Rheumatoid arthritis: The continuum of disease and strategies for prediction, early intervention and prevention

Abstract

Rheumatoid arthritis (RA) is known to include a ‘pre-RA’ stage that can be defined as the presence of familial or genetic risk factors, biomarker abnormalities (e.g. autoantibodies to citrullinated peptide antigens [ACPA]), symptoms and even abnormal imaging findings prior to the development of the onset of ‘clinical RA’ with inflammatory arthritis that is apparent on physical examination. Indeed, there are multiple completed or ongoing retrospective case-control as well as prospective observational studies to identify the key biologic drivers of disease. Furthermore, building on the predictive ability of combinations of biomarkers, symptoms and imaging for future RA, there are multiple clinical trials completed, underway or in development to identify approaches that may prevent, delay or ameliorate future clinical RA in at-risk individuals. Importantly, however, while an effective preventive intervention has not yet been identified, at-risk individuals are being increasingly identified in clinical care; and this presents a challenge of how to manage these individuals in clinical practice. This review will discuss the current understanding of the biology and natural history of RA development, nomenclature and current models for prediction of future RA, as well as evaluate the current and ongoing clinical prevention trials with the overall goal to provide insights into the challenges and opportunities in the field of RA prevention. Moreover, this review will provide up-to-date options for clinical management of individuals at-risk for RA.

Rheumatoid arthritis: overview and current management

Rheumatoid arthritis (RA) is typically diagnosed in clinical care based on the presence of symptoms and signs of active joint inflammation (i.e., a swollen joint on physical examination consistent with synovitis), as well as biomarkers such as autoantibodies and imaging findings that can demonstrate joint inflammation and/or damage. This diagnosis can be termed ‘Clinical RA’; in addition, there are established classification criteria for RA that include the 1987 American College of Rheumatology (ACR) criteria¹, and the 2010 ACR/European Alliance of Associations for Rheumatology (EULAR) criteria². There are also two ‘subclinical’ categories of clinical RA termed “seropositive” or “seronegative” defined as the presence or absence of serum elevations of autoantibodies that currently include rheumatoid factor (RF) and/or antibodies to citrullinated protein antigens (ACPA). After a diagnosis is made, treatment is typically initiated with disease-modifying anti-rheumatic drug (DMARD) therapies that have been established as effective in treating the primary disease manifestation of inflammatory arthritis (IA) in controlled clinical trials³. For the majority of individuals who are diagnosed with clinical RA, DMARD therapy results in improved well-being and function and reduced joint damage, with a subset of individuals reaching disease remission– and an even smaller subset reaching DMARD-free remission^4–6. However, for the majority of individuals who develop clinical RA, it is a disease that will require lifelong therapy, with ongoing adverse effects on their well-being (and finances), and sustained remission is infrequent (<50% of patients in some studies^6–10). As such, ultimately prevention of RA may result in substantially improved personal and public health impact.

RA begins prior to the appearance of clinically-apparent inflammatory arthritis

There is established and growing evidence that RA develops in a series of stages as outlined in Figure 1. In general, it appears that the natural history of RA begins prior to clinical RA during a period that can be termed ‘pre-RA’ where genetic and environmental factors interact to drive early breaks in immune tolerance that to date have best been identified through blood elevations of autoantibodies, including RF and ACPA^11–16, although multiple other autoantibodies including anti-CarP¹⁷, anti-malondialdehyde-acetaldehyde (anti-MAA) antibodies¹⁸ and anti-peptidyl arginine deiminase (anti-PAD) antibodies¹⁹ have been identified. These antibodies are elevated on-average ~3 to 5 years prior to a diagnosis of clinical RA (although in some cases 10 years). ^11,12,20. Furthermore, array-based technologies have also identified that ACPA undergo epitope spreading in pre-RA²⁸. ACPA have been shown to undergo variable domain (Fab) N-linked glycosylation²⁹ and glycosylations changes in ACPA have been shown to independently associate with the development of future IA in ACPA positive individuals³². Studies have also demonstrated that autoantibody elevations can be present in conjunction with altered T cell subsets²¹, B cell and interferon signatures^22,23, and a general increase in inflammation (e.g. elevated C-reactive protein, calprotectin²⁴ and cytokines and chemokines^13,25,26).

Figure 1. Model of seropositive rheumatoid arthritis development.

In this model, genetic and environmental risk factors combine to trigger initial autoimmunity, potentially initially at mucosal sites; this may progress to systemic autoimmunity that can be detectable through blood testing for autoantibodies (e.g. ACPA, RF). This may initially be in absence of articular inflammation and be considered a ‘non-articular’ stage. The process may progress to a stage of ‘subclinical’ arthritis that can be defined as articular imaging abnormalities (and may involve tenosynovium prior to diarthrodial joint involvement) in absence of physical examination findings of IA; this may progress to undifferentiated arthritis that can be seen on examination and ultimately rheumatoid arthritis that is classifiable by established criteria. Notably, a state of detectable autoimmunity in absence of clinically-apparent IA may be termed ‘pre-RA’, and a state of of clinically-apparent IA may be termed ‘clinical RA’, even if not meeting established criteria. While not well studied, there may be an asymptomatic period of early autoimmunity that is followed by development of joint symptoms (i.e. arthralgia) that can be characterized as ‘clinically suspect arthralgia;.; there may also be non-articular symptoms such as fatigue throughout the development of RA. Importantly, not all individuals may progress through all stages and indeed some individuals may stop at one stage, or regress to a prior stage; in addition, the time course of progress may differ between individuals but some studies suggest that on-average systemic autoimmunity is detectable on-average ~3–5 years prior to the diagnosis of clinical RA. Abbreviations: RA=rheumatoid arthritis, ACPA=antibodies to citrullinated protein antigens; RF=rheumatoid factor; IA=inflammatory arthritis.

Along with the expansion in autoimmunity and inflammation, symptoms may evolve, including joint pain, stiffness and swelling. There is growing evidence that while IA may not be identifiable on physical examination, features of synovitis or even joint damage may be seen on imaging and termed ‘subclinical IA’; furthermore, tenosynovitis may be one of the first manifestations of RA musculoskeletal related tissue injury^27–30. Disease can then progress to overt clinically-apparent IA that is clinically diagnosed as RA and may further meet established classification criteria^1,2. Notably, the features and nomenclature to define and label each of these stages of RA development are somewhat variable (Figure 1 and Table 1), and are best understood in seropositive RA. Notably, when referring to individuals who may be in a pre-RA state, a EULAR task force determined that the term ‘pre-RA’ only be applied retrospectively once it was known that an individual had developed clinical RA³¹; as such, the term ‘at-risk individual’ is often used to describe individuals who exhibit some risk factors for future RA but whose future development of clinical RA is unknown.

Table 1.

Terminology used to describe the natural history of rheumatoid arthritis

Terms	Definitions
Pre-RA, Pre-Clinical RA	The stage prior to the development of clinical RA. May be defined in several ways including presence of autoantibodies, abnormal symptoms and imaging findings.
At-Risk	Any state by which an individual has a higher-than-average risk of developing RA but does not have current inflammatory arthritis (i.e. being a First-Degree Relative of an RA patient, presence of autoantibodies). Notably, a EULAR task force that included individuals with clinical RA as well as individuals who are at-risk for future RA determined that the term pre-RA should only be used retrospectively once it was known that an individual developed clinical RA (Gerlag et al 201431). As such, the term ‘at-risk individual’ is often used to describe individuals who exhibit some risk factors for future RA but whose future status of clinical RA is unknown.
First-Degree Relative (FDR)	An individual who is related, by the first-degree (parent, sibling, offspring) to an individual with a specific disease. An FDR shares both genetic and often environmental risk factors for developing the disease of interest.
Clinically Suspect Arthralgia (CSA)	A combination of signs and symptoms that is suggestive of the presence and/or risk for inflammatory arthritis. Includes but not limited to morning stiffness and pain in the small joints. The pain may or may not be elicited on physical examination.
Clinically apparent inflammatory arthritis	A clinical finding of joint synovitis, in the absence of trauma, that is suggestive of the earliest stage of RA. A joint with synovitis is typically tender on range of motion/palpation, swollen with a palpable effusion and warmer than non-inflamed joints.
Palindromic rheumatism (PR)	Recurrent episodes of inflammatory arthritis, typically involving the small joints of the hands and feet, that resolves spontaneously with limited or no symptoms between events. A subset of patients with PR will develop classifiable RA.
Subclinical inflammatory arthritis	The presence of inflammation in a joint detected by imaging, typically by MRI or US that is suggestive of inflammatory arthritis and a physical examination of the involved joint that is not suggestive of synovitis (i.e. there is an absence of joint swelling).
Undifferentiated arthritis	Inflammatory arthritis that does not fulfill established classification criteria for RA or any other disease.
Clinical RA	Clinically-apparent inflammatory arthritis (with or without autoantibodies) that includes individuals who have classifiable RA, and those who do not. For example, a patient with low positive ACPA/RF and 3 swollen small joints may not have classifiable RA but still be diagnosed and treated as RA.
Classifiable RA	A patient who meets established classification criteria for RA, either 1987 or 2010 criteria (Arnett et al 19881; Aletaha et al 20102)
Seropositive RA	Classifiable RA with inflammatory arthritis and the presence of serum elevations RA-associated antibodies such as RF or ACPA.
Seronegative RA	Classifiable RA with inflammatory arthritis without the presence of RA- associated antibodies such as RF or ACPA.
Early RA	A patient who meets classification criteria for RA for a short duration, typically less than 1 year although definitions vary in the literature.

Abbreviations: RA=rheumatoid arthritis; ACPA=antibodies to citrullinated protein antigens; RF=rheumatoid factor; ACR=American College of Rheumatology; EULAR=European Alliance of Associations for Rheumatology; ESR=erythrocyte sedimentation rate; CRP=C-reactive protein; MRI=magnetic resonance imaging; US=ultrasound; CSA=clinically suspect arthralgia; FDR=first degree relative; PR=palindromic rheumatism

There are multiple genetic and environmental factors that have been associated with an increased risk for clinical RA ³², although most studies that have associated genetic and environmental risk factors with RA have been performed in a case-control fashion in individuals with established clinical RA. As such, these factors may not be truly associated with the initial generation of autoimmunity that may have occurred years prior to clinical RA. However, there are a growing number of prospective studies in at-risk populations that have identified factors that are associated with increased risk for future clinical RA (Table 2). For genetic factors, the strongest identified is a set of alleles coding certain sequences in the Human Leukocyte Antigen (HLA) region that in aggregate are called the ‘shared epitope’ (SE)³³; while findings are mixed, several studies demonstrate that SE is associated with ACPA+ as well as a transition to future clinical RA in ACPA+ individuals^34–36. Furthermore, in a nested case-control study within the Nurses’ Health Study, a score derived from 22 genetic variants predicted RA risk, suggesting genes beyond the SE are related to progression to clinical RA³⁷. For environmental factors, smoking is the most reproducibly associated with a higher future risk of RA^38,39. In addition, a number of studies have identified obesity as a risk factor for clinical RA^40,41, and that an anti-inflammatory diet and higher levels of exercise may be protective⁴². Moderate alcohol intake has also consistently been identified as a protective against future RA development, although the mechanisms behind this are not yet clear^42–45. Multiple metabolic/immunologic processes have been purported to play a role in RA development including fatty acid pathways that may be influenced by dietary intake⁴⁶. As RA is more prevalent in women, reproductive and menopausal factors have been extensively studied. Post-menopause, postpartum period and use of anti-estrogen agents have been consistently associated with RA onset^47,48; however, while one study in an at-risk population has demonstrated an association between use of oral contraceptives and decreased risk for RF positivity⁴⁹, the role of hormonal factors, including hormonal therapy, in the pre-RA period remains controversial.

Table 2.

Risk and protective factors for future RA development evaluated in prospective studies or cross-sectional studies of at-risk individuals/Pre-RA

Genetic and familial risk factors  • Shared epitope associated with higher risk for transition to RA in ACPA positive individuals at baseline (Rakieh et al 201534)  • Genetic risk score using 22 variants associated with increased risk of RA (Karlson et al 201037)  • First-degree relative status increased risk of progression to articular RA in arthralgia cohort (van de Stadt et al 201345)Certain populations have high-risk for RA, including populations indigenous to the Americas who have ~5–7 fold increased risk for RA compared to nonindigenous populations (Ferucci et al 2005113)

Sex-related factors  • Female sex given women have 2–3 fold higher risk for RA compared to men (Deane et al 201732)  • Longer duration of breast-feeding and higher parity are protective (Ren et al 2017114)  • Oral contraceptive use associated with decreased autoantibody positivity in individuals at-risk for RA (first-degree relatives) (Bhatia et al 200749)

Environmental (reviewed in Deane et al 201732 and Zaccardelli et al 201945)  Increased risk for RA   • Cigarette smoking, especially long duration and high intensity smoking   • Obesity   • Inflammatory diet  Protective against RA   • Moderate alcohol consumption   • High fatty fish intake and intake of omega-3 fatty acids

Other Mucosal and/or microbiome influences  • Lung disease (airways, parenchymal) (Demoruelle et al 2012115)  • Periodontal inflammation (Mankia et al 2019116)  • Multiple organisms including viruses and bacteria have been associated within increased risk for RA  • Stress and/or stress responses, and potentially mental health (complex and unclear if stress/mental health could drive autoimmunity, or be influenced by autoimmunity - or both)(Polinksi et al 2020 and Marrie et al 2019117,118)

Abbreviations: Pre-RA=pre-rheumatoid arthritis; RA=rheumatoid arthritis; ACPA=antibodies to citrullinated protein/peptide antigens

Importantly, it is hypothesized that the early generation of autoimmunity occurs outside of the joints during a period that can be termed a ‘non-articular’ stage of RA development (Figure 1). ^50,51. Further supporting a ‘non-articular’ stage of pre-RA, studies that have shown an absence of synovitis on synovial biopsy in individuals who have circulating ACPA or RF^52,53. As to where this autoimmunity is generated, there is some data supporting that it may be at mucosal sites including the periodontium, lung and gut ⁵⁴. In particular, there are several studies that have suggested that lung disease, including airways or interstitial lung disease may precede the onset of articular RA^55–57 and that RA-related autoantibodies may be generated in the lung in some individuals prior to clinical RA onset⁵⁸. In addition, there is growing evidence that certain microbial factors that are present at mucosal surfaces may be associated with RA-related autoimmunity^59–62.

Overall, an emerging concept is that the evolution of RA is characterized by a series of transitions, where an individual progresses from a stage of risk factors, to detectable autoimmunity, then clinical RA, with each of these transitions likely related to genetic, environmental and stochastics factors as well as some failed immunoregulatory checkpoint⁶³. However, there are several additional considerations regarding this model of RA development. First, these transitions are not necessarily consecutive or mutually exclusive. Second, not all individuals who develop detectable circulating RA-related autoimmunity develop Clinical RA indicating that the development of autoimmunity is not an indication of inevitable clinical RA. Third, the trajectories of symptoms and and onset of clinical RA differs amongst individuals; for some clinical RA may be explosive with multiple joints appearing with IA in a short time, others may have gradual accumulation IA, and others may have waxing and waning findings (i.e. termed ‘palindromic rheumatism^64,65) and the precise relationship between symptoms and autoimmunity is not clear. All of these issues will need further study.

Predicting future clinical RA

Many prospective studies have sought to develop prediction strategies to identify individuals at high risk of developing clinical RA (Table 3). Autoantibodies are the best studied predictor of RA risk, and ACPA/RF status has been shown in multiple studies to predict future RA^{13, 24–27}. Importantly, increased ACPA levels clearly confers a higher risk of RA^{26, 66}. Notably, there may be differences in the predictive ability between the various commercially-available assays for ACPA although additional research is needed to clarify the specific differences^67,68. Multiple other autoantibody systems have been identified in the pre-RA period including anti-CarP⁶⁹, anti-MAA and anti-PAD^14,18,19,70, but their overall additive predictive value needs further validation. Other blood-based biomarkers including cytokines/chemokines, and cell subsets have also been explored in RA prediction. In particular, in prospective studies, alterations of T cell subsets⁷¹, presence of B cell signatures²² and presence of a type 1 interferon signature²³ were associated in increased risk for incident clinical RA although these findings have not yet been widely replicated. Preliminary studies on serum proteomics^35–37 and sputum abnormalities^38,39 suggest that additional biomarkers may be of utility in RA prediction, but further study and validation is needed.

Table 3.

Prediction of future inflammatory arthritis and rheumatoid arthritis in prospective studies of at-risk populations

Publication	Country of origin/population	Study type	Number of subjects and incident IA/RA	Key findings
del Puente et al 1988 15	USA, Akimel O'Odham (Pima) people	Prospective cohort study	2,712 subjects, 70 (~2.6%) with incident IA/RA after up to 19 years of follow-up.	The highest rate of development of RA (48 per 1000-person years) was in subjects with baseline RF titer of >1:256.
Silman et al 1992 16	United Kingdom	Prospective cohort study	370 unaffected first-degree relatives from families with RA; 14 with incident RA	Incident RA was highest in subjects with RF positivity.
Van de Stadt et al 2012 45	The Netherlands	Prospective study of individuals presenting to rheumatology clinics	347 subjects with RF and/or ACPA positivity but no IA at baseline; 131 with incident IA/RA after a median of 12 months.	A score was developed assigning 1 point for each of the following that were present: positive FDR status, no alcohol consumption (use of alcohol was protective), symptoms starting <12 months prior, intermittent symptoms, symptoms in upper and lower extremities, visual analog pain score of >=50 millimeters, more stiffness >=60 minutes, self-reported swelling in any joint; in addition, up to 4 points were assigned if both RF and ACPA were positive. In individuals with scores of >=7, 74% developed IA/RA within 3 years.
de Hair et al 2013 39	The Netherlands	Prospective study of ACPA and/or RF positive subjects	55 subjects, 15 (27%) with incident IA after a median of 13 months.	Non-smokers and those with normal body weight had the lowest rates of progression to IA/RA.
Ramos-Remus et al 2015 76	Mexico	Prospective study of unaffected FDRs of patients with RA	819 FDRS, 17 (2.1%) with incident IA/RA over 5 years.	ACPA positivity with or without concomitant RF positivity had PPV's of 58–64% for development of RA during follow-up.
Rakieh et al 2015 34	United Kingdom	Prospective study of ACPA+ (CCP2) subjects with arthralgia referred to rheumatology clinics	100 ACPA+ individuals, 50 with incident IA/RA after a median of 7.9 months	A score was developed assigning 1 point for each of the following: tender joints, morning stiffness >30 minutes, presence of the shared epitope, high levels of RF and/or ACPA, and the presence of ultrasound power doppler findings in >=1 joint. In individuals with the highest scores (>=2), >41% developed IA/RA within 24 months, and individuals with scores of >=4, 68% developed IA within 24 months.
Burgers et al 2017 74	The Netherlands and Sweden	Prospective study of subjects with arthralgia	178 subjects with arthralgia meeting EULAR criteria for CSA at baseline, 44 (18%) developed with incident IA/RA after a median of 16 weeks.	Study to validate the EULAR definition for Clinical Suspect Arthralgia 71. The presence of 3 or more of the following factors was ~84% sensitive and had a PPV of ~30% for IA/RA within 2 years: duration of onset of symptoms <1 year, symptoms in MCP joints, morning stiffness >=60 minutes, more severe morning symptoms, having an FDR with RA, and on examination, difficulty making a fist and tenderness with an MCP squeeze. However, PPV for IA was much less if the criteria were applied by a non-rheumatologist practitioner (PPV ~3%).
Tanner et al 2019 78	Canada	Prospective cohort study of FDRs of patients with RA	374 subjects, 10.9% were ACPA+ at baseline. 18 (4.8%) developed IA/RA after ~5 years of follow up.	ACPA+/RF+ individuals at baseline developed IA at a higher rate (97.1 per 1000 yrs) than ACPA+/RF− (36.4), ACPA-/RF+ (7.2) and ACPA− /RF− (4.1). ACPA+/RF+ were just as likely to become autoantibody negative as developing IA after 5 years.
Gilbert et al 2021 79	Switzerland	Prospective cohort study of FDRs of patients with RA	1458 subjects, ~5% were ACPA+ at baseline, and ~17% were RF-IgM+.	As of 2021,16 individuals had developed incident IA and were predominately ACPA positive.
Bemis et a 2021 77	USA	Prospective cohort study of FDRs of patients with RA	1780 subjects, 304 antibody+ (17.1%) at baseline. 20 (15.3% of antibody+) developed IA after ~4.5 years of follow up.	ACPA+/RF+ individuals at screening developed IA at a higher rate (38.0%) than ACPA+ alone (15.0%) and RF+ alone (9.0%). High level ACPA (i.e. >= 2x ULN) at screening was associated with the development of IA (HR 4.1)
Bergstedt et al 2022 35	USA	Prospective cohort study of ACPA+ individuals found through health-fair screening	90 ACPA+ individuals, 26 (29%) developed IA after ~ 2 years of follow up.	Those who developed IA had higher prevalence of SE (69% vs 38%). High level ACPA (>= 2x ULN) was higher in those who developed IA (85% vs 60%). Dual RF-IgA and RF-IgM+ was associated with IA (HR 2.9).
Duquenne et al 2023 66	United Kingdom	Prospective cohort study of individuals found through clinical referrals for joint symptoms	455 ACPA+ (CCP2) subjects without baseline IA; 148 (~33%) developed clinical IA over a median follow-up of 223 weeks.	Developed scoring systems to predict IA within 1 year: 1. A simple score used to triage referrals to secondary care which included morning stiffness > 30 min, ACPA level, RF+, and ESR high. A score above the threshold had a positive predictive value of~28% for IA development within 1 year. A comprehensive score used in secondary care which included Age > 50, ever smoker, morning stiffness, ACPA level, RF+, SE+, ESR high, VAS global pain, HAQ score, US PD signal, US tenosynovitis+, US erosion+. A score above the threshold had a PPV of ~70% for IA development within 1 year.

Abbreviations: RA=rheumatoid arthritis; IA=inflammatory arthritis; Ig=immunoglobulin; RF=rheumatoid factor; ACPA=antibody to citrullinated protein antigen; CCP=cyclic citrullinated peptide; USA=United States of America; MCP=metacarpal phalangeal joints; EULAR=European Alliance of Associations for Rheumatology; SE=shared epitope; US=ultrasound; PD=power doppler; ULN=upper limit of normal.

Imaging has shown promise in RA prediction. Ultrasound (US) power doppler (PD)(which can identified increased synovial blood flow indicative of inflammation) in antibody positive individuals with arthralgia^43,44, and early erosive disease particularly in the feet⁴⁵ are predictors of RA development. Magnetic resonance imaging (MRI) detection of joint inflammation (synovial and/or tenosynovial) is associated with future RA, particularly in those who are ACPA positive^29,51. Notably, imaging evidence of tenosynovitis in the hands is also independently associated with future RA development in individuals with CSA and ACPA positivity, and may suggest that the tenosynovium is one of the first joint-related targets in RA development⁴⁶.

Joint or other musculoskeletal (MSK) symptoms of pain, stiffness and swelling, while non-specific, are important predictors of the development of future clinical RA^65,72–75. A set of symptoms and timing termed ‘clinically suspect arthralgia’ (CSA) have been developed by a EULAR task force derived on patients with joint pain without evidence of arthritis⁷². The characteristics defining CSA include timing of joint symptoms, location of symptoms in the metacarpalphalangeal (MCP) joints, morning stiffness >60 minutes, more severe symptoms in the morning, first degree relatives (FDR) with RA, difficulty making a fist and positive squeeze test of the MCPs. This was evaluated in a separate prospective study of individuals with joint/MSK symptoms that demonstrated that if >=3 parameters were present based on a rheumatologist evaluation, there was a positive predictive value of ~30% for future clinical RA; however, if this was applied in individuals prior to a rheumatology evaluation, the positive predictive value was ~3%⁷⁴.

Demographics, environmental exposures and genetic factors also contribute to prediction of future RA. These potentially include age, sex, family history of RA, smoking, dietary factors, obesity and the shared epitope^41,42,66.

Multiple studies have used combinations of factors to develop prediction models for future clinical RA. The models in general include combinations of demographic features, joint symptoms, examination findings and autoantibodies and in some cases imaging findings. The details of these studies are included in Table 3 (and the rates of development of RA in clinical trials in Table 4). In general the models impart positive predictive values (PPVs) of ~20–70% for clinical RA development within 1–5 years, with the higher PPVs found with the presence of multiple risk factors for RA; furthermore, in most models high levels of ACPA impart the highest risk for future clinical RA. Importantly, the methods to identify individuals who are at-risk for future RA differs across studies. In particular, most studies of pre-RA have focused on at-risk populations identified through clinic referrals of individuals with joint/musculoskeletal systems, and therefore may be reasonable to assume that prediction models are most applicable in those populations. However, several studies have identified at-risk individuals through methods such as testing high-risk populations (e.g. indigenous North Americans), FDRs of individuals with RA or health-fair participants for ACPA, and then followed those individuals longitudinally^{15,35,76–78}. In these studies, ACPA positivity is generally associated with rates of clinical IA development of ~20–30%, or higher, within 5 years, although one FDR cohort reported an incidence rate of RA of ~2 cases per 1000 person-years⁷⁹. These rates are similar to general rates seen in populations of individuals at-risk for RA identified through clinic referrals for joint/musculoskeletal symptoms. This could be because even individuals who participate in ‘population-based’ studies may do so because of higher degree of symptoms, although this has not been evaluated in-depth in published work. Additional studies of how the mechanism by-which individuals are identified as being ‘at-risk’ for RA should be performed to develop predictive models that are applicable in these settings.

Table 4.

Summary of published trials to prevent/delay incident clinical RA in high-risk populations^*

Study	Inclusion criteria	Study design and Intervention	Primary outcome	Results
Bos et al 2010 85	RF and/or ACPA positive SE positive; arthralgia	RCT; dexamethasone 100 mg IM x 2 doses vs placebo	Incident clinical IA	17/83 (21%) developed IA after a median follow-up of 26 months; no difference between arms (dexamethasone 21% vs placebo 20%) Dexamethasone use associated with decreased autoantibody levels
Gerlag et al 2019 (PRAIRI) 86	RF and ACPA positive; CRP >0.6 mg/L; arthralgia	RCT; rituximab 1000 mg x 1 dose (and steroid) vs placebo	Incident clinical IA	30/81 (37%) developed IA after a mean follow-up of 29 months; no significant difference in overall rates of IA between arms (rituximab 14/41 (34%), placebo 14/40 (40%)); rituximab associated with delay of onset of IA
van Boheemen et al 2021 (StapRA) 87	RF and ACPA positive or ACPA >3x ULN; arthralgia	RCT; atorvastatin 40 mg/day x 3 years vs placebo	Incident clinical IA	15/62 (24%) developed IA after a median follow-up of 14 months; no significant difference between arms (atorvastatin 9/31 (29%) vs placebo 6/31 (19%)
Krijbolder et al 2022 (TREAT EARLIER) 88	Arthralgia and MRI evidence of joint inflammation in absence of clinical swollen joint; RF/ACPA not required for inclusion although 33% of participants were RF and/or ACPA positive	RCT; methylprednisolon e 120 mg x 1 dose and methotrexate up to 25 mg/week x 1 year vs placebo; 1 year post-drug followup	RA by 2010 criteria present at 2 time points 2 weeks apart	44/236 (~19%) developed RA over the 2 years of the trial; no significant differences between arms (methotrexate 23/119 (19%), placebo 21/117 (18%); decreased measures of physical function, pain and MRI inflammation in MTX treated group. The highest rate of RA development was within ACPA+ individuals (27/54=50%), although there was no significant difference in rates between arms at 2 years.
Rech et al 2021–2022 (ARIAA) 92,93**	ACPA positive and MRI evidence of joint inflammation	RCT; abatacept 125 mg SQ weekly x 6 months vs placebo; 12 month post-drug follow-up	MRI inflammatory parameter improvement; clinical RA	In preliminary analyses, at 18 months there was significant MRI improvement in abatacept arm compared to placebo (28/49 (57%) vs 14/49 (29%)). At 18 months there was also significantly less progression to clinical RA in abatacept arm compared to placebo (17/49 (35%) vs 28/49 (57%)).
Deane et al 202291 (StopRA) **	ACPA >=2x ULN	RCT; hydroxychloroquin e 200–400 mg/day x 1 year versus placebo; 2 years post-drug followup	RA by 2010 criteria	In preliminary analyses, 43/144 (~30%) developed RA over the 3 years of the study; no significant differences between arms (24/71 (34%), placebo 26/73 (36%); trial halted and final analyses pending
Cope et al95 2023 (APIPPRA) **	ACPA+RF positive or ACPA >=3x normal and arthralgia	RCT; abatacept 125 mg weekly injection x 1 year versus placebo; 1 year post-drug follow-up	RA by 2010 criteria	In preliminary analyses, at 2 years, 65/213 (~31%) of participants developed RA by 2010 criteria; 25% in abatacept arm and 37% in placebo arm. This resulted in differences in mean arthritis-free survival time between arms of ~99 days (p=0.002).

^*While the inclusion criteria varied across studies, in all studies no subjects could have clinical RA at baseline, and the primary endpoint included the development of examination evidence of inflammatory arthritis.

^**The results from ARIAA, StopRA and APIPPRA have only been presented in abstract form.

Abbreviations: SE=shared epitope; RA=rheumatoid arthritis; ACPA=antibodies to citrullinated protein antigens; RF=rheumatoid factor; RCT=randomized controlled trial; ULN=upper limit of normal; SQ=subcutaneous; MRI magnetic resonance imaging; MTX=methotrexate; IM=intramuscular; IA=inflammatory arthritis.

RA prevention clinical trials

Given the growing understanding of the pre-RA state and prediction models for future RA multiple clinical trials aimed at preventing the first onset of clinical RA have been implemented. These trials are further underpinned by the ‘window of opportunity’ concept where earlier treatment once an individual has clinical RA leads to improve long-term outcomes and therefore moving interventions into a pre-RA period may prevent or delay the onset of clinical RA^80,81,^82,83.

Of the completed RA prevention studies (Table 4), most have used medications that are already approved for use in the treatment of clinical RA. The rationale for this is that RA-approved drugs improve clinical RA in many patients⁸⁴, thus it is conceivable that the same medications might also delay or prevent the first onset of clinical RA. Further, the safety profile of RA DMARDs is well-established and therefore regulatory agencies are more likely to approve their use in prevention trials, and many at-risk populations (e.g. FDR) who are likely to be recruited to prevention studies, are familiar with their use.

A pioneering study of RA prevention was undertaken in a clinical trial using intramuscular dexamethasone. The trial enrolled at-risk individuals with elevated autoantibodies (ACPA or RF), joint symptoms as well as the shared epitope. The development of IA, was no different between placebo and treatment groups after 2 years of follow-up⁸⁵ although a levels were significantly lower in the dexamethasone group. In the PRAIRI study, 81 ACPA and RF positive individuals (also with CRP elevation and/or imaging evidence of inflammation) without clinical IA were randomized 1:1 in a masked, placebo-controlled fashion to receive a single dose of intravenous (IV) rituximab (1000 mg) versus placebo,⁸⁶ and all subjects also received 100 mg of methylprednisolone IV. There were no significant differences in overall rates of clinical IA after 29 months of follow up between arms (rituximab 34%, placebo 40%; p=0.448), however, rituximab delayed arthritis onset by approximately 5 months (16.5 vs 11.5 months) compared to placebo. In the StapRA study, atorvastatin 40 mg daily for 3 years in ACPA positive individuals failed to show any difference in arthritis onset after 3 years of follow up⁸⁷, though the study suffered from slow recruitment and was possibly underpowered.

In the TREAT-EARLIER study, 236 individuals with CSA and MRI detectable synovitis tenosynovitis and/or osteitis (yet without clinical IA) were randomized 1:1 in a masked, placebo-controlled fashion to receive methotrexate (dosed orally up to 25 mg/week) for 1 year⁸⁸. Both autoantibody positive and negative participants were enrolled (~30% were ACPA positive). Individuals randomized to the treatment arm were also given a single dose of intramuscular methylprednisolone (120 mg). After 2 years of follow up, there was no difference in the development of clinical arthritis between the groups (methotrexate 19%, placebo 19%). However, MRI-detected joint inflammation, functional scores, pain and AM stiffness all favored the methotrexate arm at the study endpoint of 2 years.

Notably, while the VITAL study that was designed to determine the effects of Vitamin D and Omega-3 fatty acid supplementation in oncologic and cardiovascular outcomes⁸⁹ and it did not recruit individuals based on risk factors for RA, it also examined autoimmune disease onset as an exploratory outcome. The combination of supplements showed a significant reduction in incident RA after 5 years of follow-up⁹⁰, albeit in a subgroup analysis (probable incident RA HR 0.27, 95%CI 0.09 −0.80, p = 0.02).

Several other prevention studies have yet to publish formal peer-reviewed results. However, preliminary results from the StopRA study in 144 ACPA positive participants (>2x upper limit of normal [ULN]) randomized to receive hydroxychloroquine (HCQ) for 1 year (<=6.5 mg/kg/day) versus placebo show that HCQ does not delay or prevent RA onset compared to placebo⁹¹. In contrast, preliminary data from the randomized, masked, placebo-controlled ARIAA trial demonstrated that abatacept (125 mg subcutaneous [SQ] weekly) for 6 months in ACPA positive individuals who also had hand MRI synovitis/tenosynovitis resulted in reduced joint inflammation by MRI as well as reduced progression to clinical RA up to 1 year after drug cessation^92,93. In addition, preliminary data from the APIPPRA study (randomized, masked, placebo-controlled) demonstrate that 1 year of abatacept (125 mg SQ weekly) reduced rates of progression to clinical RA (by 2010 criteria) at 2 years^94,95.

Challenges and opportunities in pre-RA research

There are multiple challenges in the design, implementation and interpretation of clinical trials and observational studies in pre-RA, and there are efforts underway to harmonize such studies in order to optimize the scientific results⁹⁶. First, in terms of the prevention trials, it is noteworthy that the inclusion criteria from these trials are quite different leading to a heterogeneity of findings. This is likely primarily due to a current lack of consensus on what consititutes a high-risk state for future RA. Second, it is generally agreed that ACPA positivity is a risk factor for future RA and as such, many observational and clinical trials have focused on ACPA+ individuals (with or without additional risk factors); however, the prevalence of autoantibody positivity is typically < 5% and likely closer to 1% in the general population⁹⁷ requiring large-scale efforts to identify autoantibody positive individuals. Third, RA autoantibodies have been shown to undergo seroconversion, or reversion to a seronegative state⁷⁸; furthermore, the fluctuation of autoantibody positivity may indicate evolving biologic processes and ultimately affect risk for RA although more studies are needed to fully understand the implications on risk of autoantibody reversion to negative. Fourth, ultimately, the preferences and participation of individuals at high risk for RA will drive the performance of clinical trials as well as the actual use of preventive treatments. Indeed, many individuals, despite having relatively high risk of developing RA (i.e. ACPA positive), are not interested in participating in clinical trials⁹⁸. As such, future research is needed to better understand recruitment strategies, participants motivations and perspectives, to enhance the efficiency of RA prevention trials⁹⁹. This is critically important for populations that are at potentially the highest risk for RA development and who may most benefit from preventive interventions. This includes FDR of patients with RA who may up to 5–7% lifetime risk of developing RA¹⁰⁰ as well as indigenous communities in North America have 2–3 fold higher prevalence of RA compared to Caucasians^78,101. Finally, choosing the right intervention or study drug – and the right individual and the right time to give it during pre-RA – is a crucial consideration for RA prevention trials. This is a balancing act between several competing factors: 1) targeting the appropriate biologic process for effective prevention, 2) an individuals actual risk of developing RA, 3) the side effects/tolerability of the study drug and 4) ultimately the cost-effectiveness of an intervention. Identifying a targeted therapeutic that specifically inhibits an immunological process that is active in the pre-RA stage would be ideal and studies will need to identify which biologic pathways are most important – and feasible – to address to effect prevention ¹⁰². It is also worth considering the results from the TREAT-EARLIER study, as above, suggesting that methotrexate may not prevent RA, but modify its natural course. It is well known that treating RA in the so-called “window of opportunity” improves disease control and remission rates¹⁰³. It remains possible that at-risk individuals who start standard DMARDs such as methotrexate, may derive benefits outside of disease prevention/delay, which may still be clinically important. For those at lower risk (ACPA positive without symptoms, low-level ACPA, normal imaging, etc.), interventions with a tolerable safety profile, or lifestyle modifications, such as dietary interventions, warrant future study to determine their efficacy in delaying or preventing RA. There may also be certain biologic pathways that are important in stages of RA (e.g. potential mucosal inflammation in a non-articular stage of RA development) that may require interventions that are different from agents that may be more effective once articular inflammation has developed⁶³. These areas will need further investigation.

Current Clinical Care in Pre-RA

While research into RA prevention continues, clinicians will be evaluating individuals in various states of ‘pre-RA’ in their clinics; these can include individuals who have ACPA and/or RF testing for musculoskeletal complaints or other organ injury (i.e. lung disease) yet do not have clinical RA on examination. There is no standard of care for the management of these individuals; however, some suggestions are listed in Table 5. In general these approaches involve careful evaluation of the patient to determine if there is active clinical RA as well as counseling on the natural history and personal risk of clinical RA, signs and symptoms of clinical RA and shared decision making around the optimal approach for management.

Table 5.

Clinical care of an individual who exhibits RA-related autoimmunity or clinical features in absence of clinical RA

1) Careful assessment to ensure that clinical RA is not present (may include imaging, with careful interpretation of findings); consider evaluation for other conditions that may be related to autoantibody positivity. For example, given some association of ACPA with lung disease, consider symptom driven evaluation for lung disease; if rheumatoid factor positive, consider additional studies to ensure that autoantibody is not related to another condition (e.g. hepatitis C, hematologic malignancy).

2) Discussion of rates of progression to clinical RA; in general, studies suggest that positivity for ACPA has ~20–30% positive predictive value for the onset of clinical RA within 3–5 years although actual risk may be higher or lower depending on many factors.

3) Consideration of lifestyle interventions that may reduce risk (smoking cessation, exercise and optimal body weight, oral health). A caveat is that these interventions have not been proven to reduce RA risk in controlled trials, but may have general health benefits.

4) Counseling regarding signs and symptoms of evolving RA, and consideration for routine follow-up (perhaps annually).

5) Treatment once clinical RA develops (caution regarding overtreatment if IA is only identified by imaging and not clinical examination).

6) Monitoring for advances in prevention that will impact clinical care.

Abbreviations: RA=rheumatoid arthritis; ACPA=antibody to citrullinated protein antigen.

A major emerging consideration is the use of imaging to identify active clinical RA in the event that IA is not seen on physical examination. Indeed, ultrasound and MRI are increasingly being used in rheumatology clinics worldwide to aid in diagnosis and management of IA, and several of the clinical trials (Table 3) have used the presence of articular inflammation on MRI findings as inclusion criteria. In addition, a study by Mankia and colleagues has identified that ~70% of rheumatology clinicians would likely use a DMARD in an individual who was ACPA-positive without clinically-apparent IA yet who had an abnormal imaging finding¹⁰⁴. Furthermore, a study by Krijbolder and colleagues found that individuals who participated in the TREA-TEARLIER study who had CSA plus MRI findings of joint inflammation were willing to to be treated with DMARDs even acknowledging such treatment may only improve symptoms and not reduce progression to clinical RA¹⁰⁵. However, while research in this area is progressing, the diagnostic accuracy of imaging for ‘true’ IA has not been well established and there are known false positive as well as false negative findings that are related to inherent technologic issues as well as user/interpreter factors^106–108. Furthermore, while there are several synovial biopsy studies that have included at-risk individuals^52,109, there is not yet a broad conclusion on how imaging corresponds with biopsy findings in pre-RA. In addition, there are not yet consensus guidelines on the use and interpretation of imaging to diagnose IA in absence of physical examination findings of IA. It is also important to note that even in the TREAT EARLIER trial, only ~20% of overall participants with MRI evidence of inflammation at baseline went on to develop clinical RA. As such, until more data and guidelines are available, it will be left to individual practitioners to use imaging findings cautiously in the diagnosis and management of IA in absence of physical examination findings clinical RA – with a caution that clinicians to be wary not to overtreat individuals who we do not have clear evidence-based guidelines that support they would benefit from DMARD therapy ¹¹⁰.

Conclusion

There is increasing understanding of the pre-RA stage of disease development that has already driven the completion or development of multiple clinical prevention trials. Importantly, growing interest in this area and the huge potential for improvements in the public health impact of this disease is driving even more studies in understanding the biology of disease development and developing effective preventive interventions – with efforts also underway in other rheumatic autoimmune disease such as lupus and psoriatic arthritis^111,112. Furthermore, as a model, there is now an approved preventive intervention in type 1 diabetes – a disease that has a model of development similar to RA¹¹³. It is exciting to see the field moving forward to a point where rheumatologists can include discussions around prediction – and potentially soon approved preventive interventions - of the clinical onset of RA - and potentially other rheumatic diseases.