Rheumatoid arthritis is a chronic disease that affects almost 1% of the population and has an important effect on health, causing pain, fatigue, radiological damage, functional disability, psychological effects and reduced life expectancy.¹ Erosions develop rapidly in 10–26% of patients with rheumatoid arthritis within 3 months of disease onset, and in 75%, it develops within 2 years.^2–4 Clinical remission in early rheumatoid arthritis, obtained in 10–33% of patients in prospective international studies,^{2,5,6,7,8,9,10,11,12,13,14} is the ultimate aim for doctors treating patients. Several sets of criteria define clinical remission. The preliminary criteria of the American College of Rheumatology (ACR) include six signs and symptoms: duration of morning stiffness not exceeding 15 min, no fatigue, no joint pain by anamnesis, no joint tenderness or pain on motion, no soft tissue swelling in joint or tendon sheaths and erythrocyte sedimentation rate (ESR) <30 mm/h for a woman or <20 mm/h for a man.^15,16 Five or more of the criteria must be fulfilled for at least two consecutive months. Remission may also be defined by using a composite index taking into account joint status, patient global assessment, ESR, and Disease Activity Score (DAS).^17–19 A comparative study showed that 95% of ACR remission visits were also considered as remission by using the DAS with a cut-off value <1.6.^5,20 These two definitions of remission are based on low disease activity and ignore physical function and structural damage.

However, the final goal is to prevent radiological damage and functional disability. The relationship between disease activity and these outcome measures in early rheumatoid arthritis remains a topic of debate.^21–26 In fact, some studies suggest that structural damage can occur independently of arthritis activity.^26–32

The objective of this study was to assess the radiological damage progression over 5 years in patients with recent rheumatoid arthritis in sustained remission (DAS<1.6).

METHODS

Patients

All consecutive outpatients who were referred from primary care physicians for the purposes of a study of follow-up in early rheumatoid arthritis,^33–36 in four French centres (Montpellier, Paris-Cochin, Toulouse and Tours) and who fulfilled the ACR criteria for rheumatoid arthritis,³⁷ had a disease duration of <1 year and had not been treated previously with disease-modifying antirheumatic drugs (DMARDs) were included between March 1993 and October 1994. All patients agreed to be enrolled and provided signed informed consent. They were subsequently treated with DMARDs (usually methotrexate, sulfasalazine or a combination of both) that could be modified during the study according to efficacy and tolerance; some patients participated in a randomised, controlled, double-blind 52-week clinical trial of a combination of sulfasalazine and methotrexate compared with a single drug.^38,39 The study was approved by the ethics review board in Montpellier, France.

Clinical and biological assessment

The following evaluation data were collected at baseline: age, sex, disease duration, DAS,¹⁸ Health Assessment Questionnaire (HAQ) score,⁴⁰ ESR, C reactive protein (CRP) level, IgA and IgM rheumatoid factor positivity by anti-human Fc IgG ELISA, and anti-cyclic citrullinated protein antibody positivity by ELISA.³⁴ Human leucocyte antigen (HLA)-DRB1 and HLA-DQB1 genotyping were performed as described previously.⁴¹ Each patient was followed up by the same investigator at 6 months after inclusion, and then at 1, 3 and 5 years.

Definition of remission

Remission was defined by a DAS<1.6 at the 3-year follow-up visit, in accordance with Prevoo et al.⁵ Sustained remission was defined by a DAS<1.6 at the 3-year and 5-year evaluations.

Radiographic assessment

Hand, wrist and foot radiographs were obtained at baseline and at 3 and 5 years. They were evaluated in a blinded manner and in chronological order by two independent observers and scored according to Sharp’s method as modified by Sharp–van der Heijde Score (SHS).⁴² For each patient, an erosion score, a joint-space narrowing score and a total damage score were noted for the hands and feet. The intraclass, intraobserver and interobserver coefficients of correlation were calculated on 30 chosen pairs of radiographs of the hands and feet and were always >0.85.³⁴ No systematic differences were found in any of the scores. We used the mean of the two observers’ scores to determine the final radiographic scores for erosions, joint space narrowing and total damage. A new reading and scoring of radiographs by GC was obtained for individual analysis for patients in sustained remission.

To determine a cut-off value for changes in joint space width that would define individual radiological progression of rheumatoid arthritis unrelated to measurement errors (smallest detectable difference), we calculated the mean of the differences between two analyses.^43,44 We selected 30 pairs of radiographs of the hands and feet that were representative of the population studied. The mean (standard deviation (SD)) of the difference between the two analyses performed by the two observers was calculated. Radiological progression, according to the Outcome Measures in Rheumatoid Arthritis Trials Committee recommendations,⁴⁴ was then defined by a change in radiological scores greater than the upper boundary of the 95% confidence interval (CI) of the differences—that is, a change of at least 5, 4.9 and 4.1 in the erosions, narrowing and total damage scores, respectively. The radiographic progression was compared between patients with rheumatoid arthritis in sustained remission and patients who did not fulfil the remission criteria, at both 3 and 5 years.

Functional progression

The HAQ score was compared at baseline, 3 years and 5 years in each group. As published previously in patients in remission,^15,24 a value of 0.5 was used to dichotomise the HAQ. A score ⩽0.5 indicates hardly any difficulties and a score >0.5 indicates minor to major problems in performing activities of daily life.

Statistical methods

Wilcoxon’s signed rank test for quantitative variables and χ² tests for qualitative variables were used to assess differences between the patient groups and between baseline and 5-year assessment. The significance level was set at 5% (two-sided tests). Analyses were performed using SAS V.8.2.

RESULTS

Demographic, clinical and biological features of the patient cohort

Table 1 shows the baseline characteristics of the patients.

In total, 191 patients (140 women, 51 men) were enrolled in this study, of whom 150 (78.5%) were previously part of a randomised controlled trial.^38,39 The mean (SD) age at diagnosis was 50.5 (14.7) years and the mean (SD) disease duration at inclusion was 3.3 (2.6) months. In all, 154 (80.6%) patients were IgM or IgA rheumatoid factor positive (⩾20 IU/ml and ⩾7 units/ml, respectively) at baseline and 86 (45%) had at least one rheumatoid arthritis-associated DRB1*04 allele (DRB1*0401, 0404, 0405 or 0408). Six months after inclusion, 178 (93.2%) patients were taking DMARDs: 131 (68.6%) were taking one drug (58 methotrexate, 59 sulfasalazine and 14 other DMARDs) and 47 (24.6%) a combination of methotrexate and sulfasalazine. During the 5-year follow-up, a mean of 1.95 DMARDs (range 1–5) was prescribed (methotrexate to 175 patients, sulfasalazine to 147, intramuscular gold to 41, hydroxychloroquine to 25, D-penicillamine to 14 and ciclosporin to 1). A total of 86 patients received the same DMARD or the same DMARD combination during the 5-year follow-up, and 63 (33%) patients received prednisone treatment (5–15 mg/day) at least once during follow-up.

At 5 years, 26 (13.6) patients were lost to follow-up (6 patients died, 8 refused further follow-up and 12 moved out of the area), and at the 5-year evaluation, 31 (16.2%) had missing data and were excluded from the analysis. The baseline characteristics of these patients did not differ from those of the rest of the cohort.

Remission rate

A total of 48 (35.8%) patients fulfilled the remission criteria at the 3-year follow-up visit, 38 (28.4%) at the 5-year follow-up visit, and 30 (22.4%) at both visits; 78.9% of patients in remission at 3 years were also in remission at 5 years.

These 30 patients in sustained remission (the remission group) were compared with the 104 patients who did not fulfil the remission criteria (DAS<1.6) at both 3 and 5 years.

Of the patients in the remission group, 21 (70%) were taking DMARDs at 3 years (methotrexate, 8 patients; sulfasalazine, 3, a combination of methotrexate and sulfasalazine, 7; and other DMARDs, 3), and 17 (56.7%) at 5 years, with 7 changing the DMARD during the 2 years of sustained remission (discontinuation in 4 patients, shift from bitherapy to monotherapy in 2 patients and DMARDs changed because of side effects in 1 patient). In all, 29%, 15% and 13% of patients treated with combination methotrexate/sulfasalazine, monotherapy by methotrexate and by sulfasalazine, respectively, at 3 years were in remission (p = 0.013).

Patients in sustained remission (n = 4; 13.3%) and those with persistently active disease (n = 23; 22%) were taking corticosteroids at baseline (p = 0.22). At the third year, the cumulative dose of prednisone and the number of days of treatment by corticosteroids were higher in the non-remission group (p = 0.001 and p = 0.001, respectively).

The median (interquartile range (IQR)) period between onset of symptoms and first DMARD was 8 (4–14) and 9 (4–14) months in the remission and non-remission groups, respectively (p = 0.74). The period between diagnosis and first treatment was 3 (2–6) and 4 (2–7) months, respectively (p = 0.34).

At baseline, these two groups had significant differences described previously,³⁶ with a lower DAS (p = 0.002), CRP (p = 0.02), rheumatoid factor IgM positivity (p = 0.02), HAQ (p = 0.04) and a trend for a lower total SHS (p = 0.15) in the remission group. Table 2 shows the 5-year results for DAS, HAQ, ESR, CRP and SHS, and indicates a significant improvement for clinical and biological variables in the remission group.

Radiographic progression after 5 years of follow-up

Table 3 shows the radiographic joint damage scores.

The total SHS (median (IQR)) at baseline in the remission and non-remission groups 0.5 (0–7) and 2 (0–7), respectively (p = 0.15). At 5 years, they were 2.5 (0–14; no significant change from baseline, p = 0.18) and 13 (3–29; significant change, p<0.001), respectively in the remission and non-remission groups, with a significant difference between these two groups (p = 0.005). The progression of damage at 5 years was higher and significantly different (p = 0.0014) for patients with rheumatoid arthritis showing persistent disease activity (Δtotal SHS = 7 (1–19)) compared with those in sustained remission (Δtotal SHS = 1.5 (0–5)).

However, analysis of the radiographs showed that 10 (33%) patients with rheumatoid arthritis in sustained remission at 3 and 5 years had a significant increase in radiographic damage (smallest detectable difference = 4.1) between baseline and 5 years. This proportion was 57 of 104 (54.8%) for patients with rheumatoid arthritis not in remission. Furthermore, 5 (16.7%) patients in sustained remission had significant radiographic damage progression between the third and the fifth years; three of these patients had no treatment during this period. Erosions were found in 11 (36.7%), 16 (53.3%) and 16 (53.3%) patients in remission at baseline, 3 and 5 years, respectively. However, erosions in a previously unaffected joint developed in 6 patients (20%) between the third and fifth years.

Functional capacity progression after 5 years of follow-up

The mean (SD) HAQ scores at baseline in the remission and non-remission groups were 1.1 (0.7) and 1.4 (0.7), respectively (p = 0.04; table 4). At 3 and 5 years, HAQ scores were 0.2 (0.4) and 0.1 (0.3), respectively, for the remission group (significant variations with baseline, p<0.001); and 0.7 (0.6) at both time points for the non-remission group (significant variation with baseline, p<0.001). There was a significant difference between the two groups at 3 and 5 years (p<0.001). The HAQ score was not significantly different between 3 and 5 years in both groups (p = 0.96 for the remission group and 0.62 for the non-remission group).

A total of 28 (93.3%) patients in sustained remission had a HAQ score ⩽0.5 at 5 years v 39 (37.5%) patients with rheumatoid arthritis with persistent disease activity (p<0.001).

DISCUSSION

The main conclusion to be drawn from this study is that sustained clinical remission according to the DAS criteria <1.6 was associated with stability of radiological damage in most patients and a clear improvement of functional capacity over 5 years in cohort of patients with early rheumatoid arthritis. The significant progression at 5 years of the SHS in patients with persistent disease activity confirms that clinical remission is the absolute goal for rheumatoid arthritis treatment to avoid progressive joint destruction, deformities and disability.

Svensson et al⁴⁵and Mottonen et al⁴⁶ found similar results for 23 and 33 remitters, respectively, who had early rheumatoid arthritis treated by DMARDs with a radiographic score not significantly higher after 2 years of follow-up.

Nevertheless, 16.7% (n = 5) of patients in sustained remission had relevant radiographical progression, and 20% (n = 6) developed erosions in a previously unaffected joint between the third and fifth years. A similar result was found by Molenaar et al.²⁶ They assessed the progression of radiological damage over 2 years in 187 patients with rheumatoid arthritis in clinical remission defined with a modification of the ACR criteria by omitting the fatigue criteria.^15,16 Remission persisted in 52% of patients, and the median radiographic score progression over 2 years was null for patients with sustained remission. However, 7% (n = 13) of patients in sustained remission had relevant progression of damage, and 15% (n = 28) developed erosions in a previously unaffected joint. Thus, patients without detectable clinical joint inflammation may continue to experience joint destruction. Makinen et al⁴⁷ and Jantti et al⁴⁸ studied the frequency of remission using several sets of criteria, including the ACR criteria (excluding fatigue) and “radiographic remission” defined as no worsening of erosion (ie, Larsen Score should not worsen by more than one point), and no new erosion from baseline to 5 years. In all, 19 patients with rheumatoid arthritis fulfilled the ACR criteria for remission, but six were not in radiographic remission. Nevertheless, in this study, the difference in the Larsen Score (one point) was minimum difference usually considered clinically important lower than the minimal clinically important difference usually considered.⁴⁹ Sokka et al²⁵ found only one patient without joint tenderness and one without joint swelling with radiographic deterioration (Larsen Score progression >1) on wrist radiographs in a cohort of 58 patients with rheumatoid arthritis.

Kirwan^29,30 showed that the link between inflammation and erosion is not clear, and proposed that two pathological processes were at work simultaneously in the joint, one leading to signs and symptoms of inflammation by lymphocytic phenomena and the other leading to direct joint destruction by synovial macrophage cells. Smolen et al³¹ also indicated that patients without clinical improvement with infliximab and methotrexate showed considerable benefit with regard to the destructive process, suggesting that in such patients these two measures of diseases were dissociated. Furthermore, Garnero et al³² showed that baseline levels of urinary markers of bone were predictive of radiological progression over 4 years in patients with early rheumatoid arthritis, independently of DAS computed on 28 joints and baseline SHS, especially for those without radiological joint damage. Several lines of evidence support the idea that osteoclasts have a role in the erosion of bone in rheumatoid arthritis.⁵⁰ Kirwan et al⁵¹ suggests that synovitis is more closely related to diffuse cartilage loss than to progression of erosion. Those authors found that the link between synovitis and erosions was abolished by glucocorticoid treatment whereas the link between synovitis and cartilage loss was not, pointing to at least two different mechanisms for these observed radiological features. For McQueen and Robinson⁵², bone oedema and synovitis detected by magnetic resonance imaging (MRI) represent two separate pathological processes, which often start together but could later diverge. Bone oedema seen on MRI may represent an intraosseous process that contributes to articular damage via a pathway that is separate from synovial inflammation.⁵²

The main differences between Molenaar’s study and ours are the cohort characteristics at baseline: their patients were older (mean age 58 years), had a longer disease duration (median 7 years) and did not receive glucocorticoids.²⁶ In fact, use of low doses of prednisolone taken in addition to traditional treatment in patients with early rheumatoid arthritis has now been shown to substantially reduce radiographic progression^{2–4,53–55} and increase the remission rate.⁵⁵ Furthermore, the follow-up in our study was longer (5 v 2 years) and the definition of remission was different. Moreover, all the patients included in Molenaar’s study were in remission and followed up for 2 years, whereas our patients were in remission only at the third-year and fifth-year evaluations.

In our study, 78.9% of patients in remission at 3 years were also in remission at 5 years. This was higher than other studies (52% after 2 years in Molenaar’s study and 50% between the second and fifth years in Makinen’s study).^26,47 In our cohort, patients with rheumatoid arthritis were considered to be in sustained remission when they had a DAS <1.6 at 3 and 5 years. Furthermore, there were no pronounced therapeutic changes during this period in all patients in sustained remission, which argues against short-term flares. Nevertheless, we cannot rule out the possibility of short-term flares not sufficiently long to warrant change of a DMARD, especially in a period in which not many treatment options were available. In addition, DAS or ACR criteria for remission may not be sufficiently sensitive to rule out some degree of residual inflammation. In fact, patients can be classified by DAS as in remission despite the continued presence of tender and swollen joints, especially when patients have extremely low levels of acute-phase reactants.⁵⁶ Synovitis may also be present at a subclinical level and may cause subsequent bone erosion. Conaghan et al⁵⁷ did not find a mismatch between synovitis detected by MRI and subsequent bone damage, the number of new erosions being proportional to the amount of synovitis in a given joint.

The functional capacity was improved in both the remission and non-remission groups but considerably more so in the remission group. Nevertheless, the HAQ score was lower at baseline in the remission group, and this value was a predictive factor of the 5-year HAQ score and a prognostic factor of remission in early rheumatoid arthritis.^35,36 In previous studies, Welsing et al²¹ had showed that functional capacity was mainly associated with disease activity in early rheumatoid arthritis, and Molenaar et al²⁴ that functional disability in patients with rheumatoid arthritis in remission was most strongly related to the presence of pain and to a lesser extent to disease activity.

To conclude, our data showed that remission based on the DAS criteria is a clinically relevant goal for management of early rheumatoid arthritis despite the fact that mild radiographic progression may occur in some patients. These data suggest that radiographic damage may be partly independent of clinical joint inflammation, and that regular monitoring of radiographic damage in patients with early rheumatoid arthritis should be mandatory in addition to frequent disease activity measurement, as was recently recommended. Use of MRI and ultrasonography in the follow-up of patients with rheumatoid arthritis in remission for evaluating non-detectable clinical joint inflammation should be studied. Finally, absence of radiographic progression should be part of the remission definition in rheumatoid arthritis even though further studies are needed to decide what degree of radiological progression in patients in clinical remission may have relevance for long-term outcomes.