Minimal disease activity remission and the long-term outcomes of rheumatoid arthritis.код для вставкиСкачать
Arthritis & Rheumatism (Arthritis Care & Research) Vol. 57, No. 6, August 15, 2007, pp 935–942 DOI 10.1002/art.22895 © 2007, American College of Rheumatology ORIGINAL ARTICLE Minimal Disease Activity, Remission, and the Long-Term Outcomes of Rheumatoid Arthritis FREDERICK WOLFE,1 JOHANNES J. RASKER,2 MAARTEN BOERS,3 GEORGE A. WELLS,4 KALEB MICHAUD5 AND Objective. To determine the prevalence of minimal disease activity (MDA) and remission in patients with rheumatoid arthritis (RA), to assess the effect of anti–tumor necrosis factor (anti-TNF) therapy on MDA, and to determine the extent to which MDA status improves long-term outcomes. Methods. Using the Patient Activity Scale (PAS) as a surrogate, we assessed the prevalence of MDA and remission in 18,062 patients with RA using the newly developed Outcome Measures in Rheumatology Clinical Trials (OMERACT) criteria for MDA. Results. MDA was noted in 20.2% of 18,062 patients and persistent MDA, operationally deﬁned as having MDA during >2 consecutive 6-month observation periods, occurred in 13.5% of 7,433 patients followed longitudinally. Disease activity at remission levels was noted in 7%. Among patients with MDA, 82% received disease-modifying antirheumatic drugs or biologic agents. Following anti-TNF initiation, the cumulative probability of achieving MDA at 2 and 6 years was 4.1% and 7.6%, respectively, and persistent MDA probabilities were 2.7% and 4.5%, respectively. Regardless of RA duration, patients with MDA had substantially better outcomes, including a 10-fold reduction in work disability and an approximately 2-fold reduction in total joint replacement and mortality. Conclusion. Remission remains uncommon in RA, and the prevalence of new remission in community practice is substantially lower than noted in published trials of biologic therapy. On average, persons with MDA appear to have persistently mild RA. This might be the effect of milder RA and/or more effective treatment in early RA. The PAS had satisfactory levels of agreement with the full MDA criteria and appears suitable for use in clinical and epidemiologic research. KEY WORDS. Minimal disease activity; Remission; Tumor necrosis factor; Biologics; Rheumatoid arthritis. INTRODUCTION Remission, the absence of disease activity, is the most desirable outcome for patients with rheumatoid arthritis (RA), and recent advances in antirheumatic therapy appear to have made remission a realistic goal for many 1 Frederick Wolfe, MD: National Data Bank for Rheumatic Diseases, and the University of Kansas School of Medicine, Wichita, Kansas; 2Johannes J. Rasker, MD: University of Twente, Enschede, The Netherlands; 3Maarten Boers, MD, PhD: VU University Medical Centre, Amsterdam, The Netherlands; 4George A. Wells, PhD: University of Ottawa, Ottawa, Ontario, Canada; 5Kaleb Michaud, MS: National Data Bank for Rheumatic Diseases, Wichita, Kansas, and the Center for Primary Care and Outcomes Research, Stanford University, Stanford, California. Address correspondence to Frederick Wolfe, MD, National Data Bank for Rheumatic Diseases, Arthritis Research Center Foundation, 1035 N Emporia, Suite 230, Wichita, KS 67214. E-mail: email@example.com. Submitted for publication August 21, 2006; accepted in revised form February 5, 2007. patients (1–3). However, as simple as the deﬁnition of remission seems to be, its actual determination is difﬁcult. In 1981, the American College of Rheumatology (ACR; formerly the American Rheumatism Association) published preliminary remission criteria (4). These criteria were highly speciﬁc, but almost impossible to fulﬁll when applied strictly. Although a few studies used the ACR criteria or its modiﬁcations (5), other research opted to deﬁne remission operationally (6 –12). With the advent of improved RA treatment, the ability to measure remission accurately became more important. Among the difﬁculties in deﬁning remission are several problems. In some instances, swollen and/or painful joints may be caused by factors other than active RA. In addition, when erythrocyte sedimentation rate (ESR) or C-reactive protein level (CRP) are part of a remission deﬁnition, abnormal values unrelated to RA activity could result in exclusion of some patients from remission classiﬁcation even though they have achieved clinical remission. Remission misclassiﬁcation might also occur when persons who are in remission based on the number of swollen and 935 936 painful joints and ESR/CRP level have high levels of pain. As with many issues in rheumatology, the ability to identify remission in the individual patient must often be differentiated from the ability to identify remission in groups of patients. A modern approach probabilistically addresses remission by identifying a point on a disease activity continuum where remission is identiﬁed with the highest level of accuracy. Using the original Disease Activity Score (DAS), remission has been reported as a DAS score ⬍1.6 by Prevoo et al in a study that used the original ACR criteria (4) as an end point (6). More recently, a Disease Activity Score in 28 joints (DAS28) ⬍2.6 was suggested as a remission end point (13), a deﬁnition that is in accord with the Prevoo et al deﬁnition (6). Such deﬁnitions (14) have had wide use in the description of outcomes of clinical trials, and results of trials of biologic agents frequently describe a remission outcome. Probability-based remission criteria, however, do not eliminate misclassiﬁcation, but they do allow for its quantiﬁcation. Recently Wells et al, acting on an idea from Maarten Boers, suggested new probabilistic approaches to remission that appeared to overcome many of these problems (15). In addition, they recognized the difﬁculty in agreement regarding what is actually remission, opting instead to deﬁne minimal disease activity (MDA), which is a point on a continuum of RA disease activity that is near remission. In addition, MDA is a point on the disease activity spectrum that is of considerable interest itself (see Patients and Methods section for description of MDA deﬁnition). We used a surrogate measure of the Wells/Boers/ Outcome Measures in Rheumatology Clinical Trials (OMERACT) deﬁnition of MDA and large data sets to explore the prevalence of remission and MDA in RA. Wells et al and the OMERACT committee (15) suggested the need for examining the MDA deﬁnition in external data sets. In addition, we assessed the probability of being in an MDA state following anti–tumor necrosis factor (anti-TNF) therapy. Finally, we described the consequence of MDA on long-term RA outcomes. PATIENTS AND METHODS Patients in this study were participants in the National Data Bank for Rheumatic Diseases (NDB) longitudinal study of RA outcomes. NDB participants are recruited on an ongoing basis from the practices of US rheumatologists, and are followed prospectively with semiannual, detailed, 28-page questionnaires, as previously described (16 –18). The analyses in this report used several subsets of NDB participants, as will be described. During the NDB preenrollment process, patients with RA completed brief screening questionnaires in their rheumatologists’ ofﬁces. Patients who completed this preenrollment process (n ⫽ 28,646) were then invited to participate in the (full) NDB long-term outcome study by completing detailed semiannual questionnaires. A total of 19,052 (66.7%) patients were fully enrolled into the longitudinal study. Of the 28,646 patients, 10,584 were identiﬁed as being enrolled in safety registries for leﬂunomide and inﬂiximab. Because Wolfe et al Figure 1. Disease activity among National Data Bank for Rheumatic Diseases patients with rheumatoid arthritis (RA) as determined in a randomly selected observation of the 18,062 patients. Deﬁnitions are based on estimations from the Patient Activity Scale (PAS). DAS ⫽ Disease Activity Score; DAS28 ⫽ Disease Activity Score in 28 joints; MDA ⫽ minimal disease activity; WHO ⫽ World Health Organization; ILAR ⫽ International League of Associations for Rheumatology; OMERACT ⫽ Outcome Measures in Rheumatology Clinical Trials. these patients were selected by the enrollment process for the severity of their RA, they were excluded from analysis to avoid severity bias. The remaining 18,062 patients constituted the patient group used for prevalence estimates of MDA and the data shown in Figures 1 and 2. Of the 18,062 patients whose data were used for the MDA prevalence estimates, 12,682 continued in the longterm followup study. In the longitudinal analyses of this study, this group was further restricted to include 7,433 patients who underwent at least 3 semiannual observations and who were not receiving anti-TNF therapy at the ﬁrst observation. By this restriction we were able to determine prospectively the percentage of patients who achieved MDA status after starting anti-TNF therapy. All NDB participants provided demographic information and completed the Health Assessment Questionnaire (HAQ) disability index (19,20) and visual analog scales (VAS) for pain intensity and global severity. These 3 scales were used to calculate the Patient Activity Scale (PAS) (21). The PAS is a 0 –10 measure of RA disease activity in which higher values indicate greater RA disease activity. The PAS is computed by multiplying the HAQ by 3.33 and then dividing the sum of the VAS pain, VAS global, and HAQ by 3. The PAS may also be computed by using the shorter HAQ-II (22) instead of the HAQ, with essentially identical results. HAQ-II results were not reported in the current study. In addition to these data, patients who enrolled in the detailed longitudinal followup study reported all medication use, joint replacement surgery, and US Social Security disability awards. Analysis of disability awards was restricted to patients ⬍62 years of age. Data on deaths were obtained from family and physicians, and from yearly systematic searches of the National Death Index (23,24). We also made use of a second data set, the Rheumatoid Arthritis Evaluation Study (RAES) data set (25). The RAES data set contains examination data on 619 patients re- Minimal Disease Activity and Remission in RA 937 deﬁnitions for MDA were designed and discussed at the OMERACT VII conference in 2004. Feedback from participants and additional on-site analyses in the cross-sectional (RAES) database allowed the formulation of 2 preliminary, equivalent deﬁnitions of MDA: one based on the DAS28 index of disease activity and one based on meeting cut points in 5 of the 7 World Health Organization (WHO)/ International League of Associations for Rheumatology (ILAR) core set measures (15). We used the Wells et al deﬁnition in our analyses of MDA. In the ﬁrst method, MDA was said to be present if the patient satisﬁed at least 5 of the following 7 conditions: VAS pain ⱕ2 (0 –10), swollen joint count ⱕ1 (0 –28), tender joint count ⱕ1 (0 –28), HAQ ⱕ0.5 (0 –3), patient global severity ⱕ2 (0 –10), physician global disease activity ⱕ1.5 (0 –10), and ESR ⱕ20 mm/hour, or if the patient satisﬁed the following conditions: had no swollen joints, no tender joints, and an ESR ⱕ10 mm/hour. The deﬁnition was called the core set deﬁnition because it made use of the published core set variables for the assessment of response in patients with RA (30,31). In the second method (the DAS28 method), patients achieved MDA status if they had no swollen joints, no tender joints, and an ESR ⱕ10 mm/ hour or if their DAS28 score was ⱕ2.85. Figure 2. Percentage of patients satisfying minimal disease activity (MDA) and remission deﬁnitions according to A, duration of rheumatoid arthritis (RA) and B, duration of RA and the use of anti–tumor necrosis factor (anti-TNF) therapy. Each patient contributed a single random observation. For clarity and relevance, analyses were restricted to a subset of the 18,062 patients in Figure 1: A, the 13,128 patients with RA durations ⱕ30 years (— ⫽ MDA [core set]; – – – ⫽ remission [physician]) and B, patients with RA durations ⱕ30 years who enrolled in the National Data Bank for Rheumatic Diseases and provided treatment data at the randomly selected observation (n ⫽ 7,802) (— ⫽ MDA anti-TNF positive; – – – ⫽ MDA anti-TNF negative; ----- ⫽ remission anti-TNF positive; – 䡠 – 䡠 ⫽ remission anti-TNF negative). MDA is deﬁned as a Patient Activity Scale (PAS) value ⱕ1.625 and remission at a PAS value ⱕ0.5. cruited from the practices of Canadian and American academic and community physicians who were each asked to examine ⬃10 consecutive patients with RA in their clinics. In addition to the HAQ, VAS pain, and VAS global, the RAES data set contains information on the number of tender and swollen joints, ESR, physician global assessment, and an estimate by the physician if the patient was in remission (had no disease activity). From these data, the RA disease activity score (DAS28) was calculated. The DAS28 score is a well-recognized measure of RA disease activity (26 –29). Wells et al described the MDA deﬁnition derived from the OMERACT consensus conference (15). The consensus report made use of analyses from the RAES data set. In the Wells et al study, rheumatologists assessed 60 patient proﬁles describing real patients with RA seen in routine clinical practice. Based on their responses, several candidate Relationship between MDA and clinical measures. Because the NDB data set does not contain all of the core set or DAS28 variables, we used the PAS as a surrogate to determine MDA. Using the RAES data set, we determined that a PAS score ⱕ1.625 best identiﬁed MDA deﬁned by the core set deﬁnition. This value of the PAS correctly identiﬁed 87% of MDA cases and had a sensitivity of 69.5%, speciﬁcity of 90.6%, odds ratio (OR) of 22.0, a kappa value of 0.60 (95% conﬁdence interval [95% CI] 0.52– 0.68), and a McNemar chi-square value of 0.05 (P ⫽ 0.825). The area under the receiver operating characteristic curve (AUC) for the PAS was 0.893 (the AUC was 0.859 for the HAQ and 0.859 for VAS pain; the difference between PAS AUC and the HAQ and pain AUC was signiﬁcant at the 0.01 level). Kappa is biased downward (prevalence bias) and underestimates agreement when the percentage correct is substantially greater than 50%, and was not relevant in this context (32,33). However, kappa is presented here for completeness. The PAS score that best described the DAS28 MDA was 1.75, percentage correct was 79%, sensitivity was 54.0%, speciﬁcity was 85.9%, OR was 7.4, kappa was 0.39 (95% CI 0.31– 0.48), and McNemar chi-square was 0.19 (P ⫽ 0.662). Because PAS determination of the DAS28 MDA had substantially reduced reliability compared with the PAS-based core set estimate of the MDA, we chose to use the PAS-based core set estimate of the MDA. Because all of the PAS variables are measured in the core set and only one of the PAS elements is measured in the DAS28, it was expected that the PAS core set estimate of the MDA would have better reliability than the PAS DAS28 estimate of the patient disease activity level. We also estimated remission in the RAES data set based on rheumatologists’ clinical judgment. In the RAES study, rheumatologists reported remission by indicating a global 938 Wolfe et al disease activity score of 0 (25). For comparison purposes only, the proportion of patients in remission using the DAS28 remission recommendation of 2.6 (13), a value that had an equivalent PAS score ⱕ1.25, is also shown in Figure 1. For clarity, we summarized the 2 data sets used in this report. The ﬁrst was the 619 patients in the RAES data set. The purpose of using the RAES data set was to develop the cut points in the PAS for all of the MDA and remission criteria. The second data set, the NDB data set, made use of 2 groups. The ﬁrst group comprised 18,062 patients who were not members of safety registries and whose data were used for prevalence estimates. The second group was a subset of the ﬁrst group and comprised 7,433 patients who fully enrolled in the NDB, contributed at least 3 semiannual observations, and were not receiving anti-TNF therapy at their ﬁrst observation. Data were analyzed using Stata software, version 9.1 (StataCorp, College Station, TX). Differences between groups in cross-sectional analyses were studied by logistic regression. Longitudinal data were analyzed by Cox proportional hazards regression. Graphs of outcomes versus disease duration were constructed using Stata’s locpoly command (34) to produce kernel-weighted local polynomial smooths. An Epanechnikov kernel was used with a bandwidth of 2 for Figures 2A, 2B, and 3B, except for the MDA negative line in Figure 3B, where the bandwidth was 0.7. RESULTS MDA and remission in the RAES data set. The mean ⫾ SD age and disease duration of patients with RA in the RAES data set were 58.0 ⫾ 13.5 years and 12.5 ⫾ 10.5 years, respectively, and 73% were women. Of these 619 patients, MDA occurred in 20.8% using the core set deﬁnition and 22.1% using the DAS28-based deﬁnition. Rheumatologists’ disease activity ratings of 0 or 1 on the 0 –10 scale, which may be thought to correspond to a clinical deﬁnition of minimal activity, occurred in 22.9% of patients. In addition, remission as deﬁned by examining rheumatologists on clinical grounds occurred in 6.1%. MDA and remission in the NDB. In a randomly selected observation, MDA prevalence in 18,062 NDB patients was 20.6% and 22.5% as measured by the PAS-estimated core set deﬁnition and the PAS-estimated DAS28 deﬁnition, respectively (Figure 1). The remission prevalence was 7% by the PAS estimation (PAS ⱕ0.5), which was derived from the RAES data set remission values. When plotted against time, remission and MDA were more common in patients with shorter durations of RA (Figure 2A). Persistent MDA, deﬁned as MDA for ⱖ2 semiannual periods, occurred in 67% of patients with MDA on a single occasion, and persistent remission levels occurred in 57% of patients who had achieved remission once. Treatment, MDA, and remission. To evaluate the relationship between anti-TNF therapy and disease activity at Figure 3. A, Percentage of patients who received a Social Security disability award, had a total joint replacement, or were noted to have died subsequently, according to minimal disease activity (MDA) status. B, The relationship between cross-sectional MDA status and Social Security disability award in patients with rheumatoid arthritis (RA) ⬍62 years of age. — ⫽ MDA positive; – – – ⫽ MDA negative. Data are from a randomly selected observation from the 7,433 patients in the longitudinal analyses group. differing durations of RA, we further restricted the randomly selected study sample of 13,128 patients shown in Figure 2A to 7,820 patients who enrolled in the NDB and, therefore, had treatment data available. Enrollees generally have slightly better RA and health status than nonenrollees. As shown in Figure 2B, MDA and remission remained more common in persons with shorter durations of RA. In addition, there was almost no difference between MDA and remission percentages according to treatment status. To formally evaluate the association of anti-TNF therapy with MDA and remission, in a random observation we studied 7,433 NDB patients who had completed at least 3 full NDB semiannual surveys and who were not receiving anti-TNF therapy at their ﬁrst assessment. Of these, only 7.0% of patients with MDA received no RA-related treatments at all and 18.0% received no disease-modifying antirheumatic drug (DMARD), anti-TNF, or prednisone therapy (Table 1). MDA was less common among patients treated with prednisone (OR 0.5, 95% CI 0.4 – 0.6), antiTNF therapy (OR 0.8, 95% CI 0.7– 0.9), and leﬂunomide (OR 0.7, 95% CI 0.6 – 0.9), and more common among those treated with hydroxychloroquine (OR 1.4, 95% CI 1.3– 1.6). These results were to be expected, because they in- Minimal Disease Activity and Remission in RA 939 Table 1. The association between rheumatoid arthritis treatment and minimal disease activity (MDA)* Treatment DMARD MTX Prednisone Hydroxychloroquine Anti-TNF therapy Leﬂunomide Sulfasalazine No treatment No DMARD/biologic No DMARD/biologic or prednisone No DMARD/biologic, prednisone, or NSAID MDA neg. (n ⴝ 5,089) MDA pos. (n ⴝ 2,344) OR (95% CI)† P 75.0 51.1 36.5 24.0 13.6 11.0 7.2 76.7 49.2 22.2 30.7 11.2 8.2 8.4 1.1 (0.9–1.3) 0.9 (0.8–1.2) 0.5 (0.4–0.6) 1.4 (1.3–1.6) 0.8 (0.7–0.9) 0.7 (0.6–0.9) 1.2 (1.0–1.4) 0.093 0.064 ⬍ 0.001 ⬍ 0.001 0.002 ⬍ 0.001 0.091 22.3 17.2 5.1 22.2 18.0 7.0 0.9 (0.8–1.0) 1.0 (0.9–1.2) 1.4 (1.1–1.7) 0.193 0.509 0.004 * Values are the percentage unless otherwise indicated. OR ⫽ odds ratio; 95% CI ⫽ 95% conﬁdence interval; DMARD ⫽ disease-modifying antirheumatic drug; MTX ⫽ methotrexate; anti-TNF ⫽ anti–tumor necrosis factor; NSAID ⫽ nonsteroidal antiinﬂammatory drug. † Adjusted for availability of treatment options (secular trend). dicate that newer and/or potentially more effective therapy is administered to patients with the most severe RA. To be certain the anti-TNF results were valid with our selection methods, we re-ran the anti-TNF analyses after excluding patients who received anti-TNF therapy for ⱕ6 months. The results were similar to those described above (OR 0.8, 95% CI 0.6 – 0.9). To address the effect of anti-TNF therapy on remission, we ﬁrst excluded patients in remission at their ﬁrst observation, leaving 6,828 patients. As shown in Figure 4, the cumulative hazard, which may be interpreted as the estimated cumulative MDA at different points in time, indicates that the MDA percentage at 2 and 6 years for patients treated with anti-TNF therapy was 4.1% and 7.6%, respectively. By comparison, the MDA percentage in patients not treated with anti-TNF therapy was 5.1% and 9.3%, respectively. The patients not treated with anti-TNF therapy are shown only for comparison purposes. Because the odds of treatment prescription differed between the groups, no Figure 4. The cumulative hazard of entering the minimal disease activity (MDA) state for patients who were not in MDA and were not receiving anti–tumor necrosis factor (anti-TNF) therapy at the ﬁrst assessment. The MDA percentage at 2 and 6 years for patients treated with anti-TNF therapy was 4.1% and 7.6%, respectively. By comparison, the percentage in patients not treated with antiTNF therapy was 5.1% and 9.3%, respectively. — ⫽ patients treated with anti-TNF therapy; – – – ⫽ patients not treated with anti-TNF therapy. direct comparison of the effect of differing treatments on remission could be made. Using Cox proportional hazards regression, MDA patients in theses analyses were younger (mean ⫾ SD age 59.1 ⫾ 13.1 years versus 61.2 ⫾ 12.5 years), had more education (13.9 ⫾ 2.4 years versus 13.2 ⫾ 2.3 years), were more often male (28.0% versus 19.0%), had greater household income ($54,694 ⫾ $29,049 US versus $39,400 ⫾ $26,590 US), had fewer comorbid conditions (1.9 ⫾ 1.6 versus 2.7 ⫾ 1.9), and were more likely to be nonminorities (93.6% versus 91.2%) compared with patients who were MDA negative. We also determined the MDA percentage for patients who were MDA positive at 2 consecutive semiannual periods in order to describe the durability of the MDA state. The cumulative hazard at 2 and 6 years was 2.7% and 4.5%, respectively, for the patients treated with anti-TNF therapy and 3.2% and 5.3%, respectively, for the patients not treated with anti-TNF therapy. MDA status and RA outcome. Patients with MDA had good long-term outcomes (Figure 3A). In a random observation of the 7,433 patients in the longitudinal analyses, work disability among persons younger than age 62 years (2.4% versus 28.0%), total joint replacement (10.1% versus 22.1%), and mortality (3.3% versus 6.9%) were reduced in patients who had achieved MDA status compared with those who had not achieved MDA. To evaluate the association of RA duration with MDA status, we graphed the prevalence of US Social Security disability awards against RA duration among patients younger than age 62 years (Figure 3B). Even at RA durations as long as 30 years, work disability was ⬍8% among patients with MDA but reached levels ⬎30% in those without MDA in this cross-sectional analysis. DISCUSSION Using the proposed OMERACT WHO/ILAR core set deﬁnition, MDA was estimated to be present in 20.6% of patients with RA. The OMERACT DAS28-based recom- 940 mendation resulted in MDA in 22.5% of patients. These values were very close to those obtained in the RAES data set: 20.8% using the core set deﬁnition and 22.1% using the DAS28-based deﬁnition. In further conﬁrmation, a rheumatologist’s score of 0 –1 on a 0 –10 disease activity scale (minimal disease activity) was noted to occur in 22.9%. Thus, the data from the 3 sources are in agreement; we conclude that MDA occurs in ⬃21–22% of patients with RA. Remission is somewhat more difﬁcult to quantify. Within the 619 patients in the RAES data set, rheumatologist-deﬁned remission occurred in 7% of cases. This definition was equivalent to a PAS ⱕ0.5, and when that deﬁnition was applied to the NDB 7% of 18,062 patients were found to be in remission. We estimated that remission according to the DAS28 criterion of 2.6 occurred in 15.7%. However, Wells et al suggested that the DAS remission criteria were problematic (15), and indicated that patients with a DAS28 score ⬍2.6 can have substantial residual disease (35). Probability-based remission deﬁnitions are difﬁcult and inaccurate in individual patients, although they appear to be appropriate at the group level. These observations concerning remission lend credence to the Wells et al approach of deﬁning MDA as including remission (15). The difference between MDA and remission and the fact that MDA includes remission also indicate that it is possible to have even better clinical status than deﬁned by MDA. Makinen et al have recently reviewed remission in patients with RA (7) and noted that remission prevalence ranged from 0% to 42% in various studies. These differences might reﬂect treatment effect, patient selection, or remission deﬁnition. In considering the level of disease activity necessary to deﬁne remission and MDA, however, we did not consider persistence of low disease activity. Therefore, the MDA and remission percentages reported here may be considered optimistic. When we required MDA to occur on consecutive semiannual occasions, the MDA percentage decreased by 33% and the remission percentage by 43%. Although it is customary to think of MDA and remission in terms of persistence, the data in this study suggest that this might not be necessary. Patients with MDA on a single occasion had much better outcomes than those not in MDA (Figure 3A), including levels of work disability that were 10 times lower and levels of total joint replacement and mortality that were reduced 2-fold. Figure 3B can be interpreted to indicate that patients with MDA on a single (cross-sectional) occasion always tend to persist in this low disease activity state. Otherwise, work disability, which is a cumulative variable, would be much higher among those with MDA on a single occasion. This interpretation, that patients with low disease activity are determined early in the course of RA, has been made by us previously in a different data set (36). One important use of MDA is to assess treatment effect. The data in Table 1 indicate that only 18% of patients with MDA have a DMARD/biologic-free status, but this ﬁgure is not signiﬁcantly different from the 17.2% of patients not in MDA. If treatment-free status is related to the idea of remission, then after extrapolating from the 20.6% with Wolfe et al MDA shown in Figure 1, only ⬃3.5% of patients have treatment-free MDA. Table 1 also shows that speciﬁc drugs do not have a clear inﬂuence on MDA status, except that there is somewhat greater use of hydroxychloroquine and less use of anti-TNF therapy, leﬂunomide, and prednisone among patients with MDA. One possible interpretation to these data is that patients with less severe RA who are more likely to achieve MDA are given less potent therapies. However, we were able to study the introduction of the new therapy, anti-TNF therapy, and measure the prevalence of MDA following such therapy. Among patients not previously treated with anti-TNF therapy and not currently in MDA, the percentage of patients treated with anti-TNF therapy achieving MDA at 2 and 6 years was 4.1% and 7.6%, respectively. When consecutive semiannual periods in MDA were required, the probability of persistent MDA status at 2 and 6 years was 2.7% and 4.5%, respectively, for the patients treated with anti-TNF therapy. In contrast to the MDA prevalence noted among community RA patients described above, recent clinical trials suggest the possibility of substantial numbers of remission among patients treated with anti-TNF therapy and suggest that “the latest strategies and treatments enable remission to be achieved in many more patients than formerly” (37). For example, in the Trial of Etanercept and Methotrexate with Radiographic Patient Outcomes (TEMPO) study, the proportion of patients who achieved remission as deﬁned by a disease activity score (not DAS28) ⬍1.6 was 35% in the combination anti-TNF/methotrexate group (38,39). The Leeds group reported that 70% of patients with early, poor-prognosis RA treated with inﬂiximab plus methotrexate achieved a sustained median DAS28 score of 2.05 (remission range) (40). We did not observe improvement of this degree among NDB patients. We did, however, observe that patients with shorter durations of RA had a higher rate of MDA (Figure 2A). In Figure 2B, we observed a similar response, but this response could not be attributed to anti-TNF therapy because the early response occurred at an equal or greater proportion in patients not treated with anti-TNF therapy compared with those treated with anti-TNF therapy. Because the data of these ﬁgures are cross-sectional, we cannot state whether this represents better treatment in recent years (41– 43) or a generally better response that occurs in early RA, as we have suggested previously (36,44). In conclusion, MDA prevalence varies slightly according to different deﬁnitions, but is between 20.6% and 22.5%. The RA remission prevalence is 7.0%. These definitions are associated with profound clinical beneﬁt. MDA status appears to have trait characteristics. TNF therapy is not associated with a substantial increase in MDA or remission. PAS-based remission assessments may be a useful tool for assessing RA status and outcome. AUTHOR CONTRIBUTIONS Dr. Wolfe had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Minimal Disease Activity and Remission in RA Study design. Wolfe, Rasker. Acquisition of data. Wolfe, Michaud. Analysis and interpretation of data. Wolfe, Rasker, Boers, Wells, Michaud. Manuscript preparation. Wolfe, Rasker, Boers, Wells, Michaud. Statistical analysis. Wolfe, Rasker, Wells, Michaud. REFERENCES 1. Smolen JS, Aletaha D. Patients with rheumatoid arthritis in clinical care. Ann Rheum Dis 2004;63:221–5. 2. Pincus T, Stein CM, Wolfe F. “No evidence of disease” in rheumatoid arthritis using methotrexate in combination with other drugs: a contemporary goal for rheumatology care? Clin Exp Rheumatol 1997;15:591– 6. 3. Emery P, Salmon M. Early rheumatoid arthritis: time to aim for remission? Ann Rheum Dis 1995;54:944 –7. 4. Pinals RS, Masi AT, Larsen RA, and the Subcommittee for Criteria of Remission in Rheumatoid Arthritis of the American Rheumatism Association Diagnostic and Therapeutic Criteria Committee. Preliminary criteria for clinical remission in rheumatoid arthritis. Arthritis Rheum 1981;24:1308 –15. 5. Wolfe F, Hawley DJ. Remission in rheumatoid arthritis. J Rheumatol 1985;12:245–52. 6. Prevoo ML, van Gestel AM, van ‘t Hof MA, van Rijswijk MH, van de Putte LB, van Riel PL. Remission in a prospective study of patients with rheumatoid arthritis: American Rheumatism Association preliminary remission criteria in relation to the disease activity score. Br J Rheumatol 1996;35:1101–5. 7. Makinen H, Kautiainen H, Hannonen P, Sokka T. Frequency of remissions in early rheumatoid arthritis deﬁned by 3 sets of criteria: a 5-year followup study. J Rheumatol 2005;32:796 – 800. 8. Harrison BJ, Symmons DP, Brennan P, Barrett EM, Silman AJ. Natural remission in inﬂammatory polyarthritis: issues of definition and prediction. Br J Rheumatol 1996;35:1096 –100. 9. Eberhardt K, Fex E. Clinical course and remission rate in patients with early rheumatoid arthritis: relationship to outcome after 5 years. Br J Rheumatol 1998;37:1324 –9. 10. Reilly PA, Cosh JA, Maddison PJ, Rasker JJ, Silman AJ. Mortality and survival in rheumatoid arthritis: a 25 year prospective study of 100 patients. Ann Rheum Dis 1990;49:363–9. 11. Rasker JJ, Cosh JA. The natural history of rheumatoid arthritis over 20 years: clinical symptoms, radiological signs, treatment, mortality and prognostic signiﬁcance of early features. Clin Rheumatol 1987;6 Suppl 2:5–11. 12. Rasker JJ, Cosh JA. The natural history of rheumatoid arthritis: a ﬁfteen year follow-up study: the prognostic signiﬁcance of features noted in the ﬁrst year. Clin Rheumatol 1984;3:11–20. 13. Fransen J, Creemers MC, van Riel PL. Remission in rheumatoid arthritis: agreement of the disease activity score (DAS28) with the ARA preliminary remission criteria. Rheumatology (Oxford) 2004;43:1252–5. 14. Balsa A, Carmona L, Gonzalez-Alvaro I, Belmonte MA, Tena X, Sanmarti R, and the EMECAR Study Group. Value of Disease Activity Score 28 (DAS28) and DAS28 –3 compared to American College of Rheumatology-deﬁned remission in rheumatoid arthritis. J Rheumatol 2004;31:40 – 6. 15. Wells GA, Boers M, Shea B, Brooks PM, Simon LS, Strand CV, et al. Minimal disease activity for rheumatoid arthritis: a preliminary deﬁnition. J Rheumatol 2005;32:2016 –24. 16. Wolfe F, Anderson J, Burke TA, Arguelles LM, Pettitt D. Gastroprotective therapy and risk of gastrointestinal ulcers: risk reduction by COX-2 therapy. J Rheumatol 2002;29:467– 73. 17. Wolfe F, Flowers N, Burke TA, Arguelles LM, Pettitt D. Increase in lifetime adverse drug reactions, service utilization, and disease severity among patients who will start COX-2 speciﬁc inhibitors: quantitative assessment of channeling bias and confounding by indication in 6689 patients with rheumatoid arthritis and osteoarthritis. J Rheumatol 2002;29: 1015–22. 18. Michaud K, Messer J, Choi HK, Wolfe F. Direct medical costs and their predictors in patients with rheumatoid arthritis: a 941 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. three-year study of 7,527 patients. Arthritis Rheum 2003;48: 2750 – 62. Fries JF, Spitz PW, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 1980;23:137– 45. Wolfe F. A reappraisal of HAQ disability in rheumatoid arthritis. Arthritis Rheum 2000;43:2751– 61. Wolfe F, Michaud K, Pincus T. A composite disease activity scale for clinical practice, observational studies, and clinical trials: the Patient Activity Scale (PAS/PAS-II). J Rheumatol 2005;32:2410 –5. Wolfe F, Michaud K, Pincus T. Development and validation of the Health Assessment Questionnaire. II. A revised version of the Health Assessment Questionnaire. Arthritis Rheum 2004;50:3296 –305. Doody MM, Hayes HM, Bilgrad R. Comparability of national death index plus and standard procedures for determining causes of death in epidemiologic studies. Ann Epidemiol 2001;11:46 –50. Edlavitch SA, Baxter J. Comparability of mortality follow-up before and after the National Death Index. Am J Epidemiol 1988;127:1164 –78. Wolfe F, Michaud K, Pincus T, Furst D, Keystone E. The Disease Activity Score is not suitable as the sole criterion for initiation and evaluation of anti-tumor necrosis factor therapy in the clinic: discordance between assessment measures and limitations in questionnaire use for regulatory purposes. Arthritis Rheum 2005;52:3873–9. Van der Heijde DM, van ’t Hof MA, van Riel PL, van Leeuwen MA, van Rijswijk MH, van de Putte LB. Validity of single variables and composite indices for measuring disease activity in rheumatoid arthritis. Ann Rheum Dis 1992;51:177– 81. Van der Heijde DM, van ’t Hof MA, van Riel PL, Theunisse LA, Lubberts EW, van Leeuwen MA, et al. Judging disease activity in clinical practice in rheumatoid arthritis: ﬁrst step in the development of a disease activity score. Ann Rheum Dis 1990;49:916 –20. Van der Heijde DM, van ’t Hof M, van Riel PL, van de Putte LB. Development of a disease activity score based on judgment in clinical practice by rheumatologists. J Rheumatol 1993;20:579 – 81. Prevoo ML, van ’t Hof MA, Kuper HH, van Leeuwen MA, van de Putte LB, van Riel PL. Modiﬁed disease activity scores that include twenty-eight–joint counts: development and validation in a prospective longitudinal study of patients with rheumatoid arthritis. Arthritis Rheum 1995;38:44 – 8. Felson DT, Anderson JJ, Boers M, Bombardier C, Chernoff M, Fried B, et al. The American College of Rheumatology preliminary core set of disease activity measures for rheumatoid arthritis clinical trials. Arthritis Rheum 1993;36:729 – 40. Felson DT. Choosing a core set of disease activity measures for rheumatoid arthritis clinical trials. J Rheumatol 1993;20: 531– 4. Hoehler FK. Bias and prevalence effects on kappa viewed in terms of sensitivity and speciﬁcity. J Clin Epidemiol 2000;53: 499 –503. Feinstein AR, Cicchetti DV. High agreement but low kappa. I. The problems of two paradoxes. J Clin Epidemiol 1990;43: 543–9. Stata Corporation. Stata statistical software: release 9.1. College Station (TX): Stata Corporation; 2005. Aletaha D, Ward MM, Machold KP, Nell VP, Stamm T, Smolen JS. Remission and active disease in rheumatoid arthritis: deﬁning criteria for disease activity states. Arthritis Rheum 2005;52:2625–36. Wolfe F, Pincus T. The level of inﬂammation in rheumatoid arthritis is determined early and remains stable over the longterm course of the illness. J Rheumatol 2001;28:1817–24. Breedveld FC, Kalden JR. Appropriate and effective management of rheumatoid arthritis [review]. Ann Rheum Dis 2004; 63:627–33. Klareskog L, van der Heijde D, de Jager JP, Gough A, Kalden J, Malaise M, et al. Therapeutic effect of the combination of etanercept and methotrexate compared with each treatment 942 alone in patients with rheumatoid arthritis: double-blind randomised controlled trial. Lancet 2004;363:675– 81. 39. Choy EH. Two is better than one? Combination therapy in rheumatoid arthritis. Rheumatology (Oxford) 2004;43: 1205–7. 40. Quinn MA, Conaghan PG, O’Connor PJ, Karim Z, Greenstein A, Brown A, et al. Very early treatment with inﬂiximab in addition to methotrexate in early, poor-prognosis rheumatoid arthritis reduces magnetic resonance imaging evidence of synovitis and damage, with sustained beneﬁt after inﬂiximab withdrawal: results from a twelve-month randomized, double-blind, placebo-controlled trial. Arthritis Rheum 2005;52: 27–35. Wolfe et al 41. Sokka T, Kautiainen H, Hakkinen A, Hannonen P. Radiographic progression is getting milder in patients with early rheumatoid arthritis: results of 3 cohorts over 5 years. J Rheumatol 2004;31:1073– 82. 42. Krishnan E, Fries JF. Reduction in long-term functional disability in rheumatoid arthritis from 1977 to 1998: a longitudinal study of 3035 patients. Am J Med 2003;115:371– 6. 43. Choi HK, Hernan MA, Seeger JD, Robins JM, Wolfe F. Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study. Lancet 2002;359:1173–7. 44. Wolfe F, Hawley DJ, Cathey MA. Clinical and health status measures over time: prognosis and outcome assessment in rheumatoid arthritis. J Rheumatol 1991;18:1290 –7.