Global, regional and national trends in the epidemiology of rheumatoid arthritis from 1990 to 2021: an age-period-cohort effect analysis of the global burden of disease study 2021

Abstract

Background

Rheumatoid arthritis (RA) poses a significant global health challenge. This study examined trends in the incidence, prevalence and mortality of RA from 1990 to 2021 at global, regional and national levels, alongside age, period and cohort effects.

Methods

Age-standardised rates (per 100 000 population) on RA incidence, prevalence and mortality were extracted from the Global Burden of Disease (GBD) study 2021. An age-period-cohort model estimated the overall annual percentage change (net drift), annual percentage change within each age group (local drift), and age/period/cohort relative risks.

Results

Globally, from 1990 to 2021, the age-standardised incidence rate (ASIR) and the age-standardised prevalence rate (ASPR) of RA increased; while age-standardised mortality rate (ASMR) decreased. Low-middle sociodemographic index (SDI) region showed the most notable increases in incidence and prevalence, while high SDI region had the largest decline in RA mortality. The South-East Asia Region exhibited the most significant increases in incidence and prevalence, and the African Region had the greatest mortality decline. Ireland reported the highest ASIR and ASPR in 2021, while the Republic of Honduras had the highest ASMR. Age effects peaked at 65–69 years for incidence and 80–84 years for prevalence, with the mortality risk increased with age. Period effects indicated rising incidence and prevalence but declining mortality, while younger birth cohorts showed increasing risks of incidence and prevalence and decreasing mortality risk.

Conclusion

RA incidence and prevalence have risen globally, while mortality has declined. The regional, demographic and temporal variations underscore the need for targeted public health strategies to address the growing burden.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Rheumatoid arthritis (RA) is a significant global health issue, imposing great burden on both individuals and society, with varying epidemiology influenced by factors such as socioeconomic status, age distribution and region.

WHAT THIS STUDY ADDS

• From 1990 to 2021, global incidence and prevalence of RA increased, while mortality decreased. These trends varied across sociodemographic regions and WHO regions. Age-related effects showed distinct patterns for incidence, prevalence and mortality, while period and cohort effects indicated higher incidence and prevalence of RA, but lower mortality in more recent periods and younger cohorts.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study highlights the growing incidence and prevalence with reducing mortality of RA across age groups, periods and cohorts, emphasising the importance of region-specific and population-specific interventions for RA to ensure more equitable healthcare access and improved outcomes.

Introduction

Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease characterised by persistent inflammation of the synovial joints, leading to pain, swelling and progressive joint destruction.¹ Beyond musculoskeletal damage, RA has widespread systemic effects, contributing to significant morbidity and increased mortality.^{1 2} RA is frequently associated with a range of comorbidities, including cardiovascular disease, respiratory conditions and infections, which further elevate the risk of premature death.³ Despite advances in treatment, such as disease-modifying antirheumatic drugs (DMARDs)⁴ and biological therapies,⁵ RA remains a lifelong condition that requires continuous management.⁶ These extra-articular manifestations, coupled with the chronic and progressive nature of RA, can severely impair physical function and quality of life,⁷ making RA a major cause of long-term disability worldwide. This ongoing burden, in terms of health impact and healthcare costs,⁸ underscores the importance of understanding the global trends in RA epidemiology for long-term healthcare planning and resource allocation.

The epidemiology of RA, including incidence, prevalence and mortality, has demonstrated substantial variations across regions, populations and time periods over recent decades.^{9 10} These variations are shaped by diverse factors such as sex, socioeconomic status and access to healthcare services.¹¹ Notably, RA disproportionately affects females, with a female-to-male ratio ranging from 2:1 to 4:1 across different age groups.¹² Additionally, the burden of RA differs significantly between high- and low-income countries.¹¹ High-income countries often report higher incidence and prevalence rates, potentially reflecting better healthcare infrastructure, greater access to diagnostic tools, and more comprehensive disease surveillance.¹³ In contrast, low- and middle-income countries frequently face challenges such as delayed diagnosis and limited access to effective treatments, potentially underestimating the true burden of the disease.⁹ These disparities underscore the need for a deeper understanding of regional and national variations in trends of RA epidemiology, to address gaps in assessments for burden of RA and improve global health equity.

While many epidemiological studies have examined overall trends in RA burden,^{10 14} few have systematically accounted for the distinct contributions of age, period and cohort, which are key factors that shape disease patterns over time.¹⁵ Age effects reflect biological changes associated with ageing, while period effects capture external influences, such as medical advancements, changes in diagnostic practices or shifts in public health policies that impact all age groups simultaneously. Cohort effects reflect the unique exposures and risks experienced by specific birth cohorts, often determined by early-life factors or changing environmental conditions that can influence RA risk throughout life.¹⁶ Although ageing is a major driver of RA trends, changes in healthcare access and treatment availability (period effects) as well as generational differences in exposures (cohort effects) also play critical roles.^{9 11 17} Despite the importance of these factors, few studies have systematically disentangled age, period and cohort contributions to RA trends. The age-period-cohort (APC) framework provides a powerful tool for separating these effects, offering nuanced insights into the drivers of RA trends.¹⁸

In this study, we used data from the Global Burden of Disease (GBD) 2021 study to perform a comprehensive APC analysis of RA epidemiology from 1990 to 2021. Our primary objectives were twofold: (1) to assess trends in incidence, prevalence and mortality of RA at the global, regional and national levels, disaggregated by sex, across different sociodemographic index (SDI) regions and WHO regions from 1990 to 2021, and (2) to evaluate the independent effects of age, period and birth cohort on the overall epidemiology of RA.

Methods

Overview of GBD 2021

This study conducted a secondary analysis using data from the GBD 2021 (available at https://vizhub.healthdata.org/gbd-results/). The GBD 2021 study, coordinated by the Institute for Health Metrics and Evaluation, provides comprehensive estimates of the burden of 371 diseases and injuries across 204 countries and territories for the period 1990–2021. Detailed descriptions of GBD methodologies, data sources and statistical modelling are available in prior reports.^{19 20} The GBD 2021 complied with the Guidelines for Accurate and Transparent Health Estimate Reporting statement.²¹

Case definition

RA cases in GBD 2021 were identified and classified following the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10).¹⁹ The relevant ICD-10 codes for RA include M05–M05.9 and M08–M09.8.¹⁹

Data processing and modelling process for RA

RA epidemiological estimates (incidence, prevalence and mortality) were generated using the Bayesian meta-regression tool DisMod-MR V.2.1.²² The estimates are provided for countries and regions. Implementing a Meta-Regression-Bayesian Regularised Trimmed (MR-BRT) model, data not reported by sex were adjusted using the cause-specific pooled sex ratio within each study. Also, for data not reported by both age and sex, age-sex splitting was performed, and age-specific estimates from sources reporting by age and sex (but not by both) were adjusted using the sex ratio within each source to generate age-sex-specific estimates.¹⁹ For those aged under 5 years, it was assumed that there was no incidence, prevalence or mortality of RA.

GBD estimated RA incidence and prevalence in all countries. In cases where input data were unavailable for specific locations, estimates were generated using two main mechanisms: (1) an analytical cascade, wherein models at higher geographical levels (global, super-region and region) informed models at lower levels, and (2) predictive covariates. In regions with no direct data, estimates were informed by super-region priors.¹⁹ For regions without raw data of mortality, lacking figures could be calculated by dividing available prevalence datapoints by corresponding cause-specific mortality data, stratified by age, sex, year and location. These calculated mortality estimates were then modelled in the meta-regression tool, MR-BRT by age and sex, incorporating a prior based on the Healthcare Access and Quality Index, where an increase in this index was associated with a decrease in mortality.^{9 19} Uncertainty was propagated throughout each step of the estimation process, and uncertainty intervals (UIs) were defined as the 2.5th and 97.5th percentiles of the ordered draws from 100 model runs.¹⁹

Data extraction

For this study, we extracted the numbers and age-standardised rates (ASRs) for RA incidence, prevalence and mortality, disaggregated by age, sex, location and year, from 1990 to 2021. All estimates were presented with 95% UIs. All rates are reported per 100 000 population. The SDI is a composite measure that reflects the level of socioeconomic development within a country. It is based on lag-distributed income per capita, average years of education for individuals aged 15 and older, and the total fertility rate among women under 25. The SDI is scaled from 0 to 1, with higher values suggesting higher socioeconomic development levels.¹⁹ All countries were classified into one of five SDI quintiles: low, low-middle, middle, high-middle and high SDI regions. Furthermore, countries were grouped into six WHO regions: African Region, Eastern Mediterranean Region, European Region, Region of the Americas, South-East Asia Region, and Western Pacific Region.

Statistical analysis

Epidemiological description

To ensure comparability across populations and time periods, we used ASRs from GBD to describe the epidemiology of RA. These metrics account for differences in age structures of populations in different regions and time periods, thereby allowing for more accurate comparisons of RA epidemiology across regions and over time. The key epidemiological metrics for RA were reported as age-standardised incidence rate (ASIR), which was expressed as the number of new RA cases per 100 000 population after age standardisation. Similarly, the age-standardised prevalence rate (ASPR) reflected the total number of existing RA cases per 100 000 population at a given point in time. The age-standardised mortality rate (ASMR) was the number of deaths attributed to RA per 100 000 population per year.

Age-period-cohort analysis

We implemented an APC model to evaluate the independent effect estimates of age, period and cohort on the incidence, prevalence and mortality of RA. The results of the APC model estimated the overall time trends, that is, the net drift (% per year), which was expressed as the overall annual percentage change in incidence, prevalence and mortality, combining the influences of both period and cohort effects.²³ Additionally, local drifts provided insights into age-specific trends, capturing the annual percentage change in RA rates within each age group.²³ The model also estimated longitudinal age-specific rates, which displayed the expected age-specific rates in a reference cohort, adjusted for period effects. Period rate ratios (RRs) compared the relative risks of RA across different time periods, adjusting for age and non-linear cohort effects, while cohort RRs assessed the relative risk of RA across different birth cohorts compared with a reference cohort, adjusting for age and non-linear period effects.²⁴

For this analysis, we used data from GBD 2021 including the number of cases for RA incidence, prevalence and mortality, along with population figures, across all age groups and time periods for the globe and each region. The data were organised into consecutive 5-year periods from 1992 to 2021, excluding data from 1990 to 1991 due to incomplete periods. The analysis included 18 age groups, spanning 5–9 years to 90–94 years, with six consecutive 5-year calendar periods from 1992 to 1996 (mid-year, 1994.5) to 2017–2021 (mid-year, 2019.5). Additionally, 23 overlapping 10-year birth cohorts, from 1897 to 1907 (mid-year, 1902) to 2007–2017 (mid-year, 2012), were included. The central calendar period (2002–2006) and the central birth cohort (1952–1962) were used as reference points for calculating period and cohort rate ratios. The APC model outputs were further stratified by sex, SDI region and WHO region. All analysis was conducted in R (V.4.4.1; R Foundation for Statistical Computing, Vienna, Austria) with statistical significance set at p<0.05. The APC Web Tool provided by the National Cancer Institute was used to estimate model parameters, and the Wald χ² test was applied to assess the significance of the parameters.¹⁸

Results

Temporal trends in incidence, prevalence and mortality of RA, 1990–2021

Globally, the ASIR of RA per 100 000 population increased from 10.4 (95% UI 9.3 to 11.6) in 1990 to 11.8 (95% UI 10.6 to 13.1) in 2021 and ASPR rose from 182.5 (95% UI 161.6 to 207.5) to 208.9 (95% UI 186.3 to 236.3) per 100 000 population. Different from ASIR and ASPR, ASMR per 100 000 population decreased from 0.61 (95% UI 0.54 to 0.68) to 0.45 (95% UI 0.37 to 0.52) during the period of 1990–2021. Furthermore, the APC model estimated a global net drift in incidence, prevalence and mortality rates of RA at 0.505% (95% CI 0.483% to 0.527%), 0.561% (95% CI 0.547% to 0.574%) and −1.729% (95% CI −1.877% to −1.582%), respectively (table 1). In both 1990 and 2021, ASIR, ASPR and ASMR for RA were all higher in females than in males. However, the net drift on incidence was 0.542% (95% CI 0.515% to 0.569%) for males and 0.491% (95% CI 0.465% to 0.518%) for females; for prevalence, the net drift was 0.610% (95% CI 0.593% to 0.628%) for males and 0.555% (95% CI 0.540% to 0.571%) for females; and for mortality, it was −1.237% (95% CI −1.487% to −0.986%) for males and −1.953% (95% CI −2.137% to −1.769%) for females (online supplemental table S3).

Table 1. Temporal trends of rheumatoid arthritis incidence, prevalence and mortality at global and regional level from 1990 to 2021.

Location	Incidence			Prevalence			Mortality
	ASR 1990	ASR 2021	Net drift(% per year)	ASR 1990	ASR 2021	Net drift(% per year)	ASR 1990	ASR 2021	Net drift(% per year)
Global	10.425 (9.321 to 11.644)	11.804 (10.641 to 13.124)	0.505 (0.483 to 0.527)	182.539 (161.586 to 207.478)	208.899 (186.338 to 236.334)	0.561 (0.547 to 0.574)	0.607 (0.538 to 0.678)	0.448 (0.373 to 0.519)	−1.729 (−1.877 to −1.582)
SDI region
High SDI	15.867 (14.305 to 17.638)	16.880 (15.439 to 18.472)	0.356 (0.320 to 0.392)	263.147 (238.330 to 294.738)	282.727 (256.437 to 313.876)	0.505 (0.486 to 0.525)	0.752 (0.694 to 0.782)	0.387 (0.339 to 0.415)	−2.570 (−3.035 to −2.103)
High-middle SDI	9.364 (8.294 to 10.545)	11.663 (10.426 to 12.969)	0.884 (0.845 to 0.922)	172.087 (150.287 to 197.203)	213.277 (188.454 to 242.992)	0.798 (0.775 to 0.822)	0.425 (0.381 to 0.472)	0.363 (0.304 to 0.415)	−1.704 (−2.026 to −1.382)
Middle SDI	9.722 (8.607 to 10.952)	11.896 (10.666 to 13.376)	0.756 (0.697 to 0.815)	171.708 (150.602 to 197.151)	209.613 (185.505 to 238.440)	0.758 (0.733 to 0.782)	0.638 (0.528 to 0.726)	0.498 (0.393 to 0.573)	−1.407 (−1.621 to −1.192)
Low-middle SDI	7.876 (6.968 to 8.929)	10.488 (9.357 to 11.781)	1.038 (0.978 to 1.099)	121.212 (105.913 to 140.226)	168.372 (149.417 to 192.373)	1.197 (1.182 to 1.213)	0.663 (0.469 to 0.956)	0.650 (0.482 to 0.928)	−0.691 (−1.075 to −0.305)
Low SDI	5.630 (4.990 to 6.393)	7.039 (6.273 to 7.924)	0.839 (0.755 to 0.923)	86.900 (75.818 to 101.156)	110.976 (98.213 to 127.440)	0.900 (0.871 to 0.929)	0.357 (0.231 to 0.668)	0.356 (0.243 to 0.610)	−0.642 (−1.486 to 0.209)
WHO region
African Region	5.069 (4.509 to 5.672)	5.319 (4.743 to 5.961)	0.282 (0.184 to 0.380)	90.199 (78.929 to 103.515)	95.817 (83.675 to 110.248)	0.244 (0.214 to 0.274)	0.159 (0.111 to 0.312)	0.110 (0.089 to 0.211)	−3.109 (−3.787 to −2.425)
Eastern Mediterranean Region	5.125 (4.463 to 5.857)	5.977 (5.244 to 6.811)	0.540 (0.426 to 0.654)	99.758 (85.657 to 117.024)	118.385 (101.116 to 138.679)	0.749 (0.728 to 0.770)	0.399 (0.254 to 0.592)	0.358 (0.245 to 0.481)	−0.175 (−0.859 to 0.514)
European Region	11.456 (10.329 to 12.620)	13.566 (12.327 to 14.892)	0.618 (0.587 to 0.649)	205.485 (183.475 to 231.687)	242.586 (217.275 to 272.434)	0.642 (0.612 to 0.671)	0.600 (0.555 to 0.627)	0.342 (0.305 to 0.364)	−2.078 (−2.446 to −1.709)
Region of the Americas	13.655 (12.409 to 14.971)	16.459 (14.994 to 17.956)	0.679 (0.648 to 0.710)	236.716 (216.359 to 261.598)	290.574 (265.028 to 320.428)	0.828 (0.814 to 0.842)	0.617 (0.575 to 0.641)	0.477 (0.431 to 0.520)	−0.600 (−0.911 to −0.288)
South-East Asia Region	8.225 (7.220 to 9.389)	11.597 (10.271 to 13.173)	1.224 (1.147 to 1.302)	119.062 (103.690 to 138.808)	174.183 (153.055 to 200.332)	1.359 (1.335 to 1.384)	0.689 (0.475 to 1.003)	0.615 (0.458 to 0.901)	−1.326 (−1.784 to −0.865)
Western Pacific Region	12.141 (10.663 to 13.754)	12.994 (11.529 to 14.726)	0.356 (0.304 to 0.409)	213.483 (185.938 to 245.458)	231.094 (202.643 to 265.690)	0.441 (0.424 to 0.458)	0.712 (0.600 to 0.817)	0.491 (0.371 to 0.584)	−1.888 (−2.129 to −1.646)

Note: Parentheses for ASRs denote 95% UIs and parentheses for net drifts denote 95% CIs

ASR, age standardised rate; SDI, sociodemographic index; UI, uncertainty interval.;

In 2021, both ASIR and ASPR were highest in high SDI region, indicating 16.9 (95% UI 15.4 to 18.5) and 282.7 (95% UI 256.4 to 313.9) per 100 000 population, respectively; while ASMR was highest in low-middle SDI region (0.65 per 100 000 population; 95% UI 0.48 to 0.93). In all SDI regions, there was an increasing risk of RA incidence and prevalence from 1990 to 2021. Low-middle SDI region showed the most significant changes in incidence (net drift 1.038%; 95% CI 0.978 to 1.099) and prevalence (net drift 1.197%; 95% CI 1.182 to 1.213); while high SDI region reported the least in incidence (net drift 0.356%; 95% CI 0.320 to 0.392) and prevalence (net drift 0.505%; 95% CI 0.486 to 0.525). However, all SDI regions, except for low SDI region, saw a decreasing trend of risk on RA mortality, with net drifts ranging from −2.570% (95% CI −3.035 to −2.103) in high SDI region to −0.691% (95% CI −1.075 to −0.305) in low-middle SDI region (table 1; online supplemental tables S7–S9). Among WHO regions, the Region of the Americas reported the highest ASIR and ASPR in 2021, while the South-East Asia Region had the highest ASMR. From 1990 to 2021, all six regions faced unfavourable trends in RA incidence and prevalence. Most notably, the South-East Asia Region showed significant increases in incidence (net drift 1.224%; 95% CI 1.147 to 1.302) and in prevalence (net drift 1.359%; 95% CI 1.335 to 1.384). The African Region, European Region, Region of the Americas, South-East Asia Region and Western Pacific Region all saw reductions in RA-related mortality during this period (table 1; online supplemental tables S19–S21).

At national level, figure 1 visually presents the ASIR, ASPR and ASMR across 204 countries/territories in 2021, and online supplemental table S2 exhibited the concrete values of ASIR, ASPR and ASMR for these 204 countries and territories both in 1990 and 2021.

Figure 1. The global burden of rheumatoid arthritis across 204 countries and territories in 2021. ASIR, age-standardised incidence rate; ASMR, age-standardised mortality rate; ASPR, age-standardised prevalence rate.

Temporal trends in incidence, prevalence and mortality of RA across different age groups

Globally, all age groups exhibited increasing trends of RA incidence and prevalence, with local drifts all positive. The increasing trends in incidence and prevalence of RA increased with age in the 45–49 years to 65–69 years group; and those aged 65–69 years (local drift 0.881%; 95% CI 0.841 to 0.920) and 90–94 years (local drift 0.713%; 95% CI 0.583 to 0.844) showed the most increases for incidence and prevalence of RA separately. By contrast, all age groups from 5–9 to 85–89 years had decreasing trends in RA mortality over the same period, with local drifts ranging from −3.798% (95% CI −5.393% to −2.175%) in 5–9 years to −0.249% (95% CI −0.380% to −0.117%) in 85–89 years. Notably, the downward trends of RA mortality became less obvious as age grew (online supplemental table S3). Figure 2 visually represents temporal trends in incidence, prevalence and mortality of RA across different age groups globally.

Figure 2. Local drifts of incidence, prevalence and mortality for rheumatoid arthritis by sex from 1992 to 2021 in those over 5 years of age.

Among five SDI regions, temporal trends in different age groups for incidence and prevalence of RA were similar to global trends (online supplemental figure S1). However, in low-middle SDI region, the decreasing trends of mortality only appeared in age groups of 45–49 to 70–74 years; and in low SDI region, only in 50–54 to 60–64 years. In WHO regions, almost all age groups in the European Region, Region of the Americas and South-East Asia Region experienced upward trends in RA incidence. Almost all age groups in six WHO regions experienced higher prevalence risk and lower mortality risk for RA, except for several age groups in certain regions (online supplemental tables S7–S9, S19–S21 and figure S2).

Age, period and cohort effects on incidence, prevalence and mortality of RA

Figure 3 shows the age, period and cohort effects on RA incidence, prevalence and mortality globally. In the total population, incidence risk increased among those aged 5–69 years but decreased in those aged 70–94 years, highest in 65–69 years (34.532; 95% CI 34.250 to 34.816). The prevalence risk had a similar trend, growing fleetly since the age of 20 years and reaching the highest in those aged 80–84 years (904.752; 95% CI 898.966 to 910.576). However, the mortality risk of RA remained rising across the age range of 5–94 years, from 0.012 (95% CI 0.009 to 0.014) to 8.763 (95% CI 8.420 to 9.121). Notably, females experienced higher risk of RA incidence, prevalence and mortality across all age groups and reached the age groups with the highest incidence and prevalence risk earlier (online supplemental table S4). Similar patterns could be found in five SDI quintiles (online supplemental tables S10–S12 and figure S3). Also, age effects on RA incidence revealed a similar pattern in all WHO regions except for the Eastern Mediterranean Region (online supplemental table S22). The risk of RA prevalence and mortality were parallel with global trends (online supplemental tables S23–S24 and figure S6).

Figure 3. Age, period and cohort effects on incidence, prevalence and mortality rates of rheumatoid arthritis by sex from 1992 to 2021 in those over 5 years of age.

Period effects indicated ascending trends of RA risk on incidence and prevalence, with a descending risk of mortality. It was noteworthy that the incidence and prevalence risk of RA increased significantly in males, but the decreasing trend of mortality in males was not so pronounced in the most recent two periods (online supplemental table S5). By SDI quintiles, the incidence and prevalence of RA showed the same trends except for high SDI region (online supplemental tables S13 and S14). The period effect on RA mortality varied in five SDI regions (online supplemental table S15 and figure S4). Meanwhile, period effects showed an increasing risk of RA incidence and prevalence while a declining risk on mortality across all six WHO regions (online supplemental tables S25–S27 and figure S7).

In the 23 consecutive 10-year birth cohorts from 1897–1907 to 2007–2017, the trend of cohort risk for RA incidence and prevalence showed a generally upward trajectory, while the trend of cohort risk for RA mortality revealed a fluctuating state, decreasing since the cohort of 1912–1922 (online supplemental table S6). Generally, there were similar patterns in cohort effects across different SDI regions, with some fluctuations in several cohorts (online supplemental tables S16–S18 and figure S5). In WHO regions, cohort effects on RA incidence and prevalence also showed a clear upward trend collectively, while in the African Region, the risk remained relatively stable for cohorts born after 1952 (online supplemental tables S28 and S29). The risk of RA mortality declined progressively in younger birth cohorts, specifically noticeable in the African Region (online supplemental table S30 and figure S8).

Discussion

Our study provides the most comprehensive analysis to date of global, regional and national trends in the epidemiology of RA from 1990 to 2021 incorporating an APC framework. Globally, the ASIR and ASPR of RA have increased, while the ASMR has decreased in the past three decades. These trends varied significantly across sexes, regions and SDI levels. APC analyses further identified distinct patterns in age, period and cohort effects, shedding light on the drivers of these temporal trends. Notably, the burden of RA remains disproportionately higher in females, and regions with higher SDI levels reported the highest incidence and prevalence, whereas the highest mortality was observed in low-middle SDI regions.

From 1990 to 2021, the global incidence and prevalence of RA exhibited modest increases, as evidenced by positive net drift values. These rising trends can largely be attributed to global population ageing, improved diagnostic capabilities and heightened awareness of RA over time.^{25 26} Advances in healthcare infrastructure, particularly in high-SDI regions, have enabled earlier detection and better disease surveillance. Moreover, changes in environmental or lifestyle factors may have contributed to the observed increases. For example, studies have linked higher meat consumption and inadequate sleep to increased RA risk, likely through inflammatory pathways.^{27 28} Smoking, for instance, is a well-established risk factor for seropositive RA,^{29 30} while obesity promotes systemic inflammation, exacerbating disease onset and progression.³¹ These factors, coupled with environmental changes resulting from urbanisation, have amplified the global burden of RA.³² In contrast, RA-related mortality demonstrated a consistent global decline over the same period. This favourable trend aligns with significant advancements in RA management, particularly the widespread adoption of biological and targeted-synthetic DMARDs. These therapies have revolutionised the care of RA, reducing disease activity, complications and mortality.⁴ Improvements in healthcare access, adherence to treatment protocols and reductions in treatment costs, particularly in high-income regions, have further contributed to declining mortality rates.^{33 34}

Significant socioeconomic and geographical disparities in RA epidemiology were observed across SDI quintiles and WHO regions. High SDI regions consistently reported the highest ASIR and ASPR, likely due to better diagnostic infrastructure, greater disease awareness and higher healthcare utilisation.³⁵ However, low-middle SDI regions showed the highest net drifts in incidence and prevalence, possibly driven by rapid urbanisation, lifestyle changes and inadequate preventive strategies.^{36 37} Conversely, the highest ASMR was observed in low-middle SDI regions, and the decline in RA-related mortality has been less pronounced in low SDI regions. This lack of progress points to persistent challenges in resource-limited settings, where delayed diagnoses, inadequate access to DMARDs and insufficient management of disease complications contribute to increased mortality.³⁸ These disparities emphasise the urgent need for global efforts to reduce healthcare inequities, particularly in low-resource settings, to ensure more equitable outcomes for patients with RA worldwide.³⁹

Among WHO regions, the Region of the Americas reported the highest ASIR and ASPR in both 1990 and 2021. This may be partially explained by genetic predispositions, such as the presence of HLA-DRB1 gene variants associated with RA susceptibility in populations with Indigenous ancestry, as well as environmental and lifestyle risk factors prevalent in the region, including those in Latin America.⁴⁰ In contrast, the South-East Asia and Western Pacific regions reported the highest ASMRs, though mortality in the Western Pacific region declined significantly over time due to advancements in healthcare access and treatment strategies.^{9 41} In contrast, South-East Asia continues to face persistent challenges, including economic constraints, limited healthcare resources and inadequate public health support.⁴² These findings underscore the need for tailored interventions that address the unique challenges faced by low- and middle-SDI regions and resource-constrained WHO regions.

Consistent with previous studies,^{10 12} our findings demonstrated that the global burden of RA was larger among females than males. This disparity is likely driven by hormonal and immunological differences, as oestrogens enhance humoral immune responses while androgens suppress both humoral and cellular immunity.⁴³ Females also tend to develop RA earlier than males, as evidenced by the earlier peak in incidence and prevalence. Interestingly, while RA incidence and prevalence have increased more rapidly in males than females, the decline in RA-related mortality was more pronounced in females. This may reflect sex-specific differences in disease progression and treatment responses, as biological therapies may be more effective in females due to sex-specific differences in pharmacokinetics and pharmacodynamics.^{44 45} These findings emphasise importance of sex-specific approaches to RA management, as well as the need for further research to better understand the mechanisms underlying these disparities.

The age effects showed that RA incidence and prevalence peaked at 69 years in females and 79 years in males, consistent with the established natural history of RA as a chronic autoimmune disease that develops over time. These findings highlight the disproportionate burden of RA among older adults, particularly females, and underscore the importance of targeted screening and management strategies for this demographic. Period effects demonstrated a consistent rise in RA incidence and prevalence over time, likely driven by heightened awareness and improved diagnostic precision. Additionally, changes in environment, such as air pollution, and lifestyle factors, including increased sodium intake, are notable contributors to these trends.3146,49 The cohort effects revealed that individuals born in more recent decades faced a higher risk of RA incidence and prevalence, likely due to greater exposure to modifiable risk factors such as obesity, physical inactivity and dietary changes. However, these cohorts also experienced lower mortality risks, reflecting the benefits of improved healthcare access and advancements in RA management.^{25 34} These findings underscore the importance of public health initiatives aimed at promoting healthy lifestyles, reducing exposure to modifiable risk factors and ensuring equitable access to effective treatments.

This study underscores the urgent need for targeted interventions to address the growing burden of RA, particularly in resource-limited settings. Strengthening healthcare systems by expanding access to early diagnosis, advanced treatments and long-term disease management is critical to improving outcomes and reducing RA-related mortality. Public health initiatives should focus on reducing exposure to modifiable risk factors, such as smoking, obesity and physical inactivity, through education and lifestyle interventions. Additionally, high-risk populations, including older adults, females and individuals with a family history of RA, should be specifically targeted for early screening and proactive management to ensure timely diagnosis and effective treatment. Collaborative efforts between governments, international organisations and non-governmental organisations are essential to address healthcare inequities and ensure equitable access to effective treatments. Continued investment in research is also necessary to better understand the drivers of RA, evaluate the long-term effects of advanced therapies and guide the development of cost-effective prevention and treatment strategies.

This study has several notable strengths. To our knowledge, it is the most up-to-date and comprehensive investigation of RA epidemiology over the past three decades, providing valuable insights into global, regional and national patterns. The use of the GBD 2021 database, which is the most comprehensive global health dataset available, ensures the reliability and comparability of our findings. Additionally, incorporating an APC framework allowed for a nuanced understanding of how age, period and cohort effects have shaped RA trends, offering actionable insights for healthcare policy and planning. However, our study is not without limitations. First, the absence of subtype-specific data in GBD 2021 prevented us from analysing differences between seropositive and seronegative RA, which have distinct epidemiological features.⁵⁰ The attribution of mortality directly to RA is complicated by the presence of comorbidities, potentially introducing bias into mortality estimates. Furthermore, the quality and availability of data sources vary across regions, necessitating reliance on modelled estimates, particularly in low- and middle-income countries, which may limit the precision of non-fatal estimates.¹⁹ Lastly, while APC analysis provides valuable insights, it remains an ecological approach and cannot establish causality. Future research addressing these limitations, including subtype-specific and regionally representative analyses, is warranted to better understand the factors driving RA trends.

Conclusion

In conclusion, this study highlights significant global, regional and national trends in RA epidemiology from 1990 to 2021. While incidence and prevalence have risen, mortality has declined, reflecting advancements in treatment and healthcare delivery. However, substantial disparities persist across sexes, regions and SDI levels, emphasising the need for targeted public health interventions and equitable healthcare access. Strengthening healthcare infrastructure, addressing modifiable risk factors and investing in research and policy reforms will be essential to mitigating the global burden of RA in the coming decades.

Data availability statement

Data are available in a public, open access repository.