The global burden of osteoarthritis knee: a secondary data analysis of a population-based study

Data sources and extraction

This study was a secondary data analysis of a population-based study named the GBD 2021 Study. The annual counts of cases and the corresponding age-standardized rates (ASRs) pertaining to osteoarthritis knee, both globally and by subtypes, were retrieved from the GBD 2021 Study database¹⁷. The GBD 2021 Study, which was carried out by an independent population health research organization based at the University of Washington School of Medicine, Institute for Health Metrics and Evaluation¹⁷–¹⁹. Initiated in 1990 and regularly updated, the GBD Study has emerged as a cornerstone in comprehending the global health scenario¹⁸. Through rigorous quantification of health losses attributed to diverse diseases, injuries, and risk factors, the GBD Study furnishes a robust framework for monitoring trends, identifying disparities, and informing policy interventions¹⁷, ¹⁹. Over the past three decades, the GBD Study's scope has expanded to encompass a wider array of health outcomes, including non-fatal outcomes like disability, thereby offering a holistic perspective on the burden of diseases¹⁷, ¹⁹.

The GBD 2021 dataset encompassed data spanning 1990 to 2021 for 204 countries and territories, which were further categorized into 28 GBD super regions and 54 GBD regions based on geographical considerations¹⁷. Additionally, these countries and territories were grouped into five regions according to their socio-demographic index (SDI) values²⁰, ²¹.

For this study, we adopted the human development index (HDI), a metric introduced by the United Nations Development Programme (UNDP) in 1990²², ²³. The HDI is a composite indicator encompassing education, life expectancy, and gross national income, serving as a comprehensive measure of the economic and social development level of United Nations member states²², ²³. Since its inception, the HDI has played a pivotal role in guiding developing countries in devising appropriate development strategies²², ²³. Annually, the UNDP publishes the HDI for countries worldwide and utilizes it to assess the human development status of each country in the Human Development Report²⁴, ²⁵.

Statistical analysis

In this study, we initially reported the incidence, prevalence, and Disability-Adjusted Life Years (DALYs) attributed to osteoarthritis knee, accompanied by their respective ASRs, for the year 2021. This analysis encompassed a global perspective and was further stratified by various subtypes, encompassing age groups, sex, SDI regions, GBD regions, and individual countries.

Subsequently, we delved into investigating temporal trends in osteoarthritis knee burden, both globally and within specified subtypes, spanning from 1990 to 2021. To quantify the rate of change in this burden over the study period, we estimated the Estimated Annual Percentage Change (EAPC) utilizing a linear regression model. Based on the EAPC values obtained, we conducted a hierarchical cluster analysis to systematically categorize the 54 GBD regions into four distinct clusters: those exhibiting a significant increase, a minor increase, remaining stable or experiencing a minor decrease, and those with a significant decrease in osteoarthritis knee burden. To probe potential associations between EAPCs and socioeconomic factors, we examined the correlation between EAPCs and osteoarthritis knee-related ASRs in 1990, as well as HDI scores in 2021. Given the non-normal distribution of these variables, we adopted the Spearman correlation analysis, a non-parametric approach, to assess the strength and direction of these associations.

Furthermore, we employed two advanced time-series models to forecast the future burden of osteoarthritis knee. These models included the Age-Period-Cohort (APC) model under the maximum likelihood framework and the Bayesian APC (BAPC) model, which was integrated with nested Laplace approximations. To ensure the robustness of our predictions, we also conducted a sensitivity analysis utilizing Autoregressive Integrated Moving Average (ARIMA) and Exponential Smoothing (ES) models.

Statistical significance was set at 0.05 (two-tailed). All the analyses were performed (or conducted) using R software (version 4.0.2). The “tidyverse” package in R was used to construct linear regression models to calculate EAPC values, the “cluster” package was used for k-mean clustering analysis, the “nordpred” package was used to construct the APC model, the “BAPC” and “INLA” packages were used for the BAPC model, and the “forecast” package was used to construct the ARIMA model and ES model.

The disease burden of osteoarthritis knee in 2021

In 2021, the number of incidence cases of osteoarthritis knee was 30.85 million [95% uncertainty interval (UI): 26.53–35.19] globally. The corresponding age-standardized incidence rate (ASIR) was 353.67 (95% UI: 304.56–402.5) per 100,000 population. The number of prevalence cases was 374.74 million (95% UI: 321.86–428.35) globally. The corresponding age-standardized prevalence rate (ASPR) was 4294.27 (95% UI: 3695.04–4910.76) per 100,000 population. Moreover, the number of osteoarthritis knee-related DALYs was 12.02 million (95% UI: 5.86–23.27) globally in 2021. The corresponding ASR of DALYs was 137.59 (95% UI: 67.08–266.87) per 100,000 population (Tables 1, 2, and 3).

Gender-specific analysis revealed that in 2021, the number of incidence, prevalence, and DALYs counts, along with their respective ASRs, were higher in females compared to males (Figure S1, Tables 1, 2, and 3). An age-stratified analysis of incidence, prevalence, and DALYs in 2021 is presented in Figure S1. For the the number of cases and corresponding ASRs, there was an initial increase with age, peaking in around older years old, followed by a decrease (Figure S1, Tables 1, 2, and 3).

At the SDI region level, the high SDI region had the most ASIR at 386.58 (95% UI: 334.58–439.34), ASPR at 4656.7 (95% UI: 4033.58–5313.55), and the age-standardized DALYs rate (ASDAR) at 149.14 (95% UI: 73.09–292.89). For the number of incidence, prevalence, and DALYs cases, the disease burden was highest in the middle SDI regions (Figure S1, Tables 1, 2, and 3). Moreover, the region with the lowest disease burden was low SDI regions. The regular pattern of different SDI levels and disease burden was stable across countries and territories. From Fig. 1, the ASRs for incidence, prevalence, and DALYs showed an “J” relationship with SDI, where ASRs increase with SDI (Fig. 1).

Across the 54 GBD regions, High-income Asia Pacific ranked the top one in osteoarthritis knee-related ASIR (458.22, 95% UI: 397.65–522.59), ASPR (5573.73, 95% UI: 4841.29–6334.33), and ASDAR (180.61, 95% UI: 88.02–356.06), followed by East Asia. Meanwhile, Central Asia ranked the bottom one (ASIR: 239.29, 95% UI: 205.97–274.09; ASPR: 239.29, 95% UI: 205.97–274.09; ASDAR: 86.71, 95% UI: 42.6–169.71). For the number of cases, the top GBD region was Asia (incidence: 18.94 million, 95% UI: 16.33–21.70; prevalence: 229.88 million, 95% UI: 197.01–263.57; DALYs: 7.41 million, 95% UI: 3.58–14.27), followed by Basic Health System and World Bank Upper Middle Income. The bottom was Oceania for incidence, prevalence, and DALYs (incidence: 33635, 95% UI: 28494–39060; prevalence: 338675, 95% UI: 285507–391076; DALYs: 10965, 95% UI: 5248–21272) (Figure S1, Tables 1, 2, and 3).

The disease burden of osteoarthritis knee varied considerably across the world. For number of cases, the top number of incidence cases, prevalence cases, and DALYs cases were observed in People's Republic of China (incidence: 8.51 million, 95% UI: 7.28–9.84; prevalence: 109.58 million, 95% UI: 92.72–126.64; DALYs: 3.55 million, 95% UI: 1.72–6.84) in 2021, followed by Republic of India and United States of America. Moreover, the bottom number of cases were observed in Tokelau (incidence: 5, 95% UI: 5–6; prevalence: 66, 95% UI: 57–76; DALYs: 2, 95% UI: 1–4), followed by Republic of Niue and Republic of Nauru. As for the ASRs, the highest ASIR, ASPR, and ASDAR was observed in Republic of Korea (ASIR: 491.74, 95% UI: 427.19–560.51; ASPR: 6201.62, 95% UI: 5389.51–7093.43; ASDAR: 199.93, 95% UI: 97.42–398.67). And the lowest was observed in Republic of Tajikistan (ASIR: 218.48, 95% UI: 189.9–250.18; ASPR: 2425.49, 95% UI: 2101.17–2773.29; ASDAR: 78.21, 95% UI: 38.74–156.27) (Fig. 2, Tables 1, 2, and 3).

Temporal trend for osteoarthritis knee-related disease burden from 1990 to 2021

Globally, the number of cases of osteoarthritis knee has showed a increasing trend from 1990 to 2021. The corresponding ASRs demonstrated the same trend. The number of incidence cases increased from 14.13 million (95% UI: 12.15–16.08) to 30.85 million (95% UI: 26.53–35.19), and the ASIR increased from 330.26 (95% UI: 284.34–375.75) to 353.67 (95% UI: 304.56–402.5). The number of prevalence cases increased from 159.80 million (95% UI: 137.28–182.88) to 374.74 million (95% UI: 321.86–428.35), and the ASPR increased from 3964.75 (95% UI: 3411.86–4536.4) to 4294.27 (95% UI: 3695.04–4910.76). And the number of DALYs cases increased from 5.15 million (95% UI: 2.51–9.95) to 12.02 million (95% UI: 5.86–23.27), while the ASR of DALYs increased from 127.14 (95% UI: 62.17–246.99) to 137.59 (95% UI: 67.08–266.87) (Fig. 3, Tables 1, 2, and 3).

The trends observed in both sexes were found to be congruent with those of the overall population, suggesting a uniformity in the disease dynamics across this demographic subgroup (Figure S2, Tables 1, 2, and 3). The disease burden for most age groups also revealed same trend (Figure S2, Tables 1, 2, and 3).

At the SDI regional level, our analysis reveals a pattern in the trends of osteoarthritis knee indicators. For the ASRs and number of cases, all SDI regions showed the same trend as the overall population (Figure S2, Tables 1, 2, and 3).

To further elucidate the variability in osteoarthritis knee burden across GBD regions, we conducted a hierarchical clustering analysis. The results, presented in Fig. 4, reveal distinct clusters of regions sharing similar patterns of disease burden change. Notably, significant increases in ASRs were observed in regions spanning World Bank Upper Middle Income, World Bank Lower Middle Income, Commonwealth Middle Income, Latin America & Caribbean -WB, Limited Health System, South Asia -WB, South Asia, South-East Asia Region, and Tropical Latin America. In contrast, regions experiencing significant decreases in ASRs included Eastern Mediterranean Region, Australasia, Southern Latin America, Middle East & North Africa-WB, Northern Africa, North Africa and Middle East, Southeast Asia, and Andean Latin America (Fig. 4).

Across countries and and territories, the ASRs all showed a increasing trend except for United States of America. The most pronounced increase in ASIR and ASPR from 1990 to 2021 was observed in Republic of Equatorial Guinea [ASIR: EAPC = 0.6, 95% confidence interval (CI): 0.56–0.64; ASPR: EAPC = 0.68, 95% CI: 0.64–0.72] and Sultanate of Oman (ASIR: EAPC = 0.6, 95% CI: 0.58–0.62; ASPR: EAPC = 0.68, 95% CI: 0.66–0.70), while the most pronounced increase was only observed in Republic of Equatorial Guinea for ASDAR (EAPC = 0.71, 95% CI: 0.67–0.76) (Figure S3, Tables 1, 2, and 3).

The influential factors for EAPC

The results demonstrate a discernible correlation between EAPCs and ASRs of osteoarthritis knee in 1990, as well as HDIs in 2021, respectively (Figure S4). Specifically, the ASRs recorded in 1990 for osteoarthritis knee reflect the baseline disease burden, whereas the HDIs in 2021 serve as indicators of healthcare accessibility and a surrogate measure for the health system's maturity within each country. The positive association was observed between EAPCs and ASRs across multiple health outcomes, including incidence (P = 0.56 ρ = 0.04), prevalence (P = 0.29 ρ = 0.07), and DALYs (P = 0.79 ρ = 0.02), though the association was not significant. Furthermore, the link between EAPCs for incidence, prevalence, and DALYs and HDIs was also reach statistical correlation. However, the correlation was negative for incidence (P = 0.68 ρ = −0.09), prevalence (P = 0.47 ρ = −0.15), and DALYs (P = 0.65 ρ = −0.10) of osteoarthritis knee with the respective HDIs (Figure S4).

The predicted results from 2022 to 2046

The APC model showed that the number of incidence cases for males increased from 12.94 million to 18.45 million during the next 25 years, and it increased from 18.80 million to 25.60 million for females. The number of prevalence cases for males increased from 149.08 million in 2021 to 235.41 million in 2046, it increased from 236.39 million to 365.97 million for females. The number of DALYs cases increased from 4.82 million to 7.52 million for males, and it increased from 7.55 million to 11.55 million for females. The projected outcomes from the BAPC models also indicate an upward trend in the number of cases for both genders, spanning from 2022 to 2046. For males, the number of incidence cases increased from 16.45 million to 19.52 million, the number of prevalence cases increased from 200.10 million to 240.94 million, and the number of DALYs cases increased from 6.39 million to 7.56 million. For females, the number of incidence cases increased from 11.71 million to 13.85 million, the number of prevalence cases increased from 138.25 million to 166.31 million, and the number of DALYs cases increased from 4.47 million to 5.36 million.

For ASRs, the prediction results showed a very slight change. The APC model showed that it showed a decreasing trend for females but exhibited a trend of first increase and then decrease for males. However, for the BAPC model, the ASRs for both genders exhibits a trend of first increase and then decrease (Figs. 5, Tables 4 and 5).

Collectively, these findings underscore the persistence of a relatively substantial future disease burden, which is further corroborated by the results obtained from the ARIMA model and the ES model (Figures S5, Tables S1-S2).

As far as we know, this was the first study to comprehensively assess and quantify osteoarthritis knee-related disease burden globally. Globally, it caused a great disease burden in 2021, and significant differences in the disease burden existed between sexes and across ages, SDI regions, GBD regions, and countries. From 1990 to 2021, there was an decreasing trend for the disease burden globally. Furthermore, our predicted results showed that the disease burden would increase in the next 25 years.

The current findings highlight a notable increase in the global burden of osteoarthritis knee from 1990 to 2021, with substantial rises in both incidence and prevalence cases. Our estimates of the age-standardized incidence and prevalence rates align with previous reports, underscoring the escalating trend in this condition worldwide²⁶–²⁸. Notably, the marked escalation in DALYs further emphasizes the substantial impact of osteoarthritis knee on population health and wellbeing.

Our gender-specific analysis underscores a higher burden of osteoarthritis knee among females, both in terms of incidence, prevalence, and DALYs, as compared to males. This disparity aligns with previous research reporting similar trends, reinforcing the notion that osteoarthritis knee disproportionately affects women²⁹–³¹. The higher burden of osteoarthritis knee among females can be attributed to several social causes. Firstly, women are more likely to experience hormonal changes, particularly during menopause, which may contribute to the development and progression of osteoarthritis due to alterations in joint tissue metabolism and inflammation. Additionally, women are more prone to having specific anatomic features, such as wider hip bones and narrower femoral canals, which may increase the risk of knee osteoarthritis. Furthermore, societal norms often result in gender-specific roles and responsibilities, with women more frequently engaging in activities that place repetitive stress on their knees, such as childcare, household chores, and certain professions³², ³³. These factors, combined with the global trend of increasing obesity among women³⁴, which is a significant risk factor for osteoarthritis, contribute to the higher incidence, prevalence, and DALYs associated with knee osteoarthritis among females compared to males. The congruency in trends observed between genders further highlights the uniformity in disease dynamics across these demographic subgroups, emphasizing the need for gender-sensitive strategies in managing and preventing osteoarthritis knee. As the disease burden in women continues to escalate, it underscores the urgency for tailored interventions that address the unique risk factors and healthcare needs of this population, aiming to mitigate the disproportionate impact of osteoarthritis knee on women's health and quality of life.

The age-stratified analysis of osteoarthritis knee incidence, prevalence, and DALYs in 2021 reveals a characteristic age-related pattern, with an initial rise followed by a peak in older age groups and subsequent decline. This trend mirrors previous findings, underscoring the disproportionate impact of osteoarthritis knee on the elderly population³⁵–³⁷. This trend can be attributed to several factors, including the increased utilization of prosthesis treatment among older individuals with severe osteoarthritis, which can effectively manage symptoms and reduce the disease burden in this age group. Additionally, changes in physical activity levels with age, such as a reduction in vigorous activities among the elderly, may also contribute to the observed decline in cases. However, it is important to note that the complex interplay between genetic, environmental, and behavioral factors likely plays a role in this age-related pattern. To support this explanation, we refer to the study by Cross, which similarly observed an age-related increase in osteoarthritis cases, peaking in older age groups³⁸. Overall, these findings underscore the need for further research to fully understand the underlying causes of this age-related trend and to develop targeted interventions to mitigate the burden of osteoarthritis knee globally.

The observed disease burden across most age groups follows a similar trajectory, highlighting the importance of age-specific interventions and healthcare planning to address the unique needs of different age cohorts. As the global population ages, the need for targeted strategies to mitigate the burden of osteoarthritis knee, particularly among the elderly, becomes increasingly pressing. Our findings emphasize the significance of considering age as a critical factor in developing comprehensive prevention and management plans for this prevalent and disabling condition.

Our analysis of osteoarthritis knee burden across SDI regions reveals an interesting juxtaposition: while high SDI regions exhibited the highest age-standardized incidence, prevalence, and DALY rates, the middle SDI regions bore the greatest absolute number of cases. This paradoxical finding is consistent with prior studies highlighting the complexity of healthcare disparities across different socioeconomic strata³⁹–⁴¹. The observed "J" relationship between SDI and ASRs underscores the dual challenge faced by low-SDI regions, grappling with limited healthcare resources, and high-SDI regions, grappling with aging populations and lifestyle-related factors. Importantly, the consistent trends across SDI regions for both ASRs and absolute case numbers, mirrored in the overall population, highlight the universality of osteoarthritis knee as a significant global health issue. This reinforces the need for globally coordinated efforts to address its rising burden, tailored to the unique socioeconomic contexts of different regions. As we strive to reduce the disease burden across all SDI levels, a comprehensive approach that encompasses preventive measures, access to timely and affordable treatment, and improvements in healthcare infrastructure, becomes paramount.

Our findings highlight substantial disparities in the burden of osteoarthritis knee across GBD regions and countries. High-income Asia Pacific emerges as a hotspot, particularly in terms of ASIR, ASPR, and ASDAR, underscoring the need for targeted interventions in this region. The dominance of Asia in overall case numbers, particularly China, India, and the USA, reflects the scale of the challenge globally. Conversely, Oceania and small island nations like Tokelau, Niue, and Nauru exhibit the lowest burdens, likely due to their small populations and differing healthcare systems. The hierarchical clustering analysis reveals distinct patterns of burden change, with notable increases in ASRs across several middle-income regions, suggesting a growing impact of osteoarthritis knee in these areas. Conversely, decreases in ASRs in some regions, such as Australasia and the Middle East, may reflect improvements in healthcare or lifestyle factors. However, the exception of the USA, where ASRs showed no overall increase, underscores the complexity of disease trends and the need for nuanced approaches to prevention and management. The pronounced increases in ASIR, ASPR, and ASDAR in Equatorial Guinea and Oman are particularly noteworthy, indicating rapid changes in disease burden that warrant urgent attention. These findings align with previous studies reporting increasing osteoarthritis prevalence in developing countries⁴², ⁴³, likely driven by factors such as aging populations, obesity, and changes in physical activity patterns.

The observed correlation between EAPCs and ASRs of osteoarthritis knee in 1990, while not statistically significant, suggests a potential historical impact on disease progression. In contrast, the negative correlation between EAPCs and HDIs in 2021 indicates that higher healthcare accessibility, as reflected by HDIs, does not necessarily lead to reduced osteoarthritis knee burden. This finding contrasts with previous studies reporting positive health outcomes associated with improved healthcare indices⁴⁴, ⁴⁵. The disparity may stem from osteoarthritis's multifactorial etiology, including non-modifiable risk factors, limiting the impact of healthcare accessibility alone. Additionally, the observed trends highlight the need for a nuanced approach to osteoarthritis management, incorporating both health system strengthening and targeted interventions addressing modifiable risk factors.

Our findings align with recent projections from various models, consistently indicating a looming increase in osteoarthritis incidence, prevalence, and DALYs across genders over the next two decades⁴⁶, ⁴⁷. While ASRs exhibit minimal variation under the APC model, the BAPC model reveals a nuanced trend of initial decline followed by an increase, highlighting the complexities of disease dynamics. This discrepancy underscores the importance of considering multiple modeling approaches to comprehensively assess future trends. Our results resonate with other studies utilizing ARIMA and ES models, reinforcing the robustness of the projected burden. Collectively, these findings emphasize the urgency for targeted interventions and strategies to mitigate the substantial future osteoarthritis burden.

The present study is constrained by its reliance on the GBD database, which presents several notable limitations. A primary limitation stems from the scarcity of detailed, subnational data at the county, province, and state levels, particularly in vast countries such as China⁴⁸. This constraint necessitated a focus on broader, national-level indicators, thereby limiting the ability to provide more nuanced, region-specific insights into the burden of osteoarthritis knee. Moreover, the GBD database was an aggregated source that does not provide individual-level data, so detailed participant information such as demographics, lifestyles, or specific health metrics was not typically available at an individual level. This prevented us from conducting more in-depth studies, such as risk factor studies. Additionally, the accessibility of primary data within the GBD database poses a substantial challenge, as highlighted in previous studies⁴⁹, ⁵⁰. The reliance on estimated disease burden figures derived using the DisMod-MR, a standardized Bayesian regression framework, introduces another layer of complexity. While this approach facilitates the generation of estimates in data-sparse regions, it also necessitates the "borrowing" of information from comparable nations. This process, while valuable in filling data gaps, may introduce an element of uncertainty into the results, emphasizing the importance of cautious interpretation of the findings. Future research endeavors should strive to incorporate more comprehensive, subnational data and explore alternative methodologies to enhance the accuracy and precision of disease burden estimates, particularly in regions with limited data availability.

In conclusion, osteoarthritis knee constitutes a substantial and growing global health burden. Our research underscores the vulnerability of females and older adults as distinct high-risk groups, emphasizing the need for targeted interventions. Furthermore, our analysis identifies regions with higher SDI as high-risk areas, highlighting disparities in disease burden across different populations. Notably, the predictive models indicate a projected increase in the number of cases for both genders from 2022 to 2046, underscoring the urgency for proactive measures. Consequently, the development and implementation of rigorous mitigation and adaptation strategies are imperative to protect individuals and effectively manage the burden of osteoarthritis knee, thereby reducing its impact on global health and wellbeing.

Below is the link to the electronic supplementary material.

Supplementary material — available with the version of record.

The authors gratefully acknowledge all participants of the GBD 2021 for their contribution.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Authors’ contribution

JLR: data curation, methodology, formal analysis, software, and wrote the manuscript. JNY: data curation and supervision. WH: conceptualization, validation, reviewed and edited the manuscript.

Funding

This study did not receive any funding, grants or other financial support.

Data availability

Please get in touch with the corresponding author for more information.

Declarations

Not applicable.

All the authors have reviewed and approved the manuscript for publication.

The authors have no possible conflicts of interest to declare.