Are the benefits of a high-intensity progressive resistance training program sustained in rheumatoid arthritis patients? A 3-year followup study




Rheumatoid arthritis (RA) patients were reassessed for body composition and physical function mean ± SD 39 ± 6 months after commencing a randomized controlled trial involving 24 weeks of either high-intensity progressive resistance training (PRT) or low-intensity range of movement exercise (control) to determine whether the benefits of PRT (i.e., reduced fat mass [FM], increased lean mass [LM], and improved function) were retained.


Nine PRT and 9 control subjects were reassessed for body composition (dual x-ray absorptiometry) and function (knee extensor strength, chair test, arm curl test, 50-foot walk) approximately 3 years after resuming normal activity following the exercise intervention.


At followup, PRT subjects remained significantly leaner than control subjects (P = 0.03), who conversely had accumulated considerable FM during the study period (approximately −1.0 kg versus +2.4 kg, PRT versus controls). PRT subjects also retained most of the improvement in walking speed gained from training (P = 0.03 versus controls at followup). In contrast, the PRT-induced gains in LM and strength-related function were completely lost. Data from the controls suggest that established and stable RA patients have similar rates of LM loss but elevated rates of FM accretion relative to age-matched sedentary non-RA controls.


We found that long-term resumption of normal activity resulted in loss of PRT-induced improvements in LM and strength-related function, but substantial retention of the benefits in FM reduction and walking ability. The relatively long-term benefit of reduced adiposity, in particular, is likely to be clinically significant, as obesity is very prevalent among RA patients and is associated with their disability and exacerbated cardiovascular disease risk.


Reduced muscle mass and elevated adiposity, termed “rheumatoid cachexia,” is prevalent in rheumatoid arthritis (RA), and these adverse changes in body composition are strongly associated with RA disability, with Health Assessment Questionnaire scores inversely correlated with appendicular lean mass (ALM; a surrogate measure of muscle mass), and directly and independently related to fat mass (FM) (1). Such associations between body composition and physical function are not surprising, as they are also evident in the general elderly population, wherein classification as either muscle wasted or obese significantly exacerbates the likelihood of disability, while the coincidence of both conditions (“sarcopenic obesity”) increases the risk 12-fold in women and 9-fold in men (2).

In a randomized controlled trial (RCT) (3), we demonstrated that 24 weeks of high-intensity (HI) progressive resistance training (PRT) elicited substantial gains in ALM and losses of total and especially trunk FM, and restored “normal” levels of physical function in established RA patients. Investigation of the long-term retention of strength training benefits in RA patients who resume sedentary lifestyles has not previously been conducted. Therefore, in this 3-year followup study, we aimed to determine whether any of the PRT-induced improvements in body composition and function are retained following prolonged detraining. Additionally, reassessment of the control group provides longitudinal insight into the changes in body composition and physical function occurring in RA patients with well-controlled disease, another important addition to the literature, as longitudinal data on these aspects are lacking.


This was a followup investigation in which RA patients were reassessed mean ± SD 39 ± 6 months after commencing an RCT (3) involving 24 weeks of either supervised HI PRT or home-based, low-intensity range of movement (ROM) exercise (control). Ethical approval was received from the North West Wales NHS Trust Research Ethics Committee.

The 28 established RA patients who participated in the RCT were approached to undertake followup assessments. At followup, body mass, body composition by whole-body dual x-ray absorptiometry (DXA), maximal isometric knee extensor strength (KES), physical function by objective tests (30-second chair stand, 30-second arm curl, 50-foot walk), and disease activity by the Disease Activity Score in 28 joints (DAS28) were measured as previously described (3). At followup, subjects were also asked about their habitual physical activity (PA) at work and during leisure time (each ranked 1–4, i.e., sedentary to heavy, regular PA, and summed to provide a 2–8 aggregate score) (4), sleep patterns, current medication, and whether their usual diet had changed over the study period. To assist patients in these evaluations, responses completed at baseline were referred to.

To determine whether the followup subsamples had experienced similar benefits from PRT as the original group, outcome measures were assessed by multiple 2-way (2 × 2; treatment by time [baseline and postintervention, i.e., 24 weeks]) analyses of variance (ANOVAs). To determine whether these benefits were maintained during followup, further 2-way (2 × 2; treatment by time [baseline and followup]) ANOVAs were performed. Additionally, changes over time within a group were assessed by 1-sample t-tests. Data were analyzed using SPSS, version 14, and are shown as the mean ± SD.


Of the 28 RA patients (PRT, n = 13; controls, n = 15) eligible for recruitment for this followup study, 2 had died (both controls), 6 had moved from the area (PRT, n = 4; controls, n = 2), and 2 declined participation (both controls). The remaining 18 patients were evenly divided between PRT (8 women, 1 man) and control (6 women, 3 men) subjects. There were no differences at baseline between the groups in age, disease duration, DAS28, habitual PA, or average sleep duration, and for both groups, DAS28, habitual PA, and sleep remained similar at followup (Table 1).

Table 1. Characteristics of rheumatoid arthritis patients in the PRT and control groups*
VariablePRT group (n = 9)Control group (n = 9)P
  • *

    Values are the mean ± SD unless otherwise indicated. Followup was mean ± SD 39 ± 6 months after randomization to groups for 24 weeks of either high-intensity PRT or low-intensity range of movement exercise (control). Difference between groups was examined using independent-samples t-test or the interaction P for 2 × 2 (treatment × time) analyses of variance. PRT = progressive resistance training; DAS28 = Disease Activity Score in 28 joints.

Age at baseline, years55.7 ± 8.659.4 ± 10.80.430
Sex, women/men8/16/3
Disease duration at baseline, years6.85 ± 7.347.89 ± 7.590.805
 Baseline3.47 ± 1.413.31 ± 1.19 
 Followup3.35 ± 1.373.17 ± 0.970.969
Habitual physical activity level (range 2–8)   
 Baseline2.75 ± 1.372.72 ± 1.50 
 Followup2.73 ± 1.542.50 ± 1.000.494
Sleep duration, hours   
 Baseline6.78 ± 1.236.44 ± 1.45 
 Followup6.78 ± 1.206.89 ± 1.080.455

During the approximately 3-year followup period, 2 PRT patients had changed medication from disease-modifying antirheumatic drugs to nonsteroidal antiinflammatory drugs, whereas medication remained broadly unchanged in the control group. Five subjects from each group reported changes in habitual PA levels, and 4 from each group reported dietary changes. Typically, these alterations involved reductions in PA and carbohydrate consumption, respectively. When subjects reported an increase in PA (PRT, n = 1; controls, n = 1), it was in the form of low-intensity exercise, e.g., walking. At followup, no subjects were performing HI PRT or any other form of regular HI exercise.

The effects of PRT and subsequent detraining on body weight and body composition are detailed in Table 2, and those on physical function are shown in Table 3. In terms of the benefits of 24-week HI PRT on body composition and function, the gains for the PRT followup cohort were virtually identical to those of the original group (3): training-specific strength improved by 105% (P < 0.0001), ALM increased by 1.21 kg, total FM decreased by 3.22 kg and trunk FM decreased by 3.27 kg, percent body fat (%BF) declined by 3.92% in absolute terms, and KES, chair test, 50-foot walk, and arm curl test performance improved by 23%, 31%, 18%, and 25%, respectively (all P < 0.001; t-tests: baseline versus postintervention). Conversely, as in the original study, the control condition of low-intensity ROM exercise had no effects on body composition or physical function (Tables 2 and 3). For both groups, body weight remained stable during the exercise intervention period.

Table 2. Effects of 24 weeks of high-intensity PRT and subsequent 33-month detraining period on body composition in rheumatoid arthritis patients*
VariablePRT group (n = 9)Control group (n = 9)P
  • *

    Values are the mean ± SD. Followup was mean ± SD 33 ± 6 months after concluding 24 weeks of either high-intensity PRT or low-intensity range of movement exercise (control). P values are for group × time (baseline vs. postintervention; baseline vs. followup) interactions. PRT = progressive resistance training.

  • Significant.

Body mass, kg   
 Baseline67.83 ± 15.1971.29 ± 10.24 
 Postintervention67.39 ± 13.3671.45 ± 10.350.790
 Followup68.08 ± 15.0674.42 ± 10.420.185
Appendicular lean mass, kg   
 Baseline14.23 ± 2.1015.24 ± 4.49 
 Postintervention15.44 ± 2.2415.21 ± 4.830.030
 Followup14.15 ± 1.9214.93 ± 4.610.738
Total fat mass, kg   
 Baseline28.18 ± 14.3527.92 ± 6.71 
 Postintervention24.96 ± 12.7928.11 ± 8.560.045
 Followup27.20 ± 14.4930.34 ± 6.400.032
Percent body fat, %   
 Baseline39.76 ± 11.3439.46 ± 8.90 
 Postintervention35.84 ± 11.8139.97 ± 11.020.035
 Followup38.82 ± 11.2841.95 ± 9.220.030
Trunk fat mass, kg   
 Baseline13.94 ± 7.7513.73 ± 5.30 
 Postintervention10.66 ± 6.0214.22 ± 5.800.036
 Followup13.40 ± 7.8116.44 ± 4.800.030
Table 3. Effects of 24 weeks of high-intensity PRT and subsequent 33-month detraining period on objective physical function in rheumatoid arthritis patients*
VariablePRT group (n = 9)Control group (n = 9)P
  • *

    Values are the mean ± SD. Followup was mean ± SD 33 ± 6 months after concluding 24 weeks of either high-intensity PRT or low-intensity range of movement exercise (control). P values are for group × time (baseline vs. postintervention; baseline vs. followup) interactions. PRT = progressive resistance training; KES = maximal voluntary isometric knee extensor strength.

  • Significant.

KES, newtons   
 Baseline345 ± 78306 ± 163 
 Postintervention425 ± 95329 ± 1580.019
 Followup339 ± 71283 ± 1500.240
30-second chair stand test, reps   
 Baseline12.44 ± 4.6712.56 ± 2.74 
 Postintervention16.33 ± 4.4711.78 ± 4.660.001
 Followup12.33 ± 2.9612.00 ± 3.160.625
30-second arm curl test, reps   
 Baseline16.11 ± 5.6214.11 ± 4.65 
 Postintervention20.11 ± 4.8114.89 ± 4.010.024
 Followup15.67 ± 4.7713.56 ± 3.710.952
50-foot walk, seconds   
 Baseline9.68 ± 2.778.80 ± 2.96 
 Postintervention7.90 ± 1.388.97 ± 3.820.011
 Followup8.50 ± 1.779.06 ± 3.510.033

At followup, relative to baseline measures 39 months earlier, body weight was unchanged in the PRT group. However, this stability masked losses of LM and gains in FM since PRT cessation. In fact, during this detraining period, ALM for PRT subjects had regressed to baseline levels, while much of the reduction in total and trunk FM and %BF had been regained. Therefore, the beneficial effects of PRT on body composition were largely lost. The situation, however, was worse for the control group, as during followup they accumulated substantial FM, i.e., relative to baseline they gained 2.4 kg in total FM, 2.7 kg in trunk fat, and 2.4% (absolute) in %BF. These increases, despite concurrent gains by the PRT subjects, led to between-group differences for all measures of adiposity remaining significant at followup. In contrast, since ALM remained unchanged for the control subjects during the 39-month investigation period, and had reverted to baseline levels for the PRT subjects, there were no differences between the groups at followup for ALM.

In keeping with the maintenance of LM at followup relative to baseline, levels of objectively measured physical function did not significantly decline for either group during the followup period, although performance for all tests attenuated 3–7% for the controls. For the PRT group, whereas the training-induced gains in KES and the chair and arm curl tests were lost during detraining, most of the improvement in the 50-foot walk was retained. This retention of 66% of the improvement in walking speed subsequent to PRT meant that at followup the PRT group continued to perform this test significantly better than the controls.


The main findings of this study are that RA patients who completed 24 weeks of HI PRT approximately 3 years previously maintained significant benefit in adiposity measures and walk test performance, relative to both pretraining levels and RA patients who had previously performed low-intensity ROM exercise. In contrast, improvements in lean mass and more strength-dependent physical function measures following PRT were completely lost during the 3-year detraining period.

Due to the partial retention of FM losses by the PRT group (−1.0 and −0.5 kg for total and trunk FM, respectively, relative to baseline), and with the control group gaining fat, by followup the control group had accumulated 3.4 kg more total fat, 3.2 kg more trunk fat, and 3.5% more %BF than the PRT group.

This relative benefit is important, as obesity is very prevalent in RA, e.g., Elkan et al (5) found that among RA patients, 33% of women and >50% of men had an FM index above the 90th percentile for the general population. Additionally, the accepted associations between obesity and the classic cardiovascular disease (CVD) risk factors are also apparent in RA patients, as central obesity is linked to hypertension, insulin insensitivity, metabolic syndrome, and arterial thickening and stiffening in this population (6). Therefore, this accentuated adiposity, particularly centrally, probably contributes to the 50% increase in CVD events and mortality reported for RA patients relative to non-RA controls (6). With this in mind, recent reports that anti–tumor necrosis factor therapy increases trunk FM (7) and total FM (8) in RA patients are disturbing and further emphasize the need for safe interventions that effectively counter the adipogenic effects of disease, sedentary lifestyle, and treatment in RA. In this respect, PRT has proven efficacy.

As mentioned previously, adiposity is associated with disability in RA patients, and the substantial losses of FM that occurred following 24 weeks of PRT may have contributed to the training-related improvements in objectively measured function. However, it is notable that performances for KES and the chair stand and arm curl tests returned to baseline levels along with ALM following detraining. This observation is consistent with performance of these tests being more reliant on muscle strength than is the case for rapid walking (9). Better retention of improvements in walking may also be attributable to regular walking in daily life and hence attenuated reductions in neural activation.

Loss of strength and functional gains following PRT is inevitable once training ceases, especially in the older population (10). However, it appears that the benefits of PRT, once established (i.e., after 8–12 weeks of training), can be maintained with reduced training. In the Rheumatoid Arthritis Patients in Training study (11), gains in KES following 2 years of twice weekly HI training (including strength training) were maintained by those patients who continued exercising once per week for the subsequent 18 months, but completely lost by those who stopped exercising. Similarly, substantial retention of gains in strength and function following 6 months (12) and 2 years (13) of HI PRT is reported for RA patients who remained very physically active over the subsequent 3 years, i.e., performing on average 4 bouts of moderate-intensity to HI aerobic exercise per week (e.g., walking, cycling, cross-country skiing) and, in the latter study, not reducing the time spent exercising (4 hours/week) and often continuing strength training.

To our knowledge, the only previous longitudinal body composition study on RA patients was that performed by Westhovens during his PhD (1999), but this was never published in a scientific journal (Westhovens: PhD thesis). In this investigation, Westhovens assessed body composition changes (by DXA) over 13 months in recently-diagnosed patients. During this time, mean ALM decreased by 2.5%, whereas total FM increased by 11.3%. These changes are more dramatic than those in our 9 established and stable RA control patients (i.e., annual changes of −0.68% in ALM and +2.9% in FM), and this is consistent with our finding that muscle loss primarily occurs early in disease, probably prior to treatment (14). The annual rate of loss of LM we observed for our followup control patients (0.10 kg) is similar to that reported for age-matched individuals from the general population (men: 0.07 kg/year, women receiving estrogen: 0.08 kg/year, women not receiving estrogen: 0.16 kg/year), but the rate of FM gain is higher (0.81 kg/year versus 0.37 kg/year and 0.41 kg/year for men and women, respectively) (15). These findings, albeit on a small sample, together with our earlier results (14), suggest that the rate of muscle loss in RA patients is increased during the initial phase of disease and, provided disease is controlled, then is reduced to rates normal for sedentary individuals, while the rate of FM accumulation may be chronically elevated.

The effectiveness of HI PRT in improving body composition and physical function in RA patients is clear. Unfortunately, long-term adherence to regular training is usually poor, especially when supervision is withdrawn. Consequently, in this followup study we assessed the extent to which the benefits of 6 months of HI PRT were retained following reversion to normal activities for almost 3 years. As expected, gains in muscle mass and strength-dependent function were lost with detraining. However, significant benefit was maintained in levels of adiposity, especially trunk adiposity, and walking ability. This is particularly relevant given the high incidence of CVD and CVD risk factors in RA patients, and the efficacy of moderate-intensity to HI walking as a treatment for these. This research also highlights the need to determine how long-term adherence to exercise programs can be improved in the RA population—perhaps by incorporation of behavior-modification strategies.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Lemmey 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 conception and design. Lemmey, Marcora, Maddison.

Acquisition of data. Lemmey, Williams, Marcora, Maddison.

Analysis and interpretation of data. Lemmey, Williams, Marcora, Jones, Maddison.