1. Top of page
  2. Introduction

Since approval for rheumatoid arthritis, anti–tumor necrosis factor α (anti–TNFα) drugs such as infliximab (Remicade), etanercept (Enbrel), and adalimumab (Humira) have quickly become popular agents to combat the disease as monotherapy or in combination with methotrexate. Although these agents are frequently successful in retarding the progression of rheumatoid arthritis, concern has arisen regarding their adverse effects. Most of the attention regarding side effects has centered on reactivation of tuberculosis and other fungal diseases, injection site reactions, and upper-respiratory infections; the incidence of tuberculosis reactivation has largely been curtailed with the use of screening purified protein derivatives and chest radiographs.

Induction of autoantibodies in patients with rheumatoid arthritis treated with these anti–TNFα drugs has been well documented since 2000. In one series of 156 patients treated with infliximab with or without concomitant methotrexate, 22 patients developed a new positive anti–double-stranded DNA (anti-dsDNA) after treatment as detected by Crithidia luciliae assay (1). Anti-dsDNA was confirmed by Farr assay in 11 of these 22 patients (1). Despite this, only 1 patient developed clinical evidence of lupus. A more recent report of 53 patients with rheumatoid arthritis treated with infliximab revealed that the prevalence of patients who were positive for IgG anti-dsDNA antibodies by Crithidia luciliae assay increased from 2% at baseline to 66% at 30 weeks and 45% at 54 weeks; the IgM anti-dsDNA antibodies increased to 85% and 70% at 30 and 54 weeks, respectively (2). In this study, the prevalence of antinuclear antibodies (ANA; >1:100 dilution) by indirect immunofluorescence increased from 24% at baseline to 77% at 30 weeks and 69% at 54 weeks (2). Lower rates of induction of elevated ANA and anti-dsDNA titers have been related to use of etanercept. Interestingly, patients with Crohn's disease treated with infliximab who had positive ANAs prior to infliximab treatment were twice as likely to develop anti-dsDNA antibodies (3).

Despite the frequent induction of ANAs and antidsDNAs with infliximab and etanercept, only a handful of cases of clinical systemic lupus erythematosus (SLE) have been described with either agent (4–6). In early infliximab trials, 3 of 771 patients developed SLE symptoms that resolved within 6 months of cessation of the presumed offending drug (7). Following the report of Shakoor et al of etanercept-induced SLE (7), the Food and Drug Administration (FDA) Center of Biologics Evaluation and Research confirmed 9 other definite cases of SLE and 4 other possible cases of SLE in relation to use of etanercept (8). In the FDA reported cases, 8 patients had discoid rashes, 6 had photosensitivity, and 4 had malar rashes; none had renal or neurologic disease; 5 had anti-dsDNA antibodies and 4 had anti-Smith antibodies (8).

Case Report

We present the first published report of anti–TNFα therapy–related lupus with valvulitis in a patient with rheumatoid arthritis and secondary Sjögren's syndrome treated with infliximab for her erosive arthritis. The patient, a 34-year-old white woman, was referred for assessment of possible subacute bacterial endocarditis. The patient had erosive rheumatoid arthritis and secondary Sjögren's syndrome and was treated with oral methotrexate 10 mg/week, prednisone 10 mg/day, and infliximab infusions every 6 weeks. She had been taking infliximab for 1 year and had last received an infusion 5 weeks prior to our initial evaluation of her. The patient's ANA was positive at ≥1:640 dilution 5 years prior, but she had not exhibited any other clear clinical signs of lupus or antiphospholipid syndrome. To our knowledge, she had never been checked for extractable nuclear antigens, lupus anticoagulants (LACs), or anticardiolipin antibodies (aCL). Her laboratory findings 1 month prior to hospital admission were notable for an erythrocyte sedimentation rate (ESR) of 65 mm/hour, white blood cell (WBC) count of 6.7/mm3 with 82% polymorphonuclear cells and 13% lymphs, hemoglobin (Hgb) 11.1 gm/dl with mean corpuscular volume (MCV) 79 μm3, platelet count 349/mm3, rheumatoid factor (RF) 310 IU/ml (normal <14), anti-SSA >6 (<1.00 index negative), anti-SSB 5.68 (<1.00 index negative), and normal chemistry panel with creatinine 0.8 mg/dl.

The patient had presented to an outside hospital with a 2-week history of fever up to 103°F, dyspnea, cough, and pleuritic chest pain, and was found to have bilateral pleural effusions with cardiomegaly on imaging. A transesophageal echocardiogram (TEE) at this hospital was notable for severe mitral regurgitation, possible vegetations on the anterior leaflet of the mitral valve with a normal ejection fraction, and a small pericardial effusion. Results of repeated serial blood cultures remained negative. The patient was diagnosed with left-sided congestive heart failure and was treated medically. Given an unclear clinical picture and concern for endocarditis, the patient was transferred to our institution.

Upon transfer to our care, the patient was found to be repeatedly febrile to 103°F with a malar rash, marked parotid swelling and decreased salivary pool, synovitis of her metacarpophalangeal joints and proximal interphalangeal joints, and chronic deformities of her metatarsophalangeal joints (Figures 1 and 2). Her laboratory test results were notable for a new and marked anemia (Hgb ∼7.6 gm/dl with MCV 81 μm3). Her hemolytic anemia workup was negative. She had transient mild leukopenia (WBC 4/mm3) and thrombocytopenia (platelet count 133/mm3) as well. Her bone marrow revealed no infection or malignancy but was notable for increased megakaryocyte activity consistent with peripheral destruction or sequestration of platelets. Results of her serial blood cultures were negative.

thumbnail image

Figure 1. Malar rash on patient's face.

Download figure to PowerPoint

thumbnail image

Figure 2. A, Radiograph and B, photograph of the feet showing classic deformities of rheumatoid arthritis.

Download figure to PowerPoint

The patient's serologies were notable for an ANA >1:320 homogeneous pattern with positive dsDNA 1:2560 (subsequently supported by positive Farr assay result 43.8 IU/ml [<5 nl]), low complement levels (C3 61 mg/dl [normal 85–200], C4 2 mg/dl [normal 17–46]), ESR 105 mm/hour, RF 108 IU/ml (normal <20), anti-SSA 105 units/ml (<2.5), anti-SSB 136 units/ml (<2), aCL 8 IgG phospholipid units (0–14), aCL 53 IgM phospholipid units (0–12), LAC negative, antineutrophil cytoplasmic antibody negative, positive antihistone IgG antibody 9.7 units (>2.5 strongly positive), negative anti-RNP 0.710 units/ml (0–2.50), and urinalysis with 1+ protein with 24-hour urine 294 mg protein.

The patient underwent another TEE that revealed mitral regurgitation with a normal ejection fraction, moderate thickening of the anterior mitral leaflets, and no vegetations. The mitral valve thickening was reportedly consistent with valvulitis or Libman-Sacks endocarditis (Figure 3). In comparison, the patient had a documented transthoracic echocardiogram (TTE) as an outpatient 1 year prior to infliximab therapy that had shown only mild mitral regurgitation but no valvulopathy. Her clinical picture at the previous hospital with severe mitral regurgitation and pleural effusions that were consistent with congestive heart failure was deemed clinically consistent with the TEE-documented valvulitis at our hospital. It was likely that with institution of medical therapy at the outside hospital, she had improvement in the severity of her mitral regurgitation and now had a clinical picture of resolving heart failure.

thumbnail image

Figure 3. Transesophageal echo showing mitral valve thickening (arrow) consistent with valvulitis or Libman-Sacks endocarditis.

Download figure to PowerPoint

Given the constellation of malar rash, leukopenia, anemia, thrombocytopenia, hypocomplementemia, positive high titer ANA, positive high titer dsDNA, and presumed valvulitis on TEE, a diagnosis of SLE was made. The infliximab was permanently discontinued and the patient was treated with intravenous methylprednisolone and given 1 dose of intravenous cyclophosphamide in an attempt to control the valvulitis. After the steroid therapy, the patient's clinical condition improved rapidly with a resolution of her fevers, improvement in her malar rash, and improvement in her arthritis.


Valvulitis in association with lupus has occasionally been reported in the medical literature. To our knowledge, our patient represents the first reported case of anti–TNFα-related SLE with concomitant valvulitis.

Valvulopathy is one of the most common ways that SLE affects the heart, and the mitral valve is the most often afflicted (9). Libman-Sacks vegetations, valve thickening, and valve regurgitation are all commonly seen in SLE, with prevalence of valvular disease approaching 60–74% (9). Valvular thickening is the most commonly encountered abnormality seen on TEE (9). The pathogenesis of valvular thickening is thought to be due to immune complex and complement deposition onto the valves and the infiltration of inflammatory cells around and into valve leaflets. Despite the high frequency of valvular abnormalities, however, the presence of clinical symptoms such as congestive heart failure is rare (9). In addition, repeated hemodynamic evaluations in patients with SLE have shown an inordinate rise in systemic and pulmonary capillary wedge pressure in these patients during exercise (10). Most troubling is the associated diastolic dysfunction that even otherwise asymptomatic patients with SLE exhibit (10).

Significant valvular disease is also seen in patients with antiphospholipid syndrome. In the early 1990s, there were many reports of a link between antiphospholipid antibodies and valvular disease, but other studies failed to show a clear association (9). There were some case reports highlighting the efficacy of corticosteroids in remedying mitral valve dysfunction in patients who were antiphospholipid positive (11). Steroids were documented to not only remedy clinical symptoms attributable to heart failure from valvular dysfunction but also to rapidly decrease mitral valve leaflet thickness over time (11).

Optimal treatment of SLE-related valvulopathy remains highly debated in the literature. Some studies have shown no association between the presence of valvular disease and the severity of lupus (12). Therefore, some argue that newly diagnosed valvular disease in a patient with lupus does not require immunosuppression because it is likely that the valvular lesions will regress with time (12).

A seminal 1975 study comparing autopsy results in 36 corticosteroid-treated patients with SLE with necropsy results of SLE patients before the use of steroids revealed that Libman-Sacks–type endocardial lesions had become smaller and were fewer in number, univalvular rather than multivalvular, mainly left sided, and more likely to show healed lesions rather than active ones with the use of steroids (13). Moreover, in this study the character of pericarditis in patients with SLE changed from active fibrinous to healed fibrous, and the presence of myocarditis decreased with the use of steroids (13). The costly tradeoffs noted in patients treated with steroids were higher rates of hypertension, congestive heart failure, purulent pericarditis, and coronary atherosclerosis (13). On the whole, autopsy studies do suggest that valvular lesions have declined with the use of steroid therapy, thereby indicating that immunosuppression may indeed be helpful (9, 10). Most experts concur, however, that more studies are needed to clearly delineate the utility of immunosuppression for the specific issue of valvulopathy.

Given the difficulty in differentiating idiopathic and drug-induced lupus, debate has arisen regarding the reported cases of anti–TNFα-related lupus. Part of the argument centers around the lack of formal criteria for drug-induced lupus. Although many of these patients exposed to anti–TNFα have antihistone antibodies that are classically associated with drug-induced lupus, many also have features atypical for drug-induced lupus, such as anemia, hypocomplementemia, and malar and discoid rash (7, 14). Another confounding point is that antihistone antibodies are also seen in patients with SLE and are by no means specific for drug-induced lupus (14). Of note, however, is that antihistone antibodies in drug-induced lupus are directed against H2A–H2B or H1 and H3–H4 complex but are directed against H1 and H2B histone subunits in idiopathic SLE (14).

The presence of anti-dsDNA antibodies is also unusual in classic drug-induced lupus but is commonly seen in anti–TNFα-induced lupus. Despite this, however, only a small number of patients who develop anti-dsDNA antibodies develop clinical SLE.

Unfortunately, in some of the case reports of anti–TNFα-induced lupus, baseline ANA status was not available prior to use of the biologic agent in question. This prompted some critics to argue that without baseline ANAs one could not rule out preexisting underlying overlap syndrome. Also, in other case reports, patients with classic erosive rheumatoid arthritis who were given anti–TNFα drugs had some features that might have indicated preexisting overlap syndrome (high-titer ANA, low complements, borderline positive anti-dsDNA) (4). When exposed to the anti–TNFα drug, these patients began exhibiting even more clear signs of SLE (rash, lymphopenia and thrombocytopenia, higher-titer anti-dsDNA antibody) (4). Hence it may be argued that the biologic drug may only have unmasked or exacerbated an underlying lupus state and not necessarily created or induced it.

The counter argument that anti–TNFα drugs do indeed induce SLE in patients with rheumatoid arthritis is that some of the case reports in the literature have documented negative ANAs and negative anti-dsDNA antibodies prior to the patient being started on the biologic agent in question (3, 5, 6). In many of these reported cases, in addition to clinical improvement, other serologic abnormalities such as hypocomplementemia, high-titer anti-dsDNA antibodies, and high ESR reversed with cessation of the biologic agent.

Researchers have postulated various mechanisms of how anti–TNFα drugs may induce SLE. Some have proposed that these biologic agents may alter the milieu of apoptotic cell clearance and lead to reduced CD44 expression, which has a role in the clearance of apoptotic neutrophils (5). Others have added that these biologic agents induce apoptotic cell death, which in turn causes the release of nuclear autoantigens that drive the production of anti-dsDNA antibodies in genetically susceptible individuals (6). Biologic exposure may help lower levels of serum amyloid P, which is believed to work with complement factors to bind chromatin DNA in apoptotic bodies and help clear such DNA (2). Another theory is that exposure to biologic agents may somehow cause inhibition of cytotoxic T lymphocytes (CTLs) that suppress autoreactive B cells (2). It is believed that these CTLs suppress B cell reactivity through Fas/Fas ligand interactions and that the various biologic agents down-regulate interferon-γ–mediated reactions that support expression of Fas and Fas ligand (3).

In summary, we cannot definitively say that our patient had a case of anti–TNFα-induced lupus because it can be argued that given a preexisting high-titer ANA, she may have had an underlying overlap syndrome. However, given the temporal relation between commencement of infliximab therapy and onset of SLE, we feel that the infliximab likely contributed to the case and may have unmasked an underlying lupus state. This case also highlights the fact that such patients may have severe valvular and mild renal involvement. We raise this concern especially in light of recent pilot trials utilizing infliximab for treatment of lupus nephritis that seem to argue for the safety of infliximab in patients with lupus. We believe this case highlights long-standing concerns regarding the use of anti–TNFα agents in patients with a preexisting high-titer ANA and any clinical evidence of SLE and teaches us to be mindful of signs of cardiac dysfunction after these agents are started.


  1. Top of page
  2. Introduction