Where are we with Biomarkers in Rheumatoid Arthritis? Save

Know-it-now

• Anti-CCP remains the diagnostic gold standard, but serologies in combination may become an advantage.
• The MBDA score is one of the most studied of the putative biomarkers.
• Pathotyping of synovial tissues shows promise.
• Power Doppler ultrasonography is the most practical imaging tool for tapering decisions.
• Comorbidity surveillance can be enhanced by existing biomarkers.
• Some biomarkers are already used in routine clinical practice; but most of the putative laboratory and imaging markers are still under investigation, in search of a potential clinical application.

Biomarker interest has grown considerably in the last 2 decades, owing to advances in genetics, imaging, protein, and multiomics. Despite these advances, biomarkers as the predictive holy grail of RA therapeutics and prognostication have not yet advanced beyond rheumatoid serologies and C-reactive protein (CRP).

A current review in Nature Reviews in Rheumatology by Szekanecz et al., comprehensively appraises the current and emerging biomarker landscape in RA.  These analyses address the RA continuum — from at-risk states through established and difficult-to-treat (D2T) disease — and explicitly grade biomarkers by translational readiness.

Diagnostic Biomarkers. Anti-CCP antibodies remain both the diagnostic cornerstone (~95% specificity, ~67% sensitivity) and best performing biomarker (predicting progression and worse outcomes) in RA. ACPA should not be used for broad screening in primary care or asymptomatic populations. In one cohort of 12,000 asymptomatic adults, fewer than 1 in 10 ACPA-positive individuals developed RA within 3 years. 

Newer autoantibodies — anti-carbamylated protein (anti-CarP), anti-PAD4, and anti- scavenger receptor-A (SRA) — may be positive in some seronegative patients, these do not match the performance and utility of anti-CCP in routine practice. Most advocate for the combined use of ACPA and rheumatoid factor, with also ordering the 14-3-3η assay to improve sensitivity of early seronegative RA, up to approximately 78%.  Large multi-ancestry genome-wide association studies (GWAS) have identified more than 120 RA associated loci, the predictive use and value of polygenic risk scores in RA has been limited. Similarly, microRNAs are investigational.

Prognostic Biomarkers. ACPA and rheumatoid factor positivity — particularly at higher titers — are linked to more aggressive disease, greater erosive burden, and reduced likelihood of sustained remission.  The erythrocyte sedimentation rate (ESR) and CRP are useful systemic inflammation markers, but may be normal in 33–42% of active RA and their inflammatory metric potential may be altered by age, sex, anemia, obesity and immunoglobulin levels.  The multi-biomarker disease activity (MBDA) is an expensive ($800-1000) blood test (that is a combination of 12 biomarkers including CRP, SAA, IL-6, MMP1/3, VEGF, and others) has been shown to outperform DAS alone for predicting radiographic progression (AUC 0.767 vs. 0.521 in the BeSt trial). Yet, MBDA has yet to be prospectively tested in a clinical trial to better predict clinical or radiographic outcomes – its performance and associations have been derived from previously collected clinical trial data and samples. 

MMP3 stands out as the strongest individual predictor of future erosive disease in non-erosive early RA. On imaging, MRI-detected bone marrow edema is the strongest structural predictor of future erosions, while power Doppler ultrasonography can detect subclinical synovitis even in clinical remission. Hence, erosions predict erosions.  Synovial tissue pathotyping — characterizing synovial tissue cellular subtypes as either lympho-myeloid, diffuse-myeloid, and pauci-immune — is of great interest by either histologic or transcriptomic granularity beyond blood-based markers, which may correlate with clinical and radiographic outcomes.

Therapeutic Biomarkers. Serostatus modestly influences biologic DMARD response in a drug class-specific manner: seropositive patients show lower discontinuation rates for abatacept (HR 0.80) and rituximab (HR 0.70), with no significant association for TNF inhibitors. Anti-CarP antibodies, particularly in combination with ACPA and anti-PAD4 status, may better predict abatacept response. In the R4RA trial, synovial B cell-rich pathotype predicted rituximab response while myeloid-rich tissue predicted tocilizumab response; pauci-immune pathotype predicted non-response to both — a potentially actionable finding now being tested with machine-learning algorithms. Multi-omics approaches, including transcriptomic and DNA methylation signatures, can distinguish responders from non-responders to TNF inhibitors and other DMARDs in some trials, but these findings are inconsistent and preliminary.

DMARD Tapering. Seronegative patients are most likely to sustain DMARD-free remission. Power Doppler ultrasonography-detected subclinical synovitis is the strongest imaging predictor of flare in clinical remission, a finding supported by multiple studies and a meta-analysis. MRI data remains more equivocal. Elevated calprotectin and MBDA scores and ACPA may predict relapses during tapering.  Therapeutic drug monitoring may support selected de-escalation decisions but is not recommended for routine proactive use per EULAR recommendations.

Comorbidity Biomarkers. Standard cardiovascular risk calculators underestimate the future risk of cardiovascular events in RA; EULAR recommends a 1.5× multiplication factor. ACPA and anti-CarP antibodies are independently linked to accelerated atherosclerosis and cardiovascular mortality. Each 20 mg/L rise in CRP corresponds to approximately 1% additional 10-year cardiovascular risk. NT-proBNP >100 pg/mL confers a ~3.4-fold increased cardiovascular mortality risk in RA — modestly higher than in the general population. For RA-ILD, KL-6, SP-D, and MMP7 are the most informative serum markers; MUC5B rs35705950 identifies patients at highest risk for the usual interstitial pneumonia pattern. Bone turnover markers (P1NP, CTX, OPG/RANKL ratio) are useful for monitoring treatment effects on inflammatory bone loss.

Current biomarker use is limited by specificity, reproducibility, variable performance in a heterogeneous longitudinal disease, barriers to validation and standardization, cost and potential implementation.