Article Text

Download PDFPDF

Concise report
Patients with regular physical activity before onset of rheumatoid arthritis present with milder disease
  1. Maria E C Sandberg1,
  2. Sara Wedrén1,2,
  3. Lars Klareskog2,
  4. Ingrid E Lundberg2,
  5. Christina H Opava3,
  6. Lars Alfredsson1,4,
  7. Saedis Saevarsdottir1,2
  1. 1Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
  2. 2Rheumatology Unit, Department of Medicine, Karolinska Institutet/Karolinska University Hospital, Stockholm, Sweden
  3. 3Division of Physiotherapy, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden
  4. 4Centre for Occupational and Environmental Medicine, Stockholm County Council, Stockholm, Sweden
  1. Correspondence to Dr Maria E C Sandberg, Institute of Environmental Medicine, Box 210, Stockholm 171 71, Sweden; maria.sandberg{at}


Objectives Physical activity has been shown to decrease inflammatory markers; here we investigate the effect on the clinical presentation of rheumatoid arthritis (RA).

Methods We used the cases from the population-based EIRA study (N=617), followed in the Swedish Rheumatology Quality Register, calculating the odds of having above median level of 28-joint disease activity score (DAS28), physician assessment, pain (visual-analogue scale (VAS), VAS-pain) and activity limitation (health assessment questionnaire (HAQ)) at diagnosis, as an effect of physical activity 5 years before diagnosis, investigated both in categories and dichotomised.

Results Dose–response relationships were seen for all measures; the higher the level of physical activity, the lower the likelihood of having outcome measure above median. Further, regular physical activity associated with 42% reduced odds of having DAS28 above median (OR=0.58 (95% CI 0.42 to 0.81)). Effects were similar for VAS-pain (OR=0.62 (95%CI 0.45 to 0.86)) and physician assessment (OR=0.67 (95%CI 0.47 to 0.95)) but not for HAQ. Statistically significant effects were also found both for the combined objective components and the combined subjective components of DAS28.

Conclusions Physically active individuals seem to present with milder RA, which adds to the evidence of beneficial effects of physical activity on inflammatory diseases. The observation should be important for both health professionals and individuals seeking to reduce their risk.

  • Disease Activity
  • Early Rheumatoid Arthritis
  • Epidemiology
  • Physcial therapy
  • Rheumatoid Arthritis

Statistics from


Physical activity is crucial for human health and lack thereof is the third most important cause of preventable deaths today.1 Physical activity as a concept does not only include exercise, but is a measure of how active, or inactive, a person is, and therefore includes all intensities.2 Several studies, both in healthy individuals and patients with chronic inflammatory diseases, have shown that physical activity results in a reduction in inflammatory markers, highlighting that it influences several biological pathways (reviewed by Beavers et al).3

For a long time, physical activity was perceived as potentially harmful for patients with rheumatoid arthritis (RA). However, intervention trials have shown during the past decade that physical activity is both well tolerated and improves muscle function in established RA.4 ,5 The effect of physical activity before disease onset has, however, not been evaluated before. The aim of this study was to investigate whether leisure time physical activity before disease onset impacts the clinical presentation of RA.


Study population

The participants in this study were the patients with newly diagnosed RA, included in the Swedish EIRA study; a population-based case–control study, which has been described elsewhere.6 From year 2006, questions regarding physical activity 5 years before baseline (ie, at diagnosis) were included in the questionnaire and patients recruited from this year could thus be included in the present study (N=818). Information on disease activity and other clinical features was obtained from the Swedish Rheumatology Quality (SRQ) register, as previously described.7

Definition of exposure

Leisure time physical activity 5 years before the diagnosis was assessed by questionnaire; for definitions and examples given in the questionnaire, see table 1. This specific question has been validated in a Swedish population.8 The four categories of physical activity were also combined to a binary measure, comparing regular physical activity with no regular physical activity.

Table 1

Describing the alternative answers to the question “Which of the following alternatives fits your physical activity 5 years ago the best?”, which was used to assess exposure in this study

Outcome measures

Four clinical measures were used as outcome measures; the 28-joint disease activity score (DAS28), physician assessment of disease activity in five categories (no, low, moderate, high, maximal), health assessment questionnaire (HAQ) and pain on a visual-analogue scale (VAS-pain). We also separately analysed each of the components in the DAS28 score and further combined the objective measures of the DAS28 score; C-reactive protein (CRP), erythrocyte sedimentation rate (ESR) and 28-swollen joint count into one group, and the subjective measures; 28-tender joint count and patient global assessment, into one group.

Statistical analysis

We used logistic regression to calculate the odds of scoring above median for each of the outcome measures; complete-case analysis was used. First, we calculated the OR for each of the four categories of physical activity independently (compared with the inactive group), adjusting for sex, period of diagnosis and age at diagnosis (p value for trend was generated by including level of physical activity as a continuous variable). We then compared regular physical activity with no regular physical activity, where we additionally adjusted for smoking at diagnosis (never/past/current cigarette smokers), body mass index at diagnosis (<25, 25–30 >30 kg/m2), alcohol intake the last year before diagnosis (five levels), socioeconomic status (highest attained level of education, three levels), vegetable intake the last year before diagnosis (in quartiles) and physically demanding work 5 years before diagnosis (four levels). Effect modifiers were investigated by including interaction terms in the model.

All analyses were carried out using SAS Statistical Package V.9.3. All participants have given informed consent, and the Ethical Review Board at Karolinska Institutet, Stockholm, Sweden, approved the study.


Our final study population consisted of 617 RA patients, after excluding patients not reported to the SRQ register at baseline (diagnostic visit) (N=170), patients who lacked information on physical activity (N=7) and patients who, despite being reported to the register, lacked information on all outcomes measures (N=24). For the 617 included patients, the median symptom duration at diagnosis was 171 days and >90% of the patients had their first symptoms less than a year before diagnosis. For baseline characteristics, see table 2.

Table 2

Baseline characteristics of RA patients included in the study (N=617)

Table 3, panel A, displays the effect of each level of physical activity 5 years before RA diagnosis on the OR of having disease measures above median levels at diagnosis. We see a statistically significant trend of lower odds for patients with higher levels of physical activity for all measures (p for trend <0.01 for DAS28, VAS-pain and HAQ, p=0.02 for physician assessment), that is, the higher the level of physical activity, the lower the likelihood of having disease measures above median. For patients classifying their physical activity as ‘regular physical activity/workout’ (the highest level), the odds of having DAS28, physician assessment, VAS-pain and HAQ above median was significantly reduced by 42–59% compared with inactive patients.

Table 3

Logistic regression calculating the odds of having outcome measure (DAS28, physician assessment (0–4), pain and HAQ, respectively) above median

We then compared patients reporting regular physical activity with those reporting no regular physical activity and found that the former had a 42% reduced odds of having DAS28 above median (OR=0.58 (95% CI 0.42 to 0.81)); table 3, panel B). This was not affected by additional adjustments (OR=0.57 (95% CI 0.39 to 0–82)). Statistically significant effects of a similar magnitude were found for VAS-pain and physician assessment but not for HAQ. In these models, adjusted for sex, period and age, we also investigated potential effect measure modifiers; the effect of physical activity did not differ by anticyclic citrullinated peptide antibodies (anti-CCP) status, sex, body mass index (BMI), socioeconomic status or physically demanding work. As an additional analysis, we investigated the odds of having high disease activity, defined as DAS28 >5.1; this cut-off was similar to the median DAS28 in this population (5.24) and lead to a similar result (OR for regular physical activity=0.71 (95%CI 0.49 to 1.02)) (data not shown in table).

To investigate whether the association between physical activity and DAS28 was driven specifically by any of the measures that constitute the DAS28 score, we separately calculated the p value for each of the components and found significant associations between physical activity and ESR, CRP and patient global assessment (data not shown in table). Further, regular physical activity 5 years before diagnosis was associated with 40% decrease in odds of having all objective measures above median (OR=0.60 (95% CI 0.40 to 0.88)), and 35% decreased odds of having both subjective measures above median (OR=0.65 (95% CI 0.45 to 0.95)).


In summary, this first study of physical activity before disease onset and the clinical presentation of RA shows that regular physical activity is associated with a milder disease at diagnosis. The association was consistent across subgroups of this heterogenic syndrome and remained after adjustment for potential confounders. The beneficial effect of regular physical activity was seen for both the objective and the subjective components of the DAS28 score.

No previous studies have evaluated the influence of physical activity on the clinical presentation of RA. However, disease activity has commonly been included as an outcome in clinical trials of physical activity in established RA patients. In a review by Stenström and Minor,3 six studies reported an improvement and eight found no change in disease activity after physical activity intervention, while only a small minority of the studies included and reported improvements in VAS-pain or HAQ. The results of the present study implicate that only the highest level of physical activity is sufficient to have a beneficial effect on HAQ; in general, the knowledge of the levels and patterns of physical activity needed for beneficial effects needs to be further elucidated.

We found that CRP, ESR and patient global assessment were significantly affected by physical activity, which corresponds to research on healthy individuals, in whom physical activity resulted in decreased CRP and ESR.3 ,9 ,10 Since the odds of both the objective and the subjective groups of DAS28 components were found to be significantly decreased by regular physical activity, the effect does not seem confined to any specific component of DAS28.

Regularly physically active people are a selected subpopulation, although perhaps less so in Sweden, where 50% of the population is estimated to fulfil the recommendations of at least 30 min of physical activity every day.11 ,12 Physical activity has been shown to correlate with high socioeconomic status, younger age, non-smoking, normal BMI and high vegetable intake.9 ,13 We had the possibility to adjust for all these factors, but the estimates were not affected.

Physical activity is likely acting on the human body by multiple different mechanisms, several of them connected to chronic inflammation and its markers, as reviewed in Perandini et al.14 An example of the potency of physical activity was shown in a gene-expression analysis of patients suffering from another chronic inflammatory disease, myositis, where even a short period of just 7 weeks had profound effects on the gene-expression in the inflammatory pathways.15 Further, Barres et al showed that a single episode of exercise gave rise to transient hypomethylation, and subsequent upregulation, of genes in skeletal muscle.16 In the present study, physical activity 5 years before diagnosis is assumed to indicate a physically active lifestyle as the level of physical activity in adults has been shown to be relatively constant.17

Strengths of this study include its population-based setting and the extensive information available, which made adjustments for a multitude of potential confounders possible. In observational studies, reverse causality can never be entirely excluded. It might be particularly likely when investigating an association such as that between physical activity and RA, where it is conceivable that early RA symptoms would lead to decreased physical activity. Reassuringly, we were able to measure physical activity 5 years before diagnosis, which was, on average, 4.45 years before symptom onset. Previous studies have shown that 25% of RA patients have rheumatoid factor and/or anti-CCP antibodies present 5 years before diagnosis; we however find it unlikely that the presence of autoantibodies without symptoms would affect the physical activity.18 Finally, our exposure is self-reported, which is likely to lead to some misclassification, presumably non-differential since both exposed and unexposed are RA patients, this would in turn lead to bias towards the null and thus underestimate the true association.

In conclusion, our findings demonstrates that RA patients who are physically active before clinical disease onset present with a milder disease, both in terms of inflammation, pain and function. This adds to the growing evidence of general beneficial health effects of physical activity and is an important message for individuals at increased risk of RA, as well as for health professionals, similarly to the recent findings regarding smoking and overweight.19


We acknowledge the EIRA study group and EIRA data collectors.



  • Handling editor Tore K Kvien

  • LA and SS contributed equally.

  • Contributors All authors of this research paper have directly participated in the planning (LK, LA, SS, IEL and SW), analysis (MECS, SS and LA), interpretation (MECS, SS, LA, LK and CHP) and writing (MECS, SS, LK, LA, SW and CHP) of the study, and all authors have read and approved the final submitted version.

  • Funding This study was financially supported by grants from the Swedish council for working life and social research, Vinnova, the Swedish Foundation for Strategic Research and the Swedish Rheumatism Foundation. The funding sources had no role in the reporting of the study or in the decision to submit the manuscript for publication.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethical Review Board at Karolinska Institutet, Stockholm, Sweden.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.