Excercise and Osteoarthritis

Exercise may decrease the risk of developing knee osteoarthritis (OA) via trophic effects on cartilage and muscle strengthening.  However, exercise may also predispose to knee OA via joint loading and the potential for traumatic injury.  These predisposing factors may be more important in individuals who are overweight or obese.  Here, Felson et al (Arthritis Care & Research 2007; 57(1): 6) investigate the relationship between self-reported recreational exercise and the development of knee OA.


Participants in the Framingham Offspring Study, a cohort study investigating the inheritance of OA, without baseline knee OA underwent knee radiography and were asked questions about habitual exercise.  Knee radiography was repeated 1 to 2 years later when subjects were asked questions about interim knee injury.  Radiographs were evaluated for development of knee OA (defined as progression to a Kellgren/Lawrence (K/L) grade of 2 or higher with or without incident symptoms), the development of patellofemoral OA, and progressive joints space loss.


1,279 subjects (of which 2,259 knees were eligible) had complete radiographs with an average of 8.8 years between examinations.  56% of subjects were women with a mean age at baseline of 53 years and a mean BMI of 27.4 kg/m2.  26% of subjects reported knee pain at baseline.  215 knees (9.5%) developed incident knee OA, most knees with incident knee OA were symptomatic (80.4%), and most showed OA in the tibiofemoral compartment (84.1%).

Walking for Exercise Regular walking for exercise was reported in 48% of subjects.  There was no difference (risk or benefit) in incident OA for those walking more or less than 6 miles per week compared to those reporting no regular walking for exercise.  The risk did not differ in analyses stratified by median body mass index (BMI).  Although a small percentage of subjects reported regular jogging or running, there was not a significant risk or protective effect associated with these activities.

Intensity of Physical Activity There was no risk or benefit in incident OA observed in subjects who reported engaging in physical activity with enough intensity to work up a sweat compared to those not engaging in this intensity of activity.  These relationships did not differ in analyses stratified by BMI

Relative Physical Activity There was no significant difference in incident OA between subjects reporting higher or lower activity levels compared to peers compared to those reporting similar activity levels.  Again, there remained no significant differences in analyses stratified by BMI.


Self-reported habitual physical activity was not associated, either positively or negatively, with incident radiographic or symptomatic knee OA.  BMI level did not modify the lack of association.

Editorial Comment

These are interesting findings that suggest that while habitual exercise does not increase the risk of knee OA, it may also not hinder its development.  There are several limitations to the study that may limit the validity of these interpretations.  For one, regular exercise was assessed by patient self-report from a handful of general questions.  Second, the impact of exercise may depend more on how it is performed than merely on if it is done.  While the effect of BMI was investigated, there is no additional information on how the activities may load the knee or may promote muscle strengthening.  Despite these qualifications, the study has the advantage of a large sample size and longitudinal follow-up, even though 1 to 2 years may be insufficient to see radiographic changes.  Longer term follow-up is warranted to track the effect of physical activity on incident knee OA in this population.

Jon Giles, MD

About Jon Giles, MD