Rheumatoid arthritis (RA) is a systemic autoimmune disease characterized primarily by chronic, symmetrical synovitis of multiple joints, affecting approximately 17.6 million people worldwide (1). Its pathogenesis is complex, involving interactions between genetic susceptibility and environmental factors (2). RA not only causes joint pain, swelling, deformity, and significantly reduced quality of life but also affects multiple systems such as the cardiovascular and respiratory systems, increasing disability rates, cardiovascular event incidence, and all-cause mortality. This imposes a substantial burden on affected families and healthcare systems (3). Despite considerable advances in treatment, the etiology of RA remains incompletely understood, making the identification of modifiable risk factors crucial for primary prevention.

Among various environmental factors, obesity—a global public health issue—has garnered increasing attention due to its association with chronic inflammation and immune dysregulation (4). Obesity is now recognized not merely as excess fat storage but as a chronic low-grade inflammatory state driven by the secretion of various pro-inflammatory cytokines from adipose tissue (5). This inflammatory milieu may disrupt immune tolerance and promote autoimmune responses, providing a biological plausibility for the link between obesity and RA (6). Body mass index (BMI), a simple and commonly used indicator of weight relative to height, is widely employed to define overweight and obesity.

Numerous epidemiological studies have investigated the relationship between BMI and RA risk, yet their findings have been inconsistent. Although previous systematic reviews have synthesized evidence in this field (7–10), significant gaps remain. First, the most recent comprehensive meta-analyses included literature up to 2019, leaving nearly a decade of subsequent research un-synthesized (7–10). The inclusion of newer, larger cohort studies is essential to confirm and refine previous estimates. Second, prior analyses have predominantly focused on overweight, obesity, or continuous BMI, with the association between underweight and RA risk remaining poorly characterized and underpowered in earlier work. Third, a detailed evaluation of how study quality and the degree of confounding adjustment influence these associations is needed. Therefore, this study aims to perform an updated systematic review and meta-analysis of existing observational studies to quantitatively synthesize evidence regarding the association between different BMI categories and the risk of developing RA.

This systematic review and meta-analysis, incorporating 20 observational studies (8 cohort and 12 case–control) published up to 2025 with a total sample size of 568,889, extends and refines the existing body of evidence on BMI and RA risk (7–10). Our findings confirm and strengthen the positive association between higher BMI (overweight and obesity) and increased RA risk, previously reported (7–9), by demonstrating its robustness in the most contemporary dataset and across high-quality studies. Notably, our analysis provides novel insights into the underweight category, an area less explored in prior syntheses. While Feng et al. (8, 9) and Ohno et al. (10) primarily reported on overweight/obesity or continuous BMI, we found that underweight was not significantly associated with RA risk overall, but showed a trend towards a protective effect in methodologically rigorous subgroups, a finding that warrants future investigation. Furthermore, our stratified analyses by study quality (NOS score) and adjustment level offer a more nuanced understanding of how methodological heterogeneity may influence the observed associations, particularly for underweight.

This study confirms the promotive effect of overweight and obesity on RA onset, aligning with the biological mechanism of “obesity as a chronic low-grade inflammatory state” (4, 5). Adipose tissue, particularly visceral fat, secretes substantial pro-inflammatory cytokines and adipokines. These mediators can break immune tolerance, activate autoreactive T and B cells, and directly promote synovial hyperplasia and inflammatory infiltration—key processes in RA pathogenesis (41). The higher pooled OR for obesity compared to overweight, while not establishing a formal dose–response relationship, suggests a potential graded association between increasing BMI and RA risk. This trend is consistent with previous studies reporting that each 5 kg/m² increase in BMI is associated with an approximately 3–13% increase in RA risk (7, 8), further supporting the hypothesis that obesity-related inflammatory burden accumulates with rising BMI.

Our analysis provides a novel contribution by specifically evaluating the underweight category, which has been less clearly addressed in earlier meta-analyses (7–10). Overall, underweight was not significantly associated with RA risk. However, the significantly reduced risk observed in subgroups of studies with highly adjusted effect estimates and higher methodological quality is intriguing. This pattern may be explained by two factors. First, from a biological perspective, individuals with low body weight have reduced adipose tissue, potentially leading to lower pro-inflammatory cytokine secretion and a decreased probability of initiating autoimmune responses. Second, the potential for reverse causation bias must be considered. Undiagnosed early RA symptoms could lead to reduced physical activity and appetite loss, resulting in weight loss before formal diagnosis. Although our analysis used baseline BMI data, studies with shorter follow-up periods might still be susceptible to this bias. Furthermore, this nuanced finding highlights the importance of study quality in assessing this particular association and underscores why a dedicated analysis was needed beyond the scope of prior reviews. Finally, the finding regarding underweight is based on only 9 studies, and the limited sample size necessitates caution in interpretation. This caution is especially warranted for sex-specific interpretations. As noted in the results, the subgroup estimate for underweight in males was derived from a single study, providing insufficient evidence to characterize this association or to compare it reliably with the estimate for females. Therefore, while the overall null association for underweight appears robust, definitive conclusions regarding potential sex differences within the underweight category cannot be drawn from the current available evidence and require future studies with adequate sample sizes in both sexes.

This systematic review and meta-analysis, incorporating the most up-to-date evidence, strengthens the conclusion that overweight and obesity are modifiable risk factors for RA, particularly among females. The heightened risk observed in females may be explained by several mechanisms. During reproductive years, higher estrogen levels, while generally anti-inflammatory, can be paradoxically pro-inflammatory in the context of obesity. Adipose tissue secretes aromatase, which converts androgens to estrogen, leading to abnormally high levels that can activate the NF-κB pathway and exacerbate synovial inflammation (42). Conversely, in postmenopausal women, the sharp decline in endogenous estrogen is accompanied by a compensatory increase in pro-inflammatory cytokine secretion from adipose tissue, further amplifying the obesity-associated risk (43). In men, testosterone may confer a protective effect by activating the androgen receptor to suppress production of pro-inflammatory molecules like IL-6 and leptin while promoting anti-inflammatory factors such as IL-10 (44). This may partially counteract the inflammatory burden of obesity, potentially explaining the attenuated association between BMI and RA risk observed in the male subgroup. The regional differences, primarily the non-significant association found in the two studies conducted in China, should be interpreted with caution due to limited data from non-European/American populations and require further validation. Finally, the influence of adjustment levels on the observed associations suggests that potential interactions between BMI and confounding factors, such as smoking and alcohol consumption, warrant further investigation (45, 46).

This study has several limitations. First, as it incorporates both case–control and cohort studies, the findings are susceptible to recall and selection biases inherent in these designs. Second, the observational nature cannot eliminate residual confounding or definitively establish causality between obesity and RA. Third, BMI reflects only weight relative to height and cannot differentiate fat distribution from muscle mass, whereas abdominal obesity may have stronger pro-inflammatory effects (47). Fourth, with only two studies from Asia, it is difficult to accurately represent the true association within Asian populations. Fifth, although we employed a systematic strategy to harmonize diverse BMI categorizations across studies to the WHO standard, some degree of misclassification is inevitable. Studies using regional cut-offs or non-standard categories introduce heterogeneity in the exposure definition. Finally, as the analysis is based on published studies, the inability to access unpublished data may introduce publication bias.

This study confirms that overweight and obesity are significantly associated with an increased risk of RA, with a notable sex-specific disparity: the risk elevation was more pronounced in females, while no significant association was observed in males. Underweight showed no significant overall association with RA risk. These findings provide robust evidence for the primary prevention of RA and underscore the importance of developing sex-specific weight management strategies, laying a foundation for future research into targeted preventive interventions.