Abstract
Objective. Presence and levels of antibodies contribute to the classification of rheumatoid arthritis. We investigated the longitudinal course of anti-citrullinated protein antibodies (ACPA) and immunoglobin M (IgM) rheumatoid factor (RF) during the first year after arthritis onset in patients with very short disease duration.
Methods. Patients (aged 18–75 years) with ≥ 1 swollen joint of ≤ 16 weeks’ duration had assessments of ACPA (2nd generation anti-cyclic citrullinated peptide antibodies, anti-CCP2) and IgM RF at inclusion and after 3, 6, and 12 months. Frequencies of seroconversions (negative to positive and vice versa) and changes in antibody levels during followup were determined.
Results. A total of 281 early arthritis patients (median duration of joint swelling 32 days, 14.2% ACPA positives, 12.8% IgM RF positives) with 978 longitudinally collected serum samples were included. Only 5 patients (1.8%) negative for both antibodies at baseline turned antibody-positive during followup, while 9 antibody-positive patients (3.2%) turned antibody-negative. ACPA was more stable than RF regarding both status and levels.
Conclusion. Antibody status (ACPA/RF) is a stable phenotype in very early arthritis, as seroconversion was only found in 5% of patients. Repeated measurement of ACPA or RF during the first year after onset of arthritis does not offer major additional information.
- ANTICYCLIC CITRULLINATED ANTIBODIES
- RHEUMATOID FACTOR
- AUTOANTIBODIES
- LONGITUDINAL STUDY
- RHEUMATOID ARTHRITIS
The presence and levels of rheumatoid factor (RF) and antibodies against citrullinated proteins (ACPA) are important predictors of a poor outcome in patients with early arthritis, in terms of development of rheumatoid arthritis (RA)1, persistence of arthritis2, and development of erosive disease3. Early treatment with disease-modifying antirheumatic drugs (DMARD) improves outcome in early RA4, especially in ACPA-positive patients5.
Because presence or absence of ACPA or RF will influence classification and treatment decisions in patients with recent-onset arthritis, determination of antibody status is a central part of the initial investigation. Seroconversions from negative to positive antibody status over time are of special interest, as this will influence decision-making for diagnosis and choice of treatment. RF and ACPA seroconversion during followup in patients with early inflammatory arthritis was addressed in a recent review6. Although RF and ACPA seemed to be stable over time in the studies included in this review, no detailed reports specifically addressing this issue were identified, and the review concluded that further elaboration on this topic was needed.
Both presence and levels of ACPA and RF antibody status affect disease classification in the 2010 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria for RA7. Thus, both seroconversion and changes in antibody level from low positive [above the upper limit of normal (ULN)] to high positive (> 3 × ULN) are of interest for the classification of early RA.
The aim of our study was to investigate the longitudinal course of ACPA [measured by anti-cyclic citrullinated peptide antibodies (anti-CCP)] and immunoglobin M (IgM) RF during the first year after onset of arthritis in patients with ≤ 16 weeks’ disease duration at inclusion to the study.
MATERIALS AND METHODS
Very Early Arthritis Clinic
The Norwegian Very Early Arthritis Clinic (NOR-VEAC) study was started in 2004 as a multicenter observational study. The cohort includes patients (aged 18–75 yrs) presenting with at least 1 clinically swollen joint of ≤ 16 weeks’ duration, and patients are followed longitudinally for 2 years (visits at 0, 3, 6, 12, and 24 months). The details of the data collection have been described elsewhere8. One-year followup was our study focus, but data from the visit after 2 years were used if available. The study was approved by the regional Ethics Board and the Data Inspectorate, and patients gave written informed consent.
Laboratory markers
Sera were frozen at −70°C at baseline and used to analyze ACPA (anti-CCP2; Inova Diagnostics, San Diego, CA, USA) and IgM RF (in-house enzyme-linked immunosorbent assay) in 1 batch. As recommended by the manufacturer of the assay used for anti-CCP2, a local standard for the cutoff level determining a positive status was employed. The cutoff levels recommended by the central laboratory for positivity of the serologic markers were ACPA ≥ 25 units/ml, IgM RF ≥ 25 units/ml. These levels are also used in clinical practice, and have been used in previous reports9,10,11.
Patient selection
All patients with 1-year followup time and with ACPA and IgM RF results from baseline plus 1 or more followup visits during the first year after enrollment were included in the analyses.
Statistical analysis
ACPA and IgM RF antibody status (positive/negative) was assessed at baseline and at subsequent visits, and the numbers and proportions of patients switching antibody status from negative to positive and vice versa during followup were determined. The number and proportion of patients switching from a low positive (25–74 U/ml) to a high positive (≥ 75 U/ml) antibody status were also determined. Disease activity measures were compared between seroconverters and nonconverters with independent-samples t tests/Mann-Whitney U tests according to the distribution of data. Wilcoxon signed-rank test was used to investigate if antibody levels changed significantly from baseline to 1 year. Longitudinal changes in antibody levels were compared between patients who had received and those who had not received DMARD treatment (Mann-Whitney U tests).
RESULTS
Baseline characteristics and completeness of data
Two hundred eighty-one patients (median duration of joint swelling 32 days) with 978 longitudinally collected serum samples were included in our study. One hundred sixty-nine patients (60.1%) had complete antibody data available (baseline and all 3 followup visits) and an additional 78 patients (27.8%) had data from baseline and 2 followup visits, while the remaining 34 patients (12.1%) had antibody data only from baseline and 1 followup visit. Sixty-five of the 281 patients (23%) also had antibody data from the 2-year visit. Patient selection and reasons for missing data are shown in Figure 1.
The patient selection procedure.
Baseline characteristics and outcome after 1 year are shown in Table 1. Forty patients (14.2%) were ACPA-positive, while 36 (12.8%) were IgM RF-positive at baseline. Twenty-three patients (8.2%) were positive for both ACPA and IgM RF; 53 patients (18.9%) were positive for one of the 2.
Baseline characteristics and 1 year outcome in 281 patients with very early arthritis. Values are n (%) unless stated otherwise.
Antibody status and levels during followup
Only 5 patients (1.8%) who were negative for both antibodies at baseline turned antibody-positive during the first year of followup (Table 2). One of these patients became positive for both RF and ACPA, while the remaining 4 developed only IgM RF antibodies (Table 3). One additional patient who was already positive for IgM RF at baseline also became ACPA-positive (i.e., double-positive) during followup, while 2 initially ACPA-positive/RF-negative patients became RF-positive. Of the 65 patients with 2-year sera samples available, 1 baseline antibody-negative patient turned marginally RF-positive at 2 years.
Antibody status (ACPA/RF) at baseline and during followup in 281 patients with early arthritis. Data are number (%).
Antibody levels (U/ml) during followup, DMARD treatment, and final diagnosis in the 14 patients switching antibody status. Values representing a new status are shown in bold type. Cutoff for positive antibody status: ACPA: ≥ 25 U/ml, IgM RF ≥ 25 U/ml.
Nine antibody-positive patients (3.2%) became antibody-negative during followup (Table 2). One patient switched from positive to negative status for ACPA and RF, 4 patients switched from positive to negative RF status, and 4 switched from positive to negative ACPA status. Baseline levels in the ACPA switchers were all in the low-positive range, with levels from 27 to 37 U/ml, while baseline levels in the RF-positive to RF-negative switchers ranged from 27 to 63 U/ml (Table 3). Three additional patients switched from positive to negative RF status, but remained ACPA-positive throughout followup. The seroconversion rates were consistently low across subgroups of patients according to outcome at 1 year (Tables 1 and 3). No statistical differences could be found for measures of disease activity in converters compared to nonconverters.
If alternative cutoff levels for positivity of the serologic markers, e.g., 20 U/ml and 30 U/ml, had been used, the following total numbers of patients switching antibody status would have been found: 30 U/ml: 4 negative-to-positive switchers, 5 positive-to-negative switchers; 20 U/ml: 6 negative-to-positive switchers, 8 positive-to-negative switchers (data not shown).
ACPA levels were stable during followup, but RF levels decreased significantly in baseline RF-positive patients (Table 4). Only 2 patients switched from the 2010 ACR/EULAR criteria for RA “low positive” to the “high positive” antibody category during followup (Table 5). One of these patients would theoretically have been reclassified as RA after 1 year based on this finding; the other was already classifiable at baseline.
Baseline values* and 1 year changes in antibody levels (U/ml) in patients with samples available from the 1 year followup visit.
Antibody levels (U/ml) during followup in 2 patients switching from “low-positive” (25–74 U/ml) to “high-positive” (≥ 75 U/ml) antibody category according to the 2010 classification criteria for rheumatoid arthritis. Values representing a new status are shown in bold.
Antibody status and DMARD treatment
Ninety patients (32%) were treated with DMARD during followup: methotrexate (MTX; n = 59), sulfasalazine (n = 11), leflunomide (n = 1), sequential synthetic DMARD/DMARD combination (n = 9), or MTX plus tumor necrosis factor-α (TNF-α) antagonist (n = 10). None of the 10 patients treated with TNF inhibitors switched antibody status. Details of treatment in patients switching antibody status are given in Table 3. There were no statistically significant differences in change in levels for either ACPA or RF in patients who had received DMARD treatment compared to those who had not received DMARD treatment.
DISCUSSION
The clinician has to consider the balance between cost and usefulness when ordering specific tests for patients with recent-onset arthritis. In our study we show that patients rarely change antibody status within the first year after onset of joint swelling. Most patients with qualitative changes had marginally positive/negative samples at baseline and were thus probably not true “switchers.” Of special clinical interest is that very few antibody-negative patients became positive during followup. This finding is in accord with previous studies showing that both ACPA and RF can be found in premorbid sera in a proportion of RA patients several years before diagnosis12.
The previous evidence regarding the longitudinal course of ACPA/RF in early arthritis is limited, as recognized in a recent review6. Published studies have had relatively low sample sizes and have not addressed this issue specifically or in detail13,14,15. Two recent studies in patients with early arthritis, both with longer symptom duration than in our study, have suggested that both ACPA and RF status are stable during the first years after onset of arthritis16,17. In patients with early RA, some longitudinal studies on ACPA/RF have been published and most report a low frequency of ACPA seroconversion (4% to 7%)18,19,20. However, one early RA study reported a 15% frequency of ACPA seroconversion21. Compared to ACPA, RF status in early RA seems less stable18,19,20, which is also illustrated by the increase in IgM RF levels with age observed in patients who do not have RA7,22. Indeed, in our patients with very early arthritis, RF was also a more unstable phenotype than ACPA, and IgM RF levels decreased significantly in patients who were initially positive for these antibodies.
Some studies in RA have suggested that DMARD treatment (both synthetic and biological drugs) can lead to a decrease in antibody level20,23. An effect of treatment was not evident in our study, but the low proportion of patients treated with DMARD does not allow firm conclusions for this issue.
The majority of patients in our study had information available regarding antibody status from all 4 timepoints during the first year of followup. However, in the 267 patients who were nonconverters, 21 (7.9%) had antibody data only from baseline and 3-month followup, while 52 (19.5%) had no antibody data at 1 year (i.e., the last followup visit was at 6 months). We cannot be confident that later samples from these patients would remain unchanged with regard to antibody status. However, if these patients were excluded from the analyses, the seroconversion rate would still have been low: 6.7% instead of 5%.
The determination of which patients could be regarded as converters of antibody status is of course dependent on the cutoff level employed for positivity of the serologic markers. In our study the levels employed in clinical practice in our hospitals were used (25 U/ml for both markers). Changing the cutoff level from 25 U/ml to 30 U/ml or 20 U/ml would have changed the proportion of switchers in this cohort only marginally. This consistency strengthens the main conclusion of our study: that very few patients with early inflammatory arthritis change status for ACPA or RF in the first year after onset of arthritis.
The low frequency of seropositive patients represents a limitation to our study in studying seroconversion from positive to negative, as this could lead to underestimation of change in both antibody levels and status over time. The extremely short duration of joint swelling (median 32 days) in our study sets it apart from the existing studies regarding the longitudinal course of autoantibodies in early arthritis. This gives information on the first year after symptom onset, which is the most critical period for making a diagnosis and starting timely treatment.
Our study shows that antibody status (ACPA/RF) is a stable phenotype in very early arthritis. Repeated measurement of ACPA or RF does not offer important additional information during the first year after onset of joint swelling.
Acknowledgment
The authors thank the patients for participating in the study, the doctors and nurses for data collection, and Gro Jaaberg Talgø and Per Ivar Gaarder at the Department of Immunology and Transfusion Medicine, Oslo, Norway, for performing the serologic analyses.
Footnotes
-
Supported by the South-Eastern Norway Regional Health Authority and The Norwegian Foundation for Health and Rehabilitation.
- Accepted for publication June 21, 2011.
