Data on the burden of disease and impact on health-related quality of life (HRQOL) in hand osteoarthritis (OA) are limited. The goal of this study was to compare HRQOL in patients with hand OA with HRQOL in patients with rheumatoid arthritis (RA), healthy controls, and normative data from the general population.
A total of 190 women with hand OA were compared with 194 women with RA and 144 healthy women of the same age. Health status was measured using the Short Form 36 (SF-36), Short Form 6D (SF-6D), modified Health Assessment Questionnaire (M-HAQ), pain and fatigue visual analog scales, and grip strength. Scores were compared by analysis of variance and a multivariate analysis of covariance, adjusting for age, number of comorbidities, and years of education. Gaps between patients and population subjects were assessed by calculating S scores on all dimensions of the SF-36.
Hand OA and RA patients had worse scores on all assessed dimensions of subjective health compared with healthy controls. RA patients showed poorest general health (SF-36), poorest physical function (M-HAQ, SF-36 physical, grip strength), and highest level of fatigue compared with hand OA patients. Hand OA patients reported poorer mental health. Mean utility scores (SF-6D) in hand OA and RA were 0.64 and 0.63, respectively, with a mean difference compared with healthy controls of 0.13 in hand OA and 0.14 in RA patients. S scores confirmed a marked disparity between individuals with a rheumatic diagnosis (hand OA, RA) and population subjects.
This study illustrates that patients with hand OA experience a broad impact on HRQOL compared with healthy controls. Fatigue and physical function are worse in RA than hand OA.
Musculoskeletal diseases are increasingly recognized as a major and important group of diseases (1), which requires more attention from society and health care systems (2). Osteoarthritis (OA) is the most frequent rheumatic joint disease and its occurrence increases due to aging populations (3). Knee OA is the most common type of OA, followed by hand OA, which typically affects the distal and proximal interphalangeal joints and the carpometacarpal joint of the thumb (4). Major manifestations of hand OA are pain, stiffness, and impaired physical function (5). Hand OA is also a key manifestation in the proposed classification criteria for generalized OA (6). Several studies have explored the burden of disease, health-related quality of life (HRQOL), and the efficacy of a variety of interventions in knee OA, and recommendations for management have been developed (7, 8). Remarkably few studies have addressed similar aspects of hand OA (9, 10), even though the classification criteria were developed more than 15 years ago (11).
We have established a register of patients with hand OA to address the burden of disease in hand OA (12). Access to disease registers of a variety of rheumatic conditions in Oslo (13) offers opportunities for comparing burden of diseases over time (14) and across diseases and healthy controls (15). The goal of the present study was to compare HRQOL in female patients with hand OA with HRQOL in female patients with rheumatoid arthritis (RA), healthy female controls, and normal subjects from the female general population.
PATIENTS AND METHODS
This study included patients with hand OA, patients with RA, healthy controls, and normal subjects from the general Norwegian population. The latter 3 were included as comparison groups.
Hand OA patients.
Patients between 50 and 70 years of age with hand OA were eligible to be enrolled in the hand OA register if they did not have any other rheumatic diseases (e.g., RA or psoriatic arthritis). Of 275 invited patients, 209 (76%) consented to participate in the current data collection after receiving written and verbal information (12). For the present comparative study, we excluded all 19 male respondents. All 190 female patients (Table 1) had been referred to an outpatient rheumatology department from general practitioners within the previous 2 years according to common practice in Norway. All patients reported OA pain during the last month before screening. A total of 159 (83.7%) patients fulfilled the American College of Rheumatology (ACR) clinical classification criteria for hand OA (11), whereas 31 (16.3%) had clinical hand OA without formally fulfilling these classification criteria. These 31 patients had pain and bony enlargements in several finger joints, but not in a sufficient number or location to fulfill the ACR classification criteria. Radiographic OA abnormalities (Kellgren/Lawrence grade 2 or higher) in at least one of the finger joints were observed in 176 patients (93%) (16). Seventeen (9%) patients also fulfilled the clinical and radiographic ACR classification criteria for hip OA (17) and 112 (59%) fulfilled the clinical ACR classification criteria for knee OA (18).
Values are the mean (range) unless otherwise indicated. OA = osteoarthritis; RA = rheumatoid arthritis; NA = not applicable.
Analysis of variance.
Disease duration, years
Years of education
Number of comorbidities
Data from 194 women with RA between ages 50 and 70 years (mean age 61.1 years) included in the Oslo RA register were included in the comparative analyses (13, 19). These patients had undergone a comprehensive data collection in 1998–1999 (20). Only patients who had been routinely examined in the outpatient department during the 2 years preceding the data collection and patients without any clinical evidence of hand OA were eligible for these analyses.
The healthy controls comprised 144 healthy women (mean age 60.8 years) who had been randomly selected from the population register in Oslo and had undergone the same procedures and examinations as the RA cohort (20). From the original 249 controls included in the RA study, we excluded individuals who had clinical signs of OA, reported OA as a concomitant disease, or were outside the age range of 50–70 years (20).
The population data were extracted from a random computer draw of the National Population Register among all Norwegian inhabitants ages 50–70 years with the same digit in their social security number. Data from 384 women (mean age 59.2 years) were used for this analysis (21, 22).
HRQOL was measured by the Short Form 36 (SF-36), visual analog scales (VAS), and the modified Health Assessment Questionnaire (M-HAQ). The procedures for data collection were similar across patient groups and healthy controls. All subjects received a letter asking them to participate in the study. The patients who agreed to take part received mailed questionnaires and returned the completed forms when they subsequently attended a comprehensive clinical examination, according to the research protocol. The number of comorbidities and duration of education (years) were recorded at the patient interview. The individuals representing the general population (population subjects) completed mailed questionnaires including the SF-36, but not VAS or the M-HAQ, and returned the questionnaires by mail to the investigator (21).
Measures of HRQOL.
The SF-36 is the most widely used generic health status questionnaire, measuring 8 dimensions of health: physical function (10 items), role limitations due to physical health problems (4 items), bodily pain (2 items), general health (5 items), vitality/energy (4 items), role limitations due to emotional problems (3 items), and mental health (5 items). SF-36 scores range from 0 to 100, with low scores indicating poor health (23). The Norwegian version of the SF-36 has been validated (12, 24).
The Short Form 6D (SF-6D) is a preference-based utility instrument based on data from the SF-36, which are converted to a utility score. The score is based on 6 dimensions (physical functioning, role limitations, social functioning, pain, mental health, and vitality), each with 4–6 levels. The score range is from 0.29 to 1.0, with 1.0 representing perfect health (25).
VAS scales for joint pain and fatigue were considered generic measures of health status. Perceived symptoms were assessed on a horizontal, 0–100-mm scale with no pain/fatigue and intolerable pain/fatigue (disease activity) as the anchoring points.
The M-HAQ is an 8-item modification of the original 20-item Stanford Health Assessment Questionnaire (HAQ). Each of the 8 items represents 1 of the 8 components of the HAQ (24, 26). For this study, the M-HAQ was considered a generic measure of health status. Patients were asked about their ability to perform 8 daily activities and to indicate the difficulty from grade 1 (without any difficulty) to grade 4 (unable to do). A mean M-HAQ score is calculated from all 8 items, with a range from 1 to 4 (high values represent poor health).
Using the Jamar hand dynamometer (Jamar, Clifton, NJ), grip strength (in kg) was measured in patients with hand OA, patients with RA, and healthy controls as the best performance of 2 attempts for each hand without asking for the predominant hand (27, 28).
Ethics, analyses, and statistics.
All patients signed an informed consent form. The study was approved by the Norwegian data inspectorate and the local ethical committee. The study population was described using the mean as an index of location and the standard deviation as the index of dispersion. For continuous variables, the 3 study groups were compared by one-way analysis of variance (ANOVA) and Tukey's post hoc test (for crude values). In a multivariate analysis of covariance (ANCOVA), adjusting for age, number of comorbidities, and years of education, the least squares means were calculated. This ANCOVA analysis was performed as a first step to compare all 3 groups and then to compare hand OA versus RA.
Standard difference scores (S scores) were used to assess the difference in HRQOL between the study groups and population subjects. The S scores were calculated by subtracting the mean scores of each of the 8 SF-36 scales for the 2 patient groups and the control group from the mean values of the general population (population subject values) and then dividing these differences by the standard deviation of the general population for each score.
SPSS software, version 12.0 (SPSS, Chicago, IL) and SAS software, version 9.1.3 (SAS Institute, Cary, NC) were used in the analyses. P values equal to or less than 0.05 were considered significant. We did not correct for multiple testing.
The 2 patient groups, the control group, and the cohort drawn from the general population were similar in age and duration of education. The number of comorbidities was slightly higher in the hand OA group compared with the RA group (P = 0.05), but disease duration was longer in patients with RA (Table 1).
Patients with hand OA and RA had worse crude scores for all dimensions of health (SF-36) compared with healthy controls (Figure 1). The 3-group ANOVA statistical comparison between patients with hand OA, patients with RA, and healthy controls was significant for all measures, with a P value less than 0.005 (hand OA versus RA versus healthy controls) (data not shown). The statistically significant differences between patients with hand OA, patients with RA, and healthy controls remained after adjusting for age, number of comorbidities, and years of education (Table 2).
Table 2. Health-related quality of life in patients with hand OA and RA and in healthy controls*
Values are the mean ± SEM (adjusted for age, number of comorbidities, and years of education) unless otherwise indicated. All P values for the 3-group comparison <0.005. OA = osteoarthritis; RA = rheumatoid arthritis; SF-36 = Short Form 36; M-HAQ = modified Health Assessment Questionnaire; VAS = visual analog scale; SF-6D = Short Form 6D.
Analysis of covariance for a 2-group comparison, also adjusted for age, number of comorbidities, and years of education.
59.1 ± 1.6
48.4 ± 1.6
81.6 ± 1.9
SF-36 role limitation
36.0 ± 2.8
29.9 ± 2.8
75.3 ± 3.3
1.48 ± 0.03
1.63 ± 0.03
1.07 ± 0.04
41.3 ± 1.6
44.0 ± 1.6
69.4 ± 1.9
VAS pain (mm)
38.0 ± 1.6
35.9 ± 1.5
11.5 ± 1.9
41.7 ± 1.6
43.0 ± 1.5
60.1 ± 1.8
VAS fatigue (mm)
43.3 ± 2.0
49.7 ± 2.0
20.5 ± 2.5
69.5 ± 1.4
73.4 ± 1.3
79.3 ± 1.6
SF-36 role limitation
58.5 ± 2.9
55.4 ± 2.9
79.8 ± 3.3
69.4 ± 1.8
68.9 ± 1.8
81.3 ± 2.1
53.8 ± 1.6
46.7 ± 1.5
68.9 ± 1.9
0.64 ± 0.01
0.63 ± 0.01
0.77 ± 0.01
Grip strength right hand (kg)
19.6 ± 0.5
15.5 ± 0.5
27.5 ± 0.6
Grip strength left hand (kg)
16.9 ± 0.5
13.7 ± 0.5
24.8 ± 0.6
Patients with RA had worse scores than patients with hand OA for measures of physical function (M-HAQ, SF-36 physical, grip strength), fatigue VAS, and SF-36 general health, whereas individuals with hand OA had worse scores for SF-36 mental health. No statistically significant differences were observed between hand OA and RA for other measures (Table 2). The adjusted average difference in utility (SF-6D) between healthy controls and patients with hand OA and RA was 0.13 and 0.14, respectively, indicating that HRQOL was decreased by 13 quality-adjusted life years in 100 patients with hand OA compared with healthy controls.
Furthermore, we divided the hand OA group according to whether they fulfilled or did not fulfill the ACR clinical classification criteria for knee OA (18). Patients with knee OA involvement generally had worse health status than patients without concurrent knee OA. However, the group of patients with hand OA without knee OA had worse health than the controls, and the comparison with the RA group revealed results that were similar to the those reported in Table 2 for the entire hand OA group, with the exception of the SF-6D (0.68 for hand OA without knee involvement, P = 0.02 versus RA).
The magnitude of the difference in HRQOL between patient groups and controls versus population subjects is shown as S scores. In principle, these scores showed the same results as the comparison between patients and healthy controls, i.e., that the gaps between healthy individuals and patients with either hand OA or RA were of similar magnitude, with the exception of physical health and fatigue, which were worse in patients with RA compared with those with hand OA (Figure 2). Further, the S scores between healthy controls and the general population were ∼0 (Figure 2).
This study illustrates that patients with hand OA experience a broad impact on HRQOL compared with healthy controls. The consistency of the differences between the patients with hand OA and the 2 healthy comparison groups (Figure 1), as well as S scores around 0 for the control group (Figure 2), support the robustness of these findings.
OA is the most common type of arthritis among persons over 50 years of age (1, 29, 30). A recent report from the World Health Organization as well as several previous studies emphasize the impact of this disorder on patients and society and underline the fact that large numbers of patients are affected and that the economic burden of the disease is substantial worldwide (1, 3, 31–33). The present study demonstrates that patients with hand OA endure an impact on their HRQOL comparable with patients with RA based on data on pain (SF-36, VAS), social functioning, mental health (SF-36), and the SF-6D utility score.
These findings must be considered in the context of the high occurrence of hand OA. The prevalence of RA based on the ACR criteria in persons older than 60 years is between 1% and 2% (34), whereas the prevalence of symptomatic hand OA was found to be 5% in women and 2% in men in one study (35, 36), and another study found an estimated prevalence of 14.9% for a similar age group (35, 37). Prevalence in terms of radiographic OA/hand OA is even higher (35). An incidence rate of 100 per 100,000 person-years for radiographic and symptomatic hand OA (4) was found in one population-based study, whereas the incidence of RA (ACR criteria) is estimated to be between 25 and 40 per 100,000 individuals (19, 38).
We included RA as a comparative group in this study. RA is recognized as a severe disease with a major influence on HRQOL (39). The findings from the current analyses cannot be extrapolated to the entire patient populations with RA and hand OA, because the patients with hand OA had been referred to a rheumatology outpatient department during the 2 years preceding the study. Therefore, the examined group of patients with hand OA probably had a more severe disease than a random sample of patients with hand OA. However, for the comparative analyses we selected patients with RA who had also been examined at the same rheumatology outpatient department during the 2 years preceding the data collection.
The number of instruments in this study was limited. Still, the SF-36 is the most widely used generic health status measure, and findings on physical function, pain, and fatigue were also cross-validated with other instruments (VAS, M-HAQ, and grip strength). Ideally, the utility scores should also have been examined by the EuroQol, which has been shown to have less floor effects than the SF-6D (40).
We did not compare structural damage in this study because the patterns of radiographic abnormalities are different in RA and OA (41–46). However, more research is needed with both conventional (radiographs) and new imaging tools (ultrasonography and magnetic resonance imaging) to develop scoring systems for imaging abnormalities in hand OA and to explore the underlying pathologic processes that lead to the variety of structural abnormalities (47).
Approximately 60% of our patients with hand OA had concomitant, clinical knee OA and 9% fulfilled the ACR criteria for hip OA (17), which is consistent with the frequent occurrence of a generalized type of OA in patients with finger joint involvement (6, 48). We did not adjust for hip and/or knee OA or number of other OA sites, because our goal was not to focus only on hand OA/hand function. On the contrary, we studied HRQOL taking into account that both patients with hand OA and those with RA often have generalized joint involvement.
A limitation of this study is the low percentage of men. Men were not included in the present analyses for 2 reasons. In the original study population, only 19 of 209 patients were men, which was considered too low to give meaningful estimates of HRQOL in men with hand OA. Furthermore, the healthy control population was originally examined in a study on osteoporosis in women (20) and therefore, we did not have access to an existing cohort of healthy men.
The low percentage of men (∼10%) in our cohort is consistent with the increased occurrence of hand OA in women compared with men (49). However, we do not know why the female/male ratio was higher than in most other studies.
We did not address the issue of aesthetic damage, which has been shown to be a concern for patients with hand OA (48, 50). Aesthetic damage is an important dimension that should be included in future studies on HRQOL in patients with hand OA.
A methodologic strength of this study is that all patients with OA were examined by the same clinician, who had several years of training in rheumatology, and the same qualified study nurse. Assessments of patients with RA and controls were performed by another clinician/study nurse team. Both clinical investigators and study nurses worked and were trained in the same department. The involvement of different assessors in the patient and control groups should not, however, be a major concern because the analyses and results (with the exception of grip strength) were based on patient-reported measures. We chose to focus on mean values adjusted for baseline demographic variables (Table 2) rather than crude values (Figure 1), especially because the number of comorbidities differed between patients and controls (Table 1).
This study demonstrates that hand OA and RA are associated with a consistent burden of disease across all dimensions of health compared with population subjects. Physical functioning and fatigue were worse in patients with RA compared with those with hand OA. Our results need confirmation in other populations, new studies are needed to explore the burden of disease in men with hand OA, and longitudinal studies are required to identify the determinants of the burden of disease in patients with hand OA. More effective treatments are available for patients with RA than for those with hand OA, and more is known about disease course and predictors of disease progression. Therefore, hand OA deserves a more distinct priority in future research, with a research agenda focusing on, for example, the longitudinal outcome, prognostic markers, and potential disease-modifying therapies.
Dr. Slatkowsky-Christensen 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. Slatkowsky-Christensen, Mowinckel, Loge, Kvien.
Acquisition of data. Slatkowsky-Christensen, Loge, Kvien.
Analysis and interpretation of data. Slatkowsky-Christensen, Mowinckel, Loge, Kvien.