Measuring function in rheumatoid arthritisIdentifying reversible and irreversible components.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 54, No. 9, September 2006, pp 2784–2792 DOI 10.1002/art.22052 © 2006, American College of Rheumatology Measuring Function in Rheumatoid Arthritis Identifying Reversible and Irreversible Components Daniel Aletaha,1 Josef Smolen,2 and Michael M. Ward3 (2–<5 years) to 0.38 (5–<10 years) to 0.55 (>10 years) (P < 0.001). The reversibility of HAQ scores decreased with the duration of RA (median 100%, 83.3%, 81.9%, and 66.7%, respectively; P < 0.001). Findings were similar in patients subgrouped by quartile of radiographic scores. Conclusion. Differences in the sources of functional limitations should be considered in the interpretation of functional measures, and in their use for prediction and in cost analyses. Objective. Measurement of physical function at one point in time cannot distinguish impairment caused by the active disease process from chronic irreversible impairment. We aimed to dissect these two components of functional limitation in rheumatoid arthritis (RA) by using the disability index of the Health Assessment Questionnaire (HAQ) as the measure of function. Methods. We performed a secondary analysis of data from 6 contemporary clinical trials of RA (2,763 patients). Patients in whom remission was achieved in the trials, based on a simplified disease activity index, were identified. In an individual patient, HAQ scores at trial entry represented both reversible and irreversible impairments, while HAQ scores at the time of RA remission represented the mostly irreversible component, and the difference between these corresponded to the component related to disease activity. We tested the concept that the HAQ has a reversible and an irreversible component by associating the HAQ score during remission with 2 measures associated with the degree of accrued damage: duration of RA and radiographic severity. Results. Among patients in whom clinical remission was achieved (n ⴝ 295), average HAQ scores despite clinical remission increased progressively with the duration of RA, from 0.19 (<2 years of RA) to 0.36 Many chronic diseases cause impairment of normal function, overall or in an individual organ. Restoring normal functioning is a major therapeutic aim. However, measures of function only describe the state of the organ or the individual at the time of assessment and give no indication of the cause or duration of the dysfunction. Functional limitations may be due to acute reversible problems or chronic irreversible factors. Single measurements of function do not permit isolation of the reversible component from the irreversible component; these can only be identified by examining changes in measures of function over time or in response to treatment. This concept is commonly appreciated in clinical medicine with respect to measures of organ function. For example, a single elevated measurement of serum creatinine concentration does not allow one to know if renal dysfunction is temporary and completely reversible, if it is reflective of chronic renal damage that is irreversible, or some combination of the two. Only observations of the response to treatment and time can distinguish these possibilities. Similar concepts likely apply to the measurement of health status: functional limitations may be due to acute reversible problems or chronic irreversible ones. In contrast to organ function, physical functioning is usually not interpreted in the context of reversibility and irreversibility. When used to assess treatment responses, the irreversible component Supported in part by the Intramural Research Program of the National Institute of Arthritis and Musculoskeletal and Skin Diseases, NIH, and the Austrian Science Fund (FWF). 1 Daniel Aletaha, MD, MHS: National Institute of Arthritis and Musculoskeletal and Skin Diseases, NIH, Bethesda, Maryland, and Medical University of Vienna, Vienna, Austria; 2Josef Smolen, MD: Medical University of Vienna, and Hietzing Hospital, Vienna, Austria; 3Michael M. Ward, MD, MPH: National Institute of Arthritis and Musculoskeletal and Skin Diseases, NIH, Bethesda, Maryland. Address correspondence and reprint requests to Daniel Aletaha, MD, Department of Rheumatology, Internal Medicine III, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria. E-mail: firstname.lastname@example.org. Submitted for publication January 19, 2006; accepted in revised form May 24, 2006. 2784 MEASURES OF FUNCTIONING IN RA of functional impairment decreases overall responsiveness of functional measures; when used to predict longterm consequences, the reversible component introduces measurement error. In this study, we attempted to provide evidence of this concept of duality in functional measures by estimating the reversible and irreversible components of limitations in physical functioning in rheumatoid arthritis (RA) (1–4). In RA, reversible components of functional impairment comprise joint pain and swelling due to inflammation, or associated symptoms such as depression. Potentially irreversible components may include joint destruction, deformity, or muscle weakness, or may be related to comorbid conditions. One advantage in RA is that the active disease process can be readily characterized by joint counts, measurement of acute-phase reactants, and other measures, and that the disease-related damage can be related to the duration of RA, since it increases consistently over time (5), or assessed by joint radiography (4,6,7). We performed an analysis of individual patient data from several large RA clinical trials. Patients had to have active RA as a requirement for trial entry. Therefore, their functional limitation at entry was influenced by acute reversible inflammation, as well as any irreversible processes, such as joint damage, that might have been present. Among these patients, we examined the subset of individuals in whom clinical remission was achieved during the trials. In this subset, any change in functional limitations could be attributed to reversible RA activity or its correlates, such as depression. Any residual functional limitations could be regarded as irreversible, reflecting RA-related damage or other irreversible components, such as comorbidities. To support the concept of an irreversible Health Assessment Questionnaire (HAQ) component that is associated with RA-related damage, we compared functional limitations that remained despite clinical remission and the reversibility of functional limitations among subgroups of patients with different degrees of underlying damage, as represented by the duration of RA or by the severity of abnormalities seen on joint radiographs. PATIENTS AND METHODS Data sources. We contacted sponsors of recent clinical trials in RA and asked for their permission to obtain patientlevel data for our study. We obtained and pooled individual patient data from 6 trials: the Anti–Tumor Necrosis Factor Trial in Rheumatoid Arthritis With Concomitant Therapy (ATTRACT) (8) of infliximab and methotrexate (MTX) ver- 2785 sus placebo and MTX in patients with inadequate prior response to MTX; the Active-Controlled Study of Patients Receiving Infliximab for the Treatment of Rheumatoid Arthritis of Early Onset (ASPIRE) trial (9) of infliximab plus MTX versus MTX alone in MTX-naive patients with early RA of ⱕ3 years; the Anti–Tumor Necrosis Factor Research Study Program of the Monoclonal Antibody Adalimumab (D2E7) in Rheumatoid Arthritis (ARMADA) (10) and the DE019 (11) trials in patients with active RA despite prior MTX therapy, who received additional adalimumab versus placebo; and the European leflunomide trials, which compared leflunomide with sulfasalazine and with placebo in patients with early and established RA (12,13), or which compared leflunomide with MTX in patients with RA of ⱕ10 years’ duration (14). We were provided with data on an 80% random sample of subjects from these trials. All patients had active RA at enrollment in these studies, with requirements for ⬎6–10 swollen joints and ⬎6–12 tender joints (using a 68-joint count), or ⱖ6 swollen and ⱖ6 tender joints (using a 28-joint count in the leflunomide trials). With the exception of the ASPIRE trial, elevations in acute-phase reactants were also required (C-reactive protein [CRP] level ⱖ2.0 mg/dl and/or erythrocyte sedimentation rate [ESR] ⱖ28 mm/hour). Data on scored radiographs were available for all trials except the ARMADA trial. Study variables. The disability index of the HAQ (15) was used as the measure of functional limitations in each trial. It includes 20 questions organized into 8 categories (dressing, rising, eating, walking, hygiene, reach, grip, and errands or chores). Each question was scored using a 4-level difficulty scale ranging from 0 to 3, with 0 ⫽ no difficulty, 1 ⫽ some difficulty, 2 ⫽ much difficulty, and 3 ⫽ unable to do. The HAQ score is the mean of the highest score in each of the 8 categories, with a possible range of 0–3. The duration of RA and radiographic severity were used as 2 separate independent variables to measure chronicity and damage. Duration of RA is useful for this purpose because damage increases over time in patients with RA (4,5,16–18). For analysis, we divided the patients into clinically meaningful subgroups by RA duration: ⬍2 years, 2–⬍5 years, 5–⬍10 years, and ⱖ10 years. Radiographic scores measure the extent and severity of bone erosions and joint space narrowing in the hands and feet, and are a direct measure of joint damage. Radiographs had been graded using the original Sharp score (19) (DE019 trial and leflunomide trials) or its modification by van der Heijde et al (20) (ASPIRE and ATTRACT trials). These scores generally produce very similar results, but can be different in patients with extensive damage of joints in the feet. To diminish these differences, we defined subgroups based on quartiles of scores. Duration of RA and radiographic scores were moderately highly correlated (r ⫽ 0.66, P ⬍ 0.001, by Spearman’s rank correlation). Definition of remission. We separated the reversible RA activity–related component of the HAQ from the irreversible component by identifying patients in whom clinical remission was achieved during the trials (i.e., a state of absence of RA activity). We defined remission as a state below a threshold value from a pooled index of RA activity, the simplified disease activity index (SDAI) (21). The SDAI combines several measures of RA activity by calculating a linear sum of tender and swollen joint counts (0–28 scale), patient and evaluator global assessments of disease activity (each 0–10 cm 2786 ALETAHA ET AL Table 1. Characteristics at baseline of RA patients in whom remission was achieved during the trials* Patients grouped by duration of RA (years) (n ⫽ 295) No. of patients Female, % Age, mean ⫾ SD years Race, % Caucasian African American Other Rheumatoid factor, % positive Duration of RA, mean ⫾ SD years Disease activity Mean ⫾ SD swollen joint count, 0–28 scale Mean ⫾ SD tender joint count, 0–28 scale Mean ⫾ SD CRP level, mg/dl Mean ⫾ SD patient-reported pain, 10-cm VAS Mean ⫾ SD patient global assessment, 10-cm VAS Mean ⫾ SD evaluator global assessment, 10-cm VAS Mean ⫾ SD SDAI HAQ, 0–3 scale Patients grouped by radiographic scores (quartiles) (n ⫽ 270) All patients ⬍2 2–⬍5 5–⬍10 ⱖ10 P† ⬍2.5 2.6–8.9 9.5–22.0 ⬎22.9 P† 295 62.7 51.0 ⫾ 12.8 180 59.4 49.8 44 59.1 51.7 37 70.3 52.0 34 76.5 54.8 NA 0.19 0.18 69 62.3 44.9 66 59.1 51.9 68 63.2 52.8 67 68.7 54.4 NA 0.72 ⬍0.01 90.2 3.7 6.1 72.7 4.1 ⫾ 7.1 91.1 2.8 6.1 70.5 0.7 86.4 9.1 4.5 79.1 3.0 91.9 2.7 5.4 62.9 7.2 88.2 2.9 8.8 84.9 20.5 0.58 89.4 1.5 9.1 74.2 2.2 91.2 5.9 2.9 61.2 2.9 91.0 3.0 6.0 89.6 8.5 0.65 0.14 NA 89.9 2.9 7.3 67.7 0.9 ⬍0.01 NA 9.5 ⫾ 4.8 8.7 10.1 10.8 11.1 0.01 9.2 9.7 8.9 10.7 0.12 12.0 ⫾ 6.0 11.9 12.2 12.6 11.4 0.85 12.4 12.3 11.5 12.5 0.79 2.6 ⫾ 2.9 4.68 ⫾ 2.91 3.0 5.01 1.7 3.67 2.0 4.01 1.8 4.95 0.01 0.02 2.9 4.73 2.5 4.74 2.6 5.15 2.3 4.53 0.62 0.66 5.36 ⫾ 2.52 5.55 5.03 5.29 4.83 0.36 5.44 5.41 5.47 4.95 0.60 6.01 ⫾ 1.85 6.20 5.49 5.49 6.22 0.04 5.99 6.04 6.10 5.99 0.99 35 ⫾ 11.8 1.29 ⫾ 0.62 35.9 1.31 35.2 1.22 36.3 1.27 36.0 1.28 0.98 0.88 36.4 1.26 36.2 1.31 34.4 1.34 36.7 1.20 0.70 0.57 * Remission was defined as a simplified disease activity index (SDAI) of ⱕ3.3 for the swollen joint count, the tender joint count, C-reactive protein (CRP) level, patient global assessment, and evaluator global assessment (see Patients and Methods for details). RA ⫽ rheumatoid arthritis; NA ⫽ not applicable; VAS ⫽ visual analog scale; HAQ ⫽ disability index of the Health Assessment Questionnaire. † By analysis of variance for continuous variables, and chi-square test for categorical variables. on visual analog scales), and CRP level (in mg/dl). For remission, the sum of these measures had to be ⱕ3.3, based on the results of a recent investigation (22). The majority of patients with results below this cutoff point do not show progressive radiographic damage over time (22). Also, this definition of remission is very stringent, allowing less residual disease activity in the various component measures of RA activity than definitions based on other pooled indices (23). In a sensitivity analysis, we varied the cutoff point for remission to values of 3.0 and 3.6 to assess the effects of residual RA activity on our results. We also derived a qualitatively different definition of remission. First, we developed criteria for each individual RA activity measure and required that several of them be met simultaneously. We derived these criteria from the data from a large survey (22,24), one part of which had been used to develop the criteria for the composite index described above (22). The survey included 35 rheumatologists who were asked to rate the RA activity of 32 patient profiles. The data in each patient profile comprised swollen and tender joint counts, ESR, CRP level, patient-reported pain, as well as patient and evaluator global assessments of disease activity. The rheumatologists rated each profile as remission, low, moderate, or high disease activity (22,24). In a first step, we used classification trees to identify the measures that were most useful to discriminate patients with remission from those with more active RA, using the survey responses as a reference standard (AnswerTree soft- ware, version 3.1; SPSS, Chicago, IL). Four measures were identified in this analysis (swollen joint count, tender joint count, CRP level, and evaluator global assessment of disease activity). We then tested cutoff points for these measures in all possible combinations (n ⫽ 256): swollen joint count 0, 1, 2, and 3; tender joint count 0, 1, 2, and 3; CRP level 0.8, 1.3, 1.8, and 2.3 mg/dl; and evaluator global assessment of disease activity 10, 15, 20, and 25 mm. The level of detection of CRP was 0.8 mg/dl in all trials; therefore, this cutoff point was used as the most stringent CRP criterion. We identified all sets of criteria for which both the sensitivity and the specificity were ⱖ0.80, and from these sets, selected the set with the highest specificity for use. This set had a specificity of 0.86 and a sensitivity of 0.80 and was composed of swollen joint count ⱕ1, tender joint count ⱕ3, CRP level ⱕ0.8 mg/dl, and evaluator global assessment of disease activity ⱕ15 mm. Statistical analysis. We identified patients in whom clinical remission was achieved at the 3-month, 6-month, 9-month, or 12-month followup visit in the trials. Each patient’s first visit with remission was used for analysis. The HAQ scores at this visit were considered to reflect the irreversible component of functional limitations, which remained after effective treatment that had resulted in complete or nearcomplete resolution of active RA. We termed this quantity the “residual HAQ score,” and tested to determine whether it increased with duration of RA or quartile of radiographic damage scores, using one-way analysis of variance (ANOVA). We also tested differences in measures of RA activity across MEASURES OF FUNCTIONING IN RA these subgroups at baseline and at the time of remission as potential confounders, and adjusted the association between residual HAQ scores and duration of RA using these variables as covariates in linear regression models. Dividing the residual HAQ score by the maximum possible HAQ score of 3 provided the fraction of HAQ irreversibility in these patient subgroups. We then calculated changes in HAQ scores from trial entry to time of remission among those in whom clinical remission was achieved. For each patient, we calculated the reversibility of HAQ scores as (change in HAQ/baseline HAQ) multiplied by 100%. Since baseline HAQ scores of 0 cannot improve, we excluded these patients (n ⫽ 5) from this analysis. To determine whether reversibility was associated with the degree of expected chronic damage, we compared the proportion of reversibility among the subgroups by duration of RA and radiographic scores, using the Kruskal-Wallis nonparametric test. Analyses were first performed using the subset of patients in whom SDAI remission was achieved (including the use of different cutoff points in the sensitivity analysis), and then were repeated using the subset of patients in whom remission was achieved based on the individual RA activity measures criterion (qualitative sensitivity analysis). Analyses were performed using SAS software, version 9.0 (SAS Institute, Cary, NC). P values less than 0.05 were considered significant. RESULTS Patients. Of the 2,763 patients, 295 (10.7%) met the remission criterion based on the pooled index during the first year of the trial, 39.0% of whom were identified at the 3-month visit, 26.8% at the 6-month visit, and 34.2% at the 9-month visit. No additional patients were first identified with remission at the 12-month visit. Table 1 shows the characteristics of these patients at trial entry. Patients in different subgroups of RA duration were similar in many baseline characteristics (Table 1), but patients with a longer disease duration tended to be older (P ⫽ 0.18) and had more swollen joints (P ⫽ 0.01). For several other activity markers, at least 1 subgroup was significantly different from the others. In contrast to findings at baseline, CRP scores and pain scores at the time of remission were significantly higher in patients with more longstanding RA (P ⫽ 0.024 and P ⫽ 0.004, respectively, by ANOVA). Associations with the duration of RA. Figure 1A shows the mean residual HAQ scores at the time of clinical remission. The residual HAQ was lowest in patients with an RA duration of ⬍2 years, while it was consistently higher in patients with more established RA (2–⬍10 years), and highest in those with late RA (ⱖ10 years) (P ⬍ 0.001 by ANOVA testing for a linear trend component). This indicates that the HAQ had an irre- 2787 Figure 1. Analysis of Health Assessment Questionnaire (HAQ) scores by duration of rheumatoid arthritis (RA). A, Residual scores on the disability index of the HAQ at the time of remission in patients (n ⫽ 295) grouped by duration of RA. Residual HAQ scores increased significantly across subgroups (P ⬍ 0.001 by analysis of variance testing for a linear trend component). Use of residual HAQ scores allows for estimation of the maximum hypothetical reversibility by relating their values to the total HAQ scale of 3, e.g., a residual score of 0.5 represents one-sixth of the total scale and, conversely, the hypothetically reversible component can be estimated as five-sixths of the scale. B, Reversibility of HAQ scores. Reversibility is calculated as the relative improvement in baseline HAQ scores at the time of remission (n ⫽ 290). Data are shown as box plots. Boxes span the interquartile range. Lines inside the boxes represent the median. Whiskers indicate observations that are 1.5 interquartile ranges beyond the lower quartile. Circles indicate outliers. Reversibility of HAQ scores decreased significantly across the 4 subgroups by duration of RA (P ⬍ 0.001 by Kruskal-Wallis nonparametric test). versible component in these patients, which progressively increased with the chronicity of RA. We adjusted these associations for age and for the clinical measures that were found to be different among the 4 subgroups at baseline and at the time of remission (Table 2). The results of these adjusted models were very similar to the unadjusted results, indicating that differences in covariates among the various subgroups had no important impact on the comparison of HAQ scores among subgroups. Reversibility of the HAQ was assessed as the difference in scores from baseline to the time of clinical remission (Figure 1B). Although patient subgroups had similar average HAQ scores at baseline (Table 1), the reversibility varied with the duration of RA. HAQ scores were 100% reversible in the majority of patients with early RA (duration ⬍2 years), while the median reversibility was ⬃80% in established RA (2–⬍5 years and 5–⬍10 years), and ⬃67% in late RA (P ⬍ 0.001 by Kruskal-Wallis nonparametric test), although there was 2788 ALETAHA ET AL Table 2. Mean (95% confidence interval) residual scores on the disability index of the Health Assessment Questionnaire across subgroups, by unadjusted and adjusted generalized linear models Subgroup by RA duration ⬍2 years (n ⫽ 180) 2–⬍5 years (n ⫽ 44) 5–⬍10 years (n ⫽ 37) ⱖ10 years (n ⫽ 34) P across subgroups Unadjusted analysis Analysis adjusted for differences at baseline* Analysis adjusted for differences at the time of remission† 0.19 (0.13–0.25) 0.36 (0.24–0.48) 0.38 (0.24–0.51) 0.55 (0.41–0.68) ⬍0.001 0.18 (0.12–0.24) 0.36 (0.24–0.49) 0.41 (0.27–0.54) 0.54 (0.40–0.68) ⬍0.001 0.21 (0.15–0.27) 0.36 (0.24–0.48) 0.35 (0.22–0.48) 0.50 (0.36–0.63) 0.001 * Age and variables that were different at baseline were adjusted for by setting their levels to the cohort means in each subgroup of rheumatoid arthritis (RA) duration (i.e., age 51.0 years, baseline swollen joint count 9.5, baseline C-reactive protein [CRP] level 2.6 mg/dl, baseline pain score 4.68 cm, and baseline evaluator global assessment 6.01 cm) (see Table 1). † Age and variables that were different at the time of remission were adjusted for by setting their values to the cohort means in each subgroup of RA duration (age 51.0 years, CRP level at time of remission 0.8 mg/dl, and pain score at time of remission 0.53 cm). considerable variation among patients within each disease duration subgroup (Figure 1B). The proportion of patients who achieved a HAQ score of 0 decreased from 56.3% to 17.6% across subgroups (P ⬍ 0.001 by chisquare test). These results indicate that, among patients with the same degree of functional limitation at baseline, the reversible component of functional limitation may be very different, and decreases with the chronicity of RA. Associations with radiographic damage. Radiographic data were available on 270 patients who met the SDAI criterion for remission (Table 1). Residual HAQ scores increased across subgroups of patients with increasing radiographic damage scores (P ⫽ 0.002 by ANOVA testing for linear trend) (Figure 2A). Adjustment for differences in RA activity at baseline and at the time of remission did not alter the results of this analysis (data not shown). Reversibility of HAQ scores also decreased across quartiles of radiographic scores (P ⫽ 0.002 by Kruskal-Wallis nonparametric test) (Figure 2B). Reversibility of functional limitation was more common among patients in the lowest quartile of radiographic scores than in patients in all other quartiles. These findings parallel those found for associations with the duration of RA. Sensitivity analysis. Results of the analysis of subgroups by RA duration did not change when we used a more stringent criterion (SDAI ⱕ3.0) or a less stringent criterion (SDAI ⱕ3.6) for remission. Table 3 shows the results of the sensitivity analysis that used a definition of remission constructed from a combined set of cutoff points for individual RA activity measures. The subset of patients identified by this criterion was larger than in the analysis using the SDAI criterion for remission, and included patients with more RA activity among those with remission. Therefore, residual HAQ scores were higher and the degree of HAQ reversibility was smaller in patients in whom Figure 2. Analysis of Health Assessment Questionnaire (HAQ) scores by radiographic scores. A, Residual scores on the disability index of the HAQ at the time of remission in patients (n ⫽ 270) grouped by quartile of radiographic scores. Residual HAQ scores increased significantly across subgroups (P ⫽ 0.002 by analysis of variance testing for a linear trend component). Use of residual HAQ scores also allowed for estimation of the maximum reversible HAQ component (see legend to Figure 1). B, Reversibility of HAQ scores. Reversibility is calculated as the relative improvement in baseline HAQ scores at the time of remission (n ⫽ 265). Data are shown as box plots. Boxes span the interquartile range. Lines inside the boxes represent the median. Whiskers indicate observations that are 1.5 interquartile ranges beyond the lower quartile. Circles indicate outliers. Reversibility of HAQ scores decreased significantly across the 4 subgroups by quartile of radiographic scores (P ⫽ 0.002 by KruskalWallis nonparametric test). MEASURES OF FUNCTIONING IN RA Table 3. 2789 Analysis using remission definition based on individual RA activity measures* Subgroups by duration of RA All subgroups ⬍2 years 2–⬍5 years 5–⬍10 years ⱖ10 years Significance (P) across groups‡ Subgroups by radiographic score All subgroups 1st quartile (⬍2.5) 2nd quartile (2.6–8.9) 3rd quartile (9.5–22.0) 4th quartile (⬎22.9) Significance (P) across groups‡ Residual HAQ score, mean ⫾ SEM (n) % reversibility, median (interquartile range) (n)† 0.44 ⫾ 0.03 (410) 0.36 ⫾ 0.03 (247) 0.44 ⫾ 0.06 (64) 0.51 ⫾ 0.09 (43) 0.73 ⫾ 0.09 (56) ⬍0.001 (F ⫽ 19.1) 80.0 (50.0–100) (405) 85.7 (58.8–100) (243) 72.7 (47.8–100) (63) 66.7 (43.7–100) (43) 50.0 (29.7–89.7) (56) ⬍0.001 (2 ⫽ 22.8) 0.43 ⫾ 0.03 (384) 0.29 ⫾ 0.05 (95) 0.34 ⫾ 0.04 (98) 0.42 ⫾ 0.05 (95) 0.66 ⫾ 0.06 (96) ⬍0.001 (F ⫽ 26.5) 80.0 (50.0–100) (379) 94.7 (69.1–100) (93) 85.7 (60.0–100) (96) 75.8 (49.2–100) (94) 60.2 (32.0–92.7) (96) ⬍0.001 (2 ⫽ 27.8) * Swollen joint count ⱕ1, tender joint count ⱕ3, C-reactive protein level ⱕ0.8 mg/dl, and evaluator global assessment of disease activity ⱕ15 mm. † All patients with baseline scores of 0 on the disability index of the Health Assessment Questionnaire (HAQ) were excluded from this analysis due to the lack of ability to show improvement (n ⫽ 5 in both analyses). ‡ By analysis of variance (testing for a linear trend component) for residual HAQ scores, and Kruskal-Wallis nonparametric test for reversibility. remission was achieved by this criterion. However, using this criterion did not change the strength of association between residual HAQ score or the proportion of reversibility and either RA duration or radiographic score (Table 3). DISCUSSION In this study, we showed that physical function in RA, as measured by the HAQ, is composed of reversible and irreversible components. At any one time, HAQ scores reflect both functional limitations due to current RA activity and functional limitations due to processes that do not respond to aggressive treatment with antirheumatic drugs. These processes could include irreversible joint damage or musculoskeletal consequences of RA, or limitations due to coexisting conditions. Single measurements reflect a combination of functional limitations from these 2 sources. Although this duality has been recognized previously, these components have not been directly demonstrated (1,2,4,7,25,26). Many longitudinal studies have indicated that both RA activity and damage influence HAQ scores during the course of the disease (4,7,25). In these studies, the correlation of radiographic changes with HAQ scores increases over time (6), while RA activity is the most significant determinant of HAQ scores at all times (4,7). These studies predicted total HAQ scores, and did not separately quantify these components of the HAQ. Here, we were able to isolate the RA activity–related component and the irreversible component of functional limitations by devising a strategy to identify patients whose disease changed from active to being in remission in a short period of time. This allowed us to emphasize the ultimately quantitative nature of this duality and to demonstrate that the reversible component, at least in this patient sample, is far larger than would have been expected. Using data from clinical trials provided standardized accurate measurements of RA activity, and pooling data from several large trials of highly efficacious medications allowed us to identify sufficient numbers of patients who achieved the relatively rare event of clinical remission for our analysis. Examining patients with remission, still a rare event in the context of current treatments, enabled us to demonstrate the dual nature of the HAQ, regardless of how remission was achieved (i.e., regardless of the treatment regimen used). RA served as a useful model to test the hypothesis that both functional states and functional response are dependent on the amount of accrued damage. In contrast to other chronic diseases, such as glomerulonephritis, heart failure, and multiple sclerosis, disease activity (or, more generally, the readily reversible disease component) is easily definable in RA, and there- 2790 fore its absence can be defined. Damage caused by RA can also be readily estimated. Radiographic scores serve as simple and direct measures of joint damage, while duration of RA is a surrogate measure that captures structural changes (joint deformities, muscle weakness, debility) that may increase over time (6,18). The strengths of the study include the use of a large sample from well-designed clinical trials, consistency of results obtained using 2 different measures of chronic damage, and sensitivity analyses demonstrating that the findings are robust with different remission cutoff points, but also with a set of qualitatively different criteria for remission. One limitation of this study was the identification of remission at different time points after enrollment in the trials, which was done for reasons of sample size. However, different rates of change in the HAQ among patients would not be expected to influence the results, but the different time points of remission could potentially decrease reversibility of HAQ scores, because the underlying damage component increased from baseline until the time of the remission visit. Also, although our definition of remission was stringent, it might not represent the complete absence of RA activity in all patients. Therefore, our analysis estimated the upper limit of the latent trait “irreversible disability.” Varying amounts of residual disease activity add measurement error in our analysis, but there is no reason to believe that the degree of error would vary among the subgroups of RA duration or radiographic score. We examined only 1 measure of functional limitation (HAQ), but would expect similar findings with other measures of functional limitation. Using radiographic data from different trials introduces systematic error due to different readers and scoring methods. We used quartiles of radiographic scores to diminish the potential influence of these effects. Although neither radiography nor RA duration is a gold standard for irreversibility, associations with the latent damage trait would be expected for both measures, therefore allowing for a test of our hypothesis. While the design of this study ensured that the reversible component of the functional limitation was related to RA activity, the irreversible functional limitations may include limitations due to comorbidities as well as consequences of joint damage from RA. In our analyses, we assumed that comorbidities were similar across subgroups, for several reasons: first, comorbidity is largely age related, and ages were similar across subgroups of duration and radiographic score; second, findings were similar in analyses that adjusted for age; ALETAHA ET AL and third, patients had been enrolled in clinical trials that excluded patients with major comorbid conditions. It is important not to generalize the proportion of reversibility of the HAQ reported here to other settings, or to consider these proportions as typical of the findings expected in populations of patients with RA. These proportions were based on findings in patients with active RA enrolled in clinical trials, and will be different in patients who attend a general clinic or are enrolled in a community-based survey, because they depend in part on the degree of RA activity. Also, even the subgroup of patients with RA of the longest duration might be considered to have early disease, and the irreversible part of the HAQ might be higher in patients with more longstanding RA. Nevertheless, the importance of our findings relates to the associations of function and functional response with damage, which are likely to be similar in all RA populations. Our results have several implications. First, the dual nature of measures of functional limitations complicates attempts to identify risk factors. Risk factors for functional limitations associated with RA activity are likely to be different from risk factors for functional limitations associated with irreversible damage. Risk factors may be overlooked, or their associations inaccurately specified, if heterogeneous groups of patients are examined. Our findings suggest that risk factor analyses should be stratified to separate, as best as possible, those with functional limitations due to chronic damage from those with functional limitations due to active disease. Similar problems may complicate the prediction of future outcomes, such as work loss or mortality, using the HAQ. Second, our findings suggest that claims of improvement of disability based on improvements in functional measures in short-term clinical trials should not be interpreted to indicate alteration of functional limitations due to joint damage or the potential for progression of joint damage over time. The improvements demonstrated in trials rest on the functional consequences of active symptoms and synovitis, which are distinct from functional consequences related to chronic damage. Regulatory claims for improvement in disability in RA recognize this distinction by requiring trials of at least 2 years (27). However, even these claims will be confounded by RA activity–related improvements if the change in functional limitation from trial entry is taken as the supporting evidence, rather than maintenance of low levels of functional limitations. Third, differing sources of functional limitations may confound comparisons of treatment responses MEASURES OF FUNCTIONING IN RA across studies. Studies with a large proportion of patients who have reversible functional limitations due to active RA will have greater treatment responses than studies with a large proportion of patients whose functional limitations are irreversible, even though both cohorts may have similar mean HAQ scores at baseline. Different degrees of chronic damage will have, a priori, limited the potential reversibility of functional impairment in some patients. Fourth, methodologic studies of measures of functional impairment need to take into account the consequences of floor effects. Our results indicate that the “floor” of a measure will differ depending on the degree of irreversible damage present. These differences will affect estimates of the measure’s sensitivity to change, and other metrics based on change, such as the minimal clinically important difference (MCID). Our findings indicate that there is not likely to be a single estimate of sensitivity to change or MCID for measures of function, but that these estimates will vary depending on the degree of irreversible limitation present in the groups that are tested. This likely reduces our ability to discriminate treatment effects in patients with longstanding RA. Generally, the floor effect related to the irreversible functional limitations may also explain why the HAQ is among the measures that are least sensitive to change in many clinical trials in RA (10,28,29). Further research on the responsiveness of functional measures in different subgroups of patients will be of interest to support these conclusions. In addition, cost-effectiveness analyses that use changes in functional limitations based on short-term clinical trials as estimates of effect, and that project these estimates to model future long-term disability outcomes, may be confounded by these different components of functional disability (30–33). Changes in functional limitations observed in short-term clinical trials likely do not directly translate into similar changes in long-term disability. Therefore, modeling of longterm cost-effectiveness should ideally be based on the component of disability that is irreversible. Taken together, our findings suggest that the interpretation of functional scores, and not only in RA, should be viewed more adaptively, with the specific purpose of measurement in mind. Is the purpose to measure the total functional limitation at a specific point in time, which reflects the impact of the current and past disease process (i.e., reversible and irreversible components)? Or is the purpose to measure the activity-related component of functional limitation, which is the target of most therapeutic interventions and can be expected to 2791 be fully reversible with effective treatment? Or is the purpose to measure the damage-related component, which reflects cumulative long-term consequences of disease, irrespective of the current state of activity? In treatment decision-making, expectations of changes in functioning should take into account the degree of irreversible limitations present, to prevent overestimation or underestimation of the patient’s prognosis and potential for response to therapy. Therefore, recognition of a “damage-related” HAQ can be helpful to assess the risk of future disability and, at the same time, can be informative regarding strategies to improve current disability. We have presented a method to distinguish the contribution of reversible and irreversible components to one particular functional measure in a particular disease. Other functional measures, self-reported or not, are likely affected by similar issues. Even more important, our findings likely apply to functional measures in any chronic disease. Health status measures of functioning share with measures of organ function, such as serum creatinine concentration, forced expiratory volume, or cardiac ejection fraction, the problem that these are measures of overall functioning, that abnormalities in functioning have different causes, and that only time and effective treatment can distinguish acute reversible dysfunction from chronic irreversible dysfunction. ACKNOWLEDGMENTS We thank the rheumatologists who participated in the survey, Dr. Peter Lipsky for his help, and Abbott, Centocor, and Sanofi Aventis for kindly providing data from their trials. REFERENCES 1. Van Leeuwen MA, van der Heijde DM, van Rijswijk MH, Houtman PM, van Riel PL, van de Putte LB, et al. Interrelationship of outcome measures and process variables in early rheumatoid arthritis: a comparison of radiologic damage, physical disability, joint counts, and acute phase reactants. J Rheumatol 1994;21: 425–9. 2. Smolen JS, Aletaha D. Patients with rheumatoid arthritis in clinical care. Ann Rheum Dis 2004;63:221–5. 3. Breedveld FC, Han C, Bala M, van der Heijde D, Baker D, Kavanaugh AF, et al. Association between baseline radiographic damage and improvement in physical function after treatment of patients with rheumatoid arthritis. Ann Rheum Dis 2005;64:52–5. 4. Drossaers-Bakker KW, de Buck M, van Zeben D, Zwinderman AH, Breedveld FC, Hazes JM. Long-term course and outcome of functional capacity in rheumatoid arthritis: the effect of disease activity and radiologic damage over time. Arthritis Rheum 1999; 42:1854–60. 5. Scott DL. Radiological progression in established rheumatoid arthritis [review]. J Rheumatol Suppl 2004;69:55–65. 6. Scott DL, Pugner K, Kaarela K, Doyle DV, Woolf A, Holmes J, et 2792 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. al. The links between joint damage and disability in rheumatoid arthritis [review]. Rheumatology (Oxford) 2000;39:122–32. Welsing PM, van Gestel AM, Swinkels HL, Kiemeney LA, van Riel PL. The relationship between disease activity, joint destruction, and functional capacity over the course of rheumatoid arthritis. Arthritis Rheum 2001;44:2009–17. Lipsky PE, van der Heijde DM, St. Clair EW, Furst DE, Breedveld FC, Kalden JR, et al, for the Anti-Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy Study Group. Infliximab and methotrexate in the treatment of rheumatoid arthritis. N Engl J Med 2000;343:1594–602. St.Clair EW, van der Heijde DM, Smolen JS, Maini RN, Bathon JM, Emery P, et al, for the Active-Controlled Study of Patients Receiving Infliximab for the Treatment of Rheumatoid Arthritis of Early Onset Study Group. Combination of infliximab and methotrexate therapy for early rheumatoid arthritis: a randomized, controlled trial. Arthritis Rheum 2004;50:3432–43. Weinblatt ME, Keystone EC, Furst DE, Moreland LW, Weisman MH, Birbara CA, et al. Adalimumab, a fully human anti–tumor necrosis factor ␣ monoclonal antibody, for the treatment of rheumatoid arthritis in patients taking concomitant methotrexate: the ARMADA trial [published erratum appears in Arthritis Rheum 2003;48:855]. Arthritis Rheum 2003;48:35–45. Keystone EC, Kavanaugh AF, Sharp JT, Tannenbaum H, Hua Y, Teoh LS, et al. Radiographic, clinical, and functional outcomes of treatment with adalimumab (a human anti–tumor necrosis factor monoclonal antibody) in patients with active rheumatoid arthritis receiving concomitant methotrexate therapy: a randomized, placebo-controlled, 52-week trial. Arthritis Rheum 2004;50:1400–11. Smolen JS, Kalden JR, Scott DL, Rozman B, Kvien TK, Larsen A, et al, and the European Leflunomide Study Group. Efficacy and safety of leflunomide compared with placebo and sulphasalazine in active rheumatoid arthritis: a double-blind, randomised, multicentre trial. Lancet 1999;353:259–66. Scott DL, Smolen JS, Kalden JR, van de Putte LB, Larsen A, Kvien TK, et al. Treatment of active rheumatoid arthritis with leflunomide: two year follow up of a double blind, placebo controlled trial versus sulfasalazine. Ann Rheum Dis 2001;60: 913–23. Emery P, Breedveld FC, Lemmel EM, Kaltwasser JP, Dawes PT, Gomor B, et al. A comparison of the efficacy and safety of leflunomide and methotrexate for the treatment of rheumatoid arthritis. Rheumatology (Oxford) 2000;39:655–65. Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 1980;23:137–45. Scott DL, Symmons DP, Coulton BL, Popert AJ. Long-term outcome of treating rheumatoid arthritis: results after 20 years. Lancet 1987;1:1108–11. Welsing PM, Landewe RB, van Riel PL, Boers M, van Gestel AM, van der Linden S, et al. The relationship between disease activity and radiologic progression in patients with rheumatoid arthritis: a longitudinal analysis. Arthritis Rheum 2004;50:2082–93. Plant MJ, Jones PW, Saklatvala J, Ollier WE, Dawes PT. Patterns of radiological progression in early rheumatoid arthritis: results of an 8 year prospective study. J Rheumatol 1998;25:417–26. ALETAHA ET AL 19. Sharp JT, Young DY, Bluhm GB, Brook A, Brower AC, Corbett M, et al. How many joints in the hands and wrists should be included in a score of radiologic abnormalities used to assess rheumatoid arthritis? Arthritis Rheum 1985;28:1326–35. 20. Van der Heijde DM, van Riel PL, Nuver-Zwart IH, Gribnau FW, van de Putte LB. Effects of hydroxychloroquine and sulphasalazine on progression of joint damage in rheumatoid arthritis. Lancet 1989;1:1036–8. 21. Smolen JS, Breedveld FC, Schiff MH, Kalden JR, Emery P, Eberl G, et al. A simplified disease activity index for rheumatoid arthritis for use in clinical practice. Rheumatology (Oxford) 2003;42: 244–57. 22. 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. 23. Aletaha D, Smolen J. The Simplified Disease Activity Index (SDAI) and the Clinical Disease Activity Index (CDAI): a review of their usefulness and validity in rheumatoid arthritis. Clin Exp Rheumatol 2005;23(5 Suppl 39):S100–8. 24. Aletaha D, Machold KP, Nell VP, Smolen JS. The perception of rheumatoid arthritis core set measures by rheumatologists: results of a survey. Rheumatology (Oxford) 2006. E-pub ahead of print. 25. Escalante A, Haas RW, del Rincon I. A model of impairment and functional limitation in rheumatoid arthritis. BMC Musculoskelet Disord 2005;6:16–23. 26. Aletaha D, Ward MM. Duration of rheumatoid arthritis influences the degree of functional improvement in clinical trials. Ann Rheum Dis 2006;65:227–33. 27. US Department of Health and Human Services, Food and Drug Administration. Guidance for industry clinical development programs for drugs, devices, and biological products for the treatment of rheumatoid arthritis. URL: www.fda.gov/cder/guidance/ 1208fnl.pdf. 28. Verhoeven AC, Boers M, van der Linden S. Responsiveness of the core set, response criteria, and utilities in early rheumatoid arthritis. Ann Rheum Dis 2000;59:966–74. 29. Scott DL, Strand V. The effects of disease-modifying antirheumatic drugs on the Health Assessment Questionnaire score: lessons from the leflunomide clinical trials database. Rheumatology (Oxford) 2002;41:899–909. 30. Kobelt G, Jonsson L, Lindgren P, Young A, Eberhardt K. Modeling the progression of rheumatoid arthritis: a two-country model to estimate costs and consequences of rheumatoid arthritis. Arthritis Rheum 2002;46:2310–9. 31. Brennan A, Bansback N, Reynolds A, Conway P. Modelling the cost-effectiveness of etanercept in adults with rheumatoid arthritis in the UK. Rheumatology (Oxford) 2004;43:62–72. 32. Bansback NJ, Brennan A, Ghatnekar O. Cost effectiveness of adalimumab in the treatment of patients with moderate to severe rheumatoid arthritis in Sweden. Ann Rheum Dis 2005;64: 995–1002. 33. Kobelt G, Eberhardt K, Jonsson L, Jonsson B. Economic consequences of the progression of rheumatoid arthritis in Sweden. Arthritis Rheum 1999;42:347–56.