Global, regional, and National burden of falls among midlife women from 1990 to 2021 and projections to 2050: A systematic analysis for the global burden of disease study 2021

Abstract

Background

Declining estrogen in perimenopausal women reduces bone mineral density and increases bone fragility, elevating fall and fracture risk. This presents major challenges for patients and society, yet prior studies lack systematic analysis of this population. This study is the first to utilize mortality and disability-adjusted life years (DALYs) related to falls among perimenopausal women from the Global Burden of Disease (GBD) 2021 database, and to conduct a comprehensive and systematic analysis of the evolving burden of falls in perimenopausal women from 1990 to 2021, as well as to project trends through 2050. This study offers key guidance for optimizing healthcare resource allocation, enhancing patient management, and developing targeted prevention and intervention strategies.

Methods

This study used GBD 2021 data to systematically analyze fall-related mortality, DALYs, age-standardized rates (ASRs), and estimated annual percentage change (EAPC) among perimenopausal women, examining their associations with the Socio-demographic Index (SDI) at global, regional, and national levels. Joinpoint regression, decomposition, health inequality, and frontier analyses quantified trends, identified factors, and assessed disparities. This study also explored fall risk factors and utilized the Bayesian Age-Period-Cohort (BAPC) model to project global trends in fall burden among perimenopausal women from 2022 to 2050.

Results

Between 1990 and 2021, global mortality attributable to falls among perimenopausal women surged by 116.99% (from 1.67 to 3.63 per 100,000), while DALYs increased by 38.31% (from 116.50 to 188.85 per 100,000). In contrast, age-standardized mortality rate (ASMR) and age-standardized DALY rate (ASDR) declined by 7.00% (EAPC = − 0.30) and 14.62% (EAPC = − 0.56), respectively. Decomposition analysis identified population growth as the predominant contributor to the escalation in mortality (162.91%), while epidemiological changes were the main reason for the reduction (− 63.63%). Marked heterogeneity was observed across SDI strata: low-middle SDI regions exhibited the steepest rise in mortality (138.00%), whereas high-SDI regions achieved the most pronounced reduction in ASMR (− 21.18%). Notably, high-income North America experienced an 106.62% increase in ASMR. The 50–54-year age cohort consistently represented the highest global burden, with low bone mineral density emerging as the principal risk factor. Projections to 2050 suggest ongoing declines in ASMR and ASDR, yet the absolute burden is expected to remain elevated due to persistent demographic expansion.

Conclusions

Between 1990 and 2021, the global burden of falls among perimenopausal women has exhibited a persistent upward trend, and projections indicate that this burden will likely remain at a high level in the future. This alarming situation underscores the urgent need for targeted interventions. Identifying key risk factors for falls in perimenopausal women is essential for guiding the allocation of public health resources and formulating precise intervention strategies. It is imperative to implement nationwide, cost-effective measures, such as osteoporosis and fall risk screening, the promotion of exercise programs that enhance muscle strength and balance, and, where appropriate, consideration of pharmacological interventions (such as estrogen) to reduce fall risk. This intervention will significantly reduce the risk of falls and associated burdens among perimenopausal women.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40520-025-03210-5.

Introduction

Falls are recognized as the second leading cause of unintentional injury-related mortality worldwide, surpassed only by road traffic injuries [1], and can result in severe outcomes such as spinal cord injuries and vertebral fractures [2]– [3], significantly impacting public health and healthcare expenditures. Globally, it is estimated that 37.3 million falls occur annually, leading to approximately 17 million cases of disability and 684,000 deaths [4]. Falls are particularly prevalent among older adults [5], and the perimenopausal period represents a transitional phase for women entering later life, typically occurring between the ages of 45 and 55, and is characterized primarily by irregular menstrual cycles [6]– [7]. During this stage, ovarian function and estrogen secretion progressively decline, affecting multiple organ systems [8]. Estradiol, the most biologically active form of estrogen, exerts significant effects on nearly all musculoskeletal tissues, including bone, tendon, and muscle, and its reduction leads to bone loss and decreased muscle mass [8]– [9], both of which are well-established risk factors for increased fall incidence among perimenopausal women [10]. Falls in this population may result in severe outcomes such as fractures, mortality, disability, and functional impairment, thereby increasing the likelihood of admission to long-term care facilities and utilization of healthcare services, ultimately imposing a substantial burden on both individuals and society [4, 11]. A cohort study in Australia reported a fall incidence of up to 42% among middle-aged women, underscoring the urgent need for targeted fall prevention strategies in this demographic [12]. While previous research has addressed the disease burden associated with falls, systematic analyses focusing specifically on perimenopausal women remain scarce [13]. The GBD 2021 database provides comprehensive data on 204 countries and territories, 371 diseases and injuries, and 88 risk factors from 1990 to 2021, including information relevant to falls among perimenopausal women [14].

Drawing on data from the GBD 2021 database, this study aims to comprehensively assess the global, regional, and national burden of falls among perimenopausal women from 1990 to 2021, as well as temporal trends in this burden. Employing methodologies such as Joinpoint regression analysis and decomposition analysis, this study focuses on elucidating the dynamic changes in fall-related disease burden and its driving factors, and projects trends through 2050. To provide a robust scientific foundation for early prevention strategies within the framework of healthy aging across diverse countries and regions, to offer critical insights into the epidemiological patterns of falls among perimenopausal women worldwide, and to support public health policymakers in formulating and implementing effective prevention and intervention measures, thereby addressing the significant health challenges faced by this specific population.

Methods

Data sources

The GBD 2021 database, led by the Institute for Health Metrics and Evaluation at the University of Washington, is designed to quantify health loss attributable to diseases, injuries, and risk factors [14]. This database integrates a wide array of data sources, including demographic surveys, hospital records, vital registration systems, and published literature, thereby ensuring the scientific rigor and comprehensiveness of the analytical dataset [15]. This study utilized data from GBD 2021 to analyze and assess the burden of falls and their attributable risk factors among women aged 45–54 across 204 countries, 21 GBD regions, and 5 SDI regions globally from 1990 to 2021.The GBD study rigorously adheres to the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) statement [16].

Key definitions

This study defines the perimenopausal period as ages 45 to 54 [17]. Falls are sudden, unintentional descents to the ground caused by slipping, tripping, or other accidental movements, potentially resulting in death, disability, or injury [18]. According to the International Classification of Diseases (ICD), falls are coded as E880-E886.99, E888-E888.9, and E929.3 in the ninth revision (ICD-9), and as W00-W19.9 in the tenth revision (ICD-10) [4, 18]. The Global Burden of Disease (GBD) 2021 Study compiled data on disability-adjusted life years (DALYs) and mortality among perimenopausal women attributed to four risk factors: heavy alcohol consumption, low bone mineral density, occupational injuries, and smoking, as categorized by GBD 2021. This study will further analyze the attributable fractions of these risk factors for falls among perimenopausal women, stratified by region, age group, and temporal trends. This study evaluates fall-related metrics among perimenopausal women, specifically mortality and DALYs along with their 95% uncertainty intervals (UIs), sourced from the GBD 2021 database (https://vizhub.healthdata.org/gbd-results/). The estimated annual percentage change (EAPC) is a widely utilized indicator for assessing temporal trends in age-standardized rates (ASRs). The calculation is based on the formula: ln(y) =α + βx + ε, where (y) denotes the age-standardized incidence rate, (α) is the intercept, (x) represents the calendar year, (β) is the slope, and (ε) is the normally distributed error term. The EAPC is computed as (exp^^β−1)×100%. An EAPC with a lower bound of the 95% confidence interval (CI) greater than 0 indicates an increasing trend in the ASR, whereas an upper bound less than 0 signifies a decreasing trend [19]. This study used an Age-Period-Cohort (APC) model to assess how age, period, and birth cohort influence the burden of falls in perimenopausal women. The APC model, based on demographic characteristics, stratifies populations by birth year and follow-up period, allowing analysis of health and disease mortality and DALYs trends across cohorts. This approach clarifies temporal patterns and the relationships among age, period, and cohort effects [20]. In this study, the definitions are as follows: Period Effect: Using the 2005 ASMR and ASDR as baseline (set to 1), the period relative risk (Rate Ratio, RR) is employed to monitor temporal variations in ASMR and ASDR. An RR greater than 1 indicates a higher risk relative to the reference period, while an RR less than 1 indicates a lower risk. Cohort Effect: Using the 1962 birth cohort’s ASMR and ASDR as baseline (set to 1), cohort RR is utilized to assess trends over time. An RR greater than 1 signifies a higher risk compared to the reference cohort, whereas an RR less than 1 signifies a lower risk.

Socio-demographic index

The Socio-Demographic Index (SDI) is a composite metric for assessing national development, introduced by the Institute for Health Metrics and Evaluation at the University of Washington [21]. It integrates three key determinants: per capita income, mean years of education among individuals aged 15 and older, and total fertility rate among those under 25 [22]. This multidimensional index offers a comprehensive evaluation of a country’s development status, capturing not only economic performance but also critical social factors such as educational attainment and demographic structure. Based on SDI values, countries are stratified into five development categories: low, low-middle, middle, high-middle, and high [22].

Advanced analysis

Joinpoint regression analysis

This study utilized the Joinpoint regression model to conduct segmented trend analyses of mortality and DALYs due to falls among perimenopausal women. Key outcome measures included the annual percent change (APC) and average annual percent change (AAPC), along with their corresponding 95% confidence intervals (CIs). An APC or AAPC with a 95% CI greater than zero indicated a statistically significant upward trend, whereas a 95% CI less than zero denoted a statistically significant downward trend [19].

Decomposition analysis

This study employs decomposition analysis to investigate the determinants of the burden of falls among perimenopausal women across five global SDI regions and twenty-one GBD regions. Changes in disease burden are attributed to three distinct components: population aging, population growth, and epidemiological changes [23]. This approach quantifies the impact of each factor on overall change, thereby informing targeted interventions and effective public health strategies to mitigate the disease burden.

Health inequality analysis

The distributional inequality of the burden of falls among perimenopausal women across countries was assessed using the Slope Index of Inequality (SII) and the Concentration Index (CI) for health inequality [24]. The SII applies linear regression to the ASR weighted by the SDI, reflecting the health gap between the lowest and highest SDI groups. The Lorenz concentration curve fits the cumulative distribution of the burden of falls against the overall distribution, ranked by SDI [25]. The CI is calculated as the area under the curve, ranging from − 1 to 1. If the curve lies below the diagonal, the disease burden is concentrated in countries with higher development levels, resulting in a positive CI; conversely, a negative value indicates concentration in less developed countries.

Frontier analysis

To evaluate the association between the burden of falls among perimenopausal women and the level of sociodemographic development, this study employed frontier analysis and constructed a frontier model based on the SDI using age-standardized mortality rates (ASMR) and age-standardized disability-adjusted life year rates (ASDR). This approach established a benchmark for the burden of falls in perimenopausal women, enabling comparisons between countries and regions and the top-performing counterparts. By identifying leading countries and regions, this method establishes reference standards and targets for others. This study calculated the “efficiency gap” for each country and region, representing the disparity between the current and potential burden of falls among perimenopausal women, adjusted for SDI [26].

Bayesian Age-Period-Cohort (BAPC)

This study employed the BAPC model to project the future global disease burden of falls among perimenopausal women. The BAPC model addresses the parameter estimation challenges arising from the linear dependency among age, period, and cohort effects in traditional APC models by incorporating Bayesian priors and utilizing the Integrated Nested Laplace Approximation (INLA) algorithm [19]. The mathematical formulation of the model is: log(λij) = α + µi + βj + γk, where λij denotes mortality and DALYs, α represents the intercept, µi captures the age effect, βj the period effect, and γk the cohort effect [19]. The BAPC model applies a second-order random walk to account for temporal adjacency effects and is implemented and forecasted using the BAPC and INLA packages in R [19]. To ensure the accuracy and robustness of the predictive results, this study utilized mortality and DALY data related to falls among perimenopausal women from 1990 to 2021 to forecast the number of cases and ASR of mortality and DALY for the period 2022 to 2050.

Results

Global, regional, and National burden of falls among perimenopausal women

Between 1990 and 2021, the global mortality due to falls among perimenopausal women rose from 1.67 (95% CI: 1.36, 1.89) per 100,000 to 3.63 (95% CI: 2.86, 4.11) per 100,000, representing a 116.99% increase(Table 1). Concurrently, DALYs increased from 116.50 (95% CI: 93.56, 144.29) per 100,000 to 188.85 (95% CI: 150.78, 233.73) per 100,000, a 38.31% rise (Table S1). The ASMR declined by 7.00%, and the ASDR decreased by 14.62%, with both ASMR and ASDR showing a negative estimated annual percentage change (EAPC) of −0.30 (95% CI: −0.38, −0.22) and − 0.56 (95% CI: −0.59, −0.54) respectively, indicating an overall downward trend (Table 1, S1, and Fig. 1C, F).

Table 1.

The cases and ASR of mortality from falls among perimenopausal women in 1900 and 2021 across global, five SDI regions, and 21 GBD regions, along with the percentage change and EAPC in ASMR of falls among perimenopausal women from 1990 to 2021

Location	1990		2021		1990–2021
	Cases(×105)	ASMR	Cases(×105)	ASMR	PC(%)	EAPC
Global	1.67(1.36,1.89)	8.45(7.01,9.45)	3.63(2.86,4.11)	7.86(6.18,8.88)	−7.00	−0.30(−0.38,−0.22)
SDI
High	0.42(0.37,0.45)	6.15(5.42,6.53)	0.86(0.69,0.96)	5.44(4.49,5.96)	−11.58	−0.80(−1.03,−0.58)
High middle	0.29(0.25,0.33)	6.21(5.28,6.89)	0.54(0.39,0.67)	4.90(3.57,6.03)	−21.18	−0.76(−0.83.−0.69)
Middle	0.42(0.31,0.48)	9.38(6.83,10.57)	0.99(0.68,1.18)	8.02(5.52,9.53)	−14.50	−0.37(−0.44,−0.31)
Low middle	0.39(0.26,0.48)	14.14(10.04,17.48)	0.92(0.70,1.08)	14.54(11.28,17.39)	2.84	0.23(015,0.31)
Low	0.15(0.11,0.19)	15.41(11.67,19.29)	0.31(0.257,0.39)	15.48(12.33,19.45)	0.47	0.19(0.05,0.33)
Regions
Andean Latin America	0.005(0.004,0.006)	4.15(3.57,4.82)	0.01(0.009,0.01)	4.02(3.12,4.87)	−3.06	−0.02(−0.12,0.08)
Australasia	0.006(0.005,0.007)	4.53(3.90,4.92)	0.03(0.02,0.03)	6.86(5.46,7.72)	51.62	0.79(0.29,1.29)
Caribbean	0.01(0.009,0.01)	8.36(7.66,8.89)	0.03(0.02,0.03)	8.81(7.68,9.90)	5.34	0.90(0.54,1.26)
Central Asia	0.01(0.01,0.01)	2.99(2.83,3.14)	0.008(0.007,0.009)	1.91(1.71,2.13)	−36.01	−1.41(−1.60,−1.21)
Central Europe	0.09(0.08,0.09)	11.97(11.05,12.52)	0.07(0.06,0.07)	4.46(3.92,4.84)	−62.77	−3.10(−3.34,−2.85)
Central Latin America	0.03(0.03,0.03)	7.41(6.88,7.69)	0.04(0.04,0.05)	3.32(2.88,3.66)	−55.23	−2.91(−3.14.−2.67)
Central Sub-Saharan Africa	0.009(0.007,0.01)	10.65(8.37,14.28)	0.02(0.02,0.03)	10.71(7.91,15.18)	0.48	0.03(0.01.0.05)
East Asia	0.25(0.20,0.35)	7.37(5.82,9.75)	0.58(0.33,0.79)	5.97(3.47,8.13)	−18.96	−0.53(−0.72,−0.34)
Eastern Europe	0.05(0.05,0.05)	3.25(3.11,3.33)	0.06(0.05,0.07)	3.00(2.72,3.26)	−7.49	−0.11(−0.41,0.19)
Eastern Sub-Saharan Africa	0.04(0.03,0.05)	12.99(10.28,16.99)	0.08(0.07,0.10)	12.37(9.76,15.76)	−4.79	−0.17(−0.20,−0.14)
High-income Asia Pacific	0.03(0.02,0.03)	2.68(2.31,3.09)	0.09(0.06,0.11)	2.10(1.59,2.45)	−21.46	−1.32(−1.46,−1.18)
High-income North America	0.08(0.06,0.08)	3.22(2.74,3.45)	0.29(0.24,0.32)	6.65(5.50,7.25)	106.62	2.07(1.80,2.34)
North Africa and Middle East	0.05(0.03,0.07)	5.86(4.53,7.19)	0.08(0.06,0.09)	4.41(3.46,5.40)	−24.79	−0.74(−0.85,−0.63)
Oceania	0.0007(0.0005,0.0009)	6.85(4.84,9.61)	0.002(0.001,0.003)	6.65(4.08,9.55)	−2.91	0.06(0.02,0.11)
South Asia	0.50(0.30,0.63)	21.89(14.33,27.16)	1.30(0.95,1.54)	20.79(15.33,24.83)	−5.04	0.02(−0.10,0.13)
Southeast Asia	0.12(0.08,0.15)	10.71(6.27,13.74)	0.25(0.16,0.30)	8.64(5.40,10.63)	−19.32	−0.55(−0.66,−0.43)
Southern Latin America	0.01(0.01,0.01)	4.57(4.09,4.87)	0.02(0.02,0.02)	3.23(2.73,3.50)	−29.26	−1.31(−1.64,−0.97)
Southern Sub-Saharan Africa	0.003(0.003,0.004)	2.29(1.70,2.87)	0.005(0.004,0.007)	1.82(1.49,2.53)	−20.45	−0.63(−0.79,−0.47)
Tropical Latin America	0.02(0.02,0.02)	5.50(4.91,5.83)	0.10(0.08,0.11)	6.95(5.84,7.57)	26.33	0.81(0.53,1.09)
Western Europe	0.32(0.28,0.34)	8.38(7.36,8.94)	0.48(0.38,0.53)	6.10(4.97,6.72)	−27.27	−1.47(−1.75,−1.19)
Western Sub-Saharan Africa	0.04(0.03,0.05)	10.68(8.18,13.51)	0.08(0.06,0.10)	10.78(8.58,13.36)	0.93	0.22(0.16,0.28)

SDI, Socio-Demographic Index. ASMR, age-standardized mortality rate. PC, percentage change. EAPC, estimated annual percentage changes

Fig. 1.

(A, B) Percentage change and EAPC in ASMR for 204 countries from 1990 to 2021. (C) EAPC for ASMR from 1990 to 2021 for the global level, five SDI regions, and 21 GBD regions. (D, E) Percentage change and EAPC in ASDR for 204 countries from 1990 to 2021. (F) EAPC for ASDR from 1990 to 2021 for the global level, five SDI regions, and 21 GBD regions. SDI, Socio-Demographic Index; ASMR, age-standardized mortality rate; ASDR, age-standardized DALYs rate; PC, percentage change; EAPC, estimated annual percentage change

When analyzed by SDI levels, the increase in mortality cases was most pronounced in low-middle SDI regions at 138.00%, and least pronounced in high-middle SDI regions at 84.47%. DALYs saw the highest increase in low SDI regions at 44.27%, and the lowest in high-middle SDI regions at 26.42%(Table 1, S1). The ASMR change was most significant in low-middle SDI regions, increasing by 2.84%, while high-middle SDI regions experienced a 21.18% decrease. ASDR variation is most pronounced in the high-middle SDI region, with a decline of 20.96%, whereas the high SDI region exhibits a smaller decrease of 6.83% (Table 1, S1). The EAPC of ASMR showed the greatest increase in low-middle SDI regions at 0.23 (95% CI: 0.15, 0.31), while high, high-middle, and middle SDI regions exhibited a declining trend, with the most significant decrease in high SDI regions at −0.8 (95% CI: −1.03, −0.58). The EAPC of ASDR showed a declining trend across all SDI regions, with the most significant decrease in high-middle SDI regions at −1.01 (95% CI: −1.11, −0.90), and the least in high SDI regions at −0.17 (95% CI: −0.21, −0.13). This suggests that higher SDI regions, due to better economic and resource allocation, manage disease more effectively, resulting in a relatively lower disease burden, whereas the opposite is true for low SDI regions (Table 1, S1, and Fig. 1C, F).

Among the 21 GBD regions, the increase in mortality was highest in tropical Latin America at 340.58%, while Central Europe experienced a 25.02% decrease. DALYs increased most in Oceania at 65.90%, with Central Europe experiencing a 20.05% decrease(Table 1, S1). The ASMR change generally declined, with the most significant increase observed in high-income North America at 106.62%, and the largest decrease in Central Europe at −62.77%.The ASDR change also exhibited a general downward trend, with the highest increase observed in Oceania at 25.00% and the largest decrease in Central Latin America at −41.68% (Table 1, S1). The EAPC of ASMR showed the largest increase in high-income North America at 2.07 (95% CI: 1.80, 2.34), and the largest decrease in Central Europe at −3.10 (95% CI: −3.34, −2.85). The EAPC of ASDR showed the largest increase in high-income North America at 0.80 (95% CI: 0.67, 0.92), and the largest decrease in Central Europe at −1.63 (95% CI: −1.67, −1.58). This indicates a heavier burden in high-income North America, while regions like Central Europe show significant improvement in overall burden (Table 1, S1, and Fig. 1C, F).

Across 204 countries, Mauritius experienced the highest increase in mortality at 513.56%, although the absolute increase in numbers was not substantial. Hungary exhibited the largest decrease at 62.48%. DALYs increased most in Qatar at 513.03%, while Latvia experienced the largest decrease at 50.63%(Table S2-3). The ASMR change rate was highest in Mauritius and the United States, at 130.98% and 115.63% respectively, while Guam, Hungary, and the Czech Republic saw the largest decreases at −79.98%, −79.06%, and − 75.08% respectively. The ASDR change rate was highest in Spain at 53.50%, with Armenia and Hungary experiencing the largest decreases at −56.60% and − 53.76% respectively(Fig. 1A, D, S1, and Table S2-3). The EAPC of ASMR showed a general growth trend in most countries, with the most significant increases in Cuba, Japan, Australia, the United States, and Brazil, while a few countries like the Czech Republic and Kyrgyzstan showed notable decreases. The EAPC of ASDR showed the largest increases in the Netherlands, Cuba, Spain, Puerto Rico, and Canada, with the most significant decreases in Kyrgyzstan, Hungary, and Latvia. Overall, countries like Cuba and the Netherlands are experiencing an increasing burden, while countries like the Czech Republic and Kyrgyzstan show notable improvements in burden (Fig. 1B, E, and Table S2-3).

The trend of burden

Between 1990 and 2021, the mortality and DALYs associated with falls among perimenopausal women globally were predominantly concentrated in the 50 to 54 age group. Over time, while the ASMR for the 50–54 age group initially decreased, then slightly increased before declining again, the ASMR and ASDR for other age groups exhibited a downward trend. In the five SDI regions, the mortality and DALYs due to falls in perimenopausal women remained concentrated in the 50 to 54 age group. Over time, except for the ASMR in two age groups in the high-middle SDI regions and the 50–54 age group in the low-middle SDI regions, which showed an initial decline followed by a slight increase and then a decline, and the ASDR in the middle SDI regions, which initially decreased and then slightly increased, all other ASMR and ASDR trends showed a decline. In the 21 GBD regions, the mortality and DALYs from falls among perimenopausal women remained concentrated in the 50 to 54 age group, with noticeable upward trends over time in regions such as Australasia, high-income North America, and Oceania(Fig. 2).

Fig. 2.

(A,B) The distribution of ASMR and ASDR for falls among perimenopausal women across global, five SDI regions, and 21 GBD regions in 1990 and 2021. (C,D) The trends in ASMR and ASDR for falls among perimenopausal women globally and across five SDI regions from 1990 to 2021. (E,F) Trends in ASMR and ASDR globally and within the five SDI regions from 1990 to 2021. (G,H) Trends in ASMR and ASDR globally, across five SDI regions, and 21 GBD regions from 1990 to 2021. SDI, Socio-Demographic Index; ASMR, age-standardized mortality rate; ASDR, age-standardized DALYs rate

Joinpoint regression analysis

The analysis of the Joinpoint regression model indicates that from 1990 to 2021, the global ASMR for perimenopausal women experiencing falls underwent four significant shifts: APC_1990–2000 = −0.778 (−1.025, −0.530), APC_2000−2002 = 1.539 (−0.073, 3.177), APC_2002−2019 = −0.021 (−0.133, 0.091), and APC_2019−2021 = −0.021 (−0.133, 0.091). Similarly, the ASDR exhibited four pivotal changes: APC_1990–2002 = −0.365 (−0.393, −0.337), APC_2002−2004 = −1.216 (−1.450, −0.982), APC_2004−2015 = −0.611 (−0.643, −0.579), and APC_2015−2021 = −0.262 (−0.342, −0.182). Overall, the ASDR demonstrated a consistent downward trend, while the ASMR initially decreased, then increased, and subsequently decreased again(Table S4 and Fig. S2).The temporal trends in ASMR and ASDR across the five SDI regions are illustrated in Figure S3-7 and Table S4.

Association between ASRs and SDI

Figure 3A-D illustrates the relationship between ASMR, ASDR, and the SDI across various GBD regions and countries. Analyzing the correlation between ASMR and SDI in GBD regions, it is evident that ASMR decreases as SDI increases. The burden in South Asia significantly exceeds expectations, whereas regions such as Central Asia and Sub-Saharan Africa fall below anticipated levels, with most other regions aligning closely with projections. Examining the relationship between ASMR and SDI across 204 countries, ASMR initially declines and then rises with increasing SDI. While the majority of countries align with expected burdens, nations such as India, Palau, Cambodia, Bhutan, and Cuba surpass anticipated levels. Regarding the relationship between ASDR and SDI in GBD regions, ASDR initially rises, then falls, rises again, and finally declines as SDI increases. Similarly, South Asia’s burden is notably higher than expected, while Sub-Saharan Africa and other regions fall below expectations, with most regions aligning with projections. Across 204 countries, ASDR initially shows a slight decline followed by a significant increase as SDI rises. Most countries align with expected burdens, yet India, Andorra, Nepal, Greenland, and Belgium exceed anticipated levels.

Fig. 3.

(A) ASMR to SDI for 21 GBD regions. (B) ASMR to SDI for 204 countries. (C) ASDR to SDI for 21 GBD regions. (D) ASDR to SDI for 204 countries. (E) Relationship of EAPC with mortality and SDI. (F) Relationship of EAPC with DALYs and SDI. ASMR, age-standardized mortality rate; ASDR, age-standardized DALYs rate; SDI, Socio-Demographic Index

Factors influencing EAPCs

In 2021, across 204 countries, there was a positive correlation between female ASMR and EAPC (R = 0.35, P = 3.0 × 10⁻⁷⁾, while SDI exhibited a negative correlation with EAPC (R=−0.34, P = 8.6 × 10⁻⁷), both demonstrating significant associations with EAPC. Similarly, ASDR showed a positive correlation with EAPC (R = 0.092, P = 0.19); however, SDI maintained a negative correlation with EAPC (R=−0.063, P = 0.37), with both correlations being relatively weak(Fig. 3E-F).

Age, period, and cohort (APC) model

Figure 4 illustrate the age, period, and cohort effects on ASMR and ASDR for falls among perimenopausal women. Age Effect: Across global and five SDI regions, the ASMR and ASDR for falls are notably higher in the 50–54 age group. A progressive increase in ASMR and ASDR is observed with advancing age. The period effect reveals that post-2005, ASMR shows a declining trend globally and in middle and low-middle SDI regions, followed by an increase, indicating a rising burden, whereas other regions exhibit a consistent decline. For ASDR, except for high SDI regions which show an initial increase followed by a decrease, and middle SDI regions which show a decrease followed by an increase, other regions demonstrate a declining trend. The cohort curve indicates that, based on the 1962 birth cohort’s ASMR and ASDR as reference points, there is a discernible downward trend in both ASMR and ASDR across the global and five SDI regions.

Fig. 4.

Age-period-cohort analysis of ASMR and ASDR due to falls among perimenopausal women across the global and five SDI regions. ASMR, age-standardized mortality rate; ASDR, age-standardized DALYs rate; SDI, Socio-Demographic Index; DALYs, disability-adjusted life years

Decomposition analysis

From 1990 to 2021, the global mortality burden from falls among perimenopausal women was predominantly attributed to population growth, accounting for 162.91%, with 0.71% attributed to population aging, while epidemiological changes had a protective effect at −63.63%. For the burden of DALYs, although the percentage attribution differed, the attribution trends were consistent. In the five SDI regions, the mortality burden from falls among perimenopausal women showed a higher attribution percentage in high and high-middle SDI regions, primarily due to population growth, at 332.93% and 329.10%, respectively. In high SDI regions, aging had a protective effect at −149.69%, while in high-middle SDI regions, epidemiological changes had a protective effect at −179.01%. For the DALYs burden, despite differences in percentage attribution, the attribution trends remained consistent. In the 21 GBD regions, the mortality burden showed a higher attribution percentage in Central Asia, primarily due to population growth at 1499.63%, while epidemiological changes had a protective effect at −1192.78%. The DALYs burden was more pronounced in the high-income Asia Pacific region, mainly attributed to population growth at 1137.34%, with population aging changes having a protective effect at −746%. Globally and in most regions, the burden of falls among perimenopausal women was primarily attributed to population growth, exerting a positive effect, while epidemiological changes generally had a negative effect(Fig. 5 and Table S5-6).

Fig. 5.

(A,B) Assessment of the temporal trends in perimenopausal women-associated fall mortality and DALYs across global, 5 SDI regions, and 21 GBD regions, considering population dynamics influenced by global population growth, aging, and epidemiological shifts from 1990 to 2021. Black dots represent the aggregate change attributable to all three components. For individual components, a positive magnitude signifies a corresponding increase in perimenopausal women fall mortality and DALYs linked to that component. Negative magnitudes denote a corresponding decrease in perimenopausal women fall mortality and DALYs associated with the respective component. SDI, Socio-demographic Index. DALYs, disability-adjusted life years

Health inequality analysis

The ASMR for falls among perimenopausal women consistently decreases with increasing SDI levels, while the ASDR shows a continuous rise. In terms of absolute inequality in the burden of falls among perimenopausal women, the SII for fall mortality and DALYs in 2021 were − 5.83 (−7.22, −4.44) and 149.92 (87.87, 211.97), respectively, indicating a decline compared to 1990. In 2021, the CI of disease burden was − 0.16 and − 0.10, showing a slight decrease from 1990. Nevertheless, these figures highlight significant disparities in disease burden distribution across countries with varying SDI levels, underscoring the need to address health equity issues, particularly in low-SDI countries(Fig. 6A-D).

Fig. 6.

(A-D) Analysis of falls among perimenopausal women mortality and DALYs stratified by the sociodemographic index and corresponding concentration indices for population-level disease distribution. (E-F) Frontier analysis of ASMR and ASDR related to SDI in falls among perimenopausal women from 1990 to 2021. ASMR, age-standardized mortality rate; ASDR, age-standardized DALYs rate; SDI, Socio-Demographic Index; DALYs, disability-adjusted life years

Frontier analysis

Utilizing data from 1990 to 2021, a frontier analysis was conducted based on ASMR, ASDR, and the SDI to explore potential improvements in ASMR and ASDR for perimenopausal women falls, considering national and regional development levels (Fig. 6E-F). The solid black line represents the frontier, while the dots represent countries and regions. Blue dots indicate an upward trend, whereas red dots indicate the opposite. As the SDI value increases from 0.0 to 1.0, there is an overall decline in ASMR and ASDR for perimenopausal women falls, characterized by a year-on-year density shift from light to dark colors, indicating a general decrease in ASMR and ASDR. The frontier analysis results for 2021 clearly depict disparities among countries and regions. In terms of mortality rates, significantly higher rates were observed in 15 countries, including India, Palau, Cambodia, Bhutan, and Cuba, placing them outside the frontier. In contrast, Somalia and Niger are closer to the frontier, suggesting optimal outcomes under their SDI. In the DALYs analysis, countries like Somalia, South Sudan, and Chad have rates close to the benchmark set by the frontier; significantly higher rates were observed in India, Andorra, and Belgium, placing them outside the frontier. This analysis reveals potential areas for improvement in reducing the burden of falls among perimenopausal women across different countries and regions (Fig. 6E-F).

Proportion of mortality and dalys attributable to risk factors

Regionally, low bone mineral density was the leading cause of DALYs and mortality due to falls among perimenopausal women globally, across the five SDI regions, and within the 21 GBD regions in 2021. This was followed by occupational injuries. Age-wise stratification revealed that low bone mineral density remained the most significant risk factor for DALYs and mortality from falls in perimenopausal women, followed by occupational injuries, with a notable increase in the 50–54 age group.Temporal trend analysis indicated that, globally and across the five SDI regions, the risk factors contributing to increased mortality from falls among perimenopausal women showed minimal change, except for a slight upward trend in smoking. For DALYs, smoking showed a slight increase globally and in the five SDI regions, while heavy alcohol consumption exhibited a declining trend globally and in middle SDI regions, an initial rise followed by a decline in low-middle and low SDI regions, and an initial decline followed by a rise in high and high-middle SDI regions. Other changes were minimal(Fig. S8-9).

Projecting disease burden

From 2022 to 2050, the ASMR and ASDR for falls among perimenopausal women globally are projected to decline. Specifically, by 2050, the ASMR and ASDR for this demographic are expected to decrease to 1.84 (95% CI: 1.56, 2.12) and 6.88 (95% CI: 6.25, 7.52), respectively(Fig. 7A-D, and Table 7, and 8). Analyzing the ASMR and ASDR for the age groups 45–49 and 50–54, both are anticipated to follow a downward trend(Fig. 7A-D, and Table 9, and 10). Additionally, the global mortality for falls among perimenopausal women is projected to continue its decline, reaching 59,840.51 (95% CI: 50,655.16, 69,025.87) by 2050. The DALYs is also expected to decrease, with projections indicating a reduction to 8,311,820.54 (95% CI: 7,091,913.43, 9,531,727.65) by 2050(Fig. 7G-H, and Table 11, and 12). Analyzing the cases across different age groups, a trend of initial slight increase followed by a decline in mortality and DALYs is anticipated, with both age groups exhibiting similar patterns(Fig. 7G-H, and Table 13, and 14). The future burden of falls among perimenopausal women globally is expected to decrease.

Fig. 7.

The BAPC model analyzes and forecasts the cases and age-standardized rates of mortality and DALYs related to falls among perimenopausal women globally, from 1990 to 2050, including overall trends and age-group trends. ASMR, age-standardized mortality rate. ASDR, age-standardized DALYs rate. DALYs, disability-adjusted life years

Discussion

This study leverages data from GBD 2021 to examine the mortality and DALYs attributable to falls among perimenopausal women. Employing a multidimensional approach, this study analyzes these burdens across various geographic levels, age groups, and temporal periods. Advanced statistical methods, including decomposition analysis, frontier analysis, health inequality analysis, and the APC model, are utilized to elucidate the epidemiological patterns of falls in this population. Furthermore, this study investigates associated risk factors and applies the BAPC model to project the global disease burden of falls among menopausal women over the next three decades. Results indicate that from 1990 to 2021, despite a decline in both the ASMR and ASDR of falls in perimenopausal women globally (ASMR decreased by 7.00%, ASDR decreased by 14.62%), the absolute number of cases significantly increased (by 116.99% and 38.31%, respectively)(Table 1, S1). This is consistent with previous related research findings [27]. The core drivers of this paradoxical trend appear to be population growth (contributing 162.91% to mortality, as shown by decomposition analysis) and aging (contributing 0.71%)(Fig. 5 and Table S5-6).

Analysis across different SDI regions revealed that regions with lower SDI experienced a more pronounced increase in ASMR and ASDR, indicating a heavier burden. Conversely, regions with higher SDI showed the opposite trend(Table 1,S1). This may be attributed to advancements in economic and resource allocation in higher SDI regions, leading to improved management of low bone density (identified as the primary risk factor in risk factor analysis), fall prevention interventions (such as home environment modifications), and emergency response systems, resulting in better disease management and a lower disease burden, while the opposite is true for low-SDI regions. Furthermore, inadequate nutritional supply in low SDI regions is also one of the reasons for the relatively higher burden [28]. Further analysis of the burden’s causes revealed that in high-SDI regions, population growth was the primary contributor to mortality and DALYs. Notably, the largest increase in EAPC for ASMR and ASDR was observed in high-income North America, while the largest decrease was in Central Europe(Fig. 1). This may be closely related to lifestyle changes (sedentary behavior, increased obesity leading to decreased balance) and accelerated aging in high-income North America [28]. Analyzing the ASMR by country, the EAPC indicates an increasing trend in most nations. Cuba, Japan, Australia, the United States, and Brazil exhibit the most significant increases. Conversely, the Czech Republic and Kyrgyzstan show notable decreases. Examining the ASDR’s EAPC, the largest increases are observed in the Netherlands, Cuba, Spain, Puerto Rico, and Canada, while Kyrgyzstan, Hungary, and Latvia demonstrate the most substantial declines. Overall, countries such as Cuba and the Netherlands are experiencing an increased burden, whereas the Czech Republic and Kyrgyzstan show marked improvements(Table S2-3, and Fig. 1,S1). Furthermore, an analysis of the 45–49 and 50–54 age groups reveals that the 50–54 age group bears a heavier burden(Fig. 2).

An analysis of the relationship between ASMR, ASDR, and SDI across 21 GBD regions reveals that ASMR decreases with increasing SDI, while ASDR exhibits a trend of initial increase, followed by a decrease, and then another increase and decrease(Fig. 3). South Asia demonstrates a burden significantly exceeding expectations, whereas Sub-Saharan Africa shows a lower-than-expected burden, with most other regions aligning closely with projections. In 204 countries, ASMR initially declines and then increases with rising SDI. The majority of these nations exhibit burdens near the expected levels, although India, Palau, Cambodia, Bhutan, and Cuba significantly surpass these expectations. Conversely, ASDR demonstrates a pattern of a slight initial decrease followed by a substantial increase with rising SDI(Fig. 3). While most countries align with expected burdens, India, Andorra, Nepal, Greenland, and Belgium notably exceed these levels. The frontier analysis of fall incidence among perimenopausal women in 2021 indicates significantly elevated rates in 15 countries, including India, Palau, Cambodia, Bhutan, and Cuba, placing them beyond the frontier. Conversely, Somalia and Niger, among others, are closer to the frontier, suggesting optimal outcomes relative to their SDI. DALYs analysis reveals that Somalia, South Sudan, and Chad exhibit rates near the frontier benchmark, while India, Andorra, and Belgium demonstrate significantly higher rates, positioning them beyond the frontier. This highlights potential areas for improvement in reducing the burden of falls among perimenopausal women across different countries and regions. Health inequality analysis reveals a non-linear relationship between ASMR and SDI: ASMR is lower than expected in low-SDI countries (e.g., Sub-Saharan Africa) due to a higher proportion of young populations, but forms a “U-shaped curve” in middle-to-high SDI countries (e.g., India, Cuba) due to lagging chronic disease management and accelerated aging(Fig. 3). Frontier analysis further indicates that countries like India and Belgium are “health depressions” due to fragmented healthcare systems (e.g., disconnect between specialized services and primary care) and chronic disease comorbidities (e.g., diabetic neuropathy increasing fall risk)(Fig. 6). Cohort effects demonstrate a general decline in fall risk (RR < 1) among those born after 1962, reflecting the long-term benefits of improved childhood nutrition (reducing the risk of perimenopausal sarcopenia) and occupational safety regulations (e.g., slip-resistant work environments) globally. However, the worsening period effect (RR >1) in middle-to-low SDI regions after 2005 may be associated with the rapid urbanization process, including crowded living environments (e.g., lack of handrails in old buildings) and increased traffic injuries(Fig. 4). Attributable risk factors indicate that low bone mineral density is a core risk factor for falls in perimenopausal women, particularly in the 50–54 age group (accelerated bone flow due to estrogen decline)(Fig. S8-9). Research suggests that osteoporosis (i.e., reduced bone density) is a systemic skeletal disease characterized by decreased bone density and microstructural deterioration, leading to increased susceptibility to fractures after falls. Perimenopausal women are often more likely to be diagnosed with osteoporosis due to estrogen deficiency [29]– [30]. Furthermore, perimenopausal women frequently experience sarcopenia, which elevates the risk of falls [31]. The co-occurrence of osteoporosis and sarcopenia, termed “osteosarcopenia” [32], further exacerbates fracture risk [33]– [34]. Given that both muscle and bone cells express estrogen receptors, hormone replacement therapy (HRT) in postmenopausal women can preserve both bone and muscle mass [35]. Early menopause without exogenous estrogen treatment represents a significant risk factor for future fragility fractures [36].

Perimenopausal women, a distinct demographic, undergo a spectrum of physiological and psychological alterations [37]. Research indicates that postmenopausal women, irrespective of age, exhibit diminished overall functionality, strength, and gait speed compared to their premenopausal counterparts [38]– [39]. Falls represent a significant concern for postmenopausal women [40]. Furthermore, a bidirectional relationship exists between falls and both depression and anxiety; falls elevate the likelihood of depressive symptoms, which, in turn, heighten fall risk [41]. Given that perimenopausal women are more susceptible to depressive and anxiety symptoms, this exacerbates the risk of falls [42]. Consequently, understanding the epidemiological characteristics of falls in perimenopausal women, in conjunction with the respective economic contexts of different countries and regions, is crucial for formulating appropriate policies and management strategies. Such measures are essential for mitigating falls in perimenopausal women and, potentially, for preventing falls in postmenopausal women. Based on this study, this research projects a decline in the ASMR, ASDR, mortality, and DALYs related to falls among perimenopausal women globally from 2022 to 2050. Both age groups exhibit a decreasing trend in ASMR and ASDR. However, the corresponding case numbers for both age groups are anticipated to initially increase slightly before decreasing. Despite the overall decreasing burden, comprehensive and rational countermeasures are still warranted. A multi-dimensional analysis indicates that effective management of falls in perimenopausal women necessitates a multifaceted approach. Primarily, it is crucial to recognize that the specific risks associated with perimenopausal women are the core drivers of the disease burden. Specifically, reduced bone mineral density, directly linked to declining estrogen levels, elevates the risk of falls, particularly within the 50–54 age group. Hormone therapy, utilizing estrogen plus progestin or estrogen alone, is considered to increase net bone density, thereby mitigating the risk of falls [43–45]. In addition to established risk factors, such as excessive alcohol consumption [46]– [47] and smoking [48], the risk of falls is also elevated. Consequently, tobacco and alcohol control policies should be implemented to mitigate these risks. From a foundational health screening and prevention perspective, the following global strategies are recommended: Advocate for the WHO to incorporate bone mineral density screening for perimenopausal women into essential health service packages. Support calcium/vitamin D supplementation programs in low SDI countries, such as those in Sub-Saharan Africa, through global funding initiatives. Establish a “Fall Prevention Consortium” utilizing the GBD data platform to facilitate the sharing of best practices. At the national level, the following strategies are recommended: In low-SDI countries, integrate orthopedic and geriatric resources through a “Fracture Liaison Service” to enhance basic bone health screening, such as the utilization of dual-energy X-ray absorptiometry [49]. In middle-to-high SDI countries, establish real-time fall injury surveillance systems. In high-SDI regions (e.g., North America, the European Union), optimize menopause medication management guidelines to reduce the over-prescription of antidepressants and hypnotics.At the population level, public health policymakers can implement large-scale educational initiatives to enhance public awareness and understanding of the disease, its risk factors, and preventive strategies [29].This includes informing the target population about non-pharmacological interventions to maintain skeletal health and prevent osteoporotic fractures resulting from falls. For instance, a balanced diet, such as the Mediterranean diet, which research indicates is rich in nutrients and bioactive compounds potentially protective against muscle and bone degradation, can serve as a therapeutic tool to mitigate the onset of osteoporosis and sarcopenia [50].Furthermore, lifestyle modifications, including exercise (e.g., Tai Chi) [51], weight-bearing activities, and muscle-strengthening exercises, can be beneficial [52]. In terms of pharmacological interventions, bisphosphonates [53], vitamin D, and calcium supplementation have demonstrated efficacy [54]. Furthermore, with the rapid and ongoing advancements in medical technology, the integration of the Internet of Things (IoT) into clinical practice represents a transformative breakthrough. IoT technologies enable real-time monitoring and data collection of patient health status through interconnected smart devices and sensors [55]. In the context of managing fall risk among menopausal women, IoT can utilize wearable devices to monitor gait, balance, and activity patterns, providing timely alerts for fall risk and supporting medical interventions to enhance preventive outcomes [56].

However, this study has certain limitations, including: Estimates based on GBD data-dependent models may underestimate the burden of fall-related incidents, particularly in rural areas; the correlational analysis did not incorporate cultural factors (e.g., delayed medical consultation due to menopause stigma); BAPC predictions did not account for the potential impact of emerging technologies.

Conclusion

This study employed GBD 2021 data to systematically analyze the global disease burden of falls among perimenopausal women from 1990 to 2021. Despite a 7.00% reduction in ASMR and a 14.62% decrease in ASDR, the absolute numbers of deaths and DALYs significantly increased due to population growth and aging, highlighting the offsetting effect of demographic shifts on prevention outcomes. Regional disparities were evident: low-SDI regions experienced an increased burden due to inadequate bone health management, while high-SDI regions reduced risk through fall prevention interventions. However, high-income North America showed a rise in ASMR due to sedentary lifestyles. Low bone mineral density was identified as a core biological factor, particularly in the 50–54 age group, compounded by fragmented healthcare systems (India) and chronic comorbidities (Belgium), exacerbating health inequalities. Cohort effects indicated a risk reduction in those born after 1962 due to improved nutrition, but a rebound in risk after 2005 in rapidly urbanizing areas due to crowded living environments. Projections to 2050 suggest continued declines in ASMR and ASDR, but high case numbers persist, necessitating multi-tiered interventions: Globally: Integrate bone density screening into essential health services.Nationally: Integrate orthopedic and geriatric resources.At the population level: Promote resistance training, Mediterranean diet, and IoT-based intelligent fall monitoring.Study limitations include potential underestimation of non-fatal falls in GBD data; future research should incorporate real-world data to refine monitoring.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

DALYs

Disability-adjusted life years

BAPC

Bayesian age-period-cohort

ASRs

Age-standardized rates

ASMR

Age-standardized mortality rate

ASDR

Age-standardized disability-adjusted life years rate

PC

Percentage change

APC

Annual percentage change

AAPC

Average annual percentage change

EAPC

Estimated annual percentage change

CI

Confidence interval

Author contributions

The study was conceived and designed by LW, YSM, and ZPL. LW, XLZ, ZPL, CZ and SJZ conducted statistical analyses using R The study was conceived and designed by LW, YSM, and ZPL. LW, XLZ, ZPL, JJL, CZ, and SJZ conducted statistical analyses using R programming language, with responsibility for data interpretation and generation of all visual representations (figures) and tabular data. LW, YSM, ZPL and XLZ collaboratively prepared the initial manuscript draft. HL, JZ and TG contributed to data acquisition from public repositories and performed validation analyses. YLD served as corresponding author, providing critical revisions to intellectual content, overseeing research implementation, and securing project funding through grant acquisition. All authors participated in manuscript review processes and approved the final version for publication.

Funding

This work is supported by Development Center for Medical Science & Technology National Health Commission of the People’s Republic of China(W2018RY9).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The Research Ethics committee of the second Qilu hospital of Shandong University has determined that this study does not require ethical approval or participant consent, as it utilizes publicly available data from the GBD2021 database. Therefore, ethical review is not necessary. Additionally, informed consent from participants is not applicable in this research.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.