Investigational biologic markers in the diagnosis and assessment of rheumatoid arthritis

INTRODUCTION — Patients with rheumatoid arthritis (RA) follow a variable disease course with regard to outcome measures such as functional status and joint damage. Early aggressive intervention with disease-modifying agents is central for gaining control over the inflammatory disease. This has over the last two decades changed the outcome of RA. However, there are still patients that progress, despite an intensified treat-to-target strategy. This has led to the investigation and measurement of numerous potentially biologic "markers" in blood and joint fluids that may serve as indicators of prognosis and the response to therapy.

Many biomarkers identify mechanistically important elements in RA pathogenesis and improved our understanding of disease processes and responses to treatments. Although some of the markers under consideration are accessible in routine practice, many are in the stage of experimental evaluation and require access to specialized technology and customized reagents. Furthermore, many biomarkers reflect the time period during which they were examined. Applying a treat-to-target approach has rendered many biomarkers of historical interest obsolete as they often have low predictive value and insufficient specificity for use in guiding treatment of patients with RA [1,2].

Biologic markers that are considered investigational with respect to their clinical utility are reviewed here. Biologic markers that are clinically useful in the diagnosis and management of RA are described in detail separately. (See "Biologic markers in the assessment of rheumatoid arthritis".)

CLINICALLY USEFUL MARKERS — Among the many biologic markers that have been assessed for usefulness in estimating disease activity and prognosis of rheumatoid arthritis (RA), only a few have found a role in clinical practice. The main clinically useful biologic markers in patients with RA include rheumatoid factors (RF), anti-cyclic citrullinated peptide (anti-CCP) antibodies, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP). These tests and their clinical use in diagnosis and assessment of disease outcome are discussed in detail elsewhere. (See "Biologic markers in the assessment of rheumatoid arthritis".)

INVESTIGATIONAL MARKERS — Potential biologic markers that remain investigational can be considered in four categories:

●Immunologic (or serologic) abnormalities

●Genetic factors, such as human leukocyte antigen (HLA) class II

●The acute phase response elicited by hepatocytes as part of the inflammatory process, other than C-reactive protein (CRP) or those assessed indirectly by erythrocyte sedimentation rate (ESR)

●Macromolecules specific for joint-associated tissues that are released into the circulation or excreted in the urine as part of degenerative and reparative processes

Although not measured clinically, the pattern of cytokine activation may be important in the pathogenesis of rheumatoid arthritis (RA). This topic is discussed elsewhere (see "Pathogenesis of rheumatoid arthritis"). Furthermore, some of these markers also vary depending upon how far along the patients are in the disease course.

Immunologic abnormalities — Several immunologic abnormalities, particularly autoantibodies, are associated with the presence of rheumatoid arthritis.

Autoantibodies

Antiperinuclear factor and antikeratin antibodies — Characterization of the antigenic targets recognized by antikeratin and antiperinuclear antibodies led to the recognition of citrullinated peptides as important antigenic targets for antibodies present in the sera of patients with RA. Commercially available assays for the detection of anti-citrullinated peptide antibodies (ACPA), including antibodies to cyclic citrullinated peptides (anti-CCP), are widely used clinically in the evaluation of RA. Measurement of ACPA is especially important in relation to diagnosis of RA, whereas its ability to predict individual outcomes is of less value. (See "Biologic markers in the assessment of rheumatoid arthritis", section on 'Anti-citrullinated peptide antibodies'.)

Antikeratin antibodies (AKA) were originally identified in 1979 in patients with RA [3]. A related antibody, antiperinuclear factor antibody (APF), is also associated with RA [4]. Study of APF and AKA led to the appreciation that both types of antibodies recognize epidermal filaggrin, an intermediate filament-associated protein involved in cornification of the epidermis [5,6]. A major determinant of antibody binding to filaggrin is the presence of citrulline, formed by a post-translational modification of arginine. Citrullination is catalyzed by peptidylarginine deiminase [7]. Antibodies against this antigen had higher specificity for RA than did rheumatoid factor (RF). However, assays for these antibodies were technically demanding and never widely available.

Identification of these and other post-translationally modified peptides and proteins containing the amino acid citrulline led to the development of the more convenient and standardized assays for antibodies to citrullinated proteins, including enzyme-linked immunosorbent assays (ELISA) for anti-CCP antibodies. (See "Biologic markers in the assessment of rheumatoid arthritis", section on 'Anti-citrullinated peptide antibodies'.)

Testing for antibodies to other citrullinated proteins has been an area of investigation. Although it is interesting from a pathogenic viewpoint, it has added little to our ability to predict outcome in clinical practice. As examples, the presence of antibodies to citrullinated fibrinogen, to citrullinated vimentin (anti-Sa), to citrullinated synthetic type I or type II collagen telopeptides, and to alpha-enolase may have sensitivity and specificity similar to that of anti-CCP testing [8-15]. A commercial assay for anti-mutated citrullinated vimentin is also available. (See "Biologic markers in the assessment of rheumatoid arthritis", section on 'Anti-MCV antibodies'.)

The following discussion of APF and AKA is, therefore, of mainly historical interest. It also reflects that RA, like many autoimmune diseases, is associated with a huge variety of different autoantibodies. The pathogenic meaning of many of these antibodies remains uncertain.

The sensitivity and specificity of APF and AKA vary depending on the laboratory means of detection.

●APF detected by means of indirect immunofluorescence using buccal epithelium has a reported specificity for RA of 73 to 99 percent [4,16,17] and a sensitivity of 49 to 87 percent [4,16,18].

●The reported specificity of AKA, detected by the use of rat esophagus as a substrate [3], generally exceeds 90 percent with a sensitivity between 40 and 60 percent [16,19].

Despite the lack of widespread availability of tests to detect AKA and APF (and anti-filaggrin antibodies), identifying their presence in a given patient may have important implications.

●AKA have been reported in RF-negative RA, although the prevalence is only one-third that in RF-positive cases [16]. Similarly, APF has been reported in cases of RA that are repeatedly RF-negative [3].

●The prognostic significance of AKA and APF has been studied in three cohorts of early RA with follow-ups of three and eight years, respectively [20-22]. All patients with detectable AKA developed at least some erosions by the end of the eight-year follow-up [22].

●Similar to RF, the presence of AKA, APF, or antifilaggrin antibodies may antedate clinical RA [23-25]. However, a significant number of antibody-positive patients do not develop disease. As an example, in one of the studies cited, 12 of 70 RF-positive patients who did not develop RA tested positive for either AKA, APF, or both antibodies [24].

Anti-Sa antibodies — Anti-Sa antibodies, similar to a number of other autoantibodies linked to RA, recognize a citrulline-containing epitope located on vimentin, an intermediate filament that is widely expressed on mesenchymal cells and macrophages that is not normally citrullinated [26]. Antibodies to mutated citrullinated vimentin (anti-MCV) may be associated with more severe disease than anti-CCP antibodies and may be comparable to anti-CCP in predicting radiographic progression [11,27]. A commercial assay for antibodies to MCV is available, and the information below is presented for historical perspective. (See "Biologic markers in the assessment of rheumatoid arthritis", section on 'Anti-MCV antibodies'.)

The following observations suggested that the presence of anti-Sa antibodies may be an accurate diagnostic and predictive test for RA:

●In a review of 489 patients with inflammatory joint disease, of whom 154 had RA, anti-Sa antibodies were 68 percent sensitive and 79 percent specific for RA [28]. Anti-Sa antibodies also helped predict which patients would eventually develop late severe radiologic damage.

●A second study compared the use of anti-Sa and anti-CCP antibodies in an outpatient clinic population comprised of patients with RA, other connective tissue diseases, and spondyloarthritides [29]. Anti-Sa antibodies were much less sensitive (44 versus 72 percent) with a similar specificity (96 versus 94 percent). Anti-Sa and anti-CCP antibodies were discordant in 47 of 87 patients with RA. Thus, assays for anti-Sa antibodies were proposed as being useful when anti-CCP antibodies are negative but RA is still suspected.

Other autoantibodies — A number of other autoantibodies have also been the subject of investigation:

●ANCA – Patients with RA may have antineutrophil cytoplasmic antibodies, usually atypical (non-MPO) perinuclear ANCA (p-ANCA). Nonvasculitic aspects of RA activity, severity, and chronicity have not consistently correlated with ANCA status. One report found that patients with RA who were p-ANCA positive had an increased rate of radiographic progression of joint damage [30]. However, until this is confirmed, it is our opinion that there is little clinical utility for ANCA testing in patients with RA in whom the presence of an ANCA-associated systemic vasculitis is not suspected on clinical grounds. (See "Clinical spectrum of antineutrophil cytoplasmic autoantibodies", section on 'Nonvasculitic systemic rheumatic diseases'.)

●Gal 0 glycoforms – RA is associated with a marked increase in IgG lacking galactose (designated as Gal 0 glycoforms) and terminating in N-acetyl glucosamine in the Fc region [31]. Gal 0 levels correlate with disease severity in RA and revert to normal during the course of pregnancy-induced remissions [32]. Agalactosyl immunoglobulins (Gal 0 IgG) activate complement by means of a mannose binding protein dependent pathway in vitro [33]. More rapid radiographic progression of RA has been associated with low serum levels of mannose-binding lectin, which is determined genetically [34,35].

Gal 0 IgG and RF were measured at presentation and at a two-year follow-up in a prospective study of 60 British patients with a history of synovitis of less than one year's duration [36]. Over the study period, 39 patients developed RA as defined by the 1987 American College of Rheumatology (ACR) classification criteria for RA (see "Clinical manifestations of rheumatoid arthritis"). The initial Gal 0 IgG correctly predicted the development of RA, or otherwise, in 47 of the 60 patients. The combination of Gal 0 IgG and RF status gave a test with 90 percent sensitivity, 95 percent specificity, and 94 percent predictive value.

●Antibodies to alpha-enolase – Among 255 patients with early arthritis, elevated serum levels of antibodies to an enzyme in the glycolytic pathway, alpha-enolase, were found in 25 percent of those who were later determined to have RA [37]. One-half of those with RA and anti-enolase antibodies did not have serum RF or antifilaggrin antibodies. Among 40 patients whose arthritis remained undifferentiated after one year of observation, 10 percent had antibodies to alpha-enolase.

●Antibodies to glucose phosphate isomerase (GPI) – Antibodies to the ubiquitous enzyme, glucose phosphate isomerase, which have been implicated as pathogenically important in an animal model of inflammatory arthritis, may be associated with the presence of extraarticular manifestations of RA . These autoantibodies are more often present in those patients with RA complicated by rheumatoid vasculitis and Felty syndrome [38].

●Antibodies to PADI4 – It is known that the enzyme peptidyl-arginine deiminase 4 (PADI4) converts arginine to citrulline and that genetic variants of this enzyme are associated with RA. The presence of antibodies to PADI4 in patients with RA and in controls was examined in Chinese patients [39]. Levels of antibodies to PAD4 were assessed in 109 patients with RA, in 232 patients with related rheumatic diseases, and in 106 healthy controls. Antibodies to PAD4 were observed in 45 percent of patients with RA and in 0 to 13 percent of controls.

●Anti-carbamylated protein antibodies (anti-CarP) also belong to a group of autoantibodies directed towards post-translational modified proteins. Carbamylation is a chemical reaction mediated by cyanide in which a lysine is converted into a homocitrulline. Anti-CarP antibodies do not show the same specificity for RA as ACPA. Anti-CarP was seen in 12 percent of patients with renal disease compared with 44 percent of patients with RA [40]. The molecular structures of homocitrulline and citrulline show similarities, but ACPA and anti-CarP are distinct autoantibody classes. This has been further supported by anti-CarP being present in both ACPA-positive and ACPA-negative patients; anti-CarP has been associated with radiographic progression in patients negative for RF and ACPA [41].

Complement activation — Markers of complement activation, possibly due to immune complexes present in synovium or plasma, have been proposed as a possible indicator of RA disease activity. As an example, the concentration of covalently linked C1q-C4 complement components was correlated with disease activity in 84 patients with RA [42].

Genetic factors — The risk of developing RA is associated with carriage of particular HLA alleles. A patient's HLA DR type and other genetic factors may also play a role in determining the severity of disease. RA is positively associated with certain HLA DR alleles (particularly HLA DR4) encoding a conserved amino acid sequence in the third hypervariable region of the DRb1 chain. A number of studies have suggested the potential usefulness of HLA typing in predicting disease severity and as a guide to early aggressive therapy. This area of investigation is discussed in detail separately. (See "HLA and other susceptibility genes in rheumatoid arthritis".)

Another approach has involved RNA sequencing (RNA-seq) analysis of synovial biopsies and blood in a cohort of 90 treatment-naïve patients with mean symptom duration of 5.6 months [43]. Three different disease pathotypes were identified by synovial histopathology: fibroblastic, with minimal immune-inflammatory cell infiltration; myeloid macrophage-rich; and lymphoid-rich showing progressive accumulation of B and plasma cells. Specific synovial immune cell transcriptome signatures correlated with clinical and radiologic parameters, while peripheral blood RNA-seq showed less diversity than synovial tissue. Detectable functional peripheral blood signatures, in particular type I interferon response and toll-like receptor signaling, were associated with synovial lymphoid and myeloid pathotypes respectively, and correlated with disease activity. The authors documented their findings in an open website (https://peac.hpc.qmul.ac.uk/). This will allow other researchers access to a searchable interface for comparing a gene of interest in synovium or blood against the histologic pathotype, leukocyte subsets, CRP, ESR, CCP, RF, disease activity score derivative for 28 joints (DAS28), health assessment questionnaire (HAQ), radiology, and clinical response after 12 months of treatment.

RNA sequencing has been used to define a low or absent B cell lineage expression signature in synovial tissue of patients with an inadequate response to tumor necrosis factor (TNF) inhibition. When these patients were randomly assigned to receive treatment with either tocilizumab or rituximab, the tocilizumab group had a higher response rate (responses defined as 50 percent improvement in clinical disease activity index [CDAI]) compared with the rituximab group in the patients with synovial biopsies classified as B cell poor by RNA sequencing (63 versus 36 percent) [2].

Acute phase response — Markers of the acute phase response, such as CRP, the ESR, and interleukin (IL) 6, have been evaluated as potential markers of disease activity in RA. (See "Acute phase reactants" and "Biologic markers in the assessment of rheumatoid arthritis".)

Interleukin 6 — IL-6 has a major stimulatory effect on hepatic synthesis of acute-phase proteins [44] and is known to play a regulatory role in platelet production [45]. Furthermore, there is evidence to suggest that IL-6, as all STAT3-activating molecules, has an etiopathologic role in the anemia of chronic disease [46]. (See "Anemia of chronic disease/anemia of inflammation".)

Inflamed synovium is thought to be the principal source of plasma IL-6 in RA, since IL-6 is often detected in high concentration in the synovial fluid. Thus, it was postulated that plasma IL-6 concentrations might reflect joint inflammation better than acute-phase protein levels. However, in one study of 51 patients with early RA, there was no relationship between radiologic progression and time-integrated values of plasma IL-6 concentration even though there was a significant correlation between IL-6 and the acute phase response [47]. This finding contrasts with the relationship between radiologic findings and time-integrated values of CRP and ESR discussed above.

Tissue-specific markers — A number of biochemical markers of joint damage have been described in RA. These molecules may be synthetic or degradative, their presence in body fluids arising as a consequence of metabolism of the tissue of origin. They are predominantly derived from a single tissue such as cartilage, bone, or synovium and can be detected principally by immunoassay of joint fluid, serum, or urine. Appropriate assays are not routinely available, and joint fluids are not readily available (apart from knee joints), in comparison with serum or urine. Furthermore, there is the question of whether the spillover from small joints, versus larger joints, will only have limited influence on the plasma levels. There is, as yet, limited information regarding the usefulness of tissue-specific markers as measures of disease activity or response to therapy in RA. However, some markers may be of prognostic value.

Synovium-specific markers — Although these are, to a degree, synovial specific, to a large extent such biomarkers reflect the total "synovial mass" and thus may be influenced by treatment. Serum hyaluronan is reported to be elevated in the serum of patients with RA [48,49]. In vitro studies demonstrate that synovial lining cells of rheumatoid joints produce detectable amounts of hyaluronan, while lining cells from normal joints do not [50]. Despite a short half-life of 15 minutes or less, serum hyaluronan concentrations correlate with disease activity [49], and at least one prospective study has suggested that, in early RA, serum hyaluronan may reflect ongoing joint destruction and may even predict subsequent joint damage [51]. However, elevated serum levels of hyaluronan may be nonspecific since they may vary with physical activity independent of the degree of synovitis [52].

Other markers that may be predominantly released from the synovium are MMP-1 and MMP-3, enzymes that fragment matrix collagen. Elevated levels of MMP-3 [53] and/or MMP-1 [54] may correlate with increased radiographic joint damage. Elevated serum levels of immunoglobulin soluble receptor (Fc gamma RIIIa) in RA are thought to be due to synovial macrophages and/or NK cells [55].

An isoform of the 14-3-3 family of chaperone proteins, 14-3-3eta, which has a molecular weight of approximately 28 kDa, is present in the sera and synovial fluid of a majority of patients with early and established RA [56,57]. In synovial fluid, the levels of 14-3-3eta are more than fivefold greater than in matched serum specimens [56]. Serum 14-3-3eta may be helpful diagnostically in RA, as it may have similar sensitivity and specificity to RF and ACPA in distinguishing patients with RA from osteoarthritis, other autoimmune disorders, and healthy controls [57]. Serum levels of 14-3-3eta in inflammatory arthritis correlate with levels of MMP-1 and MMP-3, and the protein possesses ligand activity, preferentially activating cells of the innate immune system to activate proinflammatory cytokines, including IL-1, IL-6, and TNF alpha, as well as factors involved in joint injury such as MMP-9 and RANK ligand [58]. Nonetheless, serum levels of 14-3-3eta do not correlate with levels of acute phase markers or composite disease activity scores, although its presence is associated with more severe disease [57].

Cartilage-specific markers — Markers of cartilage metabolism may have some prognostic value in patients with RA, but their measurements still result in a low likelihood ratio for predicting joint destruction. For example, high serum levels of cartilage oligomeric matrix protein (COMP), a member of the thrombospondin protein family, predicted severe disease characterized by subsequent large- and small-joint destruction [59,60]. In one study, increased serum concentrations of COMP were found in all patients who developed rapid hip joint destruction [60].

The aggrecan content of synovial fluid in knee joints is also reported to be predictive of knee and hip joint destruction [61]. The chondroitin sulphate-rich region of aggrecan is most abundantly detected in synovial fluids recovered from joints with little radiologic evidence of destruction, whereas the hyaluronan binding region of core protein is released in more severely damaged joints [62].

The same study that measured serum level of COMP in patients with RA (see above) also measured serum levels of a putative marker of cartilage aggrecan synthesis, epitope 846, located on the chondroitin sulphate rich area of the aggrecan molecule. The epitope 846 levels were found to be elevated only in a group of patients with slow joint destruction, as compared with a group matched for age, sex, and disease duration but with more destructive joint disease [60]. These data indicate the presence of cartilage reparative processes in the group with a more benign course and suggest that elevated 846 epitope is indicative of a more favorable prognosis.

Measurement of crosslinked c-terminal peptides from type II collagen (CTX-II) in urine may provide some prognostic information. A correlation between the excretion of these peptides and radiographic progression up to five years was noted in a prospective study of 110 patients with early RA [63,64]. Similarly, urinary excretion of a peptide derived from the helical portion of type II collagen (HELIX-II) also correlates with radiographic progression and is independent of other variables, including baseline CRP levels, joint damage, and urinary CTX-II excretion [65]. When studied in 89 patients with early RA, patients with increased levels of both HELIX-II and CTX-II had the highest risk of radiographic progression compared with those without an elevation of either of these markers (OR 17.5, 95% CI 3.1-99).

The pericellular matrix consists mainly of collagen type IV (C4M) and has the collagen markers above. C4M has also been shown to predict outcome and association with destruction in RA [66]. The report combined two large placebo-controlled phase III trials of 687 (LITHE) and 217 (RADIATE) RA patients, respectively, showing that anti-IL-6R antibody treatment resulted in a decrease in plasma C4M and that likelihood of achieving an ACR 20 percent response rate (ACR20) response by week 16 was associated with C4M suppression exceeding the median decrease at week 4 [66].

The combination of cartilage and synovial biomarkers may have a better ability to predict radiographic progression than either alone. This was illustrated in a study of 118 patients in which several biomarkers were measured in blood samples and in which radiographs were obtained at baseline and after two years of treatment with synthetic disease-modifying antirheumatic drugs (DMARDs) but not with biologic agents; multivariable analysis indicated that the combinations of serum biomarkers that included both MMP-3 and CTX-II levels were better than any individual biomarker in predicting radiographic progression [67].

Bone-specific markers — As with cartilage, several bone-specific markers are available but have a low likelihood ratio for predicting progression in RA:

●Bone sialoprotein is an osteoblast-derived protein preferentially expressed in juxtaarticular bone. Bone sialoprotein levels in synovial fluid correlate with knee joint destruction in both RA and osteoarthritis [68]. By contrast, bone sialoprotein levels are elevated in RA serum without a correlation between concentration and joint destruction.

●Bone degradation, assessed by detection of pyridinoline cross-links in urine, correlates with disease activity in RA and diminishes after treatment with pulsed glucocorticoids and DMARDs [69].

●Immunoassays are available for measurement of cross-linked carboxyterminal telopeptides of type I collagen (ICTP), a larger serum marker for bone collagen degradation. A three-year prospective study of 66 patients with early RA found that 51 percent initially had elevated levels of serum ICTP compared with healthy controls. Throughout the follow-up, serum ICTP levels correlated with inflammatory parameters and, from the first year on, with the radiologic changes assessed annually. Initial ICTP levels correlated better than the other variables of disease activity with the subsequent erosive progression of joints, suggesting that its measurement may serve as a prognostic marker for joint damage in early RA [70]; this observation was independently confirmed in a group of 110 patients with early RA [63]. Subsequent studies found that ICTP levels in synovial fluid correlated better with prognosis than did serum levels [71] and that higher ICTP levels after six months of treatment with a combination of antirheumatic drugs was predictive of increased radiographic progression [72]. C-terminal crosslinking telopeptide of type I (CTX-I) and type II (CTX-II) collagen can be measured in the urine. Urinary CTX-I and CTX-II were examined in 155 patients with early RA. In those who did not show signs of joint damage at baseline, both u-CTX-I and u-CTX-II predict an increased risk of radiologic progression over four years [63].

●The initial ratio of serum levels of an inhibitor of osteoclast differentiation (osteoprotegerin) and a stimulatory protein (Receptor Activator of Nuclear factor Kappa B ligand [RANKL]) may be a predictor of joint destruction in patients with early RA [73].

●Serum levels of Dickkopf-related protein 1 (DKK-1) decreased, and levels of sclerostin (SOST) increased during one year's treatment with anti-TNF-antibodies in 36 RA patients [74]. SOST had a positive correlation with bone mineral density [75]. In a study of 50 RA patients, both DKK-1 and SOST were increased, and DKK-1 showed positive association with radiologic progression.

Vascular markers — Serum vascular endothelial factor (VEGF) concentrations are elevated in RA patients when compared with healthy controls and with patients with osteoarthritis [76-78]. Furthermore, in patients with early RA, there is a significant correlation between serum VEGF levels at presentation and radiographically assessed joint damage over the subsequent year [77].

Plasma proteins — In a proteomic approach, 163 plasma proteins were evaluated in 44 patients with RA to determine which were associated with disease activity [79]. Plasma proteins whose concentrations were correlated with disease activity included IL-6, oncostatin M, IL-2, macrophage colony-stimulating factor (M-CSF), TNF receptor superfamily member 9, C-C motif chemokine 23 (CCL23), transforming growth factor (TGF) alpha, and chemokine C-X-C motif ligand 13 (CXCL13). A single measurement of CXCL13 in plasma at the time of diagnosis showed a likelihood ratio of close to 5 for long-term radiographic progression in RA [80].

Glycosylated related proteins — Galectins (Gal) are important for both intracellular and extracellular mechanisms. Gal-3 has been shown to be related to disease outcome and joint destruction in ACPA-positive early RA [81]. Increased Gal-3 may serve as a serologic signature of pre-RA, while markers of synovitis and cartilage do not differ between early undifferentiated arthritis patients irrespective of subsequent clinical differentiation [82].

Multi-protein biomarker algorithms — The combined use of multiple markers may provide advantages over the use of single markers for predicting disease activity and progression. In order to develop a multi-biomarker disease activity (MBDA) test for RA, 130 candidate biomarkers were tested in feasibility studies and 25 were selected for algorithm training. Multi-biomarker statistical models outperformed individual biomarkers at estimating disease activity assessed by the DAS28 and C-reactive protein level (DAS28-CRP). The final MBDA algorithm used 12 biomarkers to generate an MBDA score between 1 and 100 [83]. Emerging data indicate that a MBDA test may provide a useful adjunct to clinical assessment to identify progression-free remission and assess subclinical disease. In a study of 163 patients with RA in the Leiden Early Arthritis Cohort, at a total of 271 visits, patients with a high MBDA score were 2.3 times more likely (95% CI 1.1-3.7) to have joint damage progression during the subsequent year [84].

Whether MBDA has a useful role in informing a clinical management decision has been more controversial. In an analysis of patients with RA from the Swedish Pharmacotherapy (SWEFOT) study with an inadequate response to methotrexate (MTX), those patients with greater MBDA score reduction with MTX were more likely to respond to triple therapy than to MTX plus infliximab [85]. However, post hoc analysis of the AMPLE ("abatacept versus adalimumab comparison in biologic-naïve RA subjects with background methotrexate") trial data has been interpreted as indicating that the MBDA score could not be reliably used to guide decision making in RA management, particularly for patients who receive abatacept or adalimumab as a first biologic agent [86].

Furthermore, as IL-6 is a contributor to the MBDA score, caution must be exercised when interpreting the MBDA response following use of IL-6 receptor (IL6R) inhibitors (eg, tocilizumab), given that blood IL-6 rises after treatment, unlike other acute phase measures and MBDA components. The increase in IL-6 levels in patients with a clinical response results in a reduced magnitude of the response of the MBDA score compared with the degree of improvement that is indicated by the DAS28-CRP [87].

The MBDA score is increased with adiposity, and a leptin-adjusted MBDA score was found to significantly add information to the DAS28-CRP and the original MBDA score in predicting radiographic progression in RA [88].

A different approach has been to divide patients into three synovial phenotypes based on the synovial histology; these include lymphoid, myeloid, and low inflammatory or fibroid-dominated inflammation [89]. Cytokine levels, primarily serum CXCL13 and ICAM-1, in patients with each of these phenotypes were examined in 69 RA patients who received anti-TNF or anti-IL-6R antibody treatment, and although there was an overlap, the best outcome of anti-TNF treatment was in patients with sICAM1-high/CXCL13-low cytokine profiles, whereas sICAM1-low/CXCL13-high patients showed the best response to anti-IL-6R.

SUMMARY AND RECOMMENDATIONS

●Early identification of patients with RA and, in particular, of those likely to assume a more rapidly destructive form of disease is important because of the possible benefit from early, aggressive intervention with disease-modifying agents. This realization has prompted the investigation and measurement of numerous biologic "markers" in blood, joint fluids, and urine that may serve as indicators of prognosis and the response to therapy. A few of these markers have therefore progressed to routine practice, but many are still in the stage of experimental evaluation. (See 'Introduction' above.)

●The main clinically useful biologic markers in patients with RA include rheumatoid factors (RF), anti-cyclic citrullinated peptide (anti-CCP) antibodies, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP). (See 'Clinically useful markers' above.)

●Biologic markers that remain investigational can be considered in four categories (see 'Investigational markers' above):

•Immunologic (or serologic) abnormalities

•Genetic factors, such as human leukocyte antigen (HLA) class II

•The acute-phase response elicited by hepatocytes as part of the inflammatory process, other than CRP or those assessed indirectly by ESR

•Macromolecules specific for joint-associated tissues that are released into the circulation or excreted in the urine as part of degenerative and reparative processes

●Several immunologic abnormalities, particularly autoantibodies, are associated with rheumatoid arthritis. Such autoantibodies include antiperinuclear factor, antikeratin antibodies, and anti-Sa, which have all been found to be directed against citrullinated peptides. These antibodies, which are now mainly of historical interest since the development of commercial tests for anti-citrullinated peptide/protein antibodies, may have both diagnostic and prognostic significance. (See 'Autoantibodies' above.)

●Other markers of interest include levels of CD154, measures of complement activation, and other autoantibodies, including (see 'Autoantibodies' above and 'Complement activation' above):

•Anti-p68 (BiP) antibodies, which target a protein that serves as a chaperone in the endoplasmic reticulum and binds to immunoglobulin heavy

•Anti-RA33 antibodies, directed against a functional component of the spliceosome and also found in systemic lupus and mixed connective tissue disease

•Antineutrophil cytoplasmic antibody (ANCA), usually atypical (non-myeloperoxidase) p-ANCA

•Agalactosyl immunoglobulins (Gal 0 IgG), which activate complement by means of a mannose binding protein dependent pathway in vitro

•Antibodies to alpha-enolase

•Antibodies to glucose phosphate isomerase

•Antibodies to peptidyl-arginine deiminase 4 (PADI4)

●Genetic factors that may influence the severity of RA include the shared epitope, matrix metalloproteinase genotype, interleukin (IL) 10 promoter genotype, and sulphoxidation status. (See 'Genetic factors' above and "Biologic markers in the assessment of rheumatoid arthritis", section on 'Genetic factors'.)

●Markers of the acute phase response, such as CRP, the ESR, and IL-6, have been evaluated as markers of disease activity in RA. (See 'Acute phase response' above.)

●A number of experimental biochemical markers of joint damage have been described in RA. These molecules may be synthetic or degradative, their presence in body fluids arising as a consequence of metabolism of the tissue of origin. They are predominantly derived from a single tissue such as cartilage, bone, or synovium and can be detected principally by immunoassay of joint fluid, serum, or urine. (See 'Tissue-specific markers' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Ravinder N Maini, BA, MB BChir, FRCP, FMedSci, FRS, who contributed to an earlier version of this topic review.