Table 3

Studies on exercise as a possible cause of knee OA, in order of quality of evidence

STUDYSTUDY TYPEN (FOLLOW-UP, %)DURATIONACTIVITYOUTCOME ON KNEE OACOMMENTS
Framingham offspring (Felson et al29)Cohort1705 (75)9 yRecalled level of activityNo increase in OA; obesity did not increase OA riskHistory, examination, and x-ray findings
Ongoing data collection reduced recall bias
Reported patellofemoral disease
Injury reported
Framingham (McAlindon et al17)Cohort473 (99)9 yRecalled level of activityIncreased incidence with heavy physical activityHistory, examination, and x-ray findings
Ongoing data collection reduced recall bias
Injury reported
Chingford (Hart et al12 )Cohort10034 yWalking, jogging, or sportsNo increased risk with greater activity levelsX-ray findings and reported pain evaluation
Women only
Self-reported exercise
Stanford University (Fries et al,28 Chakravarty et al,30 Wang et al,31 Lane et al,32 Lane et al33)Cohort113 (87)18 yRunningNo increase in OA; reduced disability; marked divergence in morbidities of all types compared with controls Careful evaluation of x-ray findings and symptomsCareful evaluation of function and disability
Runners were self-selected
Very specific for runners of high education and socioeconomic status
Panush et al34,35Cohort35 (100)8 yRunningNo increase in OADiagnosis by symptoms and x-ray findings
Cooper Clinic (Cheng et al13)Case control16 96117 yRecalled level of activityIncrease for young men exercising > 20 h/wk; increase for obese women and obese young menNo record of injury
Diagnosis of OA given to patient by another physician
Level of exercise by subjective recall
No evaluation of function
Subjects all of high socioeconomic status
Sohn and Micheli36Case control7912–55 yRunning, swimmingRunners had no increase in lower extremity pain or arthritis surgery compared with swimmersCompared hip or knee pain in runners and swimmers
Arthritis diagnosed only by history of knee or hip surgery
Lane et al37Case control863NARunningRunners showed less muscle and joint disability and increased functional capacityStudy of lower extremity disability related to knees but not specific to knee OA
Framingham (Hannan et al38)Case control1404NALevel of activityNo increase in OA with increased level of activityHistory, examination, and x-ray findings
Data collection reduced recall bias
Injury reported
Spector et al39Case control1058NAElite tennis players, runnersIncreased OA in elite athletes, but no increase in symptomsWomen only
Diagnosis based on x-ray findings only
Schmitt et al40Case control40NAElite marathon runnersNo increase in OAGood clinical and x-ray definition of OA
Kujala et al41Case control2448NAElite athletesAthletes had slightly higher risk of hospital admissionHospital admission for any sign of OA
Examination of records only
Control group was incomplete
Sutton et al27Case control1080NARecalled level of exerciseIncreased risk only with injuryDiagnosis and level of exercise self- reported
Kujala et al21Cross sectional117NASportsIncreased OA in soccer players and weight lifters; increased risk with obesity, injury, and work history of heavy liftingDiagnosis based on recorded history, physical and x-ray findings
Men only
Konradsen et al42Cross sectional30NARunningNo increase in OADiagnosis based on clinical and x-ray findings
Roos et al43Cross sectional858NASoccerIncreased OA in elite athletes onlyNo symptoms reported
X-ray diagnosis only
Szoeke et al44Cross sectional224NARecalled level of exerciseIncreased risk with obesity and in those very active in youthWomen only
Exercise self-reported
Diagnosis based on x-ray findings only
McDermott and Freyne22Cross sectional20NARunningStrong association of OA with trauma and genu varumMen only
  • Questionnaire, physical examination, and x-ray findings

  • NA—not applicable, OA—osteoarthritis.