Changes in outcome measures for impairment, activity limitation, and participation restriction over two years in osteoarthritis of the lower extremities
To describe changes in outcome measures in patients with knee and hip osteoarthritis (OA) over 2 years according to the International Classification of Functioning, Disability and Health, and to evaluate the sensitivity to change of available outcome instruments.
A total of 115 symptomatic knee or hip OA patients (mean age 60.0 years, 80% women) were followed for 2 years. Standardized knee and hip radiographs were scored for joint space narrowing (JSN) using the Osteoarthritis Research Society International Atlas. Pain intensity in knees and hips was graded during physical examination. Self-reported pain and functioning were assessed with the Western Ontario and McMaster Universities OA Index (WOMAC). Social functioning was assessed with social functioning scores of the RAND 36-item Health Survey. Standardized response means (SRMs) were calculated to evaluate sensitivity to change.
Substantial increases in JSN and pain intensity total scores over 2 years were observed (SRMs 0.43 and 0.41, respectively). Twenty-three percent of patients had an increase of at least 1 point in JSN total scores. An increase in pain intensity total scores was present in 46% of patients, whereas a decrease was observed in 19.1.% of patients. WOMAC pain and function scores showed small increases (SRMs 0.15 and 0.18, respectively). No change was seen in social functioning scores (SRM 0.01).
Objective instruments measuring impairment in body structures and function are more sensitive to change over 2 years in patients with OA than self-reported measurements of impairment in body function, activity limitation, and participation restriction. These findings encourage development of new instruments to improve measurement of disease outcome in OA.
Osteoarthritis (OA) is a chronic progressive joint disease that leads to pain and loss of physical functioning in a considerable proportion of patients. To date, treatment of OA remains symptomatic because structure-modifying drugs are not yet available. To develop new therapies that could control OA symptoms and reduce OA disability, knowledge of the progression of OA is important. Therefore, reliable, valid, and responsive instruments are needed that are sensitive to change over time.
To test disease-modifying therapies in OA, a core set of outcome measures (1) have been recommended that include clinically important patient-orientated outcomes such as pain and physical functioning. Currently, several instruments are available to assess changes in these outcomes over time. One of the instruments that has been widely used is the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), a validated and reliable self-administered questionnaire on pain and functioning in patients with hip and knee OA (2). The short-term responsiveness of the WOMAC has been demonstrated in numerous clinical trials (3). However, there have been a limited number of studies on the performance of the WOMAC over a long period of time (4–6).
For OA studies with a duration of 1 year or longer, radiographic imaging was recommended as a core outcome measure to assess structural changes (1), although in general it is considered to have limited sensitivity to change over time in OA. In most studies, radiologic change was evaluated in only 1 signal joint. However, patients with OA at multiple joint sites have an increased risk for progression of OA (7) and have more affected joints that can potentially progress. Therefore, assessment of radiologic progression in both knee and hip joints may better represent structural outcome at the patient level and improve sensitivity to change.
In recent years, it has become clear that the way in which outcomes in OA are evaluated should be improved. Conceptual frameworks such as the International Classification of Functioning, Disability and Health (ICF) (8) are often used to provide a multiperspective approach for the classification of functioning, disability, and health. In the ICF, functioning is classified in the activity and participation components. The health-related component is classified into 2 dimensions: body structures and body functions. When the ICF is applied to OA, OA can lead to impairment in body structures (e.g., joint space narrowing) and body functions (e.g., pain), which could result in limitation of activities (e.g., walking) that may restrict participation in societal activities (e.g., social involvement). To better understand the entire impact of OA on the lives of patients, all 3 ICF components should be evaluated (9). In the past, this was often not the case. In particular, the participation restriction component has rarely been evaluated, which is remarkable because that seems to be what matters most to patients with arthritis.
The main objective of the present study was to describe the changes in outcome measures in patients with knee and hip OA over a period of 2 years for all 3 components of the ICF (impairment, activity limitation, and participation restriction). We used currently available instruments for outcome assessment in OA and evaluated their sensitivity to detect change after 2 years.
PATIENTS AND METHODS
The present study was part of the Genetics, Arthrosis and Progression (GARP) study, a prospective cohort study aimed at the identification of determinants of OA susceptibility and progression that involves white sibpairs of Dutch ancestry (10). The population predominantly comprised individuals with symptomatic OA at multiple sites. The study comprised 191 white affected sibling pairs. Probands (age 40–70 years) and siblings with OA at multiple joint sites in the hands or with OA in ≥2 joint sites (hand, spine [cervical or lumbar], knee, or hip) were included after giving informed consent.
Patients with secondary OA, familial syndromes with a Mendelian inheritance pattern, and patients with a shortened life expectancy were excluded. Of the 191 sibling pairs, 104 sibling pairs in which at least 1 patient with symptomatic knee OA (but not in a radiologic end stage) or hip OA were selected for a 2-year followup study on the progression of OA.
In the present study on OA of the lower extremities, those patients from the 104 sibling pairs with symptomatic knee or hip OA in at least 1 joint at baseline were included. Patients who received a hip or knee prosthesis during the 2-year followup (n = 12) were excluded from the analyses.
For the GARP study, patients (probands) between ages 40–70 years with symptomatic OA in the hands, knees, or hips (diagnosed by rheumatologists, orthopedic surgeons, and general practitioners in Leiden, The Hague, Delft, Haarlem, and Amsterdam) were informed of the ongoing study by mail. Interested probands were subsequently sent a mailed questionnaire about demographic data, medical history, symptoms and signs of OA, OA treatment, and family history of OA. Subsequently, probands with OA at multiple sites and with a positive family history were requested to introduce a sibling “with joint complaints,” who was also sent a questionnaire. After obtaining informed consent, all sibpairs underwent a physical examination and were assessed by a single medical doctor at the outpatient clinic.
Symptomatic knee and hip OA were defined according to the American College of Rheumatology recommendations for knee and hip OA (11, 12). Knee OA was defined as pain or stiffness on most days of the preceding month in addition to osteophytes at the joint margins of the tibiofemoral joint. Hip OA was defined as pain or stiffness in the groin and hip region on most days of the preceding month in addition to femoral or acetabular joint space narrowing on radiograph. Prosthetic joints in hips or knees as a result of end-stage OA were defined as OA in those particular joints.
Assessment of outcome in body structures.
Standardized radiographs of the knees (posteroanterior [PA] weight bearing) and hips (PA weight bearing) were obtained in a single center at baseline and after 2 years. The standardized nonfluoroscopic fixed-flexion protocol was used to obtain the weight-bearing knee radiographs. Uniform anatomic alignment of the knees was facilitated using a specifically designed positioning frame. Baseline and 2-year radiographs were graded on a scale of 0–3 for joint space narrowing (JSN). The right and left medial and lateral compartments of the tibiofemoral joints and right and left hip joints were scored with the help of the Osteoarthritis Research Society International (OARSI) Atlas (13). All radiographs were scored in consensus by 2 readers. Films were read in pairs without knowledge of the chronological order. The reproducibility for JSN grading, assessed in a random sample of 20 radiographs, was excellent. Intraclass correlation coefficients were 0.88 in the knees and 1.00 in the hips.
Radiologic progression of JSN was evaluated at the patient level as well as the specific joint site level. Total JSN scores in the knees and hips were calculated by adding the scores of the left and right sides. The maximum total JSN score in a single was 18:6 in the hips and 12 in the knees.
Assessment of outcome in impairment in body functions.
During a physical examination at baseline and after 2 years, pain upon movement in the knee and hip joints was graded on a 4-point scale for intensity (where 0 = no pain, 1 = patient reported pain, 2 = patient reported pain and winced, 3 = patient reported pain and withdrew the joint). The pain intensity total score (range 0–12) is a modification of the articular index for the assessment of OA as described by Doyle et al (14).
Self-reported pain in the knees and hips was evaluated using the pain subscale (range 0–100) of the WOMAC (2). The Dutch visual analog scale version was used, where 5 pain items are rated by the patient regarding both hips and knees in the last 48 hours.
Assessment of outcome in activity limitation.
Self-reported physical functioning in the knees and hips was evaluated at baseline and after 2 years using the function subscale (range 0–100) of the WOMAC, where 17 function items are rated by the patient regarding both hips and knees in the last 48 hours.
Assessment of outcome in participation restriction.
Social functioning was assessed using the social functioning score (range 0–100) of the Dutch-validated RAND 36-item Health Survey (15). A higher score indicates better social functioning.
Data were analyzed using SPSS software, version 12.0 (SPSS, Chicago, IL). Mean changes between baseline and 2 years were reported with 95% confidence intervals. To evaluate the magnitude of the changes observed over 2 years and to compare the responsiveness of the instruments, standardized response means (SRMs) were calculated as the mean change between baseline and 2 years divided by the standard deviation of the change.
Of the 208 patients included in the GARP study, 127 were eligible for the present study. Complete followup data were available from 115 (91%) of these patients. Baseline characteristics are shown in Table 1. Symptomatic hip and knee OA was present in 76 and 53 patients, respectively, and 14 patients had a combination of hip and knee OA. At baseline, 70.4% of the patients used medication for OA, including acetaminophen (22.6%), nonsteroidal antiinflammatory agents (55.7%), tramadol (7%), and glucosamine sulfate (1.7%).
Table 1. Characteristics of 115 patients with symptomatic knee or hip OA with complete followup over 2 years*
|Age, median (range) years||60.0 (55.5–66.5)|
|BMI, median (range) kg/m2||26.0 (24.4–29.7)|
|Symptomatic knee OA||76 (66.1)|
|Symptomatic hip OA||53 (46.1)|
|Symptomatic knee and hip OA||14 (12.2)|
Changes in impairment in body structures.
Changes in JSN total scores are presented in Table 2. There was an increase in JSN total scores over 2 years and the corresponding SRM was 0.43. Of the total population of 115 patients with knee or hip OA, 27 (23%) patients had an increase of at least 1 point in JSN total score of the knees and hips (Table 3). Of the 76 knee OA patients and 53 hip OA patients, 22 (29%) and 2 (4%), respectively, showed radiographic progression of at least 1 point over 2 years.
Table 2. Changes in impairment, activity limitation, and participation restriction after 2 years in patients with symptomatic knee or hip osteoarthritis*
|Impairment|| || || || || |
| JSN total scores (0–18)||2.9 ± 2.2||3.2 ± 2.2||0.3 ± 0.7||0.2, 0.4||0.43|
| Pain intensity total scores (0–12)||2.5 ± 2.8||3.4 ± 3.2||0.9 ± 2.3||0.5, 1.3||0.41|
| WOMAC pain (0–100)||29.2 ± 21.6||32.0 ± 26.7||2.8 ± 19.2||−0.7, 6.4||0.15|
|Activity limitation|| || || || || |
| WOMAC function (0–100)||27.5 ± 22.0||30.7 ± 25.2||3.3 ± 18.7||−0.2, 6.7||0.18|
|Participation restriction|| || || || || |
| Social functioning (0–100)||77.9 ± 22.6||77.7 ± 24.8||−0.2 ± 21.9||−4.3, 3.8||0.01|
Table 3. Distribution of changes in joint space narrowing (JSN) and pain intensity total scores over 2 years in 115 patients with knee and hip osteoarthritis*
|JSN total scores (0–18)||0||0||3||85||20||5||2||0|
|Pain intensity total scores (0–12)||1||6||15||40||22||13||9||9|
In additional analyses, we examined the changes in JSN total scores of the symptomatic and nonsymptomatic patients separately. Of the total population of 115 patients with symptomatic knee and/or hip OA, 79 patients had radiologic knee OA (Kellgren/Lawrence [K/L] grade 2 or higher) at baseline. Of these 79 patients, 71 were symptomatic with knee pain. Of the 56 patients with radiologic hip OA (K/L grade 2 or higher) at baseline, 50 were symptomatic with hip pain. For both the knee and hip joints, the SRMs for changes in JSN were higher in the joints with symptomatic OA at baseline (SRM 0.51 for knee JSN and 0.20 for hip JSN) compared with the nonsymptomatic baseline OA joints (SRM 0.15 for knee JSN and 0.0 for hip JSN).
Changes in impairment in body functions.
There was an increase in pain intensity total scores over 2 years with an SRM of 0.41 (Table 2). An increase in pain intensity total scores was present in 53 (46%) patients (Table 3). Forty (34.8%) patients had no change and 22 (19.1%) had a decrease in pain intensity total scores.
Additional analyses were performed on the changes in the pain intensity total scores in the knees (0–6) and hips (0–6) of the symptomatic baseline joints versus the nonsymptomatic baseline joints. Of the 115 patients included at baseline, 90 (78%) were symptomatic with knee pain at baseline. In these 90 patients, the mean ± SD increase in pain intensity scores in the knees over the 2 years was 0.72 ± 1.3 (SRM 0.55). In the 25 patients without knee pain at baseline, the mean ± SD increase in pain intensity scores in the knees was 0.16 ± 0.47 (SRM 0.34).
Of the 115 patients included at baseline, 84 (73%) were symptomatic with hip pain. In these 84 patients, the mean ± SD increase in pain intensity scores for the hips was 0.38 ± 2.31 (SRM 0.16). In the 31 patients without hip pain at baseline, the mean ± SD increase in pain intensity scores in the hips was 0.16 ± 0.73 (SRM 0.22).
In the total population, the WOMAC pain scores showed only a small increase over 2 years, with an SRM of 0.15 (Table 2).
Changes in activity limitation.
In the total population, the WOMAC function scores also showed little change over 2 years, with an SRM of 0.18 (Table 2).
Changes in participation restriction.
The social functioning scores showed no change over 2 years with an SRM of 0.01 (Table 2).
In the present study population of patients with symptomatic knee and hip OA, an increase in impairment in body structures and functions over a followup period of 2 years was detected using radiologic and pain intensity total scores. Small increases in impairments in body functions and activity limitation were seen using WOMAC pain and function scores. Participation restriction, measured by the social functioning scores of the RAND 36-item Health Survey, did not change. The sensitivity to detect changes in outcome was remarkably higher for the radiologic progression and pain intensity total scores than for the WOMAC and social functioning scores.
Although radiographs are considered to be insensitive to changes in OA (16), radiologic progression of JSN was detected over 2 years in the present study. Radiographs may have been more sensitive to change in the present study, because OA progression was assessed at multiple joint sites including the right and left knee and hip joints of each patient. These findings may suggest that the efficacy of new structure-modifying drugs should be evaluated at several anatomic sites during OA clinical trials. This will not only broaden the opportunities for testing new drugs in OA, but will also help to understand the generalized nature of OA.
At present, there is no agreement as to which clinical measure best mirrors disease activity in OA. The current study provides results on the metrologic properties of the pain intensity total score that was obtained during the physical examination. This score was based on the articular index described by Doyle et al in 1981 (14). The articular index was developed in analogy to the Ritchie index for rheumatoid arthritis and involved 48 joints in the entire body. In the study by Doyle et al, the articular index was valid, reliable, and responsive to change. In the present study, the index was modified for usage in patients with knee and hip OA. This modified version seemed to be a sensitive instrument. Interestingly, 22 patients improved after 2 years, which may reflect how dynamic OA pain can be. Further evaluation and assessment of intra- and interobserver reliability of this index should be performed. If confirmed, this index could form a basis for the development of a measure of disease activity in knee or hip OA.
Few data on the long-term performance of the WOMAC exist. In clinical trials with a followup of 3 years, the responsiveness estimates varied between 0.11 and 0.19 for the WOMAC pain scale in patients receiving placebo treatment (6). Two prospective observational studies investigated the longitudinal performance of the WOMAC. In a study by Bruce and Fries (5), no substantial change in pain and function was found in patients with hip or knee OA after a mean of 3.2 years when measured with the WOMAC pain and function scales, with small responsiveness estimates of 0.10 and 0.05, respectively. In the latter study, a modified time frame (1 week) for the WOMAC was used. The present study demonstrated similar results, although the conventional time frame of 48 hours was used. In a study by Sharma et al (4) focusing on knee OA, a substantial number of patients (>40%) showed no change in WOMAC function scores after 3 years. Responsiveness estimates could not be calculated. Previous studies varied with respect to severity of knee pain and functional limitation at baseline. Although the present study population is representative of patients with symptomatic OA in the lower extremities, the mean reported WOMAC pain and function scores at baseline were relatively low, which could indicate a mild clinical severity in a portion of this population. We believe that this may explain the limited changes found over a 2-year period, because knee OA could take 20–40 years to progress from onset to need for total joint replacement. However, in additional analyses, we found that among patients with scores in the upper half of WOMAC pain and function scores at baseline, no substantial changes in WOMAC pain and function over 2 years were present. These results indicate that the WOMAC is not a sensitive instrument to detect changes in self-reported pain and functioning in knee and hip OA patients over a period of 2–3 years. This is important, because the WOMAC is recommended as a core measure for OA clinical trials, but its low sensitivity to change may increase the required sample size and followup time needed, increasing the costs of these clinical trials.
A possible explanation for the results concerning the WOMAC can be found in the ICF framework in which the impact of a disease in a certain patient is a dynamic interaction between the disease and environmental and personal factors. These environmental and personal factors may influence the observed changes in outcome measures in OA (17).
The participation component of the ICF framework is an important dimension that is not well represented in available OA outcome measures. None of the existing instruments such as the Arthritis Impact Measurement Scale (AIMS) and the RAND-36 uniquely assess this component and therefore cannot be used in their entirety. Subscales of these instruments, such as the social functioning subscale of the RAND-36, should be used for a more uncontaminated measurement of participation restriction (9). These subscales, however, have only a small number of items (2 for RAND-36 and 4 for AIMS). This might explain the lack of change that we found in restriction in participation when using the social functioning scores of the RAND-36. Our findings indicate that new instruments that uniquely measure participation outcome should be developed. These not only need to be included as core outcomes in OA research, but may be used to develop interventions with improving patients' levels of participation as the primary aim.
The present study had the following limitations. The OARSI and the Outcome Measures in Rheumatology Clinical Trials societies initiated a task force to improve severity scoring systems for the 3 different domains: pain, function, and structure (18). Currently, tools evaluating pain and function are available in different languages. However, the present study was performed before these were available for use. Furthermore, the radiographic protocol used in the present study for assessment of radiologic progression of knee OA was the nonfluoroscopic fixed-flexion protocol. This protocol has advantages over fluoroscopically based protocols with regard to radiation exposure and the fact that it requires less expensive radiographic equipment and less training of personnel. However, with fluoroscopic protocols, better quality and reproducibility of metatarsophalangeal alignment of the knee joint on serial radiographs may lead to an increase in sensitivity to detect change.
To conclude, the present prospective observational study provides data on the responsiveness of several instruments for outcome assessment in knee and hip OA that measure the 3 components of the ICF: impairment, activity limitation, and participation restriction. Objective instruments measuring impairment in body structures and function were most sensitive to change over 2 years. The self-reported WOMAC, which measures impairment in body function and activity limitation, was less responsive, and the participation restriction instrument had a very low sensitivity to change. These findings encourage further research and development of new instruments to improve measurement of disease outcome in OA.
Dr. Botha-Scheepers 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.
Study design. Botha-Scheepers, Rosendaal, Breedveld, Hellio Le Graverand, Kloppenburg.
Acquisition of data. Botha-Scheepers, Watt, Kloppenburg.
Analysis and interpretation of data. Botha-Scheepers, Watt, Rosendaal, Breedveld, Hellio Le Graverand, Kloppenburg.
Manuscript preparation. Botha-Scheepers, Watt, Breedveld, Hellio Le Graverand, Kloppenburg.
Statistical analysis. Botha-Scheepers, Rosendaal, Kloppenburg.
ROLE OF THE STUDY SPONSOR
This study was partially financially sponsored by Pfizer. Pfizer contributed partially to the study design. Pfizer had no influence on the decision to submit the manuscript or the content of the submitted manuscript.