210

Arl4b, m m

40

40

to
4w
ldrIalk v m

AW

ANV-Iqb v

%w4a *1

fm

10 4w l9d Ar"VI,

4:

AOIU

Clinical Significance and Interpretation of Antinuclear Antibodies

RICHARD H. WHITE, MD, and DICK L. ROBBINS, MD, Sacramento, California

"Topics in Primary Care Medicine" presents articles on common diagnostic or therapeutic problems encountered in primary care practice. Physicians interested in contributing to the series are encouraged to contact the series'
editors.
STEPHEN J. McPHEE, MD Series' Editors

BERNARD LO, MD

Primary care physicians often see patients with rheumatic complaints such as joint pain, muscle weakness and other symptoms that suggest a "connective tissue" disease. When this occurs, there is a tendency to order a panel of tests that often includes the erythrocyte sedimentation rate, a fluorescent antinuclear antibody (FANA) test and a test for rheumatoid factor. The FANA test, in particular, is usually done as a "screening" test for systemic lupus erythematosus or one of the closely related systemic rheumatic diseases. Unfortunately, most clinicians do not have a clear understanding of how to interpret the antinuclear antibody test result. Complicating this issue even further is the fact that many clinical laboratories are now offering tests for antibodies to a variety of more specific nuclear antigens, such as antibodies to native (double-stranded) DNA (nDNA). When should a FANA test be done, what does it mean and how should it be interpreted? Moreover, when should the antinuclear antibody tests that detect unique nuclear antigens, such as nDNA, be done? In this brief review we summarize what is known about the clinical utility of the various antinuclear antibody tests that are currently available and provide suggested guidelines for their use.

clinicians generally associate the presence of antinuclear antibodies with a rheumatic disease, particularly systemic lupus erythematosus, these antibodies can also be detected in persons with other immunologically mediated diseases such as Hashimoto's thyroiditis and immunopathic chronic active hepatitis. While there is some evidence that antibodies to nDNA and other antinuclear antibodies may play a direct role in the pathogenesis of nephritis in patients with lupus erythematosus, their pathogenic role in the rheumatic diseases is poorly understood. Although recent research characterizing the nature and disease associations of the nuclear antigens to which various antinuclear antibodies are directed is exciting and may eventually provide insight into the pathogenesis of various autoimmune diseases, this information is of little practical use to clinicians. Therefore, tests for antinuclear antibodies are most often used as aids in the diagnosis and management of disease.

Biology of Antinuclear Antibodies

Antinuclear antibodies are a spectrum of autoantibodies that react with various nuclear molecules, including DNA, ribonucleic acid (RNA), histones, acidic nuclear proteins or complexes of these molecular elements. The presence of antinuclear antibodies in serum appears to be associated with various factors including genetic predisposition (such as the histocompatibility locus DR3), environmental agents (ultraviolet light, viruses, certain drugs and chemicals and intravenous drug use), estrogen-androgen balance, chronic infections, neoplasms and advancing age. Aging cannot be overemphasized as a factor associated with the presence of antinuclear antibodies. One recent study noted that 18% of otherwise healthy persons older than 65 years had a positive-that is, a titer of more than 1: 10-FANA test. While

General Principles of Test Interpretation When considering the possibility of doing an antinuclear antibody test, clinicians should keep certain general principles of testing in mind. (For a more detailed discussion of these principles, see the article by Sox.) First, if the result of a test will have no effect on any patient-management decision, the test should not be done. In the case of antinuclear antibody
testing, it should be stressed that the diagnosis and management of most systemic rheumatic diseases rest primarily on clinical manifestations and the results of more routine laboratory tests. Second, if the sensitivity and specificity of a test are known, after estimating the chance (pretest probability) that a patient has a specific disease, such as systemic lupus erythematosus, it is possible to apply Bayes's theorem and calculate the effect of a positive (or negative) test result on the posttest probability (see Table 1 for definitions). Whereas few physicians think in such a quantitative fashion, two important generalizations can be easily remembered and applied:

(White RH, Robbins DL: Clinical significance and interpretation of antinuclear antibodies. West J Med 1987 Aug; 147:210-213)
From the Divisions of General Medicine (Dr White) and Rheumatology (Drs White and Robbins), Department of Internal Medicine, Primary Care Center, University of California, Davis, Medical Center, Sacramento. Reprint requests to Richard H. White, MD, Division of General Medicine, Rm 3120, Primary Care Center, 2221 Stockton Blvd, Sacramento, CA 95817.

THI

THE WESTERN JOURNAL OF MEDICINE

AUGUST 1987

147

21

211

ABBREVIATIONS USED IN TEXT

FANA =fluorescent antinuclear antibody nDNA=native DNA nRNP= nuclear ribonucleoprotein RNA=ribonucleic acid
* When a patient has a low but appreciable probability of having a particular disease, such as 5 % to 40 %, a test of high specificity, when positive, is helpful in establishing the diagnosis-that is, increasing the posttest probability. * When a patient has a moderate to high probability of having a particular disease, a test of high sensitivity, when negative, is useful in excluding this disease-that is, decreasing the posttest probability. As the sensitivity and specificity of a test diminish, the usefulness of the test also diminishes. The third principle is that any test that falls short of being 100% specific should not be used indiscriminately to screen patients who have a low pretest probability of disease, as it is likely that a positive result will be "false-positive" rather than "true-positive." Stated another way, the ability of a positive test result, by itself, to accurately predict the presence of disease (positive predictive value) is critically dependent on the prevalence (pretest probability) of the disease in the population being tested. An example that illustrates this principle is the use of the FANA test to screen patients for lupus erythematosus. Several studies have shown that in a group of "healthy" persons between the ages of 20 and 40 years, about 1 % of the population has a positive FANA test without apparent disease (the specificity for systemic lupus erythematosus = 99%), and it is known that 95 % of patients with lupus erythematosus have a positive FANA test (sensitivity = 95%). Assuming that the prevalence of the disease in this group is about 1 in 700, a positive FANA test will accurately reflect the presence of lupus in only 13 % of cases (seven false-positive results for every single true-positive result). Interestingly, if a FANA test is done on a more select group of persons, such as patients with polyarthritis, the predictive value of a positive FANA test in detecting lupus does not increase. This is because the specificity of a FANA test for the disease in this population is significantly lower (specificity =50% to 70%) because patients with rheumatoid arthritis and other inflammatory arthritides frequently have a positive FANA test. Assuming that the prevalence of systemic lupus erythematosus in patients with polyarthritis is 1 in 20, the predictive value of a positive test is still less than 15 %. The fourth principle to remember is that the initial estimates of the specificity of a test for a particular disease usually overestimate the true specificity. This is because the patients selected for these studies usually have well-defined "classic" disease, whereas the "control" cases are usually healthy persons or patients in hospital in whom the prevalence of falsepositive test results is low. A more ideal "control" group to study to define the specificity of a test would be a group of patients presenting with signs or symptoms suggesting the disease in question, such as joint pain, skin rashes and so forth. This is particularly the case for the rheumatic diseases.

which is the most widely used method of detecting antinuclear antibodies. While rat and mouse liver and kidney were commonly used cell substrates in the past, human epithelial cell lines, particularly the KB and HEp-2 cell lines, are currently the most commonly used substrates. It is important to know that methodologic differences and particularly differences in the substrates used-such as HEp-2 cells versus mouse kidney cells-may lead to differences in FANA test results from different laboratories. A report of a FANA test generally includes the pattern of staining and the highest serum dilution (titer) that has detectable antibody.
FANA liter The titer of the FANA test that should be interpreted as "positive" varies from one laboratory to another and should be noted on the report slip. In patients with a well-defined rheumatic disease, the titer of the FANA test varies widely. For example, about a third of patients with lupus erythematosus will have a FANA titer of 1:40 or less, whereas a third will have a titer of 1:360 or more. The titer is not helpful in predicting the activity of any rheumatic disease. While most false-positive FANA tests are low titer and most high-titer tests are associated with clinically obvious disease, there may be no discernible disease in some patients who have a hightiter FANA test.

Antinuclear Antibody Patterns The most common patterns of nuclear fluorescence reported and the specificity of each ofthese patterns are listed in Table 2. With the exception of the rim pattern, the pattern of nuclear staining lacks sufficient specificity to be diagnostically useful. There is often interobserver variation in pattern interpretation, and the pattern can vary as the test serum is

TABLE 1- iOn of Terms

Sens ar f a w it a who h a Perc positive test* Specdicity: Peentap of poa mt ho do nat have the disea wsewhoav anegavetestt Positive predictvWalue of a tst for a disease: (nunber of true o + ( eof false-positivs)
Ofthetest
tm

tsrst ont positive (or :eQative using the pr tpr bl of

Bayes's theorem A mathmiao

r calatng theeffetof a

*Thetrte-poeivtratX.

trs. g tt..

rate=I-

........h

rae..

~ SPM**d ~ lo

..

E.M i,

Fg

note2800:rd pa 0"MA SL,98) wf-h dMr tk O dF#eases lk

The FANA Test

Most physicians are familiar with the general principles that underlie the indirect immunofluorescence technique,

212
diluted because antibodies directed to several different nuclear antigens may be present in the same serum specimen. A rim pattern is strongly associated with antibodies to nDNA. Homogeneous fluorescence is associated with antibodies directed against nuclear histone or DNA-histone complexes. A speckled antinuclear antibody pattern is associated with antibodies to acidic nuclear proteins and RNA-protein complexes, such as nuclear ribonucleoprotein (nRNP). A
a

TOPICS IN PRIMARY CARE MEDICINE

MEDICINE

FANA test should be done. A negative test essentially ex-

cludes any of these diseases. As outlined earlier, a positive FANA test by itself is not diagnostic. In patients who have signs or symptoms compatible with either of these three disorders, however, a positive FANA test does increase the posttest probability that they have the disease.

large-speckled pattern may be reported in patients who have anticentromere antibodies. A nucleolar pattern is associated with antibodies to elements that make up this organelle. The sensitivity of a positive FANA test in various diseases is also
listed in Table 2.

Interpreting the FANA Test There are only three diseases that include or require a positive antinuclear antibody test in their diagnostic criteria. The first is systemic lupus erythematosus. It is generally accepted that greater than 95 % of patients with this disease will have a positive antinuclear antibody test and that more than 85 % will have a positive FANA test two dilutions above the lowest titer that is acceptable as being "positive." In addition, virtually 100% of patients with mixed connective tissue disease and drug-induced lupus erythematosus have a positive antinuclear antibody test. Because the sensitivity of the FANA test in detecting these diseases is high, if clinical and laboratory findings suggest the presence of lupus erythematosus, mixed connective tissue disease or drug-induced lupus,

Tests for Antinuclear Antibodies to Individual Nuclear Antigens A variety of methods is available for detecting antibodies to individual nuclear antigens. As the number of different methods used to detect specific nuclear antigens has increased, a wider spectrum of patients than previously suspected has been noted to have low or moderate titers of antibodies to one or more of these nuclear antigens. The best characterized antinuclear antibodies and the specificity and sensitivity for the most common rheumatic diseases are listed in Table 3.
Antibodies to nDNA The 1982 revised American Rheumatism Association criteria for classifying patients with systemic lupus erythematosus include, in addition to a positive FANA test, "an immunological disorder," which is satisfied if there is an elevated titer of antibody to nDNA or a positive lupus erythematosus preparation or antibodies to the Smith (Sm) antigen or a biologic false-positive serology test for syphilis. In some patients, the level of antibodies to nDNA varies directly with

TABLE 3.-Sensitivity and Specificity of Antibodies to Nuclear Antigens

Antinuclear Antfbody Sensitivity Anti-nDNA ......... SLE 60% to 700/o Antihistone ......... Drug-induced SLE 100%, RA 15% to 20%, SLE 30% Anti-Sm ......... SLE 30% to 40% Anti-nRNP ......... MCTD 100%, SLE 30%, scleroderma 20% to 30%, rheumatoid arthritis 10%, discoid lupus 20% to 30% Anti-Ro (anti-SSA) ..... SLE 30% to 40%, Sj3gren's syndrome 60% to 70%, RA 10% Anti-La (anti-SSB) .... SLE 10% to 15%, SjOgren's syndrome 50% to 60% Anticentromere ....... Scleroderma 10% to 15%, CREST 50% to 90%, Raynaud's disease 10% to 30% Scleroderma 15% to 20% Anti-Scl70 ..... Anti-PM-Scl ......*.Polymyositis 10% .. Polymyositis 30%, dermatomyositis <10% Anti-Jo, ...
erythematosus

Specificity High
Low

Disease

SLE (>90%) if present in high titer SLE 98%

High
Low

Low Low

High
High Unknown Unknown

Scleroderma or a closely related disease such

Scleroderma >95% ? Polymyositis ? Myositis
as CREST or Raynaud's disease, >95%

CREST=calcinosis, Raynaud's phenomenon, esophaga dysfunction, scierodactyly and telangiectasia; MCTD=mixed connective tissue disease; RA=rheumatold arthritis; SLE=systemic lupus

TABLE 4.-Indications for Doing Tests for Antinuclear Antibodies

Test
Indication

Fluorescent antinuclear antibody (FANA) .. .. Patient with possible SLE, drug-induced lupus or MCTD

FANA Anti-Sm .. Patient with features of SLE and a positive speckled FANA test pattem Anti-Ro (anti-SSA) ......... Woman of child-bearing age with a known connective tissue disease Patient with a negative FANA but a high pretest probabilit of SLE Anti-nRNP .... Patient with possible MCTD

Anti-nDNA*

..
.

.......

Patient with features of SLE and a positive

Usefulness A negative test virtually excludes these diseases A positive antinuclear antibody test, while nonspecific, increases the posttest probability of disease A positive test (high titer) substantially increases the posttest probability that thepatient has SLE A positive test substantially increases the posttest probability that the patient has SLE Counseling, a positive test associated with a small but real risk of neonatal:SLE and congenital heart block A positive test increases the posttest probability that the patient has SLE A negative test essentially excludes the diagnosis of MCTD A positive test in high titer, while nonspecific, increases the posttest probability that the patient has MCTD

MTM= mixed connective tissue disease, SLE = systemic hupus erythematosus *The tfter of anti-nDNA is used by some speciaists to monitor the activity of SLE; clinical features and other laboratory tests, however, are of equal or greater usefulness.

THE WESTERN JOURNAL OF MEDICINE

AUGUST 1987

147 * 2
147 *

213

213

disease activity, making this a potentially useful index to monitor response to therapy. In a given patient, however, other clinical and laboratory features such as hematologic values, a urinalysis, serum complement levels and the like may be more helpful than the level of antibodies to nDNA.

Anti-Sm and Anti-nRNP Antibodies The Sm (Smith) and nRNP antigens are closely related immunologically and are composed of protein and small RNA species. The presence of antibodies to Sm, like antibodies to nDNA, counts as a criterion for an "immunologic disorder" in the revised American Rheumatism Association criteria for lupus erythematosus. While the presence of antibodies to nRNP is not specific for patients with a mixed connective tissue disorder, the exclusive presence of a high titer of antibody to nRNP without other antinuclear antibodies appears to be unique to patients with this disorder. The absence of antibodies to nRNP virtually excludes the diagnosis of the disease.

* Anti-Ro positivity in a woman with a connective tissue disease is associated with a small but real risk of neonatal systemic lupus erythematosus and congenital heart block if the patient becomes pregnant. * Patients with "subacute cutaneous systemic lupus erythematosus," who frequently have a negative FANA test, usually have detectable anti-Ro antibodies. * Patients with Sjogren's syndrome who have anti-Ro antibodies appear to have more aggressive disease, particularly vasculitis. * Anti-Ro antibodies are commonly detected in patients who have the lupuslike syndrome associated with a deficiency ofthe second component of complement. Thus, the presence ofanti-Ro antibodies may be helpful diagnostically or prognostically in a select few patients.

Antibodies to Histone The presence of antibodies to histone is associated with several rheumatic diseases including drug-induced lupus erythematosus, rheumatoid arthritis and systemic lupus. In a patient with suspected drug-induced lupus, measuring the level of antihistone antibodies adds little diagnostic value to a positive FANA test. Antibodies to the Centromere Antigen Antibodies to the centromere antigen can usually be detected only in rapidly dividing tissue culture cell lines that include numerous mitotic figures. Occasionally they can be detected using the HEp-2 substrate. Although early studies suggested that the presence of antibodies to the centromere was specific for the CREST (calcinosis, Raynaud's phenomenon, esophageal dysfunction, sclerodactyly and telangiectasia) variant of scleroderma, further studies have shown that the presence of anticentromere antibodies is not specific for this syndrome. The exact diagnostic and prognostic significance of anticentromere antibodies remains to be defined. Antibodies to the Ro (SSA) Antigen The Ro antigen (identical to SSA, which stands for Sjogren's syndrome A antigen) is a protein-RNA complex that can be found both in the nucleus and in the cytoplasm of HEp-2 cells. It cannot be detected on mouse, kidney or liver tissue sections. As noted in Table 3, the presence ofantibodies to Ro is not specific for any disease. The real significance of anti-Ro antibodies rests with certain disease associations:

Antibodies to Other Nuclear Antigens Many different nuclear antigens including anti-La (identical to anti-SSB), anti-Scl70, anti-PM-Scl and anti-Jo, have been well characterized. The clinical value of finding elevated levels of antibodies to one of these nuclear antigens in a given patient has not been adequately defined.

Conclusion
The indications for doing specific antinuclear antibody tests are summarized in Table 4. Patients possibly having systemic lupus, drug-induced lupus or a mixed connective tissue disorder (pretest probability of disease ranging between 10% and 90%) should be tested for the presence of antinuclear antibodies. A negative FANA test is very useful in excluding these possible diagnoses. With the possible exception of a rim pattern FANA test, a positive test is nonspecific. Tests for antibodies to nDNA and the Sm antigen may be diagnostically useful when there is a moderate probability of systemic lupus erythematosus. The absence of antibodies to nRNP is helpful in excluding the diagnosis of mixed connective tissue disorder. Antibodies to Ro may be useful in a number of rather rare clinical situations.
GENERAL REFERENCES Richardson B, Epstein WV: Utility of the fluorescent antinuclear antibody test in a single patient. Ann Intern Med 1981; 95:333-338 Singsen BH, Akhter JE, Weinstein MW, et al: Congenital complete heart block and SSA antibodies: Obstetric implications. Am J Obstet Gynecol 1985; 152:655-658 Sox HC: Probability theory in the use of diagnostic tests. Ann Intern Med 1986;

104:60-66

Tan EM: Autoantibodies to nuclear antigens (ANA): Their immunobiology and medicine. Adv Immunol 1982; 33:167-240 Tan EM, Cohen AS, Fries JF, et al: The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982; 25:1271-1277