Diagnostic Accuracy for Lupus and Other Systemic Autoimmune Diseases in the Community Setting

ORIGINAL INVESTIGATION Diagnostic Accuracy for Lupus and Other Systemic Autoimmune Diseases in the Community Setting

Sonali Narain, MD, MPH; Hanno B. Richards, MD; Minoru Satoh, MD, PhD; Marlene Sarmiento, BSN; Richard Davidson, MD, MPH; Jonathan Shuster, PhD; Eric Sobel, MD, PhD; Paulette Hahn, MD; Westley H. Reeves, MD

Background: Most individuals with autoimmune and were seropositive for antinuclear antibodies but did not other immune disorders undergo initial evaluation in the have autoimmune disease. The degree of agreement for community setting. Since misdiagnosis of systemic au- referring rheumatologists (␬=0.55) was better than that toimmune diseases can have serious consequences, we for nonrheumatologists (␬=0.32). Stepwise logistic re- evaluated community physicians’ accuracy in diagnos- gression indicated that rheumatologists were 4 times more ing autoimmune diseases and the consequences of mis- likely to make an accurate diagnosis of SLE than were diagnosis. nonrheumatologists (PϽ.003). Thirty-nine patients who were seropositive for antinuclear antibodies but had no Methods: We studied the patients referred to our Au- autoimmune disease had been treated with corticoste- toimmune Disease Center for 13 months (n=476). We roids at dosages as high as 60 mg/d. estimated the degree of agreement with the final diagnosis (␬ statistic) and the accuracy indexes (sensitivity, Conclusions: Many patients with a positive anti- specificity, and predictive values) of the referring phy- nuclear antibody test are incorrectly given a diagnosis of sicians’ diagnoses. SLE and sometimes treated with toxic medications. The data support the importance of continuing medical edu- Results: We found a 49% agreement between the refer- cation for community physicians in screening for auto- ring and final diagnoses (␬=0.36). Of 263 patients re- immune diseases and identifying patients who may ben- ferred with a presumptive diagnosis of systemic lupus ery- efit from early referral to a specialist. thematosus (SLE), 125 received a diagnosis of other conditions (␬=0.34). Of those referred with SLE, 76 (29%) Arch Intern Med. 2004;164:2435-2441

HE DIAGNOSES OF SYSTEMIC reduce health care costs, the need to train lupus erythematosus (SLE), primary care physicians in rheumatologi- progressive systemic scle- cal diagnosis has been increasingly recog- rosis, and Sjögren syn- nized.9 drome are aided by well- The objective of this study was to evalu- establishedT clinical criteria.1-5 Since many ate the accuracy of community physi- patients with these potentially serious dis- cians in diagnosing autoimmune disor- eases undergo evaluation and treatment ders. We also examined the factors initially by community physicians, it is im- associated with a misdiagnosis of these portant that they be recognized so that af- conditions and the question of how fre- fected individuals can be referred to spe- quently an inaccurate diagnosis led to in- Author Affiliations: Division of cialists in a timely manner. Misdiagnosis appropriate therapy. Rheumatology & Clinical can have a serious negative impact on Immunology and Center for health as well as economic and emo- METHODS Autoimmune Disease tional consequences. The risks of overdi- (Drs Narain, Richards, Satoh, agnosis include inappropriate treatment Sobel, Hahn, and Reeves and with potentially dangerous medications, PATIENT RECRUITMENT Ms Sarmiento), Program in unnecessary referrals6,7 and laboratory AND DEMOGRAPHICS Epidemiology & Public Health tests,8 and difficulty obtaining health or life (Drs Narain and Davidson), and Four hundred seventy-six patients referred Departments of Biostatistics insurance. Conversely, underdiagnosis can between September 1, 2001, and September (Dr Shuster) and Medicine lead to a delay in appropriate therapy, cul- 30, 2002, to the University of Florida Auto- (Dr Davidson), University of minating in irreversible complications such immune Disease Center (ADC), Gainesville, Florida, Gainesville. as renal failure or pulmonary fibrosis. Be- consented to participate in this institutional Financial Disclosure: None. cause an accurate and timely diagnosis can review board–approved study. The referring

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2435

Downloaded From: https://jamanetwork.com/ on 09/29/2021 physician was asked to complete a prescreening question- tion of patients with a final diagnosis of that condition divided naire indicating the reasons for referral and a working diag- by the total number of patients undergoing analysis. nosis. Sensitivity, specificity, and positive and negative predic- Date of birth, ethnicity, education, referring diagnosis, and tive values were calculated as indicators of the ability of the re- final diagnosis were obtained from medical records and data ferring physician to diagnose autoimmune disorders. Sensitiv- collected at the ADC. Median annual household income was ity was defined as the probability that the referring physician used as an indicator of socioeconomic status. Block group incorrectly diagnosed an autoimmune disease. Specificity was the formation was obtained using the home ZIP code and matched probability that the referring physician was able to exclude the with income data provided in the Inforum software (Thom- presence of a particular autoimmune disease in the popula- son Medstat, Franklin, Tenn). The referring physician’s spe- tion found not to have that disease by the subspecialists at the cialty and year of graduation from specialty training were ob- ADC. Positive predictive value was the probability that the re- tained from the American Medical Association Web site ferring diagnosis was confirmed. (available at: http://www.ama-assn.org/aps/amahg.htm; last ac- Stepwise forward logistic regression for patients with a re- cessed August 31, 2002). Referring physicians were classified ferring diagnosis of SLE was conducted, with age, years of edu- as rheumatologists or nonrheumatologists. cation, type of insurance, median annual household income, physician’s specialty, and physician’s year of graduation as in DIAGNOSTIC CLASSIFICATION dependent variables. A diagnosis of SLE by the referring physician was scored as 1 (SLE final at ADC) or 0 (not SLE final Subjects were interviewed and examined by 1 of 4 autoim- diagnosis). Similarly, 12 or fewer years of patients’ education mune disease subspecialists (H.B.R., E.S., P.H., or W.H.R.). Di- was scored as 0, and greater than 12 years was scored as 1; me- agnostic criteria were reviewed after the evaluation by 2 sub- dian annual household income of $40000 or less was scored specialists (H.B.R. and W.H.R.). Systemic autoimmune diseases as 0 and greater than $40000 as 1. Type of insurance was scored were classified using validated criteria for SLE,1,10 progressive as 1 for privately insured and 0 for those on Medicare/ systemic sclerosis,2 Sjögren syndrome,3 rheumatoid arthritis,4 Medicaid; referring physician was scored as 1 if referred by a and polymyositis/dermatomyositis.5 rheumatologist and 0 if referred by a nonrheumatologist. Based Diagnoses were categorized as follows: (1) SLE, (2) pro- on the results of the regression procedure, we calculated the gressive systemic sclerosis, (3) Sjögren syndrome, (4) poly- odds ratio (OR) estimates of correct diagnosis and 95% confi- myositis/dermatomyositis, (5) positive for antinuclear anti- dence intervals (CI). A P value of less than .05 was accepted as bodies (hereafter referred to as positive ANA; ANA titer Ն1:40 significant. using Hep-2 substrate [The Binding Site Limited, Birming- ham, England] in the absence of autoimmune disease), (6) fi- RESULTS bromyalgia, and (7) other, including undifferentiated connective tissue disease (Ն2 clinical manifestations suggestive of connective tissue disease and the presence of Ն1 non–organ- Demographic characteristics of the cohort are shown in specific autoantibody),11 Raynaud phenomenon, Wegener granu- the following tabulation. lomatosis, antiphospholipid antibody syndrome, interstitial lung disease, or sarcoidosis. Characteristic No. (%) of Patients Female 435 (91) Referring physician AUTOANTIBODY TESTING Rheumatologists 75 (16) Nonrheumatologists 336 (71) All patients underwent extensive autoantibody testing on their Unknown 65 (13) first visit. Antinuclear antibodies were detected by means of in- Ethnicity direct immunofluorescence using Hep-2 cell substrate. Anti– White 338 (71) double-stranded (ds) DNA antibodies were detected using the African American 92 (19) Crithidia luciliae kinetoplast-staining assay (DiaSorin, Stillwa- Hispanic 19 (4) ter, Minn). Antiphospholipid antibodies were detected using Others 27 (6) a commercial IgG and IgM anticardiolipin antibody enzyme- Total 476 (100) linked immunosorbent assay (The Binding Site Limited) and/or Mean±SD age of the cohort was 44.5±14.0, and the lupus anticoagulant activity. Anti-Sm antibodies were de- ethnicity of 1 patient was unknown. Most patients were tected by means of enzyme-linked immunosorbent assay referred with the diagnosis of SLE (56%) (Table 1). Other (Varelisa; Pharmacia Diagnostics, Uppsala, Sweden), and their referring diagnoses included progressive systemic scle- presence was confirmed by means of immunoprecipitation.12 Anti–ribosomal P autoantibodies were detected by means of im- rosis (8%), Sjögren syndrome (4%), polymyositis/ munoprecipitation of the P0, P1, and P2 proteins.12 dermatomyositis (3%), fibromyalgia (1%), positive ANA (7%), and other nonrheumatological diagnoses (12%). DATA ANALYSIS Forty-seven patients with no referring diagnosis and 8 patients without a final diagnosis were excluded from fur- Agreement between the referring and final diagnoses was ana- ther analysis. lyzed using the ␬ statistic, which compares the observed with the expected number of agreements when the rater randomly AGREEMENT BETWEEN REFERRING places the cases, constrained only by the number of diagnoses AND FINAL DIAGNOSIS in each category. The observed-to-expected difference is nor- malized so that 0.0 represents pure guesswork and 1.0 represents complete agreement. Referring physicians’ diagnoses were Agreement between the referring physician’s diagnosis compared with final ADC diagnoses that were used as the gold and the final ADC diagnosis was evaluated using the ␬ standard and were based on formal diagnostic criteria. Preva- statistic. Interrater agreement for all diagnoses was 0.36 lence of an autoimmune disease was calculated as the propor- (95% CI, 0.30-0.42); 49% of the referring diagnoses

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2436

Downloaded From: https://jamanetwork.com/ on 09/29/2021 Table 1. Comparison Between the Referring and Final Diagnoses

Final Diagnosis, No. of Patients

Referring Diagnosis SLE SSC SS ANA+ PM/DM FM Other NA Total No. (%)* SLE 134 4776114274267 (56) SSC 1 23 17 103036 (8) SS 0 2 15 2000019 (4) ANA+ 5 0 1 24 0021 33 (7) PM/DM 0 0 0 1 10 100 12 (3) FM 0 0 0 1 0 1 10 3 (1) Other 12 2 2 18 2 1 19 3 59 (12) NA 12 3 3 10 5 0 6 8 47 (10) Total, No. (%) 164 (35) 34 (7) 29 (6) 139 (29) 19 (4) 17 (4) 58 (12) 16 (3) 476 (100)

Abbreviations: ANA+, seropositive finding for antinuclear antibodies without autoimmune disease; FM, fibromyalgia; NA, not available; Other, other nonautoimmune rheumatological diagnoses; PM/DM, polymyositis/dermatomyositis; SLE, systemic lupus erythematosus; SSC, systemic sclerosis (scleroderma); SS, Sjögren syndrome. *Percentages have been rounded and may not total 100.

Table 2. Diagnostic Accuracy by Specialty of the Referring Physician

Specialty

Nonrheumatologists Rheumatologists

Diagnosis Referring Diagnosis Final Diagnosis Referring Diagnosis Final Diagnosis SLE, No. (%) 192 (62) 102 (33) 45 (69) 33 (51) ANA+, No. (%) 25 (8) 107 (35) 3 (5) 12 (19) All diagnoses, No.* 311 65 ␬, mean ± SD, all diagnoses 0.32 ± 0.07 0.55 ± 0.16

Abbreviations: ANA+, seropositive finding for antinuclear antibodies without autoimmune disease; SLE, systemic lupus erythematosus. *Includes SLE, polymyositis/dermatomyositis, systemic sclerosis (scleroderma), Sjögren syndrome, ANA+, and other nonautoimmune rheumatological diagnoses.

matched the final diagnoses, and 129 patients received an incorrect diagnosis of SLE. The ␬ statistic for SLE was Table 3. Accuracy Indexes and Predictive Values for All Rheumatological Diagnoses* 0.35 (95% CI, 0.27-0.42), suggesting only a fair agreement between raters. Many of the patients receiving an Sensitivity Specificity PPV NPV incorrect diagnosis of SLE received a final diagnosis of positive ANA (76/263 with SLE as a referring diagnosis, SLE 88 52 51 89 ␬ Overall 89 34 60 74 =0.20). (all autoimmune diseases) For all diagnoses, agreement was fair for nonrheuma- ␬ tologists ( =0.32; 95% CI, 0.25-0.39) and better for rheu- Abbreviations: NPV, negative predictive value; PPV, positive predictive matologists (␬=0.55; 95% CI, 0.39-0.71). Rheumatolo- value; SLE, systemic lupus erythematosus. gists referred 45 patients with SLE as a primary diagnosis, *Data are expressed as percentages. and 33 were given a final diagnosis of SLE (␬=0.58; 95% CI, 0.37-0.77). Their diagnostic accuracy was 80% and specificity was 60%. For nonrheumatologists, the diagnosis of SLE was confirmed for only 102 of 192 patients REFERRING PHYSICIAN’S SPECIALTY AND (␬=0.31; 95% CI, 0.21-0.41) (Table 2). OTHER VARIABLES We calculated the sensitivity, specificity, accuracy, and positive and negative predictive values for the referring For patients with a referring diagnosis of SLE, stepwise physicians’ diagnoses (Table 3). Diagnostic sensitivity forward logistic regression was used to evaluate the as- was high for SLE (88%). High specificity was seen for all sociation between misdiagnosis and the patient’s educa- diagnoses except SLE (52%), and was explained by clus- tion and median annual household income and the re- tering of most of the patients in the true negative cell, as ferring physician’s specialty and year of graduation. 63% of the referrals were diagnosed as SLE and approxi- Rheumatologists were more likely to make an accurate mately 13% had diagnoses other than autoimmune dis- diagnosis of SLE than nonrheumatologists (OR, 3.9; 95% orders. The numbers in other disease categories were too CI, 1.6-9.4; PϽ.003). No other variable achieved a sig- small for meaningful analysis. nificant P value, including education (OR, 1.05; 95% CI,

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2437

Downloaded From: https://jamanetwork.com/ on 09/29/2021 A C D 100 15 20

75 15 14 10 9 53.5 8 8 11 50 7 7 10 % of Patients 30.5 No. of Patients 5 No. of Patients 6 25 5 16

0 0 0 132 <1011-2021-40 >40 Unknown 132 No. of Criteria Steroid Dosage, mg/d No. of Criteria Among Patients Receiving Steroids

B E 30 40

30 28 20 19

16 20 No. of Patients No. of Patients 10 8 11 10 5 5 33 4 2 2 2 1 1 1 0 0 Positive Positive Hemato- CNS Renal SerositisArthritisOral Photo- Malar Positive Positive Hemato- CNS Renal SerositisArthritisOral Photo- Malar Serological ANA logic Disorder Ulcers sensitivity Rash Serological ANA logic Disorder Ulcers sensitivity Rash Finding Finding 3-Criteria ANA-Positive Patients ACR SLE Criteria Among Patients Receiving Steroids

Figure. Clinical manifestations in patients with positive findings for antinuclear antibodies (positive-ANA patients) and the subset treated with corticosteroids. A, Percentage of 118 patients in the positive-ANA diagnostic criteria meeting 1, 2, or 3 of the American College of Rheumatology (ACR) criteria for systemic lupus erythematosus (SLE). B, Clinical manifestations of positive-ANA patients meeting 3 of the ACR criteria for SLE. Positive serologic findings include anti–double-stranded DNA, anti-Sm, or antiphospholipid antibodies. Other manifestations are as defined in the 1982 ACR criteria.1,10 C, Prednisone dosages of the 39 positive-ANA individuals treated with corticosteroids. D, Number of ACR criteria fulfilled by the 39 positive-ANA individuals treated with corticosteroids. Fourteen patients had a positive ANA test as the sole manifestation of SLE. E, Clinical manifestations of the 39 positive-ANA individuals treated with corticosteroids. CNS indicates central nervous system.

0.48-2.27; P=.91) and median annual household in- age was apparent in 88 (67%) of 131 patients given a fi- come (OR, 1.12; 95% CI, 0.68-1.84; P=.66). Univari- nal diagnosis of SLE. The most frequent manifestations ately, those insured by private policies were more likely in the positive-ANA subset were serological, usually an- to be correctly diagnosed as SLE (OR, 1.85; 95% CI, 1.09- tiphospholipid antibodies (11 of 16 patients). Arthritis, 3.12; P=.02). However, after adjustment for the refer- a relatively nonspecific manifestation, was the most com- ring physician as rheumatologist or not, the variable did mon clinical manifestation in this subset. not achieve an OR significantly different from 1.0 at PϽ.05 (P=.10). AUTOANTIBODY MARKERS

CLINICAL FEATURES Certain autoantibodies are highly specific for SLE,14 including anti-dsDNA, anti-Sm, and anti–ribosomal P.12 Because some diagnostic criteria are more specific for SLE They may appear years before the onset of SLE,15 thus than others,13 and to exclude the possibility that the posi- providing a second approach for evaluating whether the tive-ANA patients may have early SLE, we compared the positive-ANA group might have early lupus. Anti- clinical features of patients with SLE (Ն4 criteria) vs the dsDNA antibodies (detected by any technique at any time) 3-criteria positive-ANA group, which consisted of 16% were found in 2 of the 70 positive-ANA (6%) and 1 of of the total positive-ANA subset (Figure, A and B). The the 42 other serum samples (2%), whereas anti-Sm an- 3-criteria positive-ANA patients had a low prevalence of tibodies were found in only 1 patient with a positive- lupus-specific criteria such as malar rash, photosensitiv- ANA serum sample and not at all in the other samples. ity, renal disease, and central nervous system involve- When anti-dsDNA antibody (Crithidia assay), anti-Sm, ment (Figure, B). Overall, in the positive-ANA subset, and anti–ribosomal P (immunoprecipitation) findings major organ involvement (renal, central nervous sys- were reassessed at the time of referral, 28% of patients tem, or hematologic) was rare. In contrast, end-organ dam- with SLE were seropositive for anti-dsDNA (Crithidia)

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2438

Downloaded From: https://jamanetwork.com/ on 09/29/2021 vs 3% in the positive-ANA and 2% in the other groups. (positive predictive value, 51%). Most of these patients Anti-Sm was detected in 14% of serum samples from pa- had a positive ANA finding as the sole criterion for SLE. tients with SLE and not at all in the other 2 groups. Simi- The serological test results that were negative for larly, anti–ribosomal P was detected in 2% of serum disease-specific autoantibodies further supported the samples from patients with SLE but not in the other notion that most did not have early SLE. Autoimmune groups. Thus, lupus-specific autoantibodies generally were disease, which was rare among the positive-ANA sub- absent in the positive-ANA and other subsets, suggest- jects, frequently develops in individuals with anti- ing that these patients differed serologically as well as clini- dsDNA or anti-Sm antibodies.15,20,21 In contrast, a posi- cally from the patients with SLE. Twenty-six percent of tive ANA finding in the absence of physical signs and the positive-ANA patients had antiphospholipid auto- symptoms has limited diagnostic utility.22 Depending on antibodies. Antiphospholipid antibodies also were de- age, 3% to 13% of healthy, asymptomatic people are tected in 23% of the serum samples from patients in the seropositive for ANA.23-25 Many other conditions are other category. associated with a positive ANA test, including viral or bacterial infections, autoimmune thyroid disease, and PATIENT CARE IMPLICATIONS medication use.23,26 There is little evidence that lupus eventually develops in these individuals. It was of particular interest to determine whether the mis- Of 476 patients, 203 received a misdiagnosis from the diagnosis of systemic autoimmune diseases, especially lu- referring physicians, of the 203, 137 ultimately received pus, has health care implications. We studied this by evalu- a diagnosis of positive ANA or fibromyalgia. Of the 137 ating the prescription of inappropriate medications. patients, 39 were receiving prednisone at dosages as high Misdiagnoses were made in 203 patients (46% of those as 60 mg/d. Although disease manifestations might have referred by physicians in private, solo, or group-based been suppressed by corticosteroid therapy, we do not be- practice and 39% of those referred by university-based lieve this was the case for several reasons. First, these in- physicians). Nearly 67% of those with a misdiagnosis were dividuals did not meet lupus criteria, and only 3% of the given a final diagnosis of being positive ANA (no evi- positive-ANA patients had major end-organ damage vs dence of systemic autoimmune disease) or of having fi- 66% of the patients with SLE. Second, they did not pro- bromyalgia. Of the 68 patients with a misdiagnosis who duce disease-specific autoantibodies. Finally, during fol- were treated with steroids, 39 did not have an autoim- low-up for a 3-year period after prednisone therapy was mune disease (final diagnosis of positive ANA or fibro- discontinued, no clinical or serological manifestations of myalgia). Their prednisone dosages are shown in Figure, SLE developed in these patients. It is likely, therefore, C; 25 of 32 received moderate- to high-dosage predni- that these 39 individuals were treated inappropriately. sone (16-60 mg/d) for their presumed autoimmune con- The risks of long-term corticosteroid use include infec- dition. The SLE criteria they fulfilled are shown in Figure, tion, osteoporosis, diabetes, hypertension, cataracts, and D and E. Only 6 met 3 criteria, and major organ involve- osteonecrosis.27 Thus, in addition to possible emotional ment was rare. and financial (eg, inability to obtain health insurance) consequences, misdiagnosis of SLE may lead to inappro- COMMENT priate use of potentially risky medications. Although this is the first study to address diagnostic The objective of this study was to assess the accuracy of accuracy in a large lupus referral cohort, similar obser- diagnosing autoimmune disorders in the community vations have been made in general rheumatology popu- and the effect of misdiagnosis on medical care. Previous lations. In a Canadian study, there was poor agreement studies suggested a reluctance on the part of primary between primary care physicians and the consulting rheu- care physicians to propose a tentative diagnosis,6,7,16 an matologist for a variety of common rheumatological di- interpretation supported by the present study in which agnoses.6 That patient population, however, differed 47 patients were excluded because the referring physi- greatly from ours in the number and type of referrals. The cian did not provide a tentative diagnosis. Final diag- problem of misdiagnosis is not limited to rheumatologi- noses were arrived at by consensus among a group of cal conditions, as illustrated by the frequent misdiagno- specialists using accepted diagnostic criteria.17,18 As all sis of multiple sclerosis28 and chronic obstructive pul- patients underwent complete reevaluation, bias intro- monary disease.29 duced by knowing the referring physicians’ diagnoses The referring physician’s specialty was the only fac- was minimized. The low rate of agreement between the tor that significantly affected the rate of misdiagnosis, sug- referring and final diagnosis supports that conclusion. gesting that improving physicians’ training in diagnos- Our results indicate that autoimmune diseases, espe- ing rheumatological conditions might be beneficial. cially SLE, are overdiagnosed by community physicians. Sociodemographic factors (ie, age, education, and in- The most frequent final diagnosis for those patients come) did not affect diagnostic accuracy, suggesting that whose referring and final diagnosis differed was positive the misdiagnosis of autoimmune disease does not re- ANA. It is unlikely that these positive-ANA individuals flect lack of access to adequate medical care. This is es- had early SLE not yet meeting the diagnostic criteria, pecially germane in view of the fact that lupus afflicts mi- because they did not exhibit clinical manifestations norities more frequently than white patients.30 regarded as highly specific for SLE19 (Figure, B). Overall, Although referring rheumatologists were more likely 137 patients received an incorrect diagnosis of SLE to make correct diagnoses, their accuracy was still rela-

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2439

Downloaded From: https://jamanetwork.com/ on 09/29/2021 tively low. However, our results may have been biased pathways or algorithms to aid in determining who by the fact that rheumatologists tend to refer only those should be referred to a specialist. patients who pose diagnostic dilemmas. Another unex- pected finding was that 18 (26%) of the patients Accepted for Publication: April 26, 2004. referred with a diagnosis of SLE and an ultimate diag- Correspondence: Sonali Narain, MD, MPH, Division of nosis of positive ANA had antiphospholipid antibodies, Rheumatology & Clinical Immunology, University of which were added to the diagnostic criteria in 1997.10 Florida, PO Box 100221, Gainesville, FL 32610-0221 Various studies suggest a high prevalence (approxi- ([email protected]). mately 30%) of antiphospholipid antibodies in SLE.31,32 Funding/Support: This study was supported by re- However, primary antiphospholipid autoantibody syn- search grants AR40391 and M01 R00082 from the US Pub- drome (in the absence of lupus) is a well-known lic Health Service, National Institutes of Health, Bethesda, entity.33,34 The positive-ANA patients with antiphospho- Md, and by state funds allocated to the University of lipid antibodies will need to be followed up longitudi- Florida Center for Autoimmune Diseases. nally to see whether autoimmune disease ultimately Acknowledgment: We thank Minna Honkanen-Scott and develops.20 Krista Behney for technical assistance; Douglas Theri- This study has certain limitations. Because our group aque, MS (University of Florida General Clinical Re- has been interested in SLE for many years, there was a search Center), for computer support; Sylvia McLeran bias toward that referring diagnosis. Also, our findings for secretarial assistance; Jennifer Myles for providing the might have been strengthened by a blinded review of the Inforum data; and the nursing staff of the University of medical records. We can estimate the overdiagnosis rate, Florida General Clinical Research Center. but not the underdiagnosis rate. The key issue of whether accurate diagnosis makes a difference in the long-term REFERENCES outcome or cost of medical care could not be assessed in detail. Nevertheless, the data suggest that overdiag- 1. Tan EM, Cohen AS, Fries JF, et al. The 1982 revised criteria for the classification nosis may lead to unnecessary and potentially danger- of systemic lupus erythematosus. Arthritis Rheum. 1982;25:1271-1277. ous treatment. 2. Subcommittee for Scleroderma Criteria of the American Rheumatism Associa- It is likely that in some instances the working diag- tion Diagnostic and Therapeutic Criteria Committee. Preliminary criteria for the nosis was other than what was reported to us. For in- classification of systemic sclerosis (scleroderma). Arthritis Rheum. 1980;23: 35 581-590. stance, Calvo-Alen et al described a subset of patients 3. Vitali C, Bombardieri S, Jonsson R, et al. Classification criteria for Sjogren’s syn- referred with a diagnosis of possible SLE who actually drome: a revised version of the European criteria proposed by the American- had fibromyalgia-like symptoms and were seropositive European Consensus Group. Ann Rheum Dis. 2002;61:554-558. for ANA. We tried to assess the impact of such misrep- 4. Arnett FC, Edworthy SM, Bloch DA, et al. The American Rheumatism Associa- tion 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis resentation by recording the therapy provided to the pa- Rheum. 1988;31:315-324. tients. We believe that the decision to use potentially dan- 5. Tanimoto K, Nakano K, Kano S, et al. Classification criteria for polymyositis and gerous drugs such as corticosteroids provides a measure dermatomyositis. J Rheumatol. 1995;22:668-674. of the veracity of the referring physician’s working di- 6. Gamez-Nava JI, Gonzalez-Lopez L, Davis P, Suarez-Almazor ME. Referral and agnosis. The fact that significant numbers of patients with diagnosis of common rheumatic diseases by primary care physicians. Br J Rheumatol. 1998;37:1215-1219. misdiagnoses were treated with corticosteroids suggests 7. Gran JT, Nordvag BY. Referrals from general practice to an outpatient rheuma- that, in most cases, the working diagnosis was not mis- tology clinic: disease spectrum and analysis of referral letters. Clin Rheumatol. represented. Nevertheless, there are likely to be some in- 2000;19:450-454. stances where the working diagnosis was other than what 8. Suarez-Almazor ME, Gonzalez-Lopez L, Gamez-Nava JI, Belseck E, Kendall CJ, Davis P. Utilization and predictive value of laboratory tests in patients referred was reported to us. Our study design does not permit us to rheumatologists by primary care physicians. J Rheumatol. 1998;25:1980- to comment on whether primary care physicians or rheu- 1985. matologists were more likely to misrepresent the diag- 9. Schwartz G, Merenstein JH, D’Amico F, Agarwal A. Does curriculum make a dif- nosis. ference? a comparison of family physicians with and without rheumatology training during residency. Arch Fam Med. 1994;3:263-267. 10. Hochberg MC. Updating the American College of Rheumatology revised criteria CONCLUSIONS for the classification of systemic lupus erythematosus [letter]. Arthritis Rheum. 1997;40:1725. 11. Mosca M, Tavoni A, Neri R, Bencivelli W, Bombardieri S. Undifferentiated con- Systemic autoimmune diseases (SLE in particular) are nective tissue diseases: the clinical and serological profiles of 91 patients followed for at least 1 year. Lupus. 1998;7:95-100. overdiagnosed by community physicians, sometimes 12. Reeves WH, Satoh M, McCauliffe DP. Autoantibody testing by non-FITC methods. leading to inappropriate therapy. There may be emo- In: Rose NR, Hamilton RG, Detrick B, eds. Manual of Clinical Immunology. 6th ed. tional and financial consequences as well. Further stud- Washington, DC: American Society of Microbiology Press; 2002:933-950. ies to determine the health, emotional, and financial 13. Smith EL, Shmerling RH. The American College of Rheumatology criteria for the consequences of erroneous diagnosis of SLE seem war- classification of systemic lupus erythematosus: strengths, weaknesses, and op- portunities for improvement. Lupus. 1999;8:586-595. ranted. A large number of instances in which a referring 14. Edworthy SM, Zatarain E, McShane DJ, Bloch DA. Analysis of the 1982 ARA lu- diagnosis of SLE rested entirely on the presence of a pus criteria data set by recursive partitioning methodology: new insights into the positive ANA test in the absence of clinical manifesta- relative merit of individual criteria. J Rheumatol. 1988;15:1493-1498. tions underscores the potential benefit of continuing 15. Arbuckle MR, James JA, Kohlhase KF, Rubertone MV, Dennis GJ, Harley JB. Development of anti-dsDNA autoantibodies prior to clinical diagnosis of sys- medical education. Increased awareness of the clinical temic lupus erythematosus. Scand J Immunol. 2001;54:211-219. presentation of these diseases may help increase diag- 16. Hooker RS, Brown JB. Rheumatology referral patterns. HMO Pract. 1990;4:61- nostic accuracy. There also may be a role for diagnostic 65.

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2440

Downloaded From: https://jamanetwork.com/ on 09/29/2021 17. Bjelle A, Magi M. Rheumatic disorders in primary care: a study of two primary 26. Shiel WC Jr, Jason M. The diagnostic associations of patients with antinuclear care centres and a review of previous Swedish reports on primary care. Scand J antibodies referred to a community rheumatologist. J Rheumatol. 1989;16: Rheumatol. 1981;10:331-341. 782-785. 18. Sverdrup B, Allebeck P, Allander E. Tentative diagnoses among referrals versus 27. Boumpas DT, Chrousos GP, Wilder RL, Cupps TR, Balow JE. Glucocorticoid therapy diagnoses established at the Department of Rheumatology. Scand J Rheumatol. for immune-mediated diseases: basic and clinical correlates. Ann Intern Med. 1983;12:377-378. 1993;119:1198-1208. 19. Bombardieri S, Neri R, Tartarelli G, d’Ascanio A, Giovanelli L. The clinical rel- 28. Riaz S, Nowack WJ. Diagnostic problems in multiple sclerosis: overreliance on evance of antinuclear antibodies in connective tissue diseases. Scand J Rheu- neuroimaging. South Med J. 1998;91:270-272. matol Suppl. 1987;66:35-45. 29. Incalzi RA, Fuso L, Serra M, et al. Exacerbated chronic obstructive pulmonary 20. Arbuckle MR, McClain MT, Rubertone MV, et al. Development of autoantibodies disease: a frequently unrecognized condition. J Intern Med. 2002;252:48-55. before the clinical onset of systemic lupus erythematosus. N Engl J Med. 2003; 30. Petri M. Epidemiology of systemic lupus erythematosus. Best Pract Res Clin 349:1526-1533. Rheumatol. 2002;16:847-858. 21. Vila LM, Mayor AM, Valentin AH, Garcia-Soberal M, Vila S. Clinical outcome and 31. Love PE, Santoro SA. Antiphospholipid antibodies: anticardiolipin and the lupus predictors of disease evolution in patients with incomplete lupus erythematosus. anticoagulant in systemic lupus erythematosus (SLE) and in non-SLE disor- Lupus. 2000;9:110-115. ders: prevalence and clinical significance. Ann Intern Med. 1990;112:682-698. 22. Clegg DO, Williams HJ, Singer JZ, et al. Early undifferentiated connective tissue 32. Sebastiani GD, Passiu G, Galeazzi M, Porzio F, Carcassi U. Prevalence and clini- disease, II: the frequency of circulating antinuclear antibodies in patients with cal associations of anticardiolipin antibodies in systemic lupus erythematosus: early rheumatic diseases. J Rheumatol. 1991;18:1340-1343. a prospective study. Clin Rheumatol. 1991;10:289-293. 23. Tan EM, Feltkamp TE, Smolen JS, et al. Range of antinuclear antibodies in “healthy” 33. Hughes GR, Harris NN, Gharavi AE. The anticardiolipin syndrome. J Rheumatol. individuals. Arthritis Rheum. 1997;40:1601-1611. 1986;13:486-489. 24. Manoussakis MN, Tzioufas AG, Silis MP, Pange PJ, Goudevenos J, Moutsopou- 34. Lockshin MD. Lupus and antiphospholipid antibody syndrome: either, neither, los HM. High prevalence of anti-cardiolipin and other autoantibodies in a healthy or both [editorial]. Am J Med. 1994;96:1-2. elderly population. Clin Exp Immunol. 1987;69:557-565. 35. Calvo-Alen J, Bastian HM, Straaton KV, Burgard SL, Mikhail IS, Alarcon GS. 25. Fields RA, Toubbeh H, Searles RP, Bankhurst AD. The prevalence of anticardio- Identification of patient subsets among those presumptively diagnosed with, re- lipin antibodies in a healthy elderly population and its association with anti- ferred, and/or followed up for systemic lupus erythematosus at a large tertiary nuclear antibodies. J Rheumatol. 1989;16:623-625. care center. Arthritis Rheum. 1995;38:1475-1484.

(REPRINTED) ARCH INTERN MED/ VOL 164, DEC 13/27, 2004 WWW.ARCHINTERNMED.COM 2441

Downloaded From: https://jamanetwork.com/ on 09/29/2021