Rheumatol Int DOI 10.1007/s00296-013-2799-8
ORIGINAL ARTICLE
Clinical features and independent predictors in the further development of rheumatoid arthritis in undifferentiated arthritis Dongying Chen • Hao Li • Liuqin Liang • Youjun Xiao Ting Xu • Qian Qiu • Fan Lian • Zhongping Zhan • Yujin Ye • Hanshi Xu • Xiuyan Yang
•
Received: 9 February 2013 / Accepted: 6 June 2013 Ó Springer-Verlag Berlin Heidelberg 2013
Abstract This study aims to investigate the prognosis of undifferentiated arthritis (UA) and to estimate the putative predictors contributing to predict the development of UA into rheumatoid arthritis (RA); thus, it could improve appropriate medical intervention. A retrospective cohort study of 218 patients with an initial diagnosis of UA and 2-year follow-up monitoring was carried out. The baseline information including demographic variables, clinical features, and laboratory data was collected. A logistic regression model was used for the statistical analysis. After 2 years of follow-up, 20.18 % of UA patients evolved into RA, but 33.03 % remained undifferentiated. Meanwhile, 25.23 % went into remission, and 21.56 % developed into other connective tissue diseases. Univariate and multivariate analysis showed that the titer of antibodies to cyclic citrullinated peptide (anti-CCP), tender joint count and duration of morning stiffness were independent predictors for the development of RA. The area under the curve (AUC) of duration of morning stiffness (0.81) was largest, followed by tender joint count (0.74). The AUC of antiCCP antibodies (0.68) was higher than that of rheumatoid factor of IgM type (IgM-RF) (0.60), and the combination of these two antibodies was significantly higher than each alone (P \ 0.001). In conclusion, UA patients had variable Dongying Chen and Hao Li contributed equally to this work. D. Chen H. Li L. Liang (&) Y. Xiao T. Xu Q. Qiu F. Lian Z. Zhan Y. Ye H. Xu X. Yang (&) Department of Rheumatology and Clinical Immunology, The First Affiliated Hospital of Sun Yat-Sen University, No. 58, Zhongshan 2nd Road, Guangzhou 510080, China e-mail:
[email protected] X. Yang e-mail:
[email protected]
clinical outcomes and prognosis. Only the titer of anti-CCP antibodies, tender joint count, and duration of morning stiffness, instead of IgM-RF, could predict the development of RA. Although the anti-CCP antibody was better than the IgM-RF in predicting RA, a combined detection of them still improved the diagnostic performance. Keywords Undifferentiated arthritis Rheumatoid arthritis Antibodies to cyclic citrullinated peptide Rheumatoid factor
Introduction Undifferentiated arthritis (UA) is a form of peripheral arthritis that does not fulfill the classification criteria for a more definitive diagnosis [1]. Based on the results from several inception cohort studies, it was known that onethird of patients with UA progressed into rheumatoid arthritis (RA), while the other two-thirds remained undifferentiated, experiencing spontaneous remission or developing into other connective tissue diseases [2–4]. Rheumatoid arthritis is a chronic heterogeneous autoimmune disorder of unknown cause, resulting in inflammation in the synovium, cartilage, and bone [5, 6]. Studies showed that early use of disease-modifying anti-rheumatic drugs (DMARDs) in RA patients could delay the joint damage and help to improve the prognosis [7–10]. Therefore, early diagnosis and treatment might help to make decisions which relate to early therapeutic intervention for a better prognosis. Several risk factors of the prognosis of UA have been identified, such as presence of RF [5], anti-cyclic citrullinated peptide antibodies [11], specific human leukocyte antigens (HLA) [12], HLA–DRB1 genotypes [13, 14], and baseline radiographic joint changes [15, 16]. Antibodies to
123
Rheumatol Int
cyclic citrullinated peptide (anti-CCP) have been documented extensively over recent years as highly specific serological markers for RA diagnosis and prognosis [17, 18]. However, at present, criteria distinguish RA from other arthritic disorders, and joint damage is rarely apparent in the very early stages of disease. Moreover, clinical trials of RA treatments are hampered by lack of criteria considering study enrollment of patients at early stages of disease [6, 19]. In this article, it was aimed to find out whether there were predictive factors in UA patients which would identify the patients at the greatest risk of developing RA, so that early diagnosis and appropriate treatment could be performed in order to ameliorate their conditions.
Materials and methods Patients and methods Study population: A total of 218 patients (62 males, 156 females) presenting with UA were invited to the study. These 218 patients had been admitted to the Rheumatoid Immunology Department of the First Affiliated Hospital of Sun Yat-sen University from October, 2005 to September, 2008. Patients with UA were those who had characteristics of inflammatory synovitis in one or more joints within the previous 1 year, and they only took Chinese herbal medicine irregularly, without DMARDs, NSAIDs or Steroids. We excluded the patients with secondary causes such as infection, metabolic disorder and degeneration, and those fulfilling the classification criteria for specific rheumatic disorders [1, 20]. Patients with newly diagnosis of RA by American College of Rheumatology (ACR) 1987 criteria were excluded. Prior to the implementation, the study got approval by the hospital ethics administrative department. The informed consent from patients was also acquired before collecting data. Patients were evaluated by the same rheumatologist at baseline according to a standard protocol which included swollen joint count, tender joint count, and duration of morning stiffness in minutes. Final assessment was carried out 2 years after inclusion, by the same rheumatologist, to determine their evolving pattern, i.e., whether they remained undifferentiated, evolved to RA, developed another disease, or went into remission. Diagnosis of RA was made in accordance with the ACR 1987 revised criteria; other nonRA diagnosis was made in line with individual standard diagnostic criteria [5, 6]. Remission was defined as that symptoms disappeared for 3 months without medication [3].
morning stiffness, anti-CCP antibodies, IgM-rheumatoid factors (IgM-RF), erythrocyte sedimentation rate (ESR), and C reactive protein (CRP). The anti-CCP antibody was assessed by chemiluminescence microparticle immunoassay (Abbott Diagnostics, USA) with the cutoff set at 5 RU/ ml [21, 22]. The IgM-RF was measured by nephelometry (Siemens Healthcare Diagnostics, Germany) with the cutoff set at 20 IU/ml. CRP was measured by nephelometry, and ESR was determined by Westergren method (Oumeng Hangzhou, China). Statistical analysis Statistical Package for Social Sciences (SPSS) 16.0 statistic software was used for the analysis. Mean ± SD or the median and interquartile range was presented for continuous and ordinal data. In addition, categorical variables were presented as the absolute count and percentage, and continuous variables were evaluated by t test or Mann–Whitney U test. Moreover, categorical variables were compared by v2 test or Fisher’s exact test. Univariate and multivariate logistic regression analyses were used to predict the presence of RA. The odds ratios (OR) and 95 % confidence intervals (95 % CIs) were calculated. Variables associated with the outcome at P \ 0.20 were included in a multivariate logistic model with the forward step-wise option. Both of the sensitivity, specificity, positive predictive value (PPV), and the negative predictive value (NPV) of the antibodies for the diagnosis of RA were calculated based on the manufacturer-cutoffs and the best-cutoffs from the receiver-operating characteristic (ROC) curves, respectively. Area under the curve (AUC) and 95 % CIs were calculated for the combination model of these two antibodies and variables significant in multiple logistic regression analysis, while IgM-RF was brought into the AUC calculation no matter whether it was significant in multiple analysis. The AUC represents an overall measurement of performance of the variables for the prognosis of UA, with 1.0 as a perfect variable and 0.5 as a variable with no discriminating capacity. Wilcoxon rank-sum test was used to evaluate differences among the AUCs. Two-tailed P value of\0.05 was considered statistically significant.
Results
Baseline data collection
Patient characteristics at baseline and at the end of follow-up
Collection of patients’ baseline information involves their gender, age, tender and swollen joint count, duration of
A total of 218 patients with recent-onset UA were enrolled in the study. The baseline prevalence of anti-CCP
123
Rheumatol Int
antibodies in these UA patients was 11.93 % (26/218). The IgM-RF was positive in 25.23 % UA patients. After 2-year follow-up, 44 patients (20.18 %) evolved into RA, 72 patients (33.03 %) remained undifferentiated, 55 patients (25.23 %) went into remission, and 47 patients (21.56 %) developed into other connective tissue diseases (23 cases with systemic lupus erythematosus, 4 cases with primary Sjo¨gren’s syndrome, 6 cases with ANCA-associated vasculitis, 4 cases with adult-onset Still’s disease, and 10 cases with spondyloarthropathy). Patients with UA who developed into RA (RA group) were compared with those who did not (non-RA group). The RA group was comprised of 8 males and 36 females, with the mean (±SD) age of 50.00 ± 16.00 years old. The non-RA group consisted of 54 males and 120 females, with the mean (±SD) age of 39.00 ± 17.00 years old. A significant (P \ 0.05) difference between these two groups was found in age, tender and swollen joint count, duration of morning stiffness, positive rates of anti-CCP, and IgM-RF at baseline (Table 1). Factors predicting progression from UA to RA According to univariate logistic regression analyses, the following variables listed in Table 2 were significantly (P \ 0.05) associated with the progression from UA to RA: patients’ age, the anti-CCP antibodies, the tender and swollen joint count, and the duration of morning stiffness. In the final multiple logistic regression model, patients’ age and gender, the anti-CCP antibodies, the IgM-RF, the tender joint count, swollen joint count, and the duration of morning stiffness were under analysis. Only the anti-CCP antibodies, the tender joint count, and the duration of morning stiffness were independently associated with the development of RA (Table 3). Rheumatoid factor of IgM type which was still a commonly used method in clinical practice, as well as the combination model of IgM-RF and anti-CCP antibodies, was bought into ROC analysis, together with anti-CCP Table 1 Baseline characteristics of patients with undifferentiated arthritis
antibodies, tender joint count, and duration of morning stiffness, which were significant in multiple logistic regression analyses of progression from UA to RA. The ROC curves of IgM-RF, the anti-CCP antibodies, the combination model of these two, tender joint count, and duration of morning stiffness, in the RA and non-RA group, by logistic regression analyses were constructed, respectively (Fig. 1). The AUCROC of duration of morning stiffness was largest, followed by tender joint count. The AUCROC of the anti-CCP antibodies (0.68) was larger than the IgM-RF (0.60). Furthermore, the AUCROC of the combination of these two antibodies was significantly larger than that of each antibody tested alone (P \ 0.001) (Table 4). Based on the results of ROC analysis, with the application of the cutoffs set by manufacturers, the anti-CCP antibodies tended to have higher specificity, PPV and NPV but lower sensitivity than the IgM-RF. When a serum antiCCP of 3.55 RU/ml and the IgM-RF of 26.30 IU/ml, which were statistically generated by the ROC, were used for adjusted cutoffs; the anti-CCP antibodies yielded higher specificity, sensitivity, PPV and NPV than the IgM-RF. Combining the positive anti-CCP antibodies and/or the IgM-RF could improve the sensitivity. In addition, both of the two positive antibodies could provide the highest specificity (Table 5).
Discussion This study investigated the prognosis of UA, the distribution of levels of anti-CCP antibodies and the IgM-RF, as well as the indexes that could predict the development of UA into RA. In the study, the prevalence of anti-CCP antibodies in these UA patients was 11.93 %, and the IgM-RF was positive in 25.23 % at the baseline; after 2 years of followup, there were 20.18 % patients who evolved into RA, 33.03 % patients remained undifferentiated, 25.23 % patients went into remission, and 21.56 % patients
Patient characteristics
RA (n = 44)
Non-RA (n = 174)
P
Female patients (n, %)
36 (81.82)
120 (68.97)
0.091
Age (mean ± SD, years)
44.73 ± 15.81
39.07 ± 12.58
0.004
Swollen joint count
4.50 ± 3.00
2.00 ± 1.00
\0.001
Tender joint count
4.50 ± 3.00
2.00 ± 1.00
\0.001 \0.001
Duration of morning stiffness (h)
0.92 ± 0.45
0.57 ± 0.21
ESR (mm/h)
44.45 ± 50.00
52.50 ± 35.00
0.79
CRP (g/l)
6.25 ± 34.80
5.62 ± 40.20
0.93
The titer of anti-CCP antibodies (median, RU/ml)
1.80
\0.50
\0.001
Anti-CCP positivity (n, %)
15 (34.09)
11 (6.32)
\0.001
IgM-RF positivity (n, %)
18 (40.91)
37 (21.26)
0.007
123
Rheumatol Int Table 2 Univariate logistic regression analysis of progression from UA to RA Baseline characteristics
OR (95 % CI)
Sex
2.03 (0.88–4.65)
P
1.03 (1.01–1.06)
0.014
Anti-CCP antibodies
1.09 (1.03–1.17)
0.006
IgM-RF
1.01 (0.998–1.02)
0.13
Swollen joint count
2.54 (1.88–3.43)
\0.001
Tender joint count
2.82 (2.04–3.90)
\0.001
Duration of morning stiffness ESR
1.11 (1.08–1.15) 1.00 (0.99–1.01)
\0.001 0.79
CRP
1.00 (0.99–1.01)
0.93
Table 3 Multiple logistic regression analyses of progression from UA to RA Adjusted OR (95 % CI)
P
Anti-CCP antibodies
1.07 (1.00–1.15)
0.046
Tender joint count
2.00 (1.38–2.90)
\0.001
Duration of morning stiffness
5.16 (1.20–22.13)
0.027
Age, gender, anti-CCP antibodies, IgM-RF, swollen joint count, tender joint count, and duration of morning stiffness were brought into the multiple analysis
Fig. 1 Receiver-operating characteristic (ROC) curves for anti-CCP, IgM-RF, the combination model of these two, tender joint count, and duration of morning stiffness by logistic regression analysis
developed into other connective tissue diseases. The natural history of UA was characterized by variable clinical outcomes, thus necessitating the recognition of those UA patients was possible to develop RA so that we can initiate disease-modifying anti-rheumatic drug (DMARD) therapy at an early stage. The results showed that IgM-RF could not predict the development of RA, even after adjustment for potential confounding factors. These were, to some extent, different
123
AUCROC
Standard error
0.60
0.051
0.10
Age
Baseline characteristics
Table 4 AUCROC for anti-CCP antibodies, IgM-RF, the combination model, tender joint count, and duration of morning stiffness by logistic regression analysis
IgM-RF
P
95 % CI 0.051
0.50–0.70
Anti-CCP antibodies
0.68
0.050
\0.001
0.59–0.78
Logistic regression model
0.72
0.050
\0.001
0.62–0.81
Tender joint count
0.74
0.050
\0.001
0.65–0.84
Duration of morning stiffness
0.81
0.045
\0.001
0.72–0.90
from the foreign research and might be due to the higher positive rate of IgM-RF in our non-RA patients [23, 24]. This might be due to the generation mechanism of rheumatoid factors, likely the result of genetic predispositions and the intensity of the immune reaction, which can also be found during host defense against infectious agents and under pathologic conditions [25]. The results of multivariate analysis displayed that antiCCP antibody, tender joint count, and duration of morning stiffness were independently associated with the development of RA. Corresponding to the earlier reports, the results showed that anti-CCP antibody was an independent risk factor in the development of RA, of which the predictive value correlated with its titers [24, 26, 27]. Furthermore, our result probably suggested the rule in Chinese population, which has not reported in literature. Several other studies proved that the shared epitope (SE)containing HLA–DRB1 alleles, which was directly involved in the pathogenesis of RA via the joint erosion effect of anti-CCP antibodies, could also exist in UA patients [13, 14]. These results provided a possible explanation for the predictive ability of anti-CCP antibodies on the development of RA. As a consequence, persistent arthritis patients with positive anti-CCP antibodies, multiple joint disease, and long hours of morning stiffness should be closely followed up, although whether these patients would benefit from early initiation of DMARD therapy is still under investigation. The results showed that, when using the adjusted cutoffs, anti-CCP antibodies had higher sensitivity, specificity, PPV and NPV than IgM-RF. According to the study, the titer of anti-CCP antibodies was measured by a new sensitive automated assay named chemiluminescence microparticle immunoassay (CMIA) [21, 22]. This is considered as a procedure which shares almost the same detective sensitivity with IgM-RF but makes data from different laboratories which are comparable through reducing factitious operation errors. AUCROC is performed to evaluate the diagnostic values of RA associated with autoantibodies.
Rheumatol Int Table 5 Sensitivity, specificity, PPV and NPV of anti-CCP, IgM-RF, and the combination of these two for the clinical diagnosis of RA
Anti-CCP (?) Cutoffs set by manufacturers
Anti-CCP (?) or IgM-RF (?)
Anti-CCP (?) and IgM-RF (?)
20.00**
Specificity (%)
93.68
78.74
81.03
97.70
Sensitivity (%)
34.09
40.91
40.91
9.09
PPV (%)
57.69
32.73
35.29
50.00
NPV (%)
84.90
84.05
84.43
80.95
Adjusted cutoffs
* RU/ml, ** IU/ml
5.00*
IgM-RF (?)
3.55*
26.30**
Specificity (%)
88.51
86.78
75.86
99.43
Sensitivity (%)
45.45
36.36
65.91
15.91
PPV (%)
50.00
41.03
40.85
87.50
NPV (%)
86.52
84.36
89.80
82.38
Some other research found that IgM-RF was significant to predict the development of RA, which was different from our result [24]. So, we still brought IgM-RF into ROC analysis, as well as the combination model. AUCROC of morning stiffness duration and tender joint count was significantly larger than other variables and indicated multiple joint disease and long hours of morning stiffness predicted the development. At the same time, we focused on objective detection means. The result of ROC analysis indicated that anti-CCP antibodies alone were better than IgM-RF, while their combination was the best option. Based on the findings, two coexisting autoantibodies had higher specificity but lower sensitivity compared with the presence of one autoantibody in predicting RA. It is suggested that different tests should be taken, either singly or in combination in terms of different circumstances. In conclusion, UA patients had variable clinical outcomes and prognoses. Among all, baseline information, instead of IgM-RF, anti-CCP antibodies, tender joint count, and duration of morning stiffness, could predict the development of RA. These information might help to make decisions about early therapeutic intervention for a better prognosis. Although the anti-CCP antibodies alone were better than the IgM-RF in predicting RA, a combination of these two still improved the diagnostic value. The study was limited to retrospective data of one unit, pending further multi-center prospective study to find out the applicability of the rule in the wider population. Acknowledgments This work was supported by the Guangdong Provincial Science and Technology Funds, China (2011B080701011, 2010B080701099) and Guangdong Provincial college students’ training project of science and technology, China (1055812371).
References 1. Verpoort KN et al (2004) Undifferentiated arthritis—disease course assessed in several inception cohorts. Clin Exp Rheumatol 22(5 Suppl 35):S12–S17
2. Tunn EJ, Bacon PA (1993) Differentiating persistent from selflimiting symmetrical synovitis in an early arthritis clinic. Br J Rheumatol 32(2):97–103 3. Harrison BJ et al (1996) Natural remission in inflammatory polyarthritis: issues of definition and prediction. Br J Rheumatol 35(11):1096–1100 4. van Aken J et al (2006) Comparison of long term outcome of patients with rheumatoid arthritis presenting with undifferentiated arthritis or with rheumatoid arthritis: an observational cohort study. Ann Rheum Dis 65(1):20–25 5. Arnett FC et al (1988) The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 31(3):315–324 6. Aletaha D et al (2010) 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 69(9):1580–1588 7. Doan T, Massarotti E (2005) Rheumatoid arthritis: an overview of new and emerging therapies. J Clin Pharmacol 45(7):751–762 8. Smolen JS et al (2007) New therapies for treatment of rheumatoid arthritis. Lancet 370(9602):1861–1874 9. van der Horst-Bruinsma IE et al (1998) Diagnosis and course of early-onset arthritis: results of a special early arthritis clinic compared to routine patient care. Br J Rheumatol 37(10):1084–1088 10. van der Helm-van Mil AH, Breedveld FC, Huizinga TW (2006) Aspects of early arthritis. Definition of disease states in early arthritis: remission versus minimal disease activity. Arthritis Res Ther 8(4):216 11. Schellekens GA et al (2000) The diagnostic properties of rheumatoid arthritis antibodies recognizing a cyclic citrullinated peptide. Arthritis Rheum 43(1):155–163 12. Jawaheer D et al (2002) Dissecting the genetic complexity of the association between human leukocyte antigens and rheumatoid arthritis. Am J Hum Genet 71(3):585–594 13. Gonzalez-Gay MA, Garcia-Porrua C, Hajeer AH (2002) Influence of human leukocyte antigen-DRB1 on the susceptibility and severity of rheumatoid arthritis. Semin Arthritis Rheum 31(6):355–360 14. van der Helm-van Mil AH et al (2006) The HLA–DRB1 shared epitope alleles are primarily a risk factor for anti-cyclic citrullinated peptide antibodies and are not an independent risk factor for development of rheumatoid arthritis. Arthritis Rheum 54(4):1117–1121 15. Fathi NA et al (2008) Diagnostic performance and predictive value of rheumatoid factor, anti-cyclic-citrullinated peptide antibodies and HLA–DRB1 locus genes in rheumatoid arthritis. Int Arch Med 1(1):20 16. Kuriya B et al (2009) Validation of a prediction rule for development of rheumatoid arthritis in patients with early undifferentiated arthritis. Ann Rheum Dis 68(9):1482–1485
123
Rheumatol Int 17. Manivelavan D, Vijayasamundeeswari CK (2012) Anti-cyclic citrullinated peptide antibody: an early diagnostic and prognostic biomarker of rheumatoid arthritis. J Clin Diagn Res 6(8): 1393–1396 18. van Venrooij WJ, van Beers JJ, Pruijn GJ (2012) Anti-CCP antibodies: the past, the present and the future. Nat Rev Rheumatol 7(7):391–398 19. Zeidler H (2012) The need to better classify and diagnose early and very early rheumatoid arthritis. J Rheumatol 39(2):212–217 20. Schumacher HR (2002) Early arthritis clinics, much early arthritis is unclassified. J Rheumatol 29(11):2258–2260 21. Hwang Sang Mee et al (2010) Performance analysis of the ARCHITECT anti-cyclic citrullinated peptide antibody in the diagnosis of rheumatoid arthritis. Clin Chem Lab Med 48(2):225–230 22. Kim Sinyoung et al (2010) Evaluation of three automated enzyme immunoassays for detection of anti-cyclic citrullinated peptide antibodies in qualitative and quantitative aspects. Rheumatology 49(10):450–457
123
23. Jansen AL et al (2002) Rheumatoid factor and antibodies to cyclic citrullinated Peptide differentiate rheumatoid arthritis from undifferentiated polyarthritis in patients with early arthritis. J Rheumatol 29(10):2074–2076 24. Rantapaa-Dahlqvist S et al (2003) Antibodies against cyclic citrullinated peptide and IgA rheumatoid factor predict the development of rheumatoid arthritis. Arthritis Rheum 48(10):2741–2749 25. Do¨rner Thomas et al (2004) Rheumatoid factor revisited. Curr Opin Rheumatol 16(3):246–253 26. Forslind K et al (2004) Prediction of radiological outcome in early rheumatoid arthritis in clinical practice: role of antibodies to citrullinated peptides (anti-CCP). Ann Rheum Dis 63(9):1090–1095 27. Avouac J, Gossec L, Dougados M (2006) Diagnostic and predictive value of anti-cyclic citrullinated protein antibodies in rheumatoid arthritis: a systematic literature review. Ann Rheum Dis 65(7):845–851