To study the impact of 24 months of strength training on the physical function of patients with early rheumatoid arthritis (RA).
To study the impact of 24 months of strength training on the physical function of patients with early rheumatoid arthritis (RA).
Seventy patients were assigned to either the strength training (experimental) group (n = 35) or the control group (n = 35). Patients in the experimental group performed strength training for 24 months, and control patients were instructed to perform range of motion exercises. Maximal strength of the knee extensors, trunk flexors, and extensors, as well as grip strength were recorded with dynamometers. Disease activity was assessed by the erythrocyte sedimentation rate and Ritchie's articular index, joint damage was determined by the Larsen x-ray index, and functional capacity was assessed using the Valpar 9 test and the Stanford Health Assessment Questionnaire (HAQ). The employment status of each patient was recorded.
In the experimental group, strength training led to significant increases (19–59%) in maximal strength of the trained muscles. Such increases in the control group varied from 1% to 31%. There was a clear training effect on muscular strength in favor of the experimental group, but significant improvements in the HAQ indices as well as in the Valpar 9 test were seen also in control patients. Results of the Valpar 9 and the HAQ were statistically significantly better in patients who remained gainfully employed compared with patients who retired preterm during followup. However, compared with patients who remained in the work force, patients who retired were older, and their work was physically more demanding.
As expected, strength training led to increased muscle strength, but this increase did not correlate with improved physical function as assessed by the Valpar 9 work sample test. The increased muscle performance did not prevent a substantial proportion of patients from retiring preterm. The 2 items from the Valpar 9 test that were applied were not sensitive enough to differentiate the patients according to their working status.
Rheumatoid arthritis (RA) is a severe, progressive disease with an unpredictable course and outcome (1, 2). Although the majority of the RA population is functionally independent, a substantial decline in functional capacity is often observed over time, and health professionals are increasingly asked to make judgments regarding the degree of disability of the patients (3, 4). The functional capacity of the musculoskeletal system can be assessed by various physical measures, including grip strength, walking speed, stair climbing, and the button test (5–7). More often, self-report questionnaires such as the Stanford Health Assessment Questionnaire (HAQ) (8) and the Arthritis Impact Measurement Scales (AIMS) (9) have been applied.
Functional limitations in RA are associated with inflammation and subsequent pain and fatigue (10, 11), muscle weakness and atrophy (12, 13), and damage in joint structures resulting in limitations of joint mobility (14). Because the impairment often is associated with pain, it is sufficient to limit an individual's capacity to perform the tasks required for his or her employment. Conversely, Yelin et al in 1986 and 1987 (15, 16) showed that work-related disability in arthritis is often a function of the job itself. A job with fewer physical requirements and a higher degree of worker autonomy allows a person with arthritis to remain gainfully employed. In addition to limitations in paid work, RA may lead to limitations to satisfy one's personal needs as well as pursuits at home (17).
Patients with RA have low physical performance capacity, as shown in tests measuring aerobic capacity, muscle strength, and endurance or range of motion (ROM) of individual joints (13, 18, 19). In contrast, aerobic and dynamic exercises have shown positive influence on the physical function of patients (20–22). The factors contributing to work disability in RA have been studied extensively (3). However, little is known about the contribution of resistance strength training to the working capacity of RA patients.
The Valpar 9 Whole Body Range of Motion Work Sample test (Valpar 9) is used by occupational therapists to assess the working and home-making capacity of patients. Valpar 9 simulates light work and assesses whole-body ROM, agility, and stamina in the whole body, as well as movements of the trunk, arms, hands, and overhead reaching (23). This type of testing is functional and may give more multifaceted information than the testing of a single muscle group or ROM of a joint. Although the test is quite widely used by occupational therapists in the evaluation of working capacity of patients with musculoskeletal diseases, the search in the MEDLINE database did not present any earlier research on this topic. Thus, the purpose of this study was to examine the effects of 24-month strength training especially on the results obtained by the Valpar 9 test, and more generally on working capacity in patients with early RA.
Jyväskylä Central Hospital is the only rheumatology center in the Central Finland District, with a population of 263,000. All patients with new RA who live in the area are referred to the center for diagnostic and therapeutic purposes. Seventy patients with recent-onset RA according to the American College of Rheumatology (formerly, the American Rheumatism Association) 1987 criteria for RA (24) volunteered for the study. Originally, 35 patients were randomly allocated to the experimental group or the control group. Randomization was performed using clusters of 4 patients who had been stratified according to age (<50 years and ≥50 years) and sex, to ensure that the demographic data of the study groups remained comparable. Two patients from the experimental group and 3 from the control group withdrew after the baseline measurements were performed (2 discontinued the exercise, 1 became ill with cancer, 1 drowned, and 1 was involved in an accident that resulted in neurologic symptoms). Moreover, the diagnosis of 3 patients changed (spondylarthrosis, psoriatic arthritis, and long-standing RA), and they were also excluded from the analysis. At baseline, there were no significant differences in the demographic, strength, or clinical variables between the patients who completed the trial and those who withdrew (data not shown). Data of the other 62 patients include that of 1 patient in the experimental group who withdrew after the first year due to lack of motivation for the training (Tables 1–3).
|Variable||Experimental group||Control group|
|Sex, no. male/no. female||13/18||11/20|
|Age, mean ± SD years||49 ± 10||49 ± 11|
|Weight, mean ± SD kg||74 ± 14||72 ± 11|
|Height, mean ± SD cm||169 ± 8||167 ± 9|
|Duration of symptoms, mean ± months||10 ± 10||8 ± 12|
|Employment status at baseline|
|Employment status at posttest|
|Physical loading of the work, mean ± SD score (scale 0–7)|
|At baseline||4.0 ± 1.7||4.3 ± 2.2|
|At posttest||4.3 ± 1.4||3.8 ± 1.7|
|Parameter||Experimental group||Control group||Difference between groups (95% CI)|
|Baseline||24.4 ± 17.8||24.8 ± 15.7||−0.4 (−8.8, 8.2)|
|12 months||9.5 ± 7.5||17.3 ± 16.1||−7.8 (−14.1, −1.4)|
|24 months||10.9 ± 9.8||15.4 ± 11.5||−4.5 (−10.0, 0.9)|
|Baseline||11.8 ± 8.5||16.7 ± 9.5||−4.9 (−9.5, −0.3)|
|12 months||2.6 ± 4.6||3.9 ± 4.2||−1.3 (−3.7, 0.8)|
|24 months||2.2 ± 3.1||3.0 ± 4.7||−0.8 (−2.8, 1.2)|
|Larsen score (0–100 scale)|
|Baseline||0.9 ± 1.8||1.2 ± 2.9||−0.3 (−1.4, 0.6)|
|12 months||1.4 ± 2.9||2.3 ± 2.7||−0.9 (−1.6, 0.7)|
|24 mo||1.5 ± 3.4||3.1 ± 3.5||−1.6 (−3.1, 0.4)|
|Pain (0–100 mm scale)|
|0 months||41.7 ± 19.5||41.3 ± 27.1||0.4 (−11.6, 12.4)|
|12 months||21.1 ± 20.6||24.2 ± 22.7||−3.1 (−14.3, 8.1)|
|24 months||13.7 ± 16.2||24.9 ± 22.8||−11.2 (−21.4, −1.0)|
|Parameter||Baseline||Change from baseline to month 24||P value between difference in change†|
|Experimental group Mean ± SD||Control group Mean ± SD||Experimental group Mean (95% CI)||Control group Mean (95% CI)|
|Work sample test|
|Transfer I (s)||383 ± 77||411 ± 78||−10 (−28, 8)||−20 (−35, −6)||0.68|
|Pain during Transfer 1, mm||23 ± 28||33 ± 32||−9 (1, 23)||−9 (−1, 21)||0.24|
|Fatigue during Transfer 1, mm||26 ± 25||31 ± 25||2 (−11, 1)||−12 (10, 14)||0.18|
|Transfer II (s)||537 ± 109||552 ± 118||−42 (−70, −14)||−26 (−51, −2)||0.35|
|Pain during Transfer 1, mm||34 ± 30||46 ± 37||−4 (−9, 13)||−15 (−1, 27)||0.17|
|Fatigue during Transfer 1, mm||28 ± 24||41 ± 35||−1 (−11, 9)||−11 (−3, 25)||0.34|
|HAQ (range 0–3)||0.56 ± 0.48||0.76 ± 0.55||−0.43 (−0.61, −0.26)||−0.41 (−0.59, −0.24)||0.068|
|Muscle strength, kg|
|Trunk extension||55 ± 21||54 ± 18||8 (4, 12)||−1 (−5, 3)||<0.001|
|Trunk flexion||42 ± 14||37 ± 11||9 (6, 11)||6 (1, 10)||0.12|
|Knee extension||67 ± 30||53 ± 26||33 (26, 40)||15 (9, 20)||<0.001|
|Grip||56 ± 31||50 ± 22||18 (11, 24)||9 (3, 15)||0.012|
At baseline, there were no salient differences in physical characteristics between groups (Table 1). Therapy with disease-modifying antirheumatic drugs (DMARDs) was instituted in all patients after the initial measurements were performed. During the 24-month trial, 15 patients in the experimental group and 19 in the control group had to change their initial DMARD due to inefficacy and/or adverse events. Three patients in the experimental group and 12 control patients were treated with low-dose perioral glucocorticoids periodically during the last 24 months (2.5–7.5 mg of prednisolone daily).
Patients in the dynamic strength training group were personally instructed to perform a strength training program for 24 months at home. Loading of the strength training program was individually designed according to the present capacity of each patient. A physiotherapist with lengthy experience guided the patients during their 5-day inpatient period. Strength training included exercises for all main muscle groups of the body, using elastic bands and dumbbells as resistance. Subjects were programmed to exercise twice weekly with moderate loads of 50–70% of the repetition maximum, 2 sets per exercise, 8–12 repetitions per set. The intensity of the strength training was reevaluated (according to the strength measurements) every six months, during visits to the clinic. In addition, patients were encouraged to engage in recreational physical activities, such as walking, cycling, skiing, and swimming, an average of 2–3 times per week. Patients in the control group were instructed to perform ROM and stretching exercises twice weekly, without any additional resistance, in order to maintain their joint mobility. They were free to continue their recreational physical activities, with the exception of strength training of any kind. All patients completed training diaries during the 2-year followup period. Diaries were mailed to the investigators every second month for evaluation.
Maximal unilateral concentric strength of the knee extensors was measured using the David 200 dynamometer (25), and isometric grip strength was measured by a Digitest dynamometer (26). For the results, the sum of the right and left side of knee extension and grip strength were used. The maximal isometric force of the trunk flexors and extensors was measured using an isometric strain-gauge dynamometer (27).
The Valpar Whole Body Range of Motion Work Sample (Valpar 9) was used to measure a patient's capacity to handle various-sized objects while standing, stooping, crouching, and reaching overhead (28). In the present study, 2 of the 4 transfers of the Valpar 9 test were used. Transfer 1 was performed moving 3 different objects from the eye-level panel to the overhead panel (Figure 1), activating mostly the upper extremities. Transfer 2 was performed from the overhead panel to the knee-level panel, with vision occluded while stooping (Figure 2), which demands more stamina and flexibility of the lower extremities. These 2 transfers require working in extreme positions. The height of the test panel was adjusted individually for each patient. Pain and fatigue experienced during the transfers were measured using a 100-mm visual analog scale (VAS). The time in seconds used to move objects from one panel to another was registered. In addition, the percentage reference values of the manual were applied: the relative result >112.5% was graded to exceed the normal working capacity (87–112.5%) to meet the normal working capacity and <87% below the normal working capacity (23). The Stanford Health Assessment Questionnaire (HAQ) was used to assess subjectively perceived functional capacity (8).
Physical loading of work was evaluated with a self-administered questionnaire. The questionnaire included a 7-point scale accompanied by more detailed illustrations and descriptions of various types of work corresponding to each scale point. The scale ranged from 1 (not at work) to 7 (very heavy manual work) (29). Each patient's employment status over the 24-month study period was recorded.
Disease activity was measured using the erythrocyte sedimentation rate (ESR) and Ritchie's articular index (30). Radiographs of the hands and feet were obtained at the time of diagnosis and at 12 and 24 months thereafter. The Larsen score (0–100) was applied to grade the structural damage of the first through fifth metacarpophalangeal joints, the wrist, and the second through fifth metatarsophalangeal joints (31). Pain was assessed using a 100-mm VAS (32).
Results are expressed as the mean and standard deviation (SD) or the mean with 95% confidence intervals. Statistical evaluation between the groups was performed by analysis of covariance (Pillai's trace criterion), using baseline values and age as covariates. The crude Spearman's correlation coefficient was used to determine the relationships between the variables. The alpha level was set at 0.05 for all tests.
The local ethics committee approved this study, and participants gave written informed consent.
At baseline, 28 patients in the experimental group were engaged in paid employment, 2 were retired, and 1 was unemployed. Corresponding numbers in the control group were 24, 6, and 1 (Table 1). At the end of followup, 18 patients in the experimental group and 16 in the control group were gainfully employed. The mean physical loading of the work of those patients able to continue paid work remained constant. During the 24-month study period, clinical parameters improved in favor of the experimental group, but pain was the only variable that reached a statistically significant difference between groups (Table 2).
In the experimental group, the reported compliance with the exercise program averaged 1.5 times per week during the first 12 months and 1.4 times per week during months 13–24, instead of the planned 2 times per week. The respective mean ± SD times used for various types of physical exercises (including strength training in the experimental group) during the first year were 240 ± 124 minutes per week for the experimental group and 205 ± 103 minutes per week for the control group. During the second year, the corresponding figures were 249 ± 121 and 187 ± 107 minutes per week. Three male patients in the experimental group started to exercise in the gym instead of performing in-home exercises with the rubber bands.
During the 24-month followup period, statistically significant improvements in the results of the Valpar 9 (except transfer 1 in the experimental group) and the HAQ were observed in both groups (Table 3). At baseline, the mean ± SD performance scores in the experimental group were 94% ± 22% for transfer 1 and 83% ± 15% for transfer 2 compared with the reference values of healthy persons. The corresponding percentages for the control group were 80% ± 19% and 88% ± 19%. During the 24-month followup period, improvements in transfers 1 and 2 were 8% ± 24% and 7% ± 13%, and 5% ± 10% and 9% ± 11% in the experimental and control groups, respectively. The intergroup differences were not statistically significant. Changes in pain and fatigue experienced during the tests were not statistically significant in either group.
However, when comparing the Valpar 9 test results of patients who were still gainfully employed (workers) at 24 months (n = 34) and those who retired during followup (n = 16), the respective mean ± SD scores were 87% ± 19% and 71% ± 8% in transfer 1 at baseline and 99% ± 23% and 87% ± 19% in transfer 2 (P < 0.001 for both transfers). At month 24, transfer 1 and transfer 2 scores of 93% ± 22% and 104% ± 22%, respectively, for the workers remained significantly better than the scores of 73% ± 12% and 84% ± 9% for patients who retired (P < 0.001 for both transfers).
At baseline, the 16 patients who retired were statistically significantly older compared with workers (56.3 ± 4.6 years versus 43.8 ± 9.1 years; P < 0.001). The mean physical loading scores for the job were 3.9 ± 2.0 and 4.4 ± 1.7 (P not significant) for workers and patients who retired, respectively. When changes in the Valpar 9 test results between workers and retired patients were adjusted for age, the differences between groups still existed (P < 0.049 for transfer 1 and P < 0.033 for transfer 2). In comparison, the baseline HAQ indices were 0.65 ± 0.51 and 0.73 ± 0.56 for workers and patients who retired, respectively. During the 24-month followup period, the HAQ score improved in favor of workers, with respective indices of 0.15 ± 0.34 and 0.39 ± 0.40) (P < 0.020). In contrast, the values in muscle strength, ESR, and Ritchie's index did not differ between groups with regard to employment status.
Improvements in the values for muscle strength favored the experimental group, although for trunk flexion the difference between groups did not reach statistical significance (Table 3). When all patients were evaluated as a group, the changes in transfer 1 correlated with the changes in trunk extension (r = −0.47, 95% confidence interval [CI] −0.66, −0.21) and flexion strengths (r = −0.59, 95 % CI −0.74, −0.36). Also, the combined HAQ index correlated with transfer 1 as well as with transfer 2, both at baseline (r = 0.41, 95% CI 0.17, 0.60 and r = 0.38, 95% CI 0.14, 0.58) and after followup (r = 0.37, 95% CI 0.12, 0.58 and r = 0.47, 95% CI 0.23, 0.66).
Strength training, performed in the home for 24 months, led to significant increases (19–59%) in maximal strength of the trained muscles. Corresponding changes in the control group varied from 1% to 31%. Both groups also demonstrated significant improvements in physical function, as assessed by either the Valpar 9 test or the HAQ index. In contrast, although muscle strength is an important part of the endurance type of activity, the Valpar 9 test was unable to detect improvements in muscle strength. Further, although Valpar 9 test results were, both at baseline and after followup, statistically significantly lower in patients who retired during the 24-month trial compared with those who remained in the work force, the test cannot be used to predict preterm retirement—information that is indispensable for preventive purposes.
Against our expectations, the increases in muscle strength values did not reflect improvements in Valpar 9 test performance or disease activity parameters. Only increases in abdominal and back muscle strength correlated with transfer 1, performed from eye-level panel to overhead panel. Stronger trunk muscles apparently are an advantage for postural control during this phase of the test (which takes 6–7 minutes), which also places demands on muscle endurance. Both parts of the applied Valpar 9 test also require fine-motor skill of the fingers, while other parts of the body have primarily a stabilizing role. In contrast, a patient's ability to perform the tests did not correlate with improvements in maximal grip strength. Obviously, in addition to hand movement, the test tasks require accuracy in other key components of upper extremity function as well as in eye–head and eye–hand coordination. The learning effect as well as motivational factors of the patients must be considered when interpreting the results of the repeated test. Thus, it seems apparent that the results obtained by the Valpar 9 test poorly reflect changes in muscle strength, one of the most important components of physical performance.
Another unexpected finding was that although patients could improve their performance both in muscle strength and Valpar work sample tests, 8 of 28 patients in the experimental group and 8 of 24 patients in the control group retired preterm during the trial. Nevertheless, the mean physical loading of the work of those who continued at paid employment remained constant. When these workers (n = 32) were examined as a group, regardless of their training background, they obtained significantly better results in the Valpar 9 tests compared with patients who retired preterm. Among workers, the performance percentages of 93% and 104% in transfers 1 and 2, respectively, met the limits of normal working capacity in the reference population (87–112%) given in the manual. Among retired patients, the corresponding values (73% and 84%) mirrored satisfactory working capacity. Thus, it seems that the applied Valpar 9 test could be used to determine which functional skills of patients are required to be able to continue at paid employment. However, although we do not have reference values according to age, the older age of the patients who retired preterm may have contributed to their poorer Valpar 9 test results.
The Valpar 9 test has some limitations that should be taken into consideration when using and interpreting the test results. First, the Valpar 9 test does not assess the quality of the task performed, as subjects are free to choose the way to perform the test. Second, the test does not reveal the reason for decreased functional capacity. Finally, use of Valpar work samples assumes that patients engaged in different types of physical work should be individually tested using 1 of the 19 standardized tests available. This, however, is a complicated and unrealistic requirement in a clinical setting. Our occupational therapist spent almost an hour testing one patient, even though only 2 of the 4 transfers of the Valpar 9 test chosen for this study were applied. Furthermore, at baseline, as well as at the outcome, both transfer 1 and transfer 2 results correlated with those obtained by the much more easily instrumented patient self-report, the HAQ. Thus, the 2 items from the Valpar 9 that were used were not useful predictors of work status in the patient sample studied, and the additional information provided by the test was limited. In contrast, the degree of functional disability of the patients was rather low, and further studies are needed to obtain information regarding the suitability of the test for assessing patients with more severe disability.
The results confirmed that the working capacity of patients with RA is at risk from the very start of the disease, as evidenced by the fact that 31% of our patients retired preterm during the 2-year followup period. This result is consistent with that of other prospective studies concerning work disability in patients with early RA (33–36). Although our results, due to the small sample size, should be interpreted with caution, older age and a physically more demanding job obviously predict preterm retirement. Nevertheless, work disability is a complex process in which the circumstances and qualities of the work as well as the characteristics of the patient or the disease process may variably play a major role. Finally, all of our patients were actively treated with DMARDs, and a proportion of the patients received low-dose glucocorticoids as well. Most probably, the applied drug treatment policy contributed little to improvements in muscle strength, and its accurate role in the observed functional improvement of the patients remain uncertain.
In conclusion, the results of the present study indicate that strength training for 24 months resulted in salient improvement in muscle strength. However, muscle strength is only 1 component of physical function, and improvement in muscle strength poorly reflected the results obtained by the Valpar 9 test and a patient's ability to remain gainfully employed. Furthermore, regardless of working status, the 2 items applied from the Valpar 9 work sample test did not distinguish between the physical functional skills of patients with a rather low degree of disability better than the more simple HAQ. Thus, we do not recommend use of the Valpar 9 test as a routine tool in clinical practice.