Pattern on the antinuclear antibody–HEp-2 test is a critical parameter for discriminating antinuclear antibody–positive healthy individuals and patients with autoimmune rheumatic diseases

Abstract

Objective

To identify features of antinuclear antibody (ANA)–HEp-2 test results that discriminate ANA-positive healthy individuals and patients with autoimmune rheumatic diseases (ARDs).

Methods

We sequentially retrieved data on 918 healthy individuals and 153 patients with ARDs after clinical assessment. ANA-positive healthy individuals for whom data were available were reevaluated after 3.6–5.0 years. An ANA–HEp-2 test result was considered positive when a clear ANA pattern was observed at 1:80 dilution in 2 distinct commercial HEp-2 slides by 2 blinded independent observers.

Results

ANAs were present in 118 healthy individuals (12.9%) and 138 patients with ARDs (90.2%). The ANA titer was higher in patients with ARDs than in healthy individuals (P < 0.001). The ANA pattern profile was distinct in the 2 groups. Nuclear homogeneous, nuclear coarse speckled, and nuclear centromeric patterns appeared exclusively in patients with ARDs. The nuclear dense fine speckled pattern occurred only in healthy individuals. The most frequent ANA pattern in both groups was the nuclear fine speckled pattern, which occurred at lower titer in healthy individuals than in patients with ARDs (P < 0.001). Anti–extractable nuclear antigen was present in 1 healthy individual (anti-SSA/Ro) and in 52 patients with ARDs (37.7%). None of the 40 reevaluated healthy individuals developed ARDs, and 29 (72.5%) remained ANA positive. All healthy individuals who became ANA negative had an ANA titer of 1:80 at baseline.

Conclusion

Our findings suggest that the titer, and especially the pattern, on the ANA–HEp-2 test strongly enhances our ability to discriminate ANA-positive healthy individuals and patients with ARDs.

Antinuclear antibodies (ANAs) are considered a hallmark of autoimmune rheumatic diseases (ARDs), and the indirect immunofluorescence (IIF) assay on HEp-2 cells (ANA–HEp-2 test) is the standard method for ANA detection (1). However, a positive ANA–HEp-2 test result at a 1:80 dilution has been reported in up to 13.3% of healthy individuals (2–9). Therefore, ANA–HEp-2 testing outside a proper clinical framework may yield a sizable portion of ANA-positive individuals with no consistent evidence of autoimmune disease, causing some concern and anxiety in patients and physicians. This becomes even more crucial with the perception that autoantibodies may precede the clinical onset of ARD for many years (10, 11). Therefore, it would be useful to identify intrinsic features of a positive ANA–HEp-2 test result that would allow us to discriminate subjects with and those without autoimmune disease.

It is generally accepted that individuals without autoimmune disease would present lower autoantibody serum levels than those with autoimmune disease (12). Accordingly, ANA titers are usually thought to be low in subjects without autoimmune disease and with a positive ANA–HEp-2 test result. Less emphasis has been given to another intrinsic feature of the ANA–HEp-2 test, namely, the IIF pattern. The IIF pattern reflects the topographic distribution of the target autoantigens and therefore may convey significant information about their nature. In fact, several IIF patterns have been shown to bear tight association with certain autoantibody specificities. This is the case for antibodies to proliferating cell nuclear antigen (13), p80-coilin (14), nuclear mitotic apparatus 1 (NuMA-1) and NuMA-2/HsEg5 (15, 16), CENP-F (17), DFS70/lens epithelium–derived growth factor (LEDGF)–P75 (18, 19), and DNA topoisomerase I (20). Considering that not all autoantibody specificities are tightly associated with autoimmunity, it would be expected that subjects with and those without autoimmune disease and having a positive ANA–HEp-2 test result would display different IIF pattern profiles. This reasoning is supported by recent demonstrations that the nuclear dense fine speckled pattern, which is correlated with anti–DFS70/LEDGF-P75 autoantibody (18, 19), is preferentially displayed in samples from individuals with no evidence of active systemic autoimmune disease (8, 19, 21, 22).

The aim of this study was to look for features within positive ANA–HEp-2 test results that differentiate healthy individuals and patients with defined ARD. Additionally, the prognostic significance of a positive ANA–HEp-2 test result in apparently healthy individuals was assessed by reevaluating a series of healthy individuals with positive ANA–HEp-2 test results after an average time period of 4 years.

PATIENTS AND METHODS

Subjects.

A total of 918 healthy individuals (634 women and 284 men, mean ± SD age 32.2 ± 10.3 years [range 18–66 years]) from São Paulo, Curitiba, and Foz do Iguaçu (large cities in southeast Brazil) were included after being considered healthy according to a clinical questionnaire that was administered to investigate current or past ARD, serious chronic infections, and neoplasia. All subjects fulfilled the following inclusion criteria: 1) age ≥18 years; 2) negative findings on serologic tests for infection with human immunodeficiency virus, hepatitis B virus, and hepatitis C virus; and 3) no regular use of glucocorticoids, immunosuppressive agents, or antiinflammatory or antimicrobial drugs. Patients with ARDs and healthy individuals belonged to the same ethnic blended Brazilian background, composed mainly of European, African, and American Indian ancestry. All participants signed the consent form approved by the Ethics Committee at Universidade Federal de São Paulo. From August 2002 to December 2003, all healthy individuals donated 10 ml of peripheral blood. Serum was separated immediately and frozen at –20°C until use. The control group included 153 patients with ARDs (87 with systemic lupus erythematosus [SLE], 45 with systemic sclerosis [SSc], 11 with Sjögren's syndrome [SS], and 10 with idiopathic inflammatory myopathy) enrolled according to established criteria for those diseases (23–26). All samples from patients with ARDs were selected from patients regularly followed up at the Outpatient Rheumatology Division.

After a 4-year period, the 66 blood donors living in São Paulo who had a positive ANA–HEp-2 test result were invited by letter and/or telephone for clinical reevaluation and for a new blood draw. Twenty subjects had changed their address and telephone number and could not be found, and 5 refused to participate. Forty-one individuals (62.1%) agreed to participate in the reevaluation, and 40 of them agreed to have additional blood drawn. The mean ± SD time interval between the 2 samples was 3.9 ± 0.3 years (range 3.6–5.0 years).

ANA–HEp-2 testing and autoantibody determination.

Serum samples from all participants were subjected to the ANA–HEp-2 test using 2 commercial HEp-2 cell slides (Bio-Rad and Bion). Serum samples were diluted in 0.15M NaCl and 10 mM phosphate buffered saline (PBS), pH 7.4, and incubated with HEp-2 cells for 30 minutes at room temperature in a moist chamber. After washing twice in PBS for 10 minutes, cells were incubated with fluorescein isothiocyanate–conjugated goat anti-human Ig (IgG heavy and light chains) for another 30 minutes in the dark. After washing twice as before, slides were assembled with buffered glycerol, pH 9.5, and coverslips. ANA titer was determined by testing successive 2-fold dilutions of the serum up to 1:5,120. Analysis was performed by 2 independent expert observers (SHB and LECA) using an Olympus BX 50 microscope under 400× magnification. In the first round, samples from the 918 healthy individuals and 153 patients with ARDs were processed in a nonblinded manner by both independent observers. Samples were classified as ANA–HEp-2 positive if a well-defined IIF pattern was identified at 1:80 dilution in both substrates and by both observers. The requirement for a positive reaction in 2 commercial HEp-2 slides was a precaution to avoid a falsely high frequency, since some HEp-2 brands are extremely sensitive. The requirement for agreement between both observers aimed to minimize the influence of subjectivity.

In a second round, all samples from healthy individuals and patients with ARDs with a positive ANA–HEp-2 test result were processed simultaneously in a blinded manner for definition of ANA titer and pattern. Discrepant cases were segregated in a “waiting line” and were reprocessed for a second blinded analysis by both observers. The observers were aware that those samples had been discordant but did not know the nature of the findings or how each sample had previously been rated. Most samples became concordant in the second round reading. There were 2 samples that remained discrepant (borderline positive), and these were considered negative. For the followup comparison, the 2 samples from each patient (baseline and followup) were tested in the same assay and were read in a blinded manner by the same observers.

All sera positive on the ANA–HEp-2 test were screened for antibodies against extractable nuclear antigens (ENAs: Sm, U1 RNP, SSA/Ro, SSB/La) by double immunodiffusion against calf spleen extract as the antigen source, according to the Ouchterlony technique as previously described (27). Secondary standards derived from the Centers for Disease Control and Prevention primary standards were used for identification of the antigen specificity. All samples were screened for anti–native DNA at a 1:10 dilution by IIF on Crithidia luciliae as previously described (28). Samples with the nuclear homogeneous and nuclear quasihomogeneous patterns were processed for antinucleosome and antihistone antibodies according to the instructions of the manufacturers (Inova Diagnostics and HUMAN, respectively).

Western blotting.

Sera displaying the nuclear fine speckled and the nuclear dense fine speckled IIF patterns at titers ≥1:320 were analyzed by Western blotting with HEp-2 whole cell extract that was separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to nitrocellulose as previously reported. (19). Briefly, serum samples were tested at 1:50 dilution in 5% skim milk–0.05% Tween 20 in PBS for 1 hour at room temperature, with shaking. After two 15-minute washing steps in 0.05% Tween 20–PBS, nitrocellulose strips were incubated with horseradish peroxidase–labeled goat anti-human IgG antibodies (Bio-Rad) diluted to the ratio 1:1,500 in 5% skim milk–0.05% Tween 20 in PBS for 1 hour at room temperature in the dark, with shaking. After washing with 0.05% Tween 20–PBS for 15 minutes and with PBS for 15 minutes, strips were incubated with chromogenic solution (6 mg 4-chloro-1-naphthol in 2 ml methanol, 10 ml PBS, and 20 ml 30% H₂O₂). The reaction was stopped with water.

Statistical analysis.

Categorical variables, such as sex, age group, and ANA–HEp-2 IIF pattern, were analyzed by the chi-square test. The Mann-Whitney U test was used to compare ANA–HEp-2 titers between groups. The ANA–HEp-2 titer in samples collected 4 years apart was analyzed by Wilcoxon's paired sample test. Receiver operating characteristic (ROC) curve analysis was used to evaluate the accuracy of the ANA–HEp-2 test in distinguishing healthy individuals and patients with ARDs. All data were analyzed using Excel Microsoft 2007 and SPSS for Windows 15.0. P values less than 0.05 were considered significant.

RESULTS

A positive ANA–HEp-2 test result was observed in 118 healthy individuals (12.9%), with no difference in ANA prevalence according to sex (13.8% in women versus 10.5% in men; P = 0.2) or age (P = 0.43). However, there was a trend toward higher frequency of positive ANA–HEp-2 test results in individuals ages 51–66 years (20.8%; n = 48) compared with individuals ages 18–30 years (13.4%; n = 461), individuals ages 31–40 years (11.0%; n = 236), and individuals ages 41–50 years (12.9%; n = 154). There was a preponderance of low-titer reactivity among the ANA–positive healthy individuals. This was in frank contrast to samples from patients with ARDs, which had a 90.2% frequency of positive test results and a skewed distribution toward high-titer ANAs (Figure 1A and Table 1). However, it should be emphasized that high-titer ANA reactivity was also observed in a sizable fraction of healthy individuals (Table 1). The frequency of positive ANA–HEp-2 test results was 96.5% in SLE, 88.8% in SSc, 70% in SS, and 63.6% in polymyositis/dermatomyositis.

Figure 1. A and B, Distribution of antinuclear antibody (ANA)–positive healthy individuals (HI) and patients with autoimmune rheumatic disease (ARD) according to titer for all patterns combined on the ANA–HEp-2 test (A) and for the nuclear fine speckled pattern on the ANA–HEp-2 test (B). Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the 10th and 90th percentiles. + indicates outliers. C, Receiver operating characteristic (ROC) curve analysis of the ANA–HEp-2 test for discriminating 918 healthy individuals and 153 patients with ARDs. The area under the curve = 0.923 (P < 0.001).

Table 1. Distribution of ANA-positive healthy individuals and ANA-positive patients with ARDs according to the titer and prevalent patterns on the ANA–HEp-2 test*

Titer	All patterns			Nuclear fine speckled pattern			Nuclear dense fine speckled pattern (healthy individuals only) (n = 39)
	Healthy individuals (n = 118)	Patients with ARDs (n = 138)	P†	Healthy individuals (n = 54)	Patients with ARDs (n = 58)	P†
* Values are the number (%) of subjects. ANA = antinuclear antibody; ARDs = autoimmune rheumatic diseases. † By chi-square test.
1:80	54 (45.8)	10 (7.2)	<0.001	33 (61.1)	9 (15.5)	<0.001	10 (25.6)
1:160	9 (7.6)	5 (3.6)	0.259	5 (9.3)	5 (8.6)	0.831	1 (2.6)
1:320	15 (12.7)	23 (16.7)	0.477	8 (14.8)	12 (20.7)	0.573	4 (10.3)
1:640	21 (17.8)	0 (0)	<0.001	5 (9.3)	0 (0)	0.056	15 (38.5)
1:1,280	9 (7.6)	31 (22.5)	0.002	0 (0)	11 (19.0)	0.02	5 (12.8)
1:2,560	3 (2.5)	1 (0.7)	0.507	1 (1.9)	1 (1.7)	0.515	1 (2.6)
≥1:5,120	7 (5.9)	68 (49.3)	<0.001	2 (3.7)	20 (34.5)	<0.001	3 (7.7)

ROC curve analysis showed good performance of the ANA–HEp-2 test for discriminating patients with ARDs and healthy individuals, with an area under the curve of 0.923 (P < 0.001) (Figure 1C). As expected, the ability to discriminate patients with ARDs and healthy individuals varied according to the dilution cutoff level. At a screening dilution of 1:80, the ANA–HEp-2 test had sensitivity of 90.2% and specificity of 87.1% with a high (98.1%) negative predictive value (NPV). At a dilution of 1:160 (recommended by the ANA Subcommittee of the International Union of Immunological Societies Standardization Committee [6]), the sensitivity was 83.7% and the specificity was 93.0%. At the 1:5,120 dilution cutoff level, the test had sensitivity of 44.4% and specificity of 99.2% with a moderate (90.6%) positive predictive value (PPV). At an intermediate 1:1,280 dilution cutoff level, the test had sensitivity of 65.4% and specificity of 97.9% with a high (94.6%) NPV and a moderate (84.9%) PPV.

The ANA pattern distribution among the 118 ANA-positive healthy individuals and 138 ANA-positive patients with ARDs is shown in Table 2. It should be noted that some samples from both groups had more than 1 pattern (multipattern). The morphologic characteristics of the most relevant nuclear patterns are shown in Figure 2. The nuclear homogeneous pattern was characterized by smooth texture staining of the whole interphase nucleus and bright hyaline staining of the metaphase chromatin plate. The nuclear quasihomogeneous pattern was defined by an extremely fine grainy texture staining of the whole interphase nucleus with similar staining of the metaphase chromatin plate. The nuclear fine speckled pattern was defined by a variably grainy texture staining of the interphase nucleus and no staining of the metaphase chromatin plate, usually sparing the nucleolar domains. The nuclear coarse speckled pattern was characterized by several larger and brighter speckles standing out against a coarse grainy texture staining of the whole interphase nucleus with no staining of the metaphase chromatin plate. The nuclear dense fine speckled pattern was characterized by tightly packed speckled staining of the whole interphase nucleus with heterogeneity in brightness of the several speckles and similar staining pattern in the metaphase chromatin plate. The centromere pattern was characterized by numerous discrete speckles spread throughout the interphase nucleus and aligned in an orderly manner at the metaphase chromatin plate.

Table 2. Distribution of 118 ANA-positive healthy individuals and 138 ANA-positive patients with ARDs according to the pattern on the ANA–HEp-2 test*

Pattern on ANA–HEp-2 test	ANA-positive healthy individuals†	ANA-positive patients with ARDs†	P‡
* Values are the number (%) of subjects. See Table 1 for definitions. † Total percentage in each column is >100% because some samples had >1 pattern. ‡ By chi-square test. § Exclusively cytoplasmic staining. ¶ One patient with exclusively cytoplasmic staining and 4 patients with additional patterns (2 with nucleolar pattern, 1 with nuclear fine speckled pattern, and 1 with nuclear coarse speckled pattern).
Nuclear fine speckled	54 (45.8)	58 (42.0)	0.636
Nuclear dense fine speckled	39 (33.1)	0 (0)	<0.001
Nuclear coarse speckled	0 (0)	36 (26.1)	<0.001
Nuclear homogeneous	0 (0)	10 (7.2)	0.008
Nuclear centromeric	0 (0)	11 (8.0)	0.005
Nuclear quasihomogeneous speckled	5 (4.2)	19 (13.8)	0.017
Nucleolar	8 (6.8)	18 (13.0)	0.148
Cytoplasmic	12 (10.2)§	5 (3.6)¶	0.065
Other	8 (6.8)	7 (5.1)	0.604

Figure 2. Morphologic characteristics of the most relevant nuclear patterns found on indirect immunofluorescence assay on HEp-2 cells, performed on samples from healthy individuals and patients with autoimmune rheumatic diseases. Patterns are as follows: A, homogeneous; B, quasihomogeneous; C, nuclear fine speckled; D, nuclear coarse speckled; E, nuclear dense fine speckled; and F, nuclear centromeric. Sera were diluted 1:80. Bars = 5 μm in A and B; 13 μm in C–F.

The most frequent ANA patterns in healthy individuals were the nuclear fine speckled pattern, which was present in 54 subjects (45.8% of ANA-positive healthy individuals), and the nuclear dense fine speckled pattern, which was observed in 39 subjects (33.1% of ANA-positive healthy individuals). Interestingly, the frequency of the nuclear fine speckled pattern was stable across 4 age strata (43.5% in those ages 18–30 years, 38.5% in those ages 31–40 years, 50% in those ages 41–50 years, and 50% in those ages 51–65 years), but the frequency of the nuclear dense fine speckled pattern dropped significantly in healthy individuals >50 years of age (32.3%, 42.3%, 35.5%, and 10%, respectively). The most frequent ANA pattern in patients with ARDs was also the nuclear fine speckled pattern, which was present in 58 subjects (42.0% of ANA-positive patients with ARDs), followed by the nuclear coarse speckled pattern, which was observed in 36 patients (26.1% of ANA-positive patients with ARDs). No healthy individual had the nuclear coarse speckled pattern, and no patient with ARD had the nuclear dense fine speckled pattern. Other patterns exclusively observed in patients with ARDs were the nuclear homogeneous (7.2%), nuclear centromeric (8.0%), and cytoplasmic dense fine speckled (2.8%) patterns.

The nuclear quasihomogeneous pattern (Figure 2) was predominantly observed in samples from patients with ARDs. This pattern is distinct from the nuclear homogeneous pattern in that its texture is finely grainy as opposed to the plain smooth texture of the homogeneous pattern. This difference is particularly appreciated at the chromatin mass in mitotic cells, where the homogeneous pattern assumes a hyaline appearance, in contrast to the quasihomogeneous pattern, which keeps its finely grainy texture. In addition, samples with the quasihomogeneous pattern had a different autoantibody profile from the homogeneous pattern. No quasihomogeneous pattern sample from patients with ARDs or from healthy individuals was reactive against native DNA. Furthermore, no quasihomogeneous pattern sample from healthy individuals reacted with nucleosome or histone, and only 60% of the quasihomogeneous samples from SLE patients reacted with nucleosome and/or histone. In contrast, all 10 samples with the homogeneous pattern reacted with components of the chromatin system. In fact, all reacted with nucleosome: 3 reacted only with nucleosome, 4 reacted with nucleosome and histone, 1 reacted with nucleosome and native DNA, and 2 reacted with nucleosome, histone, and native DNA.

Since the nuclear fine speckled pattern was the most frequently observed pattern in healthy individuals and in patients with ARDs, we decided to analyze the titer of the nuclear fine speckled pattern in both groups. Healthy individuals presented a significantly lower titer nuclear fine speckled pattern than did patients with ARDs (P < 0.01) (Figure 1B). On the other hand, the nuclear dense fine speckled pattern was observed exclusively in healthy individuals and mostly at high titer (Table 1). In fact, in healthy individuals, samples with the nuclear dense fine speckled pattern had higher titers than did samples with the nuclear fine speckled pattern. The nuclear fine speckled pattern occurred at a titer ≤1:320 in 46 of the 54 healthy individuals with that pattern (85.2%), and the nuclear dense fine speckled pattern occurred at a titer ≥1:640 in 24 of the 39 healthy individuals with that pattern (61.5%) (P < 0.001). The antigenic specificity of the 2 ANA patterns observed most frequently in healthy individuals, the nuclear fine speckled pattern and the nuclear dense fine speckled pattern, was investigated by Western blotting. As expected, all samples displaying the nuclear dense fine speckled pattern recognized a polypeptide with an estimated molecular weight of 75 kd (18, 19, 22) (Figure 3). On the other hand, no specific band was observed among the samples displaying the nuclear fine speckled pattern (Figure 3).

Figure 3. Western blot analysis of antibody specificity associated with nuclear dense fine speckled pattern and nuclear fine speckled pattern in sera from healthy individuals. Whole HEp-2 cell extract was separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and probed with sera diluted at a ratio of 1:50. Lanes 1–18, Samples that showed the nuclear dense fine speckled pattern on the antinuclear antibody (ANA)–HEp-2 test; lane 19, anti-SSA/Ro control; lanes 20–29, samples that showed the nuclear fine speckled pattern on the ANA–HEp-2 test; lane 30, negative control. Arrows indicate proteins migrating with mobility compatible with 52 kd and 75 kd.

Another distinctive feature differentiating ANA-positive samples from healthy individuals and ANA-positive samples from patients with ARDs was the presence of anti-ENA antibodies in the samples from patients with ARDs. Fifty-two samples from ANA-positive patients with ARDs (37.7%) had anti-ENA reactivity (35 with anti-SSA/Ro [25.4%], 13 with anti-SSB/La [9.4%], 3 with anti-Sm [2.2%], and 20 with anti–U1 RNP [14.5%]). As expected, some samples had more than 1 antibody specificity. The frequency of anti-ENA antibodies in each ARD was 42.5% in SLE, 13.3% in SSc, 50% in SS, and 36.3% in polymyositis/dermatomyositis. In contrast, only 1 ANA-positive healthy individual (0.85%) had anti-ENA antibody (anti-SSA/Ro). As expected, anti-DNA antibody was absent in healthy individuals and exclusively found in 12 SLE patients.

We were able to reassess 41 of the ANA-positive healthy individuals 3.6–5 years after the initial evaluation. None had symptoms suggestive of an ARD, except for 1 woman who presented with mild hand pallor after exposure to cold; however, the clinical examination and nailfold capillaroscopy performed at the reassessment yielded normal findings. Among the 40 ANA-positive healthy individuals who agreed to donate blood at the reassessment, there was slight variation in titer in both directions, as shown in Figure 4, with no significant variation in ANA pattern between paired samples. Twenty-nine healthy individuals (72.5%) remained ANA positive without significant variation in ANA titer (P = 0.795). All 11 ANA-positive healthy individuals who became ANA negative had significantly lower initial titer (1:80) than those who remained ANA positive (P = 0.002).

Figure 4. Longitudinal findings from antinuclear antibody (ANA)–HEp-2 testing in 40 ANA-positive healthy individuals at the first evaluation and at reassessment within an average 4-year interval. Solid lines indicate stable or increased titer; dashed lines indicate decreased titer; Ø indicates negative.

DISCUSSION

Taking into consideration that the ANA–HEp-2 test yields positive results in a sizable portion of the general population, it is critical for the accurate interpretation of the test that we uncover differential characteristics of the results in healthy individuals and in patients with ARDs. In the present study, we were able to identify distinctive ANA profiles in healthy individuals and patients with ARDs by analyzing samples from 918 healthy individuals and 153 patients with ARDs. Positive ANA–HEp-2 test results in healthy individuals involved predominantly low-to-moderate titers, and the most common ANA patterns were the nuclear fine speckled pattern and the nuclear dense fine speckled pattern. In contrast, positive ANA–HEp-2 test results in patients with ARDs involved predominantly moderate-to-high titers and exhibited a distinct ANA profile characterized mainly by the absence of the nuclear dense fine speckled pattern and the exclusive occurrence of the nuclear coarse speckled, nuclear homogeneous, and nuclear centromeric patterns.

The frequency of positive ANA–HEp-2 test results in the present series of healthy individuals (12.9%) was within the range established by previous studies (3, 6–9). Several studies reported in the literature found a higher prevalence of ANAs in healthy women than in healthy men (3–5, 7, 29). In fact, sex-related factors, such as the hormone profile, fetal microchimerism, and strategic genes located at sex chromosomes, may be involved in the female predominance in autoimmune diseases (30). Interestingly, we could not find any difference in the frequency of positive ANA–HEp-2 test results between women and men despite the substantial sample size for both sexes. We observed a higher prevalence of ANAs in elderly healthy individuals, as shown by previous studies in nonagenarians and centenarians (31, 32). However, the difference did not reach statistical significance, which may be related to the limited number of healthy individuals older than 65 years in the present study.

Our study indicates that ANA titers of 1:80 and 1:160 are suitable dilutions for screening for ARDs. The high NPV for ARD diagnosis found in the present series is consistent with findings of Slater et al, who evaluated 1,010 consecutive samples in which ANA–HEp-2 testing was ordered (33). These results indicate that a negative ANA–HEp-2 test result at 1:80 screening dilution is unlikely in patients with ARDs, especially those with SLE, SSc, and SS. The observed predominance of low-titer ANAs (≤1:160) among healthy individuals is also consistent with previous reports (2, 6–8). On the other hand, we did find a substantial fraction of healthy individuals with a moderate titer (1:320 and 1:640) and a smaller fraction with a high titer (≥1:1,280). ROC curve analysis and traditional diagnostic performance parameters (sensitivity, specificity, and predictive values) showed that in a scenario of indiscriminate requests for ANA–HEp-2 tests, there was relevant gain in specificity and PPV for discriminating healthy individuals and patients with ARDs up to a titer of 1:1,280. Therefore, although the 1:80 and 1:160 dilutions are appropriate for screening, it is advisable that clinical laboratories proceed with testing up to a 1:1,280 dilution. Altogether, these findings suggest that the titer of the ANA–HEp-2 reaction is a relevant but limited parameter in discriminating ANA-positive healthy individuals and patients with ARDs.

This limitation may be partially overcome by the analysis of the pattern on the ANA–HEp-2 test. In fact, our findings suggest that this pattern is a critical parameter for the interpretation of a positive ANA–HEp-2 test result. In the present study, the pattern on the ANA–HEp-2 test was comparatively scrutinized in samples from patients with ARDs and healthy individuals. Some patterns, such as the nuclear coarse speckled, nuclear homogeneous, and nuclear centromeric patterns, were observed solely in samples from patients with ARDs. This is probably because these patterns are typically associated with disease-restricted autoantibodies, such as anti-Sm/anti–U1 RNP, anti–double-stranded DNA/antinucleosome, and anticentromere antibodies, respectively (34–37). On the other hand, the nuclear dense fine speckled pattern was not observed in any sample from a patient with an ARD, but was found exclusively in samples from healthy individuals. Confirming previous reports, all tested samples with a nuclear dense fine speckled pattern in this series reacted with a 75-kd protein compatible with the LEDGF/p75 antigen (8, 18, 19).

There are not many studies focusing on the pattern on the ANA–HEp-2 test in healthy individuals, and the majority of them were performed before the characterization of the nuclear dense fine speckled pattern (2, 4, 5, 38–40). Given the high prevalence of the nuclear dense fine speckled pattern among healthy individuals, it is reasonable to assume that this pattern might have been mistaken for some other pattern in those studies. In the present series, the nuclear dense fine speckled pattern was the second most frequent ANA pattern among healthy individuals. Of special importance, the majority of samples with this pattern had a high ANA titer, which might mislead the clinical judgment of physicians not acquainted with the ANA–HEp-2 test. Similar findings have been previously reported by Watanabe et al among 597 healthy hospital workers in Japan (8) and by our group in a study with 13,641 samples determined to be positive by an ANA–HEp-2 test performed in a Brazilian general clinical laboratory that was certified on-site by the US College of American Pathologists (19). However, it should be noted that the nuclear dense fine speckled pattern and anti-LEDGF/p75 antibodies have also been observed in patients with a variety of unrelated conditions, such as interstitial cystitis, asthma, Hashimoto thyroiditis, and atopic dermatitis (18, 19, 21, 22).

The combined analysis of pattern on the ANA–HEp-2 test and ANA titer was especially relevant for tests showing the nuclear fine speckled pattern. This was the most frequent ANA pattern in healthy individuals and in patients with ARDs, and our results indicated that a low-titer nuclear fine speckled pattern was preferentially observed in healthy individuals, whereas a high-titer nuclear fine speckled pattern was preferentially observed in patients with ARDs. Interestingly, the samples with a nuclear fine speckled pattern from healthy individuals failed to elicit any reactivity on Western blotting, and this may indicate some peculiar features of the autoantigen(s) associated with this pattern in healthy subjects. One other useful element in the interpretation of a positive ANA–HEp-2 test result with the nuclear fine speckled pattern is the presence of antibodies to ENAs. In the present series, antibodies to SSA/Ro and SSB/La were largely restricted to samples with the nuclear fine speckled pattern from patients with ARDs. In contrast, only 1 healthy individual had anti-SSA/Ro antibody. In addition, antibodies to U1 RNP and Sm, which are usually associated with the nuclear coarse speckled pattern, were observed solely in samples from patients with ARDs.

At the followup visit, none of the 41 ANA-positive healthy individuals had developed signs and symptoms of ARDs, and 72.5% of the 40 individuals who agreed to donate blood at the reassessment remained ANA positive after a mean followup period of 3.9 years. Our findings also indicate that those with low-titer ANAs on the ANA–HEp-2 test are the ones with the highest probability of future normalization of the result. The extended persistence of ANA positivity in healthy individuals has been previously reported by Xavier et al, who found that 78% of 23 elderly healthy individuals continued to have a positive ANA–HEp-2 test result after an average followup period of 4 years (40). In another study, Dinser et al also found a low PPV for a positive ANA–HEp-2 test result at titers >1:320 for the development of ARD after a 3-year followup period (41). Therefore, a positive ANA–HEp-2 test result in apparently healthy individuals seems to be a sustained phenomenon in ∼75% of the cases and appears to have a low PPV for the development of a future ARD. However, this assumption should be made with caution, because the followup in the present study included only samples (from healthy individuals) with patterns on the ANA–HEp-2 test shown not to be associated with ARD. In the unusual scenario of a pattern on the ANA–HEp-2 test associated with ARD (e.g., nuclear homogeneous or nuclear coarse speckled) in an apparently healthy individual, a careful followup is advised, since it has been demonstrated that specific autoantibodies may antedate signs and symptoms of SLE for years (10).

In conclusion, the ANA–HEp-2 assay offered distinctive titer and pattern profiles for ANA-positive healthy individuals and for patients with ARDs. The ANA pattern on the test seemed to be more consistent than the ANA titer for discriminating ANA-positive healthy individuals and patients with ARDs. The nuclear dense fine speckled pattern was observed exclusively in healthy individuals and tended to appear at high titer and to be stable over the years. Some other patterns on the ANA–HEp-2 test (nuclear coarse speckled, nuclear homogeneous, nuclear centromeric) were restricted to patients with ARDs. The nuclear fine speckled pattern was equally frequent in both groups and showed different titer distributions in healthy individuals (predominantly low titer) and patients with ARDs (predominantly high titer). ANA-positive healthy individuals tended to keep ANA reactivity but did not develop evidence of ARD after a 4-year followup period. Given the critical role of the pattern on the ANA–HEp-2 test, future efforts shall address the reproducibility of ANA–HEp-2 test interpretation among different ANA experts and among different ANA–HEp-2 test slide brands.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Andrade had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Mariz, Sato, Andrade.

Acquisition of data. Mariz, Barbosa, Rodrigues, Dellavance, Andrade.

Analysis and interpretation of data. Mariz, Sato, Dellavance, Andrade.

Acknowledgements

We acknowledge Gilda Ferreira, MD, PhD, for providing some of the serum samples for the group of patients with ARDs, Valdecir Marvoulle, PhD, for expert advice with statistics, and Edward Chan, PhD, and Eng M. Tan, MD, for helpful comments.