Progression of radiologic damage in patients with rheumatoid arthritis in clinical remission

Abstract

Objective

To assess whether radiologic progression occurs during clinical remission in patients with rheumatoid arthritis (RA).

Methods

One hundred eighty-seven patients with RA in clinical remission were followed up clinically and radiologically for 2 years. Clinical remission was defined according to a modification of the American College of Rheumatology criteria (i.e., the criterion of fatigue was omitted, and patients had to fulfill 4 of the 5 remaining criteria). Radiologic joint damage was assessed by the Sharp/van der Heijde method.

Results

After 2 years of followup, remission persisted in 52% of patients. The median radiologic score for the total group of patients increased from 21 (interquartile range [IQR] 5, 65) at the time of entry to 25 (IQR 7, 72) after 2 years (P < 0.001). The median score for radiologic progression between baseline and 2 years was 0.5 (IQR 0, 2.5). Among patients with an exacerbation of RA (n = 86), the median score for progression over 2 years was 1.0 (IQR 0, 4.5) (P < 0.001), and in patients with a persistent remission (n = 93) it was 0 (IQR −0.5, 2.0) (P < 0.001). Clinically relevant progression of damage was more frequent in patients with exacerbation (23%) than in those with persistent remission (7%) (P = 0.001). However, in 15% of patients with persistent remission, an erosion developed in a previously unaffected joint. In the logistic regression analysis, the area under the curve of the Disease Activity Score, a continuous measure, was related to the chance of radiologic progression, regardless of the absolute disease activity level. Results were similar when other definitions of remission were used.

Conclusion

Although rare, clinically relevant progression of joint damage does occur in patients with RA in prolonged remission. This suggests the need for markers that predict progression during periods of low disease activity and for drugs that prevent damage that is independent of disease activity.

Rheumatoid arthritis (RA) is characterized by the presence of disease activity and bone destruction. The pattern of disease activity varies considerably between patients (1, 2). Disease activity can be persistent, but exacerbations and periods of low disease activity or even remission can also occur. In the majority of patients, RA causes progressive joint destruction, deformities, and disability. Only a minority of patients achieve a permanent remission (3). The process of bone destruction in RA is correlated with arthritis activity (4), although evidence is accumulating that bone destruction can occur independently of arthritis activity (5–7).

Two sets of criteria for clinical remission are currently being used: the criteria proposed by the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) (8) and the criteria based on the Disease Activity Score (DAS) (9). In general, it is estimated that remission occurs in 10–20% of patients with RA (3, 9–12). The occurrence of clinical remission has been studied in patients with early RA and short-term followup (9, 12) and in long-term cohort studies (10). All of these studies were undertaken in groups of patients with various levels of disease activity, and in particular those with high levels of disease activity. However, only a relatively small number of patients with RA in clinical remission have been studied prospectively, and knowledge about factors that are possibly associated with remission is limited. Stopping use of second-line antirheumatic drugs during remission increases the chance of a disease flare (13).

Achieving clinical remission is an important milestone, but such remission is unsatisfactory if joint destruction is not halted. Whether the absence of arthritis activity prevents further joint damage is still a matter of debate. To our knowledge, longitudinal studies investigating the relationship between joint damage and clinical remission are lacking. The present study was undertaken over a 2-year followup period in a cohort of patients with RA in clinical remission, in order to determine whether radiologic joint damage worsens during remission.

PATIENTS AND METHODS

Patients.

The study was designed as a prospective cohort study consisting of 187 patients with RA (14) in clinical remission. Patients were recruited between November 1996 and December 1998. The majority of patients (n = 135) were recruited from the Jan van Breemen Instituut, which is a large outpatient clinic covering an area in the Amsterdam region with ∼1 million inhabitants. Approximately 2,400 patients with RA are treated in this outpatient clinic. An additional 52 patients were recruited from neighboring outpatient clinics in the western part of The Netherlands.

The first selection of patients was done using a clinical database and by chart review: all patients at the Jan van Breemen Instituut who had a diagnosis of RA and an erythrocyte sedimentation rate (ESR) of <20 mm/hour (men) and <30 mm/hour (women) for a period of at least 6 months were selected. Subsequently, patients were evaluated for fulfillment of the ACR criteria for clinical remission (8). According to these criteria, clinical remission is present when 5 of the following 6 requirements are fulfilled: 1) duration of morning stiffness not exceeding 15 minutes, 2) no fatigue, 3) no joint pain, 4) no joint tenderness or pain on motion, 5) no soft tissue swelling in joints or tendon sheaths, and 6) ESR (Westergren method) <30 mm/hour for a female or <20 mm/hour for a male.

Among patients in the present study, fatigue had not been assessed routinely. Therefore, clinical remission was defined by a modification of the ACR criteria; i.e., by omitting the fatigue criterion, a patient was considered to be experiencing clinical remission when 4 of the remaining 5 ACR criteria were fulfilled at the time of inclusion in the study, as assessed by one investigator (EM). In order for a patient to be included in the study, his or her RA had to have been in clinical remission for at least the previous 6 months. Patients were allowed to take disease-modifying antirheumatic drugs (DMARDs) and nonsteroidal antiinflammatory drugs. The present study was part of a study on bone markers (15), and because of the possible effect of steroids on such markers, patients receiving treatment with glucocorticoids were excluded.

During the 2-year followup period, disease activity was assessed every 3 months by the same investigator (EM), in particular for the occurrence of an exacerbation. Patients who did not meet our criteria for clinical remission at a certain time point were classified as having an exacerbation. Because the assessments were performed every 3 months, we decided that patients should also be classified as having an exacerbation if DMARD therapy was changed by the treating rheumatologist because of worsening of arthritis activity.

The study protocol was approved by the local medical ethical committees of each participating hospital.

Demographic and clinical characteristics.

The following demographic characteristics were recorded at baseline: age, sex, duration of RA, treatment with DMARDs, number of DMARDs ever prescribed, and presence of increased rheumatoid factor (RF) concentrations. Clinical characteristics, which were documented at baseline and at each followup visit, were as follows: use of DMARDs, morning stiffness lasting more or less than 15 minutes, number of swollen joints (maximum 66), number of tender joints (maximum 68) (16), Ritchie articular index score (17), visual analog scale (VAS) scores for the patient's impression of general health and pain, VAS score for the physician's impression of disease activity, and the score on a Dutch version of the Health Assessment Questionnaire (18). Also, the DAS was computed. The DAS is a composite index consisting of a count of 44 joints assessed for swelling, the Ritchie score, the ESR, and the VAS score for general health (19). Finally a DAS area under the curve (AUC) was calculated, which represented the mean of the DAS as measured every 3 months.

Laboratory assessments.

The levels of IgA and IgM RF were measured at baseline. The ESR and the C-reactive protein level were measured during each visit. HLA typing was performed to assess shared epitope–positive alleles (20).

Radiologic assessments.

Radiographs of the hands and feet were obtained at baseline, at 1 year, and at 2 years. Radiologic damage of the hands and feet was assessed according to the Sharp/van der Heijde method (21). The radiographs were scored in random order without knowledge of the clinical data, by 2 experienced observers (EM and HD). Intraclass correlation coefficients (22) for interobserver and intraobserver variability were 0.98 and 0.99, respectively.

Statistical analysis.

Wilcoxon's signed rank test was used to assess differences in Sharp/van der Heijde joint scores between baseline and 1- and 2-year followup. Subsequently, the relationship between arthritis activity and progression of radiologic joint damage was evaluated. Clinically relevant progression was defined as an increase of the Sharp/van der Heijde score of ≥5 after 2 years. The cutoff point of 5 was chosen based on results of a preliminary analysis (by the same investigators involved in the present study) of the 1-year followup data in patients with RA in remission. That analysis demonstrated that progression of >5 points during 1 year of followup should be interpreted as progression beyond measurement error (23). In a logistic regression analysis, we evaluated whether the observed radiologic progression could be explained by baseline variables such as disease duration, disease activity as expressed by the DAS, the Sharp/van der Heijde score, IgM RF, IgA RF, or use of DMARDs. The analysis was performed twice, by defining radiologic progression as a Sharp/van der Heijde score of >1 and a Sharp/van der Heijde score of >5.

RESULTS

Demographic and clinical data at baseline.

The study subjects were predominantly female (65%), with mean (±SD) age of 58 ± 13 years and a median disease duration of 7 years (interquartile range [IQR] 3–12 years). The majority of the patients (69%) were ever RF positive, but only a minority (39%) had increased IgM RF titers at baseline. Increased levels of IgA RF at baseline were observed in 63% of the patients, and erosions were observed in 92%. Shared epitope alleles were carried by 69% of the patients. Disease activity in the whole group was low, as measured by joint counts, acute-phase reactants, and VAS scores. Seventy percent of the patients had ever been treated with a DMARD at the time of study entry. The majority of patients (53%) had been treated with at least 1 DMARD, while only one-fifth of patients had been treated with steroids. The most commonly used DMARDs at the time of inclusion were sulfasalazine, intramuscular gold, methotrexate, and antimalarials.

Of all 187 patients with RA in clinical remission according to our (modified) definition, 110 fulfilled the original ACR definition because they met 5 criteria (fatigue was not measured), 152 patients fulfilled the DAS remission criteria (i.e., DAS of <1.6) (9), and 107 fulfilled both sets of criteria. During the 2-year followup period, 2 patients died because of cardiovascular disease, and 5 were lost to followup for several reasons. Remission persisted in 52% of the 187 patients meeting our definition of clinical remission, in 59% of the patients fulfilling the ACR criteria for remission, and in 42% of those experiencing remission according to the DAS.

During the first study year, 140 patients had a (time-averaged) DAS AUC below 1.6. Of those 140 patients, 95 continued to experience clinical remission of disease during the second study year. Of the 40 patients with a DAS AUC of ≥1.6 in the first study year, only 7 continued to experience clinical remission during the second study year.

Radiologic followup.

Radiologic followup data were available for 179 patients (Tables 1–3. Progression of joint damage occurred in the total group of RA patients but was most prominent in the subgroup of patients with disease exacerbation. However, even in patients with RA in persistent clinical remission, a slight but statistically significant progression in radiologic abnormalities was evident. Significant progression of joint damage between baseline and 2 years was also observed in those patients meeting the ACR and DAS criteria for remission (data not shown).

Table 1. Radiologic joint damage scores in patients with RA in clinical remission at baseline*

Patient group	Sharp/van der Heijde score
	Baseline	1-year followup	2-year followup	Δ, baseline and 1-year followup	Δ, baseline and 2-year followup
* Except where indicated otherwise, P values were determined by Wilcoxon's signed rank test. RA = rheumatoid arthritis. † P < 0.05, versus baseline. ‡ P < 0.001, versus baseline. § P = 0.004, versus baseline. ¶ P < 0.001 by Mann-Whitney U test, versus patients with exacerbation.
Complete cohort (n = 179)
Mean ± SD	49 ± 68	50 ± 69	51 ± 69	1.2 ± 2.9	2.2 ± 4.7
Median (range)	21 (0, 365)	23 (0, 368)†	25 (0, 368)‡	0 (−6, 16)	0.5 (−5, 28)
Interquartile range	5, 65	6, 67	7, 72	0, 1.5	0, 2.5
Persistent remission (n = 93)
Mean ± SD	52 ± 72	52 ± 72	53 ± 72	0.8 ± 2.7	1.2 ± 3.8
Median (range)	21 (0, 338)	21 (0, 338)†	23 (0, 338)§	0 (−5, 11)¶	0 (−5, 18)‡
Interquartile range	7, 67	7, 70	−8, 73	0, 0.9	−0.5, 2.0
Exacerbation of RA (n = 86)
Mean ± SD	46 ± 64	47 ± 64	51 ± 69	1.5 ± 3	3.2 ± 5.3
Median (range)	23 (0, 365)	23 (0, 368)†	25 (0, 368)‡	1.5 (−6, 16)	1.0 (−2.5, 28)‡
Interquartile range	4, 61	5, 67	5, 70	0, 2	0, 4.5

Table 2. Radiologic joint damage scores in patients with RA in clinical remission as defined by the ACR criteria*

Patient group	Sharp/van der Heijde score
	Baseline	1-year followup	2-year followup	Difference between baseline and 1-year followup	Difference between baseline and 2-year followup
* Except where indicated otherwise, P values were determined by Wilcoxon's signed rank test. RA = rheumatoid arthritis; ACR = American College of Rheumatology. † Radiologic data were available for only 104 patients. ‡ P < 0.05, versus baseline. § P < 0.05, versus baseline and 1-year followup. ¶ P < 0.05, versus 1-year followup. # P < 0.05 by Mann-Whitney U test, versus patients with exacerbation.
Complete cohort (n = 110)†
Mean ± SD	44 ± 62	46 ± 64	43 ± 58	0.7 ± 2.3	1.4 ± 3.3
Median (range)	17 (0, 273)	20 (0, 270)‡	21 (0, 272)§	0 (−6, 11)	0.5 (−5, 18)
Interquartile range	4–48	6–53	6–52	0–1	0–2
Persistent remission (n = 65)
Mean ± SD	44 ± 60	46 ± 60	46 ± 60	0.7 ± 2.5	1.2 ± 3.6
Median (range)	21 (0, 273)	22 (0, 271)‡	23 (0, 272)¶	0 (−5, 11)	0 (−5, 18)#
Interquartile range	7–46	7–53	7–55	0–0.5	0–2
Exacerbation of RA (n = 39)
Mean ± SD	43 ± 66	47 ± 71	39 ± 52	0.8 ± 1.9	1.7 ± 2.3
Median (range)	14 (0, 267)	15 (0, 267)‡	16 (0, 224)§	0.5 (−6, 6)	1.0 (−3, 9)
Interquartile range	4–48	5–53	5–51	0–1.5	0–4

Table 3. Radiologic joint damage scores in patients with RA in clinical remission as defined by the DAS criteria*

Patient group	Sharp/van der Heijde score
	Baseline	1-year followup	2-year followup	Difference between baseline and 1-year followup	Difference between baseline and 2-year followup
* Except where indicated otherwise, P values were determined by Wilcoxon's signed rank test. RA = rheumatoid arthritis; DAS = Disease Activity Score. † Radiologic data were available for only 144 patients. ‡ P < 0.05, versus baseline. § P < 0.05, versus baseline and 1-year followup. ¶ P < 0.001 by Mann-Whitney U test, versus patients with exacerbation.
Complete cohort (n = 152)†
Mean ± SD	44 ± 59	45 ± 60	46 ± 60	0.9 ± 2.7	1.7 ± 4.3
Median (range)	19 (0, 338)	22 (0, 338)‡	22 (0, 338)§	0 (−6, 16)	0.5 (−5, 28)
Interquartile range	5–57	6–61	6–63	0–1	0–2
Persistent remission (n = 85)
Mean ± SD	47 ± 65	48 ± 65	48 ± 65	0.7 ± 2.6	1.1 ± 3.5
Median (range)	19 (0, 338)	22 (0, 338)‡	22 (0, 338)§	0 (−5, 11)¶	0 (−5, 18)¶
Interquartile range	7–60	7–65	7–69	0–0.5	0–1.5
Exacerbation of RA (n = 59)
Mean ± SD	39 ± 51	40 ± 52	42 ± 53	1.3 ± 2.8	2.6 ± 5.1
Median (range)	20 (0, 208)	20 (0, 212)	25 (0, 214)§	0.5 (−6, 16)	1.0 (−3, 28)
Interquartile range	4–58	4–58	5–59	0–2	0–4

Patients with an exacerbation during followup were much more likely to have relevant progression of damage (Δ in Sharp/van der Heijde score of ≥5) than were patients with persistent remission (23% versus 7%; P = 0.001) (Figure 1). Similar differences in relevant progression rates were seen when the DAS criteria for remission were used (17% versus 6%; P = 0.017), while only a trend was observed when the ACR definition was applied (10% versus 7%; P = 0.053). Likewise, the DAS AUC was also significantly higher in patients with relevant progression (1.7 [IQR 0.7–2.8]) than in patients with low or no progression (1.3 [IQR 0.4–2.8]) (P = 0.001).

Figure 1. Individual progression scores of patients with rheumatoid arthritis in persistent clinical remission (n = 93) and those with an exacerbation (n = 86) during 2-year followup. SHS = Sharp/van der Heijde score.

In the logistic regression analysis, no baseline variable had an important influence on the observed association between persistent remission and joint destruction (data not shown). However, the DAS AUC was a stronger predictor of radiologic progression than was the absence of persistent remission.

The development of an erosion in a previously unaffected joint was evaluated in patients with RA in persistent clinical remission (according to our criteria) for 2 years. Such erosions were found in the same joint in 14 patients, by both observers and independently of each other. Of those 14 patients with RA in persistent remission according to our criteria, 11 fulfilled the ACR criteria for clinical remission and 9 fulfilled the DAS criteria for clinical remission for 2 years. Erosions in a previously unaffected joint developed in 15% of patients (5 patients in the first year of the study, 6 patients in the second year, and 3 patients in both years of followup).

DISCUSSION

The main conclusion to be drawn from the present study is that radiologic progression can occur during a state of persistent remission as defined by our remission criteria, but also as defined by the ACR and DAS criteria.

Studies with large numbers of patients with RA in clinical remission are limited and give rise to several intriguing questions (10, 13). The first question is whether a state of clinical remission is persistent. In about half of our patients, clinical remission persisted in the second year of the study. However, remission persisted in 68% of patients with a DAS AUC of <1.6 over the first year. As expected, this confirms that repeated observations are more reliable than a single observation (e.g., at baseline) as an indicator of a stable low level of disease activity.

The radiologic progression observed over 1 year in our group of patients with RA in remission is consistent with that in patients with active RA treated with combination therapy or tumor necrosis factor–blocking agents. In the Combinatietherapie Bij Reumatoïde Artritis (COBRA) study, patients treated with sulfasalazine had, after 1 year, a median score for progression of 6 (range 0–54), whereas patients treated with combination therapy of methotrexate, sulfasalazine, and prednisone had a median score of 2 (range 0–43) (24). In the Anti–Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy (ATTRACT) study, the mean (±SD) score for progression over 1 year was between −0.7 ± 3.8 and 1.6 ± 8.5 in patients treated with infliximab, whereas in patients treated with methotrexate only it was 7.0 ± 10.3 (25). In the etanercept study, the mean score for progression over 1 year was 1.0 in patients treated with etanercept, whereas in patients treated with methotrexate it was 1.6 (26).

It is interesting to note the differences between the frequency of patients fulfilling the ACR remission criteria and the DAS remission criteria, both at baseline and after followup. Although one of the ACR remission criteria (fatigue) was not evaluated in our patients, and comparison between the remission criteria used may therefore be somewhat difficult, the observed differences stress the need for consensus on the use of remission criteria in the near future.

Stratification according to disease activity revealed that progression of joint damage was present in both the group of patients with an exacerbation of RA and the group of patients with RA in persistent clinical remission. In fact, the major part of radiologic progression can be attributed to disease activity, but some patients had considerable progression despite persistent remission. The relationship between disease activity and radiologic progression is consistent with that observed in previous studies (4, 27). However, the unique finding in this study of radiologic progression during persistent clinical remission supports the hypothesis that synovitis and joint destruction are 2 different and independent processes (5–7). Other support for this hypothesis comes from studies on bone markers, showing that levels of bone markers are increased in patients with RA in remission (15), and that a reduction of the levels of bone markers is associated with a reduction in long-term joint damage, independent of arthritis activity (28, 29).

The results of the present study impact the whole concept of remission, including radiologic parameters. The ultimate goal of therapy is to achieve a state of inactive disease, both clinically and structurally. Perhaps structure should be included in a new definition of remission. Also, more attention should be focused on the evolution of remission during followup, to determine the influence of exacerbations and smaller fluctuations in disease activity on structure and long-term outcome.

In conclusion, half of patients with RA in clinical remission can expect a relapse within 2 years, which often is preceded by an increase in the DAS. Radiologic progression occurs in relation to disease activity, but persistent remission is not fully protective. It is proposed that structure is an important dimension of the concept of remission.

Acknowledgements

We thank the following rheumatologists for their participation: Dr. J. M. W. Hazes, Leiden University Medical Centre, Leiden, and Erasmus University Medical Centre, Rotterdam; Dr. A. J. Peeters, Reinier de Graaf Gasthuis, Delft; Drs. J. P. Terwiel and C. Mallee, Spaarne Ziekenhuis and Kennemer Gasthuis, Haarlem; and Drs. M. L. Westedt and I. Speijer, Bronovo Ziekenhuis, Den Haag.