Physical activity, physical fitness, and general health perception among individuals with rheumatoid arthritis


  • The members of the PARA Study Group are listed in Appendix A.



To describe self-reported physical activity and physical fitness and to identify correlates of physical activity and general health perception.


Data on self-reported physical activity, physical fitness, activity performance, and disease activity were collected from a sample of 298 patients with rheumatoid arthritis (RA).


Forty-seven percent of our sample reported physical activity behaviors that did not comply with public health recommendations. A majority of the patients had decreased lower-limb muscle function (72%), grip force (94%), joint motion (94%), and functional balance (68%). Correlations between self-reported physical activity and other variables were rs = 0.25 or less. Variation in general health perception was explained (total adjusted R2 = 0.65) by pain and activity performance.


Our findings indicate that there is a case for recommendations on and support for healthy physical activity behaviors among people with RA.


Rheumatoid arthritis (RA) is a chronic, inflammatory, and systemic disease often associated with impairments, activity limitations, and social participation restrictions. Moreover, mortality rates are increased (1) and tend to be associated with high levels of inflammation (2), activity limitation (3), and comorbidity mainly from cardiovascular and cerebrovascular diseases (4). RA also seems to be associated with reduced levels of physical activity (5–8).

Physical activity, exercise, and physical fitness are interrelated factors influencing health. Physical activity is defined as “any bodily movement produced by skeletal muscles that results in energy expenditure” (9), whereas exercise is a subset of physical activity defined as “planned, structured, and repetitive bodily movement done to improve or maintain one or more components of physical fitness” (9). In the present article, physical activity includes everyday physical activity as well as planned exercise. Physical inactivity denotes a level of activity less than that needed to maintain good health (10). Physical inactivity is a general health problem in the Western world (10, 11). Epidemiologic and experimental studies indicate that physical activity reduces risk factors for coronary heart disease (12–14) and other physical and psychological health problems (12). It is therefore recommended that each individual accumulates 30 minutes or more of moderate-intensity physical activity, such as brisk walking, on most days of the week (10, 12, 15, 16). This recommendation also applies to individuals with arthritis (17).

Physical fitness is “a set of attributes that people have or achieve that relates to the ability to perform physical activity” (9). One of the health-related fitness components is cardiorespiratory endurance, also known as aerobic fitness (18). In the present article, aerobic fitness is used and quantified as estimated maximum oxygen uptake (VO2max) based on submaximal tests and expressed as milliliters per kilogram body weight. Other health-related fitness components are muscular endurance, muscular strength, body composition, and flexibility (9). Poor aerobic fitness is a significant risk factor for all-cause mortality (19, 20). Reduced aerobic fitness has been found in studies involving individuals with RA (5, 6, 21–23), as have decreased muscular strength (6, 21, 22, 24, 25) and endurance (6), limited flexibility (5, 26), and poor standing balance (27, 28).

Although correlations between physical activity, overall physical fitness, and general health perception have been found in population-based studies (29), our knowledge is still limited on whether this is also true for subgroups with various health conditions, such as arthritis. The need for more research in this area is indicated (30).

The purpose of the present investigation was to describe self-reported physical activity and outcomes of physical fitness tests in patients with RA, to identify correlates of self-reported physical activity and general health perception, and to compare the self reports and test results with norm data.



Patients (n = 484) with a confirmed diagnosis of RA (31), disease duration ≤6.5 years, and included in the Swedish RA register were asked to perform 5 physical fitness tests and complete 2 questionnaires in connection with an ordinary medical evaluation at an out-patient clinic. The Swedish RA register, including almost every rheumatology unit in the country, was started by the Swedish Society for Rheumatology in 1997 for epidemiologic purposes. All individuals diagnosed with RA are systematically reported to a central database as regards disease activity and treatment. All 42 connected rheumatology units all over the country were invited to participate in the present descriptive study of physical activity and physical fitness. Seventeen of them agreed.

Physical therapists at each unit recruited patients from the physicians' waiting lists during 2 years, February 1999 to January 2001. All patients received oral and written information about the study and informed consent forms were signed. Two ways of recruitment, depending on practical circumstances at each participating unit, were possible. Either each patient included in the national RA register and scheduled for a physician visit was asked to participate, or a certain number of these patients were asked. In the latter case, random selection was sought.

Two hundred ninety-eight (62%) of the 484 patients with median age 57 years (range 19–90 years) agreed to participate. Two hundred twenty-five (76%) were women with median age 55 years (range 19–90 years) and 73 (24%) were men with median age 63 years (range 27–84 years). All were able to perform at least 3 of 5 physical fitness tests (listed below) and to complete 1 of 2 questionnaires (one on physical activity described below and one that will be reported elsewhere).

There were several reasons for nonparticipation among the other 186 patients, 136 (73%) women and 50 (27%) men with median age 63 years (range 19–89 years). Ninety-three individuals did not answer the invitation, said they did not want to participate, or just did not appear at a scheduled visit. Fifty-eight were not included due to problems with logistics, mainly that physician visits were scheduled or rescheduled without notifying the physical therapist. Another 35 patients were generally willing to participate but were unable to perform the fitness testing or answer the questionnaires in accordance with participation requirements. The reasons were physical (acute infections, comorbidity, recent surgery, or high disease activity), psychological (confusion, dementia), or a combination.



The questionnaire outlined below was used specifically for the purpose of the present study. A written self-reported measure of physical activity, which was part of a larger questionnaire previously used in a Swedish epidemiologic survey on physical activity and health habits in the general population, was used (29). It includes 8 questions (total score 0–32): 3 questions on the frequency of low-, moderate-, and high-intensity exercise for at least 20 minutes during a typical week, with 5 alternative answers from “never” to “≥3 times per week” (0–20); 4 questions on the duration of everyday physical activity related to seasonal variations, with 3 alternative answers from “less than half an hour per day” to “≥3 hours per day” (0–6); and 1 question about the intensity of daily work during a typical week, with 4 alternative answers from “not hard at all (mostly sitting all day)” to “very hard work (heavy manual work)” (0–6). Sum scores ≥12.5 indicate the amount of physical activity recommended to maintain good health. Each individual can also be classified into 1 of 5 levels of physical activity: very low = 0–3.5, low = 4–9.5, average = 10–14.5, high = 15–19.5, and very high = 20–32. Very low equals <1 hour of everyday physical activity and physical inactivity for the rest of the day. Very high equals high-intensity exercise at least 3 times a week plus 3 hours of physical activity per day plus hard daily work (29). Norm data from a representative sample of the Swedish population, age 20–65 years, is available and was used for comparison with the present patients age 20–65 years (158 women and 38 men). In all other parts of our study, physical activity levels are reported for the entire sample, including another 55 women and 28 men age 66–90 years.

Physical fitness.

The following tests of physical fitness were specifically used for the purpose of the present study.

Submaximal aerobic fitness tests were performed on the treadmill or bicycle ergometer to estimate VO2max. A prediction equation including age, sex, self-selected walking speed (km/hour), and work heart rate was used to calculate the outcome of the treadmill test for each participant in ml/kg/minute (31, 32). Age, sex, body weight, work load (watts), and work heart rate were used to calculate the outcome for the bicycle test for each participant in ml/kg/minute (33). The prediction equation for the bicycle ergometer test was limited to individuals <75 years of age and, thus, older individuals had therefore to be considered as internal dropouts.

Aerobic fitness classification based on VO2max values in ml/kg/minute obtained from either of the 2 submaximal tests mentioned above was used for comparing aerobic fitness across individuals independent of sex or age. It was based on 4 age groups between 20 and 65 years for women and 5 age groups between 20 and 69 years for men. Thus, age and sex were taken into account when assigning each individual to 1 of 5 aerobic fitness classes: low = 1, fair = 2, average = 3, good = 4, and high = 5 (33). The aerobic fitness classification was originally derived from ergometer bicycle tests of 76 women (students and housewives) and 129 men (physically active and draymen) (33). The classification system did not include women >65 years or men >69 years of age. Thus, 36 women older than 65 years were assigned to the 50–65 year age group and 14 men older than 69 years were assigned to the 60–69 year age group.

The timed-stands test was used to assess lower-limb muscle function (34). The time needed to rise 10 times from a standard chair is recorded and norm data are available for men and women age 20–89 years (35). One woman younger than 20 years was assigned to the 20–24 age group and 1 woman older than 89 years was assigned to the 85–89 age group.

The Grippit was used to measure grip force peak in Newtons and average values of sustained force during 10 seconds (N). Norm data are available for men and women age 20–69 years (36). Because the norm data do not include women and men younger than 20 or older than 69 years, 1 woman younger than 20 years was assigned to the 20–29 age group and 46 women and 26 men older than 69 years were assigned to the 60–69 age group in the present study.

The Escola Paulista de Medicina-Range of Motion scale was used to estimate general range of motion (0–30; 0 = motions with no restrictions), based on 10 bilateral motions in 7 joints measured with a goniometer (37).

Functional balance was determined while walking in a figure-of-eight with inner circles of 1.5 m diameter and outer circles of 1.8 m diameter. Subjects walked with shoes on, 2 circuits along the 150-mm wide track. The touches and number of oversteps were counted (28). More than 2 oversteps or touches were considered as decreased functional balance.

Disease activity.

The following measures from the minimum core set of disease activity recommended by the European League Against Rheumatism (38) were available from the Swedish RA register.

Self-reported pain rated on a visual analog scale (0–100, 0 = no pain). The Health Assessment Questionnaire (HAQ) functional disability index was used to assess the patients' estimated activity performance (0–3, 0 = performance without difficulty) the previous week (39, 40).

Disease activity measures included C-reactive protein (mg/liter) and Disease Activity Score in 28 joints (DAS28) based on a calculation of erythrocyte sedimentation rate (mm/hour), number of swollen and tender joints (28-joint index), and self-reported general health perception scored on visual analog scale (0–100, 0 = totally fine) (41). The DAS28 is scored 0–10 with scores <3.3 indicating low and those >5.1 high disease activity.


The physical fitness tests and questionnaires were administered by a physical therapist no more than 1 week prior to or after a physician visit, during which data on disease activity, activity performance, and general health perception were collected for the Swedish RA register. These 2 data sets were then matched. The patient sample may be regarded as one of convenience because recruitment was influenced by practical circumstances, such as the physical therapists' working hours, the patients' availability, and other logistics at the participating clinics. Similarly, the selection of aerobic fitness test method for each patient was not randomized, but rather determined by practical circumstances, mainly related to actual access to equipment at the participating units. At some units, with possibilities for both bicycle and treadmill testing, the conditions and desires of the patients might sometimes have determined the choice of test method. Thus, the presence of systematic errors in this part of the data collection cannot be excluded entirely.


Descriptive results are presented as medians and ranges. Mann-Whitney's U test was used to analyze sex differences in self-reported physical activity and general health perception. Correlations between these 2 variables and other variables were analyzed with Spearman's rank order correlation coefficients. Coefficients ≤0.25 were considered very low or nonexistent (42) and omitted from further analyses. Multiple regression models were used to explain the total variation in general health perception.


The Regional Ethics Research Committee at Karolinska Institutet approved the design of the study. Informed consent was obtained from each study participant.


The median sum score for self-reported physical activity in our total sample was 12.5 (range 2–30). One hundred forty-seven patients had scores ≥12.5, indicating that they reached the physical activity levels recommended to maintain good health. One hundred thirty-two patients had scores <12.5 and 19 did not answer the questionnaire completely enough to determine their scores. Individuals age ≤65 years scored somewhat higher (median 13, range 2–30) than the older ones (median 11, range 3–25.5). The proportions of men and women scoring ≥12.5 were 56% and 52% in the total sample, 58% and 59% among those age ≤65 years, and 58% and 32% among those >65 years. Most patients were categorized as performing average physical activity and the self reports were fairly normally distributed over the 5 levels of physical activity (Figure 1). Both women and men with RA age 20–65 years estimated their levels of physical activity somewhat high compared with norm data for the same age groups from the representative sample of the Swedish population (Figure 2).

Figure 1.

Self-reported levels of physical activity in 213 women and 66 men with rheumatoid arthritis.

Figure 2.

Self-reported levels of physical activity in A, 158 women, and B, 38 men with rheumatoid arthritis (RA), age 20–65 years, compared with norm data.

The vast majority of the patients were assigned to the low, fair, and average aerobic fitness classes. Generally, men were mainly assigned to the average aerobic fitness class and women to the low or fair classes. Very few individuals were assigned to the good or high classes (Figure 3). Aerobic fitness (VO2max) in patients with RA age 20–65 years was similar to norm data for the same age groups from a representative sample of the Swedish population (Figure 4).

Figure 3.

Aerobic fitness classification in 184 women and 53 men with rheumatoid arthritis.

Figure 4.

Mean value of aerobic fitness (VO2max) in A,147 women, and B, 35 men with rheumatoid arthritis (RA), age 20–65 years, compared with norm data (640 women and 557 men).

Descriptive data for lower-limb muscle function, peak and average grip force, range of motion, and functional balance are shown in Table 1. A majority of the patients had decreased physical fitness compared with norm data (Table 2). Descriptive data indicated mainly low-to-moderate disease activity and moderate impact on general health perception, pain, and activity limitations. Variations were generally wide between individuals (Table 3).

Table 1. Results of physical fitness testing of 298 individuals with RA*
 All participants (n = 298)Female (n = 225)Male (n = 73)
  • *

    Data presented as median (range). RA = rheumatoid arthritis; EPM-ROM = Escola Paulista de Medicina-Range of Motion.

Timed-stands test, seconds25 (11–86)25 (11–86)25 (11–72)
Grippit; grip force peak, Newtons284 (20–1,108)228 (20–683)500 (55–1,108)
Grippit; average value, Newtons222 (27–1,184)193 (27–668)417 (50–1,184)
EPM-ROM scale, 0–305 (0–19)5 (0–18)6 (0–19)
Figure-of-eight; oversteps, no.5 (0–102)4 (0–102)6 (0–68)
Table 2. Results of physical fitness testing of 298 people with rheumatoid arthritis (225 women and 73 men)*
 Normal n (%)Decreased n (%)Inability n (%)Missing n (%)
  • *

    Results classified into normal capacity for age and sex, decreased capacity, or inability to perform the tests.

Lower-limb muscle function
 All83 (28)200 (67)15 (5)0 (0)
 Female67 (30)145 (64)13 (6)0 (0)
 Male16 (22)55 (75)2 (3)0 (0)
Grip force peak
 All13 (4)274 (92)5 (2)6 (2)
 Female8 (4)207 (92)5 (2)5 (2)
 Male5 (7)67 (92)0 (0)1 (1)
Average grip force
 All17 (6)270 (91)5 (2)6 (2)
 Female12 (5)203 (90)5 (2)5 (2)
 Male5 (7)67 (92)0 (0)1 (1)
General range of motion
 All18 (6)279 (94)0 (0)1 (3)
 Female16 (7)208 (93)0 (0)1 (4)
 Male2 (3)71 (97)0 (0)0 (0)
Functional balance
 All97 (33)193 (65)8 (3)0 (0)
 Female82 (36)137 (61)6 (3)0 (0)
 Male15 (21)56 (77)2 (3)0 (0)
Table 3. General health perception, pain, activity limitation, and disease activity for 298 individuals with RA*
 All (n = 298)Female (n = 225)Male (n = 73)
  • *

    Data presented as median (range). RA = rheumatoid arthritis; HAQ = Health Assessment Questionnaire functional disability index; CRP = C-reactive protein; DAS28 = Disease Activity Score in 28 joints; ESR = erythrocyte sedimentation rate.

General health perception, 0–10030 (0–96)37 (0–96)22 (0–83)
Pain, 0–10032 (0–96)37 (0–96)20 (0–85)
HAQ, 0–30.63 (0–2)0.75 (0–2)0.30 (0–2)
CRP, mg/liter10 (0–124)10 (0–110)10 (0–124)
DAS28, 0–103.51 (0–7.37)3.61 (0–6.80)2.92 (0–7.37)
ESR, mm/hour13 (0–102)14 (0–102)10 (0–96)
28-joint index, 0–28   
 Swelling, no.2 (0–27)3 (0–17)2 (0–27)
 Tenderness, no.2 (0–27)2 (0–27)1 (0–27)

Self-reported physical activity did not differ significantly (P > 0.05) between men and women. Correlations between physical activity and physical fitness, demographic factors, general health perception, pain, activity performance, and disease activity were all (although statistically significant) very low (rs < 0.26, P < 0.001–0.05) and therefore omitted from further analyses.

General health perception differed significantly (P < 0.01) between men and women. Several correlations between general health perception and other variables were greater than rs = 0.25: peak grip force (rs = −0.39, P < 0.001), average grip force (rs = −0.39, P < 0.001), pain rating (rs = 0.80, P < 0.001), activity performance measured by the HAQ (rs = 0.63, P < 0.001), number of swollen joints (rs = 0.33, P < 0.001), and number of tender joints (rs = 0.45, P < 0.001). When those 7 variables were included in a multiple regression model, general health perception was explained (total adjusted R2 = 0.65) by pain (adjusted R2 = 0.63, P < 0.001) and activity performance (adjusted R2 = 0.02, P < 0.001). None of the other variables contributed significantly in the model.


Physical activity levels reported in our sample resembled those from a Swedish epidemiologic study (29). Almost half (47%) of our sample failed to meet the recommendations on healthy physical activity behaviors. Thus, our results support previous findings of low physical activity levels among individuals with RA (6, 7, 11). One important finding of our study was that women age >65 years seemed to be less physically active than either men of the same age or younger women. This finding accords with previous results indicating lower rates of both moderate and vigorous leisure-time physical activity among women and a larger decline with age for women than for men (18, 43, 44). One reason for these findings might be that older women, also those with RA, are satisfied with an inactive lifestyle and are not motivated to become more physically active (45), which represents a challenge to society and health professionals.

Also as regards aerobic fitness, our classification of the majority of the patients as having low, fair, or average aerobic fitness is in accordance with the results from a previous Swedish population-based survey (29) and previous descriptions of individuals with RA (5, 6, 21, 22). However, the absence of correlation between physical activity and physical fitness in our study was surprising because it contradicts results of population-based studies (13, 29). The overall low explanation of physical activity levels in our study by traditional demographic and disease activity measures might indicate that other factors are involved. The influence of cognitive–behavioral factors, such as motivation (45), self efficacy (46, 47), and beliefs regarding health benefits of physical activity (18), have previously been found to relate to physical activity levels and need to be included in future studies of physical activity and fitness in RA.

The relationship between general health perception and pain found in our study supports previous findings (48). However, the relatively strong relations previously found for the general population between physical activity/physical fitness and general health (12, 13, 29, 49) were not present in our sample of patients with RA. Possibly, factors related to the disease might be more important for general health perception in this population.

The questionnaire for self-reporting physical activity used in the present study was chosen mainly because it had previously been used in a Swedish population-based survey providing norm data that could be used for comparison with our sample. Its validity and reliability have not been described, but because many experts were involved in the previous national survey, there is reason to believe that it has at least appropriate face and content validity. However, the questionnaire has not previously been used for individuals older than 65 years, and this could be assumed to bias our findings and explain some of the low correlations found in our study. All analyses were also run separately for patients up to age 65 in our sample, with no major differences compared with the results of the entire sample (data not shown). Another issue of concern might be that the questionnaire has not been validated for people with arthritis. However, the questionnaire was developed for the general population, which includes individuals with arthritis, long-standing pain, and other chronic diseases. Nevertheless, it cannot be excluded that the answers from patients with arthritis might be biased either by social desirability or overestimation of their physical activity (50). For example, our patients were urged to take into account only their cardiovascular exertion when filling out the questionnaires, but pain, stiffness, and fatigue might still have influenced their estimation and thus biased the results toward a higher estimated physical activity level. Unfortunately, such possible bias is hard to exclude entirely from any epidemiologic research relying on self reports.

Several issues related to the selection of our sample need to be discussed as regards their influence on the external validity of our results. No systematic bias in the selection of the participating clinics seems likely because they represent university, county, and local hospitals with both urban and rural catchment areas. Furthermore, although reasons may vary among those who declined to participate, no major differences may be suspected between them and the participants. Those who declined could be assumed either to be healthier and thus busier or unhealthier and less willing to do the tests. The median age of those excluded for administrative reasons did not differ from that of the sample studied. However, the 35 patients who were willing to participate but were excluded because they were unable to perform the test protocol were generally older (median age 72 years) and, by definition, less fit. It is thus important to bear in mind that our results cannot be generalized to older patients with RA with severely impaired body functions.

In conclusion, self-reported physical activity behaviors in our sample of patients with RA were similar to those in the general population, with almost 50% failing to meet recommendations on healthy physical activity. Poor physical fitness and poor congruency between self-reported physical activity and physical fitness indicated possible bias in self reports. Our findings indicate that there is a case for recommendations on and support for healthy physical activity among people with RA. This is particularly important because the disease might represent extra barriers to physical activity and because these people risk comorbidity and premature death.



Physical therapists Britt Andersson, Karolinska Hospital, Stockholm; Anneli Bilberg, Sahlgrenska University Hospital, Göteborg; Åsa Björklund, Västerås Hospital, Västerås; Ingrid Carlquist, Uppsala University Hospital, Uppsala; Hedvig Cronstedt, Danderyd Hospital, Danderyd; Ylva Ericsson, Malmö University Hospital, Malmö; Sofia Hagel, Lund University Hospital, Lund; Ann-Marie Holmén-Andersson, Sahlgrenska University Hospital, Mölndal; Eva Liberg, Skövde Hospital, Skövde; Karina Malm, Spenshult Rheumatology Hospital, Spenshult; Eva Matton, Gävle-Sandviken Hospital, Gävle-Sandviken; Helene Nilsson, Örebro University Hospital, Örebro; Anne-Marie Norén, Huddinge University Hospital, Huddinge; Lena Olin, Falun Hospital, Falun; Kerstin Oxelbark, Borås Hospital, Borås; Viveka Pålsson, Trelleborg Hospital, Trelleborg; and Eva Teintang, Linköping University Hospital, Linköping, Sweden.