Connective tissue disorders (CTDs) comprise a heterogeneous group of systemic autoimmune diseases characterized by immune-mediated inflammation affecting multiple organ systems. Major entities include systemic lupus erythematosus (SLE), systemic sclerosis (SSc), Sjögren’s syndrome (SS), idiopathic inflammatory myopathies (IIM), and mixed connective tissue disease (MCTD) [1,2]. These disorders often present with nonspecific and overlapping symptoms such as fatigue, arthralgia, Raynaud phenomenon, mucocutaneous manifestations, sicca symptoms, or mild cytopenias, making early diagnosis challenging [3,4].

Serologic testing for autoimmune markers forms a cornerstone of diagnostic evaluation in suspected CTD. Antinuclear antibodies (ANA) are widely used as a screening tool due to their high sensitivity in diseases such as SLE and systemic sclerosis [5]. More specific autoantibodies, including anti–double-stranded DNA (anti-dsDNA), anti-Smith (anti-Sm), anti-U1-ribonucleoprotein (anti-RNP), anti-Ro/SSA, anti-La/SSB, anti-Scl-70, and anti-centromere antibodies, contribute to disease classification and prognostication [6,7]. Additionally, rheumatoid factor (RF), anti-cyclic citrullinated peptide (anti-CCP), anti-neutrophil cytoplasmic antibodies (ANCA), and complement levels (C3, C4) are frequently assessed in the evaluation of systemic autoimmune conditions [8].

However, the interpretation of autoimmune markers is complicated by their presence in individuals without definitive autoimmune disease. Low-titer ANA positivity has been reported in 20–30% of healthy individuals, particularly in women and the elderly [9,10]. Furthermore, transient autoantibody positivity may occur in infections, malignancy, and other inflammatory states [11]. This creates diagnostic uncertainty, especially in patients presenting with nonspecific symptoms who do not fulfill established classification criteria at initial evaluation.

A subset of patients presenting with suggestive clinical features but lacking sufficient criteria for classification is often labeled as having “undifferentiated connective tissue disease” (UCTD) or remains formally undiagnosed [12]. Longitudinal studies suggest that a proportion of such individuals eventually evolve into classifiable CTDs, whereas others remain stable or experience symptom resolution [13,14]. Identifying which patients are at higher risk of progression remains a key clinical challenge.

Although autoimmune markers are frequently ordered in this diagnostic grey zone, their true prevalence and predictive value in undiagnosed suspected CTD populations have not been comprehensively synthesized. Previous studies have focused primarily on established disease cohorts or diagnostic performance in confirmed cases rather than on individuals without a definitive diagnosis [15,16]. As a result, clinicians lack clear evidence regarding the utility of broad serologic screening in undiagnosed populations.

A systematic review and meta-analysis focusing specifically on adults evaluated for suspected but initially undiagnosed CTD is therefore warranted. Quantifying pooled seroprevalence and evaluating diagnostic and prognostic accuracy of these markers for subsequent disease development may help refine testing strategies, reduce overdiagnosis, and improve cost-effective clinical decision-making.

The present study aims to (1) estimate the pooled prevalence of autoimmune marker positivity in adults with suspected but undiagnosed connective tissue disorders, and (2) determine the diagnostic performance of these markers in predicting subsequent confirmed CTD during longitudinal follow-up.

METHODOLOGY

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines [17]. For studies reporting diagnostic accuracy outcomes, recommendations from PRISMA-DTA were additionally followed [18]. The protocol was developed a priori to minimize selection bias and ensure methodological transparency.

Eligibility criteria were defined using a modified PICOS framework. We included studies enrolling adults (≥18 years) evaluated for suspected connective tissue disorder (CTD), defined as individuals presenting with clinical features suggestive of systemic autoimmune disease—such as arthralgia, Raynaud phenomenon, mucocutaneous manifestations, sicca symptoms, inflammatory myopathy symptoms, or unexplained cytopenias—but without a definitive CTD diagnosis at baseline. Studies exclusively involving established CTD cases were excluded unless data for undiagnosed participants were reported separately. Eligible index tests included antinuclear antibodies (ANA), extractable nuclear antigen (ENA) antibodies (anti-Ro/SSA, anti-La/SSB, anti-Sm, anti-RNP, anti-Scl-70, anti-Jo-1), anti–double-stranded DNA (anti-dsDNA), rheumatoid factor (RF), anti-cyclic citrullinated peptide (anti-CCP), anti-neutrophil cytoplasmic antibodies (ANCA; MPO and PR3), and complement levels (C3 and C4). No restrictions were placed on assay methodology, including indirect immunofluorescence, ELISA, or multiplex immunoassay; however, positivity thresholds were recorded. For diagnostic accuracy analyses, the reference standard was a confirmed CTD diagnosis during follow-up according to recognized classification criteria (e.g., ACR/EULAR) or diagnosis established by a qualified rheumatologist.

Primary outcomes included pooled baseline seroprevalence of autoimmune markers and pooled sensitivity and specificity for predicting subsequent CTD diagnosis. Secondary outcomes included positive and negative predictive values, likelihood ratios, time to diagnosis, and association between antibody titer and disease progression. Eligible study designs comprised prospective and retrospective cohort studies, diagnostic accuracy studies, and cross-sectional studies with longitudinal follow-up. Case reports, case series with fewer than 10 participants, pediatric-only populations, and review articles were excluded.

A comprehensive literature search was conducted in MEDLINE (via PubMed), Embase, Web of Science, and Cochrane CENTRAL from database inception to December 2025, without language restrictions. The search strategy combined controlled vocabulary and free-text terms related to connective tissue diseases, autoimmune markers, and undiagnosed or suspected disease states. Reference lists of included studies and relevant reviews were manually screened to identify additional studies.

All identified records were imported into reference management software and duplicates were removed. Two reviewers independently screened titles and abstracts for eligibility, followed by full-text review of potentially relevant articles. Disagreements were resolved through discussion or consultation with a third reviewer. Data extraction was performed independently by two reviewers using a standardized and piloted extraction form. Extracted variables included study characteristics (author, year, country, design, clinical setting), participant demographics (sample size, mean age, sex distribution), presenting clinical features, type and methodology of autoimmune testing, positivity thresholds, baseline seroprevalence, duration of follow-up, number of participants progressing to confirmed CTD, and raw 2×2 contingency data when available. Corresponding authors were contacted for clarification or missing data when necessary.

Risk of bias for diagnostic accuracy studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool [19], while cohort and prevalence studies were evaluated using the Newcastle–Ottawa Scale (NOS) [20]. Domains assessed included patient selection, index test conduct and interpretation, outcome ascertainment, follow-up adequacy, and selective reporting.

Statistical analyses were performed using R. Pooled seroprevalence estimates were calculated using a random-effects model based on the DerSimonian–Laird method [21], with logit transformation of proportions to stabilize variance. Heterogeneity was assessed using Cochran’s Q test and quantified using the I² statistic, with values above 50% considered indicative of substantial heterogeneity [22]. For diagnostic accuracy outcomes, pooled sensitivity and specificity were estimated using a bivariate random-effects model [23], and summary receiver operating characteristic (SROC) curves were constructed. Positive and negative likelihood ratios were calculated from pooled sensitivity and specificity estimates. Subgroup analyses were conducted based on ANA titer thresholds, assay methodology, clinical setting, and duration of follow-up when sufficient studies were available. Publication bias was assessed using funnel plot asymmetry and Egger’s regression test for prevalence analyses [24], and Deeks’ funnel plot asymmetry test for diagnostic accuracy studies [25]. The overall certainty of evidence was evaluated using a modified GRADE approach adapted for diagnostic test accuracy research [26].

RESULTS

Study Selection

The electronic database search identified 1,284 records. After removal of 312 duplicates, 972 titles and abstracts were screened. Of these, 146 articles underwent full-text review. Thirty-two studies met inclusion criteria for qualitative synthesis, and 24 studies comprising 6,842 participants were included in quantitative meta-analysis. The study selection process is summarized in the PRISMA flow diagram.

Figure 1. PRISMA 2020 flow diagram of study selection process. Adapted from the PRISMA 2020 statement: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71.

Study Characteristics

The 32 included studies were published between 2001 and 2024 and were conducted across Europe (n=12), Asia (n=9), North America (n=7), and other regions (n=4). Most studies were prospective cohort designs (n=18), followed by retrospective cohorts (n=10) and diagnostic accuracy studies (n=4). Sample sizes ranged from 54 to 612 participants.

The mean age of participants ranged from 32 to 58 years, with a female predominance (range: 61–84%). The majority of studies were conducted in tertiary rheumatology clinics, while 6 studies were performed in secondary care settings.

Follow-up duration ranged from 12 months to 8 years (median 36 months). During follow-up, 1,124 of 6,842 participants (16.4%) developed a confirmed connective tissue disorder.

Risk of Bias Assessment

Using QUADAS-2 and the Newcastle–Ottawa Scale, 19 studies were classified as low risk of bias, 9 as moderate risk, and 4 as high risk. The most common sources of bias were referral bias due to tertiary care recruitment and lack of standardized antibody titer thresholds. Outcome ascertainment was generally robust, as most studies applied established classification criteria for CTD diagnosis.

Pooled Prevalence of Autoimmune Marker Positivity

Antinuclear Antibody (ANA)

Twenty-one studies (n=5,934 participants) reported baseline ANA positivity. The pooled prevalence of ANA positivity was 38% (95% CI: 32–44%), with substantial heterogeneity (I² = 78%).

Subgroup analysis demonstrated that ANA positivity was higher in tertiary care cohorts (44%) compared with secondary care settings (26%).

High-titer ANA (>1:320) was reported in 12 studies, with a pooled prevalence of 14% (95% CI: 10–18%).

Extractable Nuclear Antigen (ENA) Antibodies

Fifteen studies reported ENA panel positivity. The pooled prevalence of any ENA antibody was 18% (95% CI: 14–23%; I² = 65%).

Among specific antibodies:

• Anti-SSA/Ro: 9% (95% CI: 6–12%)
• Anti-RNP: 6% (95% CI: 4–9%)
• Anti-Scl-70: 4% (95% CI: 2–6%)
• Anti-Sm: 3% (95% CI: 2–5%)

Anti-dsDNA

Ten studies evaluated anti-dsDNA antibodies, yielding a pooled prevalence of 7% (95% CI: 4–10%; I² = 52%).

Rheumatoid Factor and Anti-CCP

RF positivity was reported in 13 studies, with a pooled prevalence of 15% (95% CI: 11–20%; I² = 60%). Anti-CCP positivity was lower at 6% (95% CI: 4–9%).

ANCA

Eight studies assessed ANCA, demonstrating a pooled prevalence of 5% (95% CI: 3–8%), with moderate heterogeneity (I² = 49%).

Progression to Confirmed Connective Tissue Disease

Across included studies, 16.4% of initially undiagnosed individuals developed a confirmed CTD during follow-up. The most common subsequent diagnoses were:

• Systemic lupus erythematosus (34%)
• Sjögren’s syndrome (22%)
• Systemic sclerosis (18%)
• Mixed connective tissue disease (11%)
• Idiopathic inflammatory myopathies (9%)
• Other CTDs (6%)

Longer follow-up duration (>3 years) was associated with higher cumulative diagnosis rates (pooled progression 21% vs 11% for ≤2 years).

Diagnostic Performance of Autoimmune Markers

ANA

Seventeen studies provided 2×2 data for ANA in predicting future CTD diagnosis. The pooled sensitivity was 82% (95% CI: 75–88%), and pooled specificity was 55% (95% CI: 47–63%).

The positive likelihood ratio (LR+) was 1.82, and the negative likelihood ratio (LR−) was 0.33.

High-titer ANA (>1:320) demonstrated improved specificity (78%) but reduced sensitivity (61%).

Anti-SSA/Ro

Anti-SSA positivity demonstrated pooled sensitivity of 41% (95% CI: 33–50%) and specificity of 88% (95% CI: 82–92%).

LR+ was 3.42 and LR− was 0.67.

Anti-dsDNA

Anti-dsDNA showed pooled sensitivity of 29% (95% CI: 21–38%) and specificity of 94% (95% CI: 89–97%).

LR+ was 4.83, indicating stronger rule-in value compared to ANA.

Anti-RNP

Anti-RNP demonstrated moderate specificity (90%) but low sensitivity (24%).

Summary Receiver Operating Characteristic (SROC) Analysis

Hierarchical SROC analysis demonstrated that ANA had moderate overall discriminative ability, with area under the curve (AUC) of 0.74. Anti-dsDNA demonstrated higher rule-in performance (AUC 0.81), though sensitivity remained limited.

Subgroup and Meta-Regression Analyses

Meta-regression identified ANA titer threshold (p=0.02) and clinical setting (p=0.03) as significant contributors to heterogeneity. Assay method (IFA vs ELISA) did not significantly affect pooled sensitivity but influenced specificity estimates.

Publication Bias

Funnel plot asymmetry was observed for ANA prevalence studies. Egger’s regression test indicated mild publication bias (p=0.04). Deeks’ funnel plot for diagnostic accuracy studies did not show significant asymmetry (p=0.18).

Table 1. Characteristics of Included Studies

ANA = Antinuclear antibody; ENA = Extractable nuclear antigen; RF = Rheumatoid factor; anti-CCP = Anti-cyclic citrullinated peptide; ANCA = Anti-neutrophil cytoplasmic antibody; IFA = Indirect immunofluorescence assay; ELISA = Enzyme-linked immunosorbent assay; CTD = Connective tissue disorder.

Table 2. Pooled Prevalence of Autoimmune Marker Positivity

Table 3. Diagnostic Performance of Autoimmune Markers for Predicting Future CTD

Values in parentheses represent 95% confidence intervals.

Table 4. Subgroup Analyses for ANA

Table 5. Risk of Bias Assessment Summary

Table 6. Progression to Specific CTD Diagnoses

Figure 2. Forest plot showing pooled prevalence of antinuclear antibody (ANA) positivity in adults with suspected but initially undiagnosed connective tissue disorders. Individual study estimates are presented with 95% confidence intervals. The vertical line represents the pooled random-effects estimate. Heterogeneity was assessed using the I² statistic.

Figure 3. Forest plot of pooled sensitivity estimates for selected autoimmune markers in predicting subsequent confirmed connective tissue disorder. Squares represent individual study estimates with 95% confidence intervals. The pooled estimate was calculated using a bivariate random-effects model.

Figure 4. Summary receiver operating characteristic (SROC) curve illustrating the overall diagnostic performance of autoimmune markers for predicting progression to confirmed connective tissue disorder. Each point represents an individual study, and the curve reflects the hierarchical bivariate model. The area under the curve (AUC) indicates overall discriminative ability.

DISCUSSION

This systematic review and meta-analysis synthesizes current evidence regarding the prevalence and prognostic value of autoimmune serologic markers in adults presenting with suspected but initially undiagnosed connective tissue disorders (CTDs). The findings demonstrate that while seropositivity—particularly ANA—is common in this population, its specificity for predicting future CTD diagnosis is limited. Conversely, disease-specific antibodies such as anti-dsDNA and anti-SSA exhibit stronger rule-in value but lower sensitivity. These results underscore the complexity of interpreting autoimmune markers in the diagnostic grey zone between health and classifiable autoimmune disease.

ANA positivity was observed in approximately one-third of undiagnosed individuals, consistent with prior population-based studies reporting background ANA prevalence ranging from 20–30%, particularly among women and older adults [27,28]. However, our pooled specificity of 55% for ANA in predicting subsequent CTD confirms its limited discriminative ability in isolation. This aligns with previous diagnostic evaluations demonstrating that ANA serves primarily as a sensitive screening tool rather than a confirmatory test [29]. Importantly, high-titer ANA (>1:320) demonstrated improved specificity, suggesting that quantitative thresholds may offer incremental prognostic value.

The progression rate of 16.4% from undiagnosed suspected CTD to confirmed disease during follow-up is clinically meaningful. This figure is comparable to longitudinal studies of undifferentiated connective tissue disease (UCTD), where progression rates between 10–30% have been reported over 3–5 years [30,31]. The heterogeneity observed across studies likely reflects differences in referral patterns, baseline symptom burden, and follow-up duration. Notably, longer follow-up was associated with higher cumulative diagnosis rates, reinforcing the importance of longitudinal surveillance in high-risk patients.

Disease-specific autoantibodies demonstrated superior specificity. Anti-dsDNA and anti-SSA antibodies showed specificities exceeding 85%, consistent with their established role in SLE and Sjögren’s syndrome classification criteria [32,33]. However, their limited sensitivity suggests that absence of these antibodies does not reliably exclude future CTD development. This pattern reflects the broader principle that systemic autoimmune diseases are immunologically heterogeneous and may evolve over time, with autoantibody profiles expanding prior to overt clinical classification [34].

The moderate area under the SROC curve for ANA (AUC 0.74) further highlights its intermediate predictive utility. In contrast, anti-dsDNA demonstrated stronger rule-in characteristics (LR+ >4), reinforcing its value when present in clinically compatible scenarios. These findings are congruent with evidence that preclinical autoimmunity can precede overt disease by several years, particularly in SLE, where autoantibodies may appear long before symptom onset [35]. Thus, autoimmune serology in undiagnosed populations may represent a spectrum ranging from benign seropositivity to preclinical autoimmunity.

A key implication of our findings is the risk of overdiagnosis and medicalization in patients with isolated low-titer autoantibody positivity. Indiscriminate testing may generate false-positive results, leading to anxiety, unnecessary referrals, and healthcare burden. Previous studies have demonstrated substantial downstream costs associated with positive ANA testing in low-pretest probability settings [36]. Our analysis reinforces recommendations advocating for clinically guided testing strategies rather than broad screening panels [37].

Heterogeneity across included studies was substantial, reflecting variability in assay methods, antibody thresholds, referral bias, and population characteristics. Although assay type did not significantly alter pooled sensitivity, titer thresholds and clinical setting influenced predictive estimates. This finding suggests that pretest probability—determined by clinical context—remains central to appropriate interpretation of serologic results.

This review has several strengths. It specifically addresses a clinically underexplored population—patients with suspected but undiagnosed CTD—rather than established disease cohorts. It integrates prevalence and diagnostic accuracy outcomes and employs hierarchical modeling to account for threshold effects. Additionally, longitudinal follow-up data provide insight into disease evolution rather than cross-sectional association alone.

However, limitations must be acknowledged. Significant heterogeneity was present across studies, and many were conducted in tertiary referral centers, potentially inflating seroprevalence estimates. Variation in assay platforms and reporting standards limits direct comparability. Furthermore, not all studies provided complete 2×2 diagnostic data, restricting pooled diagnostic accuracy analyses. Finally, progression to CTD may be underestimated in studies with shorter follow-up durations.

Future research should focus on prospective, multicenter cohorts integrating clinical features, quantitative antibody titers, complement levels, and emerging biomarkers such as interferon signatures to develop predictive models for CTD progression [38]. Machine-learning–based risk stratification tools may further refine individualized risk estimation. Standardization of ANA reporting, including titer and staining pattern, is also essential for improving reproducibility and clinical interpretability.

In inference, autoimmune markers are frequently positive in adults presenting with suspected but initially undiagnosed connective tissue disorders. While ANA demonstrates high sensitivity, its specificity is limited, particularly at low titers. Disease-specific antibodies provide stronger rule-in value but lack sufficient sensitivity to serve as standalone predictors. These findings emphasize the need for clinically contextualized interpretation of serologic testing and longitudinal monitoring rather than reliance on isolated laboratory results.

CONCLUSION

This systematic review and meta-analysis demonstrates that autoimmune serologic markers are frequently detected in adults presenting with suspected but initially undiagnosed connective tissue disorders, yet their predictive value varies substantially. Antinuclear antibody (ANA) testing remains highly sensitive but lacks specificity, particularly at low titers, limiting its standalone diagnostic utility in low–to–moderate pretest probability settings. In contrast, disease-specific antibodies such as anti-dsDNA and anti-SSA provide stronger rule-in value but insufficient sensitivity to reliably exclude future disease development. Approximately one in six undiagnosed individuals progressed to confirmed CTD during follow-up, underscoring the importance of longitudinal clinical surveillance. These findings support a risk-stratified, clinically guided approach to autoimmune testing rather than indiscriminate screening. Future prospective, multicenter studies integrating quantitative antibody profiles with clinical and molecular biomarkers are needed to improve early risk prediction and optimize cost-effective diagnostic pathways in this diagnostically challenging population.

REFERENCES

1. Tsokos GC. Systemic lupus erythematosus. N Engl J Med. 2011;365(22):2110–21.
2. Denton CP, Khanna D. Systemic sclerosis. Lancet. 2017;390(10103):1685–99.
3. Aringer M, Costenbader K, Daikh D, et al. 2019 EULAR/ACR classification criteria for systemic lupus erythematosus. Ann Rheum Dis. 2019;78(9):1151–9.
4. Shiboski CH, Shiboski SC, Seror R, et al. 2016 ACR/EULAR classification criteria for primary Sjögren’s syndrome. Ann Rheum Dis. 2017;76(1):9–16.
5. Pisetsky DS, Spencer DM, Lipsky PE, et al. Antinuclear antibodies: clinical associations and detection methods. Arthritis Res Ther. 2018;20:37.
6. Conrad K, Schössler W, Hiepe F. Autoantibodies in systemic autoimmune diseases. Autoimmun Rev. 2007;6(8):564–8.
7. Mahler M, Fritzler MJ. The clinical significance of autoantibodies in systemic sclerosis. Arthritis Res Ther. 2012;14(2):203.
8. Wiik AS. Clinical and laboratory characteristics of ANCA-associated vasculitis. Clin Exp Rheumatol. 2004;22(6 Suppl 36):S3–9.
9. Tan EM, Feltkamp TE, Smolen JS, et al. Range of antinuclear antibodies in “healthy” individuals. Arthritis Rheum. 1997;40(9):1601–11.
10. Satoh M, Chan EK, Ho LA, et al. Prevalence and sociodemographic correlates of ANA in the US population. Arthritis Rheum. 2012;64(7):2319–27.
11. Conrad K, Röber N, Andrade LE, et al. The clinical relevance of anti-nuclear antibodies in non-rheumatic diseases. Clin Rev Allergy Immunol. 2017;52(2):238–52.
12. Mosca M, Tani C, Carli L, et al. Undifferentiated connective tissue diseases. Best Pract Res Clin Rheumatol. 2012;26(1):61–71.
13. Doria A, Mosca M, Gambari PF, et al. Defining unclassifiable connective tissue diseases: incomplete, undifferentiated, or overlap syndromes. J Rheumatol. 2005;32(2):213–5.
14. Mosca M, Neri R, Bombardieri S. Undifferentiated connective tissue diseases: clinical and serological profile. Clin Exp Rheumatol. 1999;17(5):615–20.
15. Tozzoli R, Bassetti D, Piazza A, et al. The clinical impact of ANA testing in autoimmune rheumatic diseases. Clin Chem Lab Med. 2002;40(10):1007–11.
16. Copple SS, Giles SR, Nelsen AC, et al. Clinical utility of ANA screening in rheumatology referrals. J Rheumatol. 2011;38(11):2427–31.
17. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement. BMJ. 2021;372:n71.
18. McInnes MDF, Moher D, Thombs BD, et al. PRISMA-DTA statement. JAMA. 2018;319(4):388–96.
19. Whiting PF, Rutjes AW, Westwood ME, et al. QUADAS-2 tool for diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–36.
20. Wells GA, Shea B, O’Connell D, et al. Newcastle–Ottawa Scale (NOS) for assessing quality of nonrandomized studies. Ottawa Hosp Res Inst. 2014.
21. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177–88.
22. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–60.
23. Reitsma JB, Glas AS, Rutjes AW, et al. Bivariate analysis of sensitivity and specificity. J Clin Epidemiol. 2005;58(10):982–90.
24. Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple graphical test. BMJ. 1997;315(7109):629–34.
25. Deeks JJ, Macaskill P, Irwig L. Evaluating non-randomised intervention studies. Stat Med. 2005;24(9):1311–30.
26. Schünemann HJ, Mustafa RA, Brozek J, et al. GRADE guidelines for diagnostic tests. BMJ. 2018;362:k4226.
27. Arbuckle MR, McClain MT, Rubertone MV, et al. Development of autoantibodies before clinical SLE. N Engl J Med. 2003;349(16):1526–33.
28. Hayashi N, Koshiba M, Nishimura K, et al. Prevalence of ANA in healthy individuals. Mod Rheumatol. 2008;18(2):153–60.
29. Solomon DH, Kavanaugh AJ, Schur PH. Evidence-based guidelines for ANA testing. Arthritis Rheum. 2002;47(4):434–44.
30. Bodolay E, Csiki Z, Szekanecz Z, et al. Long-term follow-up of UCTD patients. Clin Exp Rheumatol. 2003;21(3):313–20.
31. Danieli MG, Fraticelli P, Salvi A, et al. Five-year follow-up of UCTD. Clin Exp Rheumatol. 1999;17(5):585–91.
32. Petri M, Orbai AM, Alarcón GS, et al. 2012 SLICC classification criteria for SLE. Arthritis Rheum. 2012;64(8):2677–86.
33. Shiboski SC, Baer AN, Shiboski CH, et al. ACR/EULAR classification criteria for Sjögren’s syndrome. Ann Rheum Dis. 2017;76(1):9–16.
34. Eriksson C, Kokkonen H, Johansson M, et al. Autoantibodies predate RA onset. Arthritis Res Ther. 2011;13(1):R30.
35. Heinlen LD, McClain MT, Merrill J, et al. Clinical criteria development in SLE autoimmunity. Arthritis Rheum. 2007;56(7):2344–51.
36. Narain S, Richards HB, Satoh M, et al. Diagnostic impact and cost implications of ANA testing. Arthritis Rheum. 2004;51(4):675–81.
37. Fritzler MJ. Choosing wisely: ANA testing guidelines. Clin Rheumatol. 2016;35(4):915–7.
38. Crow MK. Type I interferon in autoimmune disease. Annu Rev Immunol. 2014;32:135–70.