Misdiagnosis of Lupus

Many people who are told they have or might have SLE actually do not have the disorder. Hochberg and associates⁵ noted that only one third of patients who were told they had lupus by a physician actually fulfilled the American College of Rheumatology (ACR) classification criteria for SLE. Of the 149 patients referred to the University of Alabama for the management of and/or consultation for suspected SLE, 60% (90 patients) met the 1982 revised ACR classification criteria, whereas 15% had a diagnosis of clinical SLE but did not meet the ACR criteria.⁶ Another 25% of patients had fibromyalgia-like manifestations, tested positive for ANAs, and very likely did not belong to the SLE spectrum.⁶ Of the 263 patients referred to the University of Florida Autoimmune Disease Clinic between 2001 and 2002 with a working diagnosis of SLE, a 49% agreement rate with the referring physicians’ working diagnosis of SLE was observed.⁷ Of all the referring physicians, rheumatologists were four times more likely to make the correct diagnosis of SLE than non-rheumatologists.⁷ Treatment with steroids was administered to 15% of the patients (as high as 60 mg/day) without any autoimmune disease but with positive ANA testing.⁷ Misdiagnosing lupus leads to unnecessary, toxic, and expensive treatments, the stigmatization of patients, and pointless lifestyle and dietary restrictions, and affects family relationships and reproductive planning.

Positive Antinuclear Antibody Testing: How Often Is it Systemic Lupus Erythematosus?

To rule out SLE, rheumatologists all too often are referred patients who feel good or have vague symptoms but have a positive ANA test. The ANA test is sensitive for the diagnosis of SLE (95% to 100% patients with SLE have ANAs), but this test should be ordered only if the pretest possibility of this diagnosis is high. A positive ANA test is often found in patients with other disorders and in seemingly healthy patients. The prevalence of positive ANAs depends on the patient’s sex and age; older persons, particularly women older than 65 years of age, more commonly have ANAs. With the use of the HEp-2 substrate, approximately 20% of healthy individuals have an ANA titer of 1:40 or higher and 5% have an ANA titer of 1:160 or higher.⁸ Relatives of 25% to 30% of patients with connective tissue disease have titers equal to or higher than 1:40.² A positive ANA test has an 11% positive predictive value for SLE.² Of note, solid-phase methods, which are being increasingly used for assessing ANAs, have increased false-negative results for SLE; the ACR task force for ANA supports immunofluorescence testing as the “gold standard” method.⁹ Of the 471 patients with SLE, ANA levels were the first ACR criterion accrued among 20% of patients.¹⁰ In a study using donated sera from 130 soldiers who were being inducted into the armed forces and who ultimately developed SLE,¹¹ 115 had at least one SLE autoantibody (78% had ANAs) a mean 3.3 years before diagnosis.

Antinuclear Antibody–Negative Lupus

A positive ANA test is only 1 of 11 criteria that are used to define SLE according to the ACR classification. Of the 11 criteria, 4 must be present to make a diagnosis, but ANA positivity is so central to the current concepts of SLE that many rheumatologists find it inconceivable for SLE to be present without it. Several reports have documented the delayed appearance of ANAs in patients suspected of having SLE. In the view of the authors of this text,^12,13 documenting a mean of 3 to 4 years between onset of symptoms and time of diagnosis is not surprising. Case series of patients with lupus nephritis in whom negative ANAs persisted for years before becoming positive are known.^14,15 Persillin and Takeuchi¹⁶ found ANAs in the urine and pleural fluid of a patient with diffuse proliferative nephritis and nephrotic syndrome for some time before serum ANAs were present. Low antibody concentrations in the serum secondary to a loss in body fluids can be present as noted by Ferreiro and associates.¹⁷

The concept of ANA-negative SLE was introduced in 1976¹⁸ and has been said to be typical of patients with photosensitivity and antibodies to Sjögren syndrome antigen A (anti-SSA/Ro).¹⁹ With the use of modern microscopes and HEp-2 cells as a source of nuclear antigens, the existence of ANA-negative SLE has been questioned.²⁰ Reichlin²¹ has stated that with a Ketjen Black (KB) or Hep-2 substrate, 98% of all patients with SLE are ANA positive, because non–DNA-containing antigens such as SSA/Ro are better represented when these cell lines are studied. Unfortunately, human cell lines are less specific, although they are more sensitive.

Technical inaccuracy, specimen collection and storage issues, variations in microscope quality, ANAs hidden within circulating immune complexes (CICs), in vivo binding of ANAs by tissues, types of fixation on substrate slides, antiimmunoglobulin-conjugate characteristics (e.g., isotype specificity, fluorescein isothiocyanate–to–protein ratio, antibody-to-protein ratio, specific antibody content and working dilution), and problems with reference sera are causes of negative ANAs in patients with SLE.⁸ Wide variations in the reproducibility of ANA tests and difficulties in standardization are also problems that remain unresolved.⁸ It needs to be stressed, however, in view of increasing use of quantitative, automated high-volume solid-phase assays, that if the clinician strongly suspects SLE and ANAs are negative, then checking with the laboratory concerning the assay used to assess ANAs is important. The use of an enzyme-linked immunosorbent assay (ELISA) or coated beads increases false-negative rates as do solid-phase substrates that use a limited number of autoantigens.⁹ In a study comparing the frequency of ANAs using immunofluorescence and ELISA (BioPlex) among 192 patients with SLE, the latter fared worse than immunofluorescence (75.5% versus 81.3%).²² BioPlex sensitivity and specificity for SLE were 78.9% and 38.9%, respectively.²² Therefore confirming with the laboratory which assay was used to measure the ANAs is important; if the ELISA was used and the physician’s suspicion for SLE is strong, then a repeat of the ANA test with immunofluorescence assay is recommended.

Undifferentiated Connective-Tissue Disease

The term was first coined by LeRoy in 1980 and is used to refer to a group of systemic autoimmune diseases with signs and symptoms that are not sufficiently evolved to fulfill the accepted classification criteria for the defined connective tissue diseases.⁴ In a series of papers written by Williams and colleagues,²³ 213 patients with early UCTD were followed for over 10 years. At 10 years they reported that the presence of malar rash, serositis, or discoid lupus in patients with UCTD suggested an eventual diagnosis of SLE. In another follow-up study of 83 patients with UCTD, 18 patients (22%) developed SLE at a mean period of 54 months,⁴ and the presence of anticardiolipin (aCL) antibody was associated with the development of SLE. The authors concluded that the rate of evolution to a connective tissue disease is greatest in the first years of follow-up and then decreases over time. Bodolay and others²⁴ followed 746 patients with early UCTD and noted the development of SLE in 4.2% and a resolution of symptoms among 12% of the patients. The highest probability of development of a defined connective tissue disease was noted in the first 2 years; 12% of the patients underwent remission and 64.5% remained in the UCTD category. Most of the UCTD studies note that the disease is mild, and most patients remain in the UCTD category; those cases that evolve further into a defined connective tissue disease category do so early on.^4,24,25

Incomplete Lupus

Incomplete lupus is a misleading term that has appeared in the literature to denote patients who are thought to have SLE but do not fulfill four ACR criteria. Hence, this term could be inappropriately applied to UCTD or to patients with clinical SLE. These individuals range from having biopsy-documented nephritis to idiopathic thrombocytopenic purpura with positive ANAs to fibromyalgia. In a study of 28 Swedish patients with incomplete SLE, 57% developed complete SLE after a median period of 5.3 years.²⁶ As in the patients with UCTD, the presence of aCL antibodies was found to be a predictor for SLE. Swaak and colleagues²⁷ noted skin, musculoskeletal, and leukopenia-related disease activity in 122 patients of European descent with incomplete lupus; 27 patients met the full criteria shortly after study entry, and 3 patients met the criteria over the following 3 years. Patients with incomplete SLE have been found to have less frequent disease flares²⁸ and a good prognosis.²⁷ Too many patients who carry this label undergo unnecessary treatments and become stigmatized and medicated; hence, this term should not be used in patients meeting the criteria for a UCTD diagnosis.

Rheumatoid Arthritis

Juvenile Idiopathic Arthritis

Like healthy adults, healthy children may have ANA positivity. A 2004 study on ANA prevalence among 214 healthy children from Brazil between the ages of 6 months and 20 years (mean age 8.7 years) demonstrated that ANA was observed in 13% of children using the immunofluorescence technique. Although no differences in ANAs were noted by gender, an association of higher frequency of ANA titers (≥1:80) was observed among children between 5 and 10 years of age. Of the 27 healthy children with positive ANAs, 8 were reevaluated 36 months later, and none of them developed any rheumatic diseases, although the sera remained positive in 2 of them. ANAs were present in 42 of 116 patients (36.2%), and the authors concluded that ANA determination should be required only in individuals with clinical signs and symptoms suggestive of autoimmune disease.⁴⁹

Children with JIA also have positive ANAs. Several large-scale studies have observed ANAs in approximately 60% of patients with JIA, particularly oligoarticular disease in young girls with uveitis.^50,51 The following rates of ANA positivity and titers were noted among 153 children of Korean descent with JIA⁵²: 1:40 dilution in 33% of patients, >1:40 in 70%, >1:80 in 2%, >1:160 in 16%, >1:320 in 2%, and >1:640 in 10%. ANA seropositivity was associated with female sex, negative HLA-B27, and a persistently elevated erythrocyte sedimentation rate (ESR) at follow-up.⁵² Recently, Ravelli and colleagues⁵³ proposed that ANA-positive patients with JIA by current International League of Associations for Rheumatology (ILAR) criteria constituted a distinct yet homogeneous subgroup of patients characterized by younger age at disease onset, female preponderance, asymmetric arthritis, iridocyclitis, relatively fewer joints affected over time, and a lack of hip involvement, as compared with patients who were ANA-negative with JIA.⁵³ Another study that sought to describe the patterns and time course of arthritis in patients with ANA-positive JIA in 195 patients found that 72% (including most of those who later developed polyarthritis) had monoarthritis at disease onset.⁵⁴ Similar results were noted by Guillaume and associates in 2000.⁵⁵ Among patients with oligoarticular onset, polyarticular extension occurred in approximately 50% of patients within the first 3 to 4 years after disease onset and tended to be less likely thereafter.⁵⁴

Vasculitis

Although polyarteritis nodosa (PAN) is relatively rare, it can be mistaken for SLE. In contrast to patients with SLE, those with PAN are usually men and include all age groups equally. A patient with SLE may develop PAN-like vasculitis of renal arteries.^56,57 Cutaneous vasculitis may be more prominent, as may eosinophilia, wheezing, and nerve and bowel symptoms. The ANAs are often negative. Hypersensitivity angiitis may mimic SLE at first but, ultimately, can be distinguished by a self-limited course, an absence of ANAs, and rarity of severe visceral involvement. Ordering antineutrophilic cytoplasmic antibody testing can often differentiate lupus from microscopic polyangiitis and Wegener granulomatosis.⁵⁸ One case of SLE in a child with Kawasaki disease has been reported.⁵⁹

Behçet disease (BD) is a condition complicated by recurrent painful oral and genital ulcers, eye manifestations of uveitis, and complications and skin manifestations ranging from a positive pathergy test (rarer in the Western world and more common in Japan and around the Mediterranean) and erythema nodosum to other papulopustular lesions.⁶⁰ Some cases also are complicated by the occurrence of vasculitis of small and medium vessels, and venous thrombosis occasionally occurs, more often in men than in women. Arthropathy (seen in 16% to 84% of patients with SLE) is generally mild, mostly with arthralgias; when arthritis (often nonerosive and monoarticular) are the most frequent clinical features of the disease.⁶¹ ANAs are frequently negative. Considering that SLE is a female-predominant disease with painless oral ulcers and positive ANAs in almost 99% of patients, the distinction is relatively easy to make; however, in the rare event that a confusion exists, HLA-B51 testing might help.⁶² Among 4800 patients with BD and 16,289 controls from 78 independent studies,⁶² the pooled odds ratio (OR) of HLA–B51/B5 allele carriers to develop BD compared with noncarriers was 5.78 (95% confidence interval [CI] of 5.00 to 6.67). The lack of any diagnostic test for BD further complicates the picture, and one case of co-existent disease has been reported.⁶³

SLE is a disease of the small arteries and medium-sized arterioles, but it can rarely affect larger-caliber vessels. Large-vessel vasculitis is not associated with autoantibody formation. Older people more commonly develop polymyalgia rheumatica and giant-cell arteritis, however, and SLE is occasionally included in the differential diagnosis because musculoskeletal symptoms are present and age-related positive ANAs may be found.⁶⁴ A true concurrence of giant-cell arteritis and SLE has been reported twice.^65,66 Takayasu pulseless arteritis is found in young women who are mostly Japanese but also in other Asian and Hispanic women. Several people have reported case reports and case series of patients with either SLE or Takayasu arteritis that later developed the other disease^67–73; however, co-existent SLE and Takayasu arteritis is rare.

Polymyositis and Dermatomyositis

In contrast to SLE, patients with polymyositis and dermatomyositis are less often women⁷⁴ and rarely have an autoimmune family history. In addition, dermatomyositis and polymyositis are easier to distinguish because they usually exhibit overt muscle weakness, mostly in middle-aged women, and evident elevations of muscle enzymes, with or without skin changes such as Gottron papules, heliotrope lid, shawl sign, periungual erythema, and cuticular thickening. Different co-existing malignancies may also occur; serositis is rare, and nephritis, liver inflammation, and hematologic abnormalities are absent. ANAs overall are observed in approximately 30% of patients with inflammatory myopathies.

A low-grade myositis with muscle enzyme levels two to three times normal may be seen in SLE that responds to low doses of corticosteroids; however, it may be as severe as inflammatory myositis.⁷⁴ Amyopathic dermatomyositis may be present with symptoms that closely resemble those of SLE, especially if the classical dermatomyositis rash is not present. Consider a female patient with fever, alopecia, oral ulcers, myalgias, arthritis, and a few macular erythematous plaques on her leg, and laboratory testing has confirmed lymphopenia and anemia but without overt muscle weakness. She would fulfill the ACR criteria for SLE; however, if an ANA test is negative, then diagnostic considerations would point toward another connective tissue disorder. In a review article by Gerami and associates⁷⁵ of 291 patients with adult-onset clinically amyopathic dermatomyositis, 63% of the patients had positive ANAs.

Scleroderma and Other Fibrosing Syndromes

Although ANAs are present in most patients with scleroderma, other serologies associated with SLE are observed in a small minority of those with scleroderma. These include antiphospholipid antibodies^76,77 and other nuclear antigens.⁷⁸ Anti-topoisomerase antibodies are usually associated with scleroderma but can be found in up to 7.7% of patients with SLE; however, the titers are higher in scleroderma.⁷⁹ Scleroderma and SLE fit within the same spectrum of interferon-mediated disease, and a subset of patients with scleroderma, as well as anti-topoisomerase and anti–U1 ribonucleoprotein (anti–U1-RNP) antibodies, show a lupus-like high interferon-inducible–gene expression pattern.⁸⁰

In contrast to SLE, familial occurrence of scleroderma is rare. Clinically, sclerodactyly, telangiectasias, calcinosis, and malignant hypertension with acute renal failure are almost unheard of in patients with SLE. Differentiating SLE, MCTD, and scleroderma is important, because the last rarely is responsive to steroids or cytotoxic agents.

Case reports have appeared of autoimmune hemolytic anemia,^81,82 high levels of anti–native DNA (anti-nDNA), lupus nephritis,⁸³ and discoid lupus^84–86 occurring in patients with scleroderma. Patients with anti-scleroderma (anti-Scl) antibodies probably have lupus rather than scleroderma if anti–double stranded DNA (anti-dsDNA) is present.⁸⁷ Scleroderma may evolve into SLE and vice versa⁸⁸; morphea⁸⁹ and linear scleroderma can be seen with SLE.⁹⁰ Neonatal SLE with morphea⁹¹ and eosinophilic fasciitis with SLE have also been reported. Although digital ulcers are rarely seen in SLE, they may be the initial manifestation.⁹² Similarly, patients with SLE may also develop calcinosis cutis,⁹³ although this condition is usually seen in scleroderma and myositis.

Retroperitoneal fibrosis in SLE has been noted,^94,95 although rarely. Nephrogenic fibrosing dermopathy is usually seen in patients with impaired renal function, and its occurrence has been reported in SLE.⁹⁶

Serum Sickness

Serum sickness was initially described as a clinical syndrome characterized by fever, lymphadenopathy, cutaneous eruptions, and arthralgias often in association with proteinuria, but without other evidence of glomerulonephritis. Hence, one can see how the features of serum sickness can mimic SLE. Serum sickness has a self-limiting course unlike SLE, and the skin findings are most commonly described as a morbilliform eruption that tends to begin on the trunk as a patchy erythema before spreading to involve the extremities. ANAs are usually negative, as evidenced by the paper by Lawley and colleagues,⁹⁷ who studied the immunologic features of serum sickness in 12 patients and found none of the 9 patients who were tested for ANAs to be positive. C3 and C4 levels were both low in all patients, with C4 being significantly reduced, compared with C3.⁹⁷ Furthermore, glomerulonephritis is less likely, although patients with severe glomerular and tubular damage as a result of serum sickness have been reported.⁹⁸

Kikuchi Disease

Kikuchi-Fujimoto disease (KFD) is a benign and usually self-limited form of histiocytic-necrotizing lymphadenitis first described by Kikuchi and others and usually affecting young adult women. Clinical features at presentation can include lymphadenopathy involving one or more predominantly posterior cervical lymph nodes and may be associated with fever, myalgias, arthralgias, weight loss, diarrhea, chills, sweating, and/or, less commonly, hepatosplenomegaly.^99,100 Associations with SLE have been reported,¹⁰⁰ and common features are fever, myalgias, and leukopenias. Reports of KFD evolving into SLE have also been reported.^101,102 The diagnosis of KFD is usually based on lymph node histologic findings; therefore obtaining a lymph node biopsy when lymphadenopathy is the dominant feature in a young woman is important. The clinical course of this disease is usually favorable, with spontaneous remission in less than 4 months in almost all patients.

Fibromyalgia

Patients with SLE not in frequently experience chronic fatigue,^103,104 and many symptoms are indistinguishable from fibromyalgia. The prevalence of secondary fibromyalgia in SLE is common and observed in approximately 22% of patients with SLE.^105–107 A study on the prevalence of fibromyalgia in three SLE groups—African Americans, Caucasians, and Hispanics—found a negative correlation with African American ethnicity.¹⁰⁸ Tender points, altered sleep patterns, and a lack of restorative sleep with the absence of objective parameters of active disease—normal C3, C4, anti-dsDNA, and ESR—makes it difficult to differentiate secondary from primary fibromyalgia. Some of these symptoms may be a result of rapid tapering of steroids or emotional and/or physical stress.

Patients with primary fibromyalgia are commonly referred to the rheumatologist because of weakly positive ANAs. Patients may research these finding over the Internet and convince themselves that they have early SLE. Frequently when ANAs are positive, primary care physicians may also find making the correct diagnosis a challenge. ANAs can be seen in fibromyalgia in low titers; however, their presence is not a predictor of future development of connective tissue disease.¹⁰⁹

Wolfe and colleagues¹¹⁰ compared the various co-morbid conditions in RA, SLE, fibromyalgia, and noninflammatory rheumatic diseases (NIRDs) and found depression in 34% of patients with fibromyalgia and SLE. Self-reported diagnosis was a limitation to this study. The LUMINA study group¹¹¹ studied the effect of body mass index (BMI) on fibromyalgia in patients with SLE. Obesity and SLE have a component of inflammation with increased levels of tumor necrosis factor–alpha (TNF-α), interleukin (IL)-1 and C-reactive protein (CRP) in both conditions.^112,113 The association of increased BMI and fibromyalgia has been described¹¹⁴ and was confirmed in this study, but obesity and fibromyalgia had no effect on organ damage accrual in SLE.¹¹¹ The association with fibromyalgia, however, results in increased symptoms with poor response to treatment.¹¹¹ Factors associated with the quality of life and fibromyalgia seem to have a greater influence on the severity of reported fatigue than the level of SLE disease activity.¹¹⁵ Increased number of fibromyalgia tender points is also likely to correlate with poor health status.¹¹⁶ These complaints may be erroneously interpreted as active SLE and treated inappropriately with increased steroids and immunosuppressive agents.

Sleep studies,¹¹⁷ performed on patients with SLE who had fatigue and daytime sleepiness showed a pattern similar to that observed in fibromyalgia. Patients with disabling tiredness were found to have a poor quality of sleep. Some of these patients were depressed, but daytime sleepiness in SLE was not necessarily associated with depression. Patients with SLE with the worst depressive symptoms complained of tiredness but were not objectively sleepy.

Fatigue in SLE is multidimensional, and a depressed mood contributes to physical and mental tiredness. Physical fatigue has also been associated with poor aerobic capacity in SLE.¹¹⁸ The treatment needs to be individualized in these patients. Optimizing pain control and managing SLE flares may take care of the physical fatigue, but the mental fatigue may require focusing on alleviating the depressed mood.¹¹⁸

Crystal-Induced Arthropathies

Although 29% of patients with SLE are hyperuricemic (usually secondary to nephritis, diuretics, or chemotherapy), clinical gout is rare, which could be the result of a predominance of menstruating women among those with active SLE.

Until 2000, fewer than 20 cases had been described in the literature^119–121; most were men taking diuretic medications. Recently, however, two groups examined the clinical features of 15 and 10 patients with gout and SLE, respectively. Over 90% had nephritis, many had received transplants, some were taking diuretics and cyclosporine, and the lupus was almost always inactive.^122,123 Wallace and colleagues¹²⁴ reviewed the negative association between gout and RA. Only 3 of the 464 patients with idiopathic SLE had clinical gout, including a 25-year-old woman with nephritis who had tophaceous deposits. One report reviewed three young women with SLE and tophaceous deposits; all were underexcretors of uric acid.¹²⁰ It has been proposed that patients with SLE (who often have decreased synovial fluid complement levels) have a natural barrier to gout, because urates require the presence of near-normal synovial fluid complement levels to induce inflammation.¹²⁵ The rarity of pseudogout in patients with SLE has been reviewed by Rodriquez and colleagues.¹²⁶

Dermatitis Herpetiformis

The association of dermatitis herpetiformis (DH) and SLE is rare but reported.¹²⁷ Thomas and Su¹²⁸ found nine patients with concomitant DH and SLE who were followed at the Mayo Clinic from 1950 to 1981 and reviewed the literature. Penneys and Wiley¹²⁹ reported four patients with SLE with lesions histologically resembling DH, but immunofluorescence testing, was typical of SLE in three of the four patients. Five other reports have appeared, the most important of which are those of Aronson and associates¹³⁰ and Davies and colleagues.¹³¹

Sarcoidosis

SLE and sarcoidosis share many immunologic features. Both manifest hyperglobulinemia, decreased skin test and lymphocyte responsiveness, lymphopenia, impaired antibody-dependent cellular cytotoxicity, and increased levels of CICs. Cryoglobulins and antilymphocyte antibodies may be present in both disorders, and up to 29% of patients with sarcoidosis may have positive ANAs.¹³² In a retrospective study of 34 patients with sarcoidosis, 10 were ANA positive, and of these, two had dsDNA in addition.¹³² SLE did not develop in these patients in the 10- to 15-year follow-up period. Differential diagnosis can be a problem,¹³³ but despite these similarities, co-existence is infrequent.^134–137

Amyloid

Patients with SLE would be expected to have an increased incidence of amyloid, as do those with RA or ankylosing spondylitis. Cathcart and Scheinberg¹³⁸ enumerated many reasons why SLE and amyloid should co-exist. For example, the two have a common pathogenetic pathway, and polyclonal B-cell proliferation is observed in both. Benson and Cohen¹³⁹ found serum levels of amyloid protein A to be elevated in 25 patients of active SLE, although these were one half the levels observed in an RA group. This alpha-globulin is a precursor of the major protein constituent of secondary amyloid fibrils. The serum amyloid P component can also be deposited in lupus tissues without evidence of clinical amyloid¹⁴⁰ and may be protective against lupus.¹⁴¹ Despite this, fewer than 20 patients with co-existing SLE and amyloidosis have been reported.^142–144 Primary cutaneous nodular amyloidosis has been reported to occur in a patient with SLE.¹⁴⁵