The musculoskeletal system in SLE

Nonerosive arthritis

The most common pattern of lupus arthritis has historically been described as a nonerosive, nondeforming inflammatory arthritis. Insidious in onset, it is an early disease manifestation causing inflammatory joint symptoms such as erythema, tenderness, and/or effusion. Usually symmetric with a predilection for hands, wrists, and knees, it can be confused with early rheumatoid arthritis (RA). Therefore it is vital during the diagnostic evaluation of patients with inflammatory arthritis to include lupus serologies.

Patients with lupus arthritis do not develop voluminous synovial effusions. When obtained, the fluid is clear to mildly opaque, and viscosity may be normal or slightly reduced. The white cell count of synovial fluid is low, rarely exceeding 10,000 cells/ml. Lymphocytes and monocytes predominate, although neutrophils may be abundant in fluids with marked leukocytosis. Serologic tests on synovial fluid are of no diagnostic value, but it is imperative to culture the fluid if circumstances warrant. Radiographs may demonstrate soft tissue swelling, but erosive changes and joint space narrowing are typically absent.

The concept of nonerosive arthritis in lupus has been challenged recently with the advent of sensitive imaging techniques, such as high-resolution US with power Doppler and MRI. The development of consensus-driven definitions of musculoskeletal pathology by the OMERACT 7 US special interest group allows reproducibility and comparison between study results that was not before possible. Recently it has been shown that US has a high specificity (98.7%) but only fair sensitivity (35.9%) in detecting erosions in lupus patients compared to CT. The sensitivity of US in this study was found to be better at MCP joints (57.1%), compared to wrist.

In a recent meta-analysis of US studies of lupus patients performed by Lins and Santiago, a total of 1091 joints, representing 610 SLE patients, were studied sonographically. Hands and wrists represented approximately 80% of the studies. Effusions were observed in 55% of joints, synovitis in 20%, and synovial hypertrophy in 14%. Joint synovitis on US was correlated with patient’s symptoms or tender/swollen joints on physical exam. Notably, erosions were present in a small but measurable subset, with a range of 26% to 47%, despite removing patients with serologic markers typical of RA from the analysis. Importantly, the correlation of sonographic features of lupus with various standard disease activity indices has been inconsistent between studies.

Yet unanswered is whether subclinical synovitis is predictive of future arthropathy, can be used as a biomarker of flare, and its prognostic implications of long term outcome measurements. Yoon et al. utilized gray-scale US and power Doppler to identify subclinical inflammatory changes in the joints of 48 SLE patients without clinically overt musculoskeletal involvement. Synovitis, found in 58.3% of patients, was noted at the wrist (33.3%), second MCP joint (29.2%), and the third MCP joint (31.3%). New musculoskeletal symptoms subsequently developed in 22.9% of patients, and among predictive risk factors was a higher baseline US synovitis index. The degree, or grade, of synovitis in many US studies is not standardized and/or reported and therefore must be interpreted with caution as low-grade synovitis can be found in noninflammatory arthritis.

The prevalence of erosions identified by US in historically nonerosive disease sets is supported by data from MRI from several studies. Chiara et al. reported MRI findings in 50 SLE patients with arthritis who underwent hand and wrist MRI’s without contrast. Bone marrow edema was observed in two patients in the hand (4%) and in 15 in the wrist (13%). Erosions were observed in the hands in 24 patients (48%) and in the wrists in 41 (82%) patients. They concluded that involvement of the wrist in SLE is similar in frequency to RA, but involvement of the hand in SLE is significantly less frequent compared to RA. Similarly, in their study of 34 patients with SLE arthritis, Ball et al. demonstrated the presence of erosions at the wrist in 93% of subjects and at the MCP’s in 61% using contrast-enhanced MRI. In total, 93% of patients had at least grade 1 synovitis at one or more MCP joints, and wrist synovitis was present in all.

In contrast to RA, synovium from patients with lupus does not demonstrate exuberant inflammatory changes. Natour et al. examined 30 percutaneous synovial knee biopsies from lupus patients and noted: (1) synoviocyte hyperplasia, (2) scarce inflammatory infiltrates, (3) vascular proliferation, (4) edema and congestion, (5) fibrinoid necrosis and intimal fibrous hyperplasia of blood vessels, and (6) fibrin on the synovial surface. Insight into the mechanism by which lupus patients with arthritis are generally spared from significant erosions was provided by Mensah et al. They postulated that an interferon-rich milieu, as is seen in most lupus patients, supports the differentiation of myelomonocytic precursors to myeloid dendritic cells as opposed to osteoclasts.

Erosive arthritis (rhupus)

An uncommon subset of lupus arthritis, referred to as “rhupus”, is marked by articular features of RA, namely erosive arthritis, and serologic features of lupus. Findings may include synovitis, synovial proliferation, rheumatoid nodules, and malalignment. Predictors for the development of erosive arthritis are poorly understood. Amezcua-Guerra et al. demonstrated a median C-reactive protein (CRP) concentration in erosive arthritis of 14.5 mg/L compared to 0.8 in nonerosive arthritis. AntiCCP antibodies were also significantly associated with erosive arthritis, but serum IL-6, interferon-gamma, IL-4 and IL-10, while numerically greater in erosive arthritis, were not statistically different.

Similar observations were made by Chan et al. who compared the presence of antiCCP antibodies in 104 SLE patients with either erosive or nonerosive/no arthritis. While only 6 patients had major erosive arthritis, 4 (67%) of those patients did had antiCCP2 antibodies; only 4 of the other 98 patients were CCP2 antibody positive. Kakumanu et al. compared the frequency of antiCCP antibodies (>1.7 units) in patients with RA (68%) to those with SLE (17%). Their presence was over two-fold higher in SLE patients with erosive arthritis (38%). AntiCCP antibodies were less commonly observed (8.8%) in the SLE cohort described by Ball et al.

Jaccoud’s arthropathy

Francois-Sigismond Jaccoud, a Swiss physician living in Paris in the 1800s, originally described an arthritis associated with rheumatic fever. Characterized by reducible deformities on exam, there is a notable absence of erosions on plain radiographs in patients with JA. Bleifeld and Inglis published their observations of the hands of 50 lupus patients. Abnormalities included laxity (50%) and reducible deformities (38%), but radiographs failed to demonstrate erosive changes. Prevalence figures for JA from two SLE cohorts were 3.47% and 6.1%. The most frequent joint deformities include swan neck, thumb subluxation (z-thumb), ulnar deviation, hallux valgus, and boutonniere deformity. Spronk et al developed a JA index that has been used to standardize classification of patients. Clinical features of lupus in JA do not differ from patients with other types of lupus arthropathy. In fact, SLE disease activity scores are comparable. Interestingly, Spronk et al. found that CRP, which classically does not increase in lupus disease flares, was significantly higher in JA compared with nonJA patients.

JA could be mistaken for RA as these patients share clinical features, such as ulnar deviation and subluxation at the MCP’s. Features that distinguish JA from RA include the reducibility of the deformities and the lack of erosions on x-ray although more recent studies have demonstrated a small rate of erosions utilizing US, MRI, and CT. Lins et al. demonstrated at least 1 US finding at the hand and wrist in 50% of their 40 patients. In this cohort of SLE patient with only JA, synovial hypertrophy was found in 48%, tenosynovitis in 23%, and bony erosions in 5%. More current studies have suggested erosion rates of greater than 50% when including the fifth MCP joint, an area not typically studied in earlier reports or using MRI. Erosions viewed with these advanced imaging studies in JA have a characteristic hook-shaped appearance.

Severe JA may compromise hand function as a result of contractures or ulnar deviation. Severe contractures may lead to maceration of the palmar skin. The pathogenesis of joint laxity in JA is not understood. Though rarely obtained, biopsy specimens have shown normal synovium. Whereas JA is generally observed in the hands, it has been reported in feet and knees.

Tendons and entheses

The presence of tendon involvement particularly in asymptomatic patients has been increasingly recognized through US. Tendon involvement largely consists of tenosynovitis, though tendon dislocation, tendon tear, tendonitis, enthesitis, and tendon thinning also occur. The presence of tenosynovitis ranges between 20% and 65% in US studies and is often found in patients without active joint symptoms. Recent US studies therefore suggest that utilizing patients’ symptoms or the physical exam alone may underrecognize musculoskeletal pathology. At least one author has proposed rescoring a C on MS-BILAG, but with positive US power Doppler signal on tendon or joints, as a B and considered having intermediate active musculoskeletal disease.

Spontaneous tendon rupture, albeit rare, occurs in patients with lupus. Ruptures may be acute and are sometime bilateral with the most commonly affected tendons being the patellar and Achilles tendons. Predisposing factors include trauma and steroids. Histology, reported in a limited number of cases, has shown variable degrees of inflammatory changes, ranging from no inflammation to exuberant synovial proliferation.

Treatment

Arthritis activity dictates how aggressively the clinician will need to intervene. Nonsteroidal antiinflammatory drugs and low dose corticosteroid (e.g.., prednisone 5–10 mg daily) are often effective for mild symptoms. For those failing minor interventions, hydroxychloroquine or immunosuppressives may be warranted. Wong and Esdaile noted inconsistent effects in their review of methotrexate in controlled and uncontrolled studies. Fortin et al. demonstrated the steroid-sparing effects of methotrexate in a randomized, double-blind placebo-controlled study of 86 SLE patients. While the study did not focus solely on articular manifestations, 90% of participants had musculoskeletal involvement at baseline. Azathioprine represents yet another treatment option.

A trial comparing the efficacy of mycophenolate to cyclophosphamide in lupus nephritis yielded data on extra-renal responses. Of those with baseline BILAG A or B musculoskeletal domain scores, over 85% of the patients in both arms had improvements in their domain scores at 24 weeks. However, despite the greater use of mycophenolate for both renal and extra-renal disease, lupus experts polled in a 2015 survey recommended sequential use of hydroxychloroquine, methotrexate, and rituximab for the treatment of lupus arthritis.

In post hoc analyses of the phase 3 belimumab data, Manzi et al. published that those subjects who entered BLISS-52 and BLISS-76 studies with musculoskeletal domain activity had statistically significant improvement when belimumab, as opposed to placebo, was added to standard of care. While most rheumatologists shy away from off-label use of tumor necrosis factor (TNF) inhibitors in lupus because of fear of disease exacerbations, anecdotal experience suggests that some patients with lupus arthritis benefit. It must be remembered that TNF inhibitors can promote the synthesis of antiDNA antibodies, thus confusing the picture for those clinicians whose assessment of disease activity is guided by serologies.

Favorable effects on musculoskeletal disease activity, albeit not statistically significant, have been demonstrated with tocilizumab and abatacept. Laquinimod, in development for multiple sclerosis, was studied in lupus arthritis; however, study results were never presented. Rituximab, while not approved for SLE, is favored by some for refractory arthritis.

Splints will correct the maligned digits in a patient with JA. However, it is the authors’ experience that deformities recur once splints are removed. Similarly, pharmacologic interventions, while capable of reducing inflammation in the more typical subset of lupus arthritis, do not affect deformities associated with JA. On occasion, surgical intervention of the hand is required. Reported results have been variable.

The traditional teaching that lupus arthritis is not associated with erosions has been contested in recent years with the introduction of sensitive imaging modalities. These findings further confound the classification of joint disease in SLE. Ball and Bell emphasized that lupus arthritis remains largely understudied in comparison to RA. It is this lack of understanding of SLE arthritis pathogenesis and classification that may thwart therapeutic advances for this disease.

Fibromyalgia

The prevalence of fibromyalgia in lupus is 6.2%. In their recent study, Torrente-Segarra and coworkers showed that disease duration of more than 5 years was associated with slightly higher prevalence rates. Furthermore, patients with SLE-FMS had more depression, secondary Sjogren’s, photosensitivity, and oral ulcers compared to their nonFMS matched controls.