching 333 million cases in 2021 (95% UI: 223,351,181.88 to 465,397,251.84). The global age-standardized prevalence rate (ASPR) of BVL in this population was 17,203.68 per 100,000 (95% UI: 9,712.29 to 27,661.60) (Table 1). The ASPR showed a significant increase across all SDI regions, particularly in high-middle and middle SDI regions. However, the low-middle and low SDI regions still bore the highest burden of BVL. According to the age–period–cohort analysis, the global net drift of BVL prevalence was 1.72% (95% CI: 1.50 to 1.94) per year. Notably, the net drift in low SDI countries was considerably lower than that observed in the low-middle, middle, and high-middle SDI groups, and was closer to that of the high SDI group, which exhibited the lowest net drift overall. This pattern may reflect a combination of limited temporal changes in population age structure, slower epidemiological transition, and potential underdiagnosis or underreporting of BVL in low SDI settings, where access to gynecological health services and diagnostic resources remains constrained.

Table 1.

Global and SDI trends in BVL prevalence in WCBA and PMW from 1992 to 2021.

Among the 204 countries and territories worldwide, 40 reported more than one million prevalence number of BVL-related diseases in 2021 (Fig. 1, Table S1). India, China, Brazil, and Indonesia have the highest prevalence of BVL, with more than ten million individuals affected in each country. In addition, 25 countries exhibit ASPR at or above the global average, while four countries—South Africa, India, Nepal, and Nigeria—show prevalence rates exceeding 1.5 times the global average. From 1992 to 2021, the relative change of ASPR showed an increase in 73 countries and a decrease in 131 countries. Furthermore, the net drift values estimated from the APC model for the majority of countries globally followed a similar trend. China showed the most significant increase in ASPR (120%), with a net drift of prevalence at 2.35% per year (95% CI: 1.81 to 2.89). The number of YLDs in WCBA and PMW showed a similar trend in different countries (Table S2).

Fig. 1.

The ASPR in 2021 and net drift of prevalence from 1992 to 2021 for BVL in WCBA and PMW in 204 countries and territories. (A) World map of ASPR in 2021. (B) World map of the Net drift of prevalence from 1992 to 2021.

Temporal trends in BVL prevalence in WCBA and PMW across different age groups

Figure 2 illustrates the annual percentage change in ASPR of BVL among women of reproductive and perimenopausal age across different age groups. Globally, the prevalence of BVL in WCBA and PMW exhibited an overall upward trend, with the annual percentage increase in prevalence rising with age (Fig. 2A). Based on the local drift estimates, positive APC values were observed across all age groups in Middle, Low-middle, and Low SDI regions, indicating an overall increasing trend in BVL prevalence. In contrast, the High SDI region exhibited negative local drift values in 20–35 age groups, suggesting a declining trend during those ages.

Fig. 2.

Local drift and age distribution of BVL prevalence in WCBA and PMW in different SDI regions from 1992 to 2021. (A) Local drift in the incidence of BVL in WCBA from 1992 to 2021 for 7 age groups. Dots and shaded areas indicate local drift. (B) Age distribution of BVL incidence in WCBA and PMW over time from 1992 to 2021.

We further examined the temporal changes in the age distribution of BVL prevalence among women (Fig. 2B). The prevalence is notably higher in middle-aged and older women, with more than 50% of the global and regional prevalence concentrated in women aged 40–54 years, particularly in those within the perimenopausal period. From 1992 to 2021, the age distribution of BVL prevalence has increasingly concentrated in middle-aged and older women globally. This trend is particularly pronounced in high, high-middle, and middle SDI regions. Furthermore, regression analysis demonstrated a stronger negative association between SDI and BVL prevalence among women of reproductive and perimenopausal age, indicating that higher levels of sociodemographic development were associated with a lower BVL burden in these populations. (Figure S1).

Effect of age, period, and birth cohort on the prevalence of BVL in WCBA and PMW

Figure 3 illustrates the effects of age, period, and birth cohort on BVL prevalence in WCBA and PMW, as derived from the APC model. The age effect exhibited a similar pattern across SDI regions, with BVL prevalence progressively increasing with age in women (Fig. 3A). Simultaneously, the overall prevalence was lower in high SDI regions compared to other SDI regions. The period effect demonstrated a consistent global trend of increasing BVL prevalence in WCBA and PMW. The period effect risk was lower in high SDI regions, where it remained relatively stable and close to 1, while other SDI regions primarily exhibited unfavorable trends characterized by a continuous increase in period-related relative risk over successive periods (Fig. 3B).

Fig. 3.

APC model predictions of the effects of age, period and birth cohort on the prevalence of BVL in WCBA and PMW. (A) Age effects are shown by age-specific longitudinal ratios adjusted for variation between birth cohorts to account for period-specific bias. (B) Period effects are shown by the relative risk of BVL prevalence in different periods, calculated as age-specific ratios for the period 1992–1996 to 2017–2021, with the base period set to 1992–1996. (C) Birth cohort effects are indicated by the relative risk of cohort prevalence, calculated as age-specific ratios from the 1942–1951 cohort to the 1997–2006 cohort, with the reference cohort set to 1967–1976. Points and shaded areas indicate incidence rates or ratios and their corresponding 95% CI.

Regarding birth cohort effects, the prevalence risk for successive cohorts initially increased and then decreased globally (Fig. 3C). All consecutive birth cohorts in regions other than high SDI regions demonstrated a progressive increase in prevalence risk. Globally, compared with the reference cohort of individuals born between 1967 and 1976, individuals born prior to this cohort had a lower risk of BVL in WCBA and PMW, while those born after it exhibited a relatively higher risk. However, individuals born after the 1967–1976 cohort exhibited progressively lower risks in high SDI regions.

Age group differences in different BVL-related eye diseases in WCBA and PMW

As of 2021, the prevalence number and YLDs for six categories of ocular diseases—near vision loss, refractive disorders, other vision loss, cataracts, AMD, and glaucoma—showed an increase with age in women, peaking at 50–54 years. Near vision loss and refractive errors contributed significantly to the absolute incidence of YLDs, particularly among middle-aged and older women (Fig. 4A). Similarly, the number of women with cataracts and glaucoma—two common blinding eye diseases—along with YLDs, also increased significantly with age (Fig. 4B). It should be noted that although glaucoma and AMD are major causes of irreversible vision loss at older ages, their absolute prevalence and YLDs remain relatively low in women of childbearing and perimenopausal age. Consequently, these conditions appear less prominent in Fig. 4 when absolute numbers are displayed.

Fig. 4.

Prevalence number and YLDs of different BVL-related eye diseases in 2021. (A) Prevalence number (B) Number of YLDs.

Global burden of disease projections for BVL to 2050

A Bayesian Age-Period Cohort (BAPC) model was used to project future trends in the prevalence of BVL in women of different ages. The total prevalence number is projected to increase steadily after 2021, reaching approximately 400 million by 2035, and continuing to rise to nearly 800 million by 2050 under current trends (Fig. 5A). Notably, a substantial rise is already evident by 2035, suggesting a considerable future e public health burden in the medium term. For women of all ages, increases of varying magnitudes are shown, especially for middle-aged and older women aged 45–54 years (Fig. 5B). ASPR show the same trend, with a more substantial increase expected for the 45-54-year-old female population (Fig. 5C).

Fig. 5.

Projects the prevalence and ASYR of BVL in WCBA and PWM from 2022 to 2050. (A) global. (B) Different SDI regions. (C) ASPR.

Discussion

The central contribution of this study is the demonstration that the rising burden of BVL among WCBA and PMW is driven by concurrent and independent age, period, and cohort effects rather than demographic aging alone. The results indicate that while the BVL burden among WCBA and PMW varies significantly across regions, the overall trends and regional patterns highlight key public health challenges and guide policy interventions.

The study found a substantial increase in the prevalence of BVL in WCBA and PMW from 120 million cases in 1992 to 333 million cases in 2021, with a global ASPR of 17,203.68 per 100,000 population. Notably, the most pronounced increase in ASPR occurred in high-middle and middle SDI regions, while low-middle and low SDI regions bore the heaviest burden of BVL. This trend is likely linked to disparities in economic development, healthcare resource distribution, and medical conditions across regions. By 2021, 40 countries and regions globally had more than one million people with BVL in WCBA and PMW, with India, China, Brazil, and Indonesia having the highest numbers, indicating the urgent need for focused attention in these populous countries. Additionally, 25 countries had ASPR levels at or above the global average, particularly Niger, India, and South Africa, where the prevalence exceeded 1.5 times the global average. This suggests the need for future health policies to develop targeted interventions based on national and regional disease burden characteristics.

Furthermore, our APC model analysis revealed a consistent age effect across SDI regions: BVL prevalence in women increased significantly with age, especially in women aged 40–54 years, who accounted for more than 50% of the prevalence, with the condition becoming more concentrated in middle-aged and older women. This may reflect the elevated risk in perimenopausal women due to hormonal changes, metabolic abnormalities, and other factors. Numerous studies have shown that the prevalence of BVL in women, particularly during the perimenopausal stage, exceeds that in men of the same age, with this difference potentially linked to disrupted sex hormone levels. The significant decrease in estrogen levels during perimenopause may exacerbate the development of various eye diseases, such as cataracts, glaucoma, and AMD^10,22. Clinical studies have observed reduced prevalence and severity of cataracts in postmenopausal women receiving hormone replacement therapy or in younger women using oral contraceptives. Additionally, several studies have shown that estrogen can prevent cataracts by counteracting the effects of TGF-β^23,24. Estrogen’s role in retinal neuroprotection is well-documented, and its reduction may lead to retinal ganglion cell damage and degeneration of visual function⁸. A previous study also showed that postmenopausal women with reduced circulating estrogen levels had a significantly increased incidence of glaucoma, with intraocular pressure increasing by approximately 1.5–3.5 mmHg compared to premenopausal women of the same age^10,11. Furthermore, the researchers found that topical estradiol administration effectively prevented retinal ganglion cell death in an animal model of glaucoma in rats. Researchers have also found that endogenous estrogen exposure and hormone replacement therapy may reduce the risk of developing AMD^9,25. Estrogen deficiency may exacerbate ocular disease progression through mechanisms such as oxidative stress, inflammatory responses, and increased apoptosis. Perimenopausal women are also at higher risk for metabolic disorders, such as insulin resistance and abnormal lipid metabolism, which are strongly associated with the development of major BVL-related diseases like cataracts and glaucoma^24,26–29. On the other hand, the period effect revealed a persistent upward trend in BVL prevalence in WCBA and PMW globally, especially in low SDI areas, which may be influenced by insufficient health resources and limited screening efforts. Regarding the birth cohort effect, although some high and high-middle SDI regions showed relative stabilization or slight attenuation of cohort-specific risk in more recent cohorts, low and low-middle SDI regions continued to demonstrate increasing cohort-associated risk.

Importantly, projection analyses suggest that these interacting dynamics are likely to persist through 2050, indicating structural epidemiologic expansion rather than a transient fluctuation. Using Bayesian prediction modeling, the prevalence number of BVL in WCBA and PMW is projected to grow further to 800 million cases globally by 2050, with a particularly significant growth rate in women aged 45 to 54 years. These projections are consistent with the observed cohort patterns, reinforcing that without substantial improvements in prevention and eye-care access, a natural cohort-driven decline in the coming decades appears unlikely in this demographic. The global burden of BVL in this population also poses a significant challenge to WHO targets as the population ages and the burden of disease continues to increase. Therefore, future health resource allocation and policy interventions need to be further skewed toward high-risk populations, especially in low and middle SDI countries, to strengthen disease prevention, control, and health education efforts.

This study provides the first comprehensive picture of global trends and regional characteristics of the burden of BVL in WCBA and PMW through detailed APC analysis and future projections. However, this study has several limitations, including insufficient surveillance data in some low SDI countries and incomplete data for diseases such as glaucoma and AMD, which may lead to an underestimation of the results. At the same time, the GBD prediction models also have certain limitations. Their accuracy highly depends on the original data, the assumptions of the model, and the accuracy of the predicted covariates, which may introduce inherent biases. Additionally, the prediction models do not account for factors such as developments in medical technology or changes in policies, which may also lead to bias. Future studies should further explore the potential mechanisms of environmental, hormonal, and socioeconomic factors on the burden of BVL. In summary, this study highlights the importance of addressing the visual health issues of WCBA and perimenopausal women globally and provides a scientific basis for the development of regional health strategies and the achievement of the Sustainable Development Goals. More broadly, our findings underscore the need to translate epidemiologic evidence into proactive, equity-oriented public health action to mitigate the expanding burden of vision loss in this critical demographic group.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

X.X. and R.R. conceptualized and reviewed the manuscript. Y.L. and Y.Z. wrote the manuscript. All authors have read and approved the final manuscript.

Funding

This study was financially supported by grants from the National Clinical Key Specialty of Ophthalmology, the National Natural Science Foundation of China (No. 82401261), the Scientific Research Program of FuRong Laboratory (No. 2024PT5107), the China National Postdoctoral Program for Innovative Talents (BX20230434), the China Postdoctoral Science Foundation (No.2023M743952), and the Youth Science Foundation of Xiangya Hospital (No. 2023Q12).

Data availability

Data utilized in this study are derived from the Global Burden of Disease Study 2021 (GBD 2021) (https://vizhub.healthdata.org/gbd-results/).

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

The Ethics Committee of Xiangya Hospital of Central South University determined that the study did not need approval because it used publicly available data.