Rheumatology 2001; 40: 432-437
© 2001 British Society for Rheumatology
Physical exercise and risk of severe knee osteoarthritis requiring arthroplasty
Department of Public Health and General Practice, University of Kuopio, Kuopio
1 Department of Epidemiology and Biostatistics, Finnish Institute of Occupational Health, Helsinki
2 National Public Health Institute, Helsinki
3 Department of Surgery, Kuopio University Hospital, Kuopio, Finland
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
Objectives. To examine the association between physical exercise and the risk of severe knee osteoarthritis requiring arthroplasty.
Subjects and methods. A case–control study was carried out in which the cases were men (n=55) and women aged 55–75 yr (n=226) receiving knee arthroplasty for primary osteoarthritis at the Kuopio University Hospital. Controls (n=524) were selected randomly from the population of Kuopio province. Lifetime physical exercise was assessed retrospectively. Cumulative exercise hours were calculated and divided into two classes by mean (low/high).
Results. The risk of knee osteoarthritis requiring arthroplasty decreased with increasing cumulative hours of recreational physical exercise. After adjustment for age, body mass index, physical work stress, knee injury and smoking, the odds ratios (with 95% confidence intervals) of knee arthroplasty were 0.91 (0.31–2.63) in men with a low number of cumulative exercise hours and 0.35 (0.12–0.95) in those with a high number of cumulative exercise hours, with a history of no regular physical exercise as the reference. For the women, the corresponding odds ratios were 0.56 (0.3–0.93) and 0.56 (0.32–0.98).
Conclusion. Moderate recreational physical exercise is associated with a decrease in the risk of knee osteoarthritis.
KEY WORDS: Case–control study, Sports, Physical work stress, Risk factors.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
Physical exercise may have either injurious or beneficial effects on the joints. The harmful effect of intensive physical exercise on articular cartilage is well documented, whereas there is less evidence of the beneficial effect of recreational or moderate physical exercise on the joints.
Increased risk of osteoarthritis in the knee has been found among soccer players [1–2] and weight-lifters [2]. It has also been speculated that longstanding endurance sports accelerate cartilage degeneration. Lane has reviewed studies on runners and concluded that the risk of knee osteoarthritis may be increased by running at a competitive but not at a recreational level [3]. Recreational running did not show an increased risk of knee osteoarthritis in a prospective study [4–5]. Similarly, in the Framingham cohort no increased risk of knee osteoarthritis was observed in those practising habitual [6] or light or moderate [7] physical activity.
The purpose of this case–control study was to explore the association between cumulative physical exercise during different life periods and the risk of severe knee osteoarthritis requiring knee arthroplasty.
|
Subjects and methods |
|---|
|
TopAbstractIntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
Subjects
The base population included all 55- to 75-yr-old inhabitants of the province of Kuopio in Finland. The cases were all men and women who had received their first knee arthroplasty due to primary osteoarthritis in the Kuopio University Hospital during the years 1992 and 1993. The 383 cases were identified through the Finnish Registry of Arthroplasties [8]. Seven hundred and seventy-four controls were selected at random from the population of Kuopio province and were frequency-matched according to age (±2.5 yr) and sex.
One of the patients had died. Some of the subjects were lost for the following reasons: no telephone number was found (n=180); not reached by telephone (n=45); severe health problems (n=25); on a long journey (n=2); refused to cooperate (n=30). In all, a total of 874 subjects were interviewed. The interviewers considered 17 of the interviews to be unreliable and these were excluded from the study. A total of 53 subjects were considered to have secondary osteoarthritis and were therefore excluded. The reasons for exclusion were cerebral palsy (n=6), rheumatoid arthritis (n=30), other lower extremity disease (n=11), such as status post rachitis, osteomyelitis or tuberculosis, and length difference in the lower extremities (n=6). Subjects with an injury-induced disease of a lower extremity were included. In the final analyses there were 805 (70%) subjects (610 women and 195 men). The subjects comprised 281 cases and 524 controls. There was no statistical difference in the percentage who failed in the interview between the cases (26.7%) and the controls (32.3%) or the percentage failing because of chronic diseases of lower extremity.
|
Methods |
|---|
|
TopAbstractIntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
An introductory letter was sent to all cases and controls. A few days later they were contacted by telephone. Those agreeing to participate in the study were interviewed by means of a computer-assisted telephone interviewing system (CATI; Berkeley, CA, USA). The interviews were carried out in 1994. The interviewers did not know who was a case and who was a control. The subjects were asked about, inter alia, their work history, recreational physical exercise, knee injuries, chronic diseases, malformations in lower extremities, weight, height and smoking habits.
In the interview, lifetime recreational physical exercise was assessed with the following question: ‘have you sometimes during your lifetime regularly performed some physical exercise for at least two years?’ Physical exercise at school was excluded. Those who answered affirmatively were asked to indicate the amount of physical exercise for different types of sport. First, the subjects were asked how many hours per week, how many months per year and for how many years they had participated in the following sports: running, cross-country skiing, cycling, gymnastics, skating, volleyball, badminton, ice-hockey, weightlifting, wrestling, motor sports, walking, swimming, track and field athletics, and other sports. The ages when physical exercise started and stopped were requested. For each subject, cumulative hours of physical exercise were calculated for the different types of sports and for all physical exercise. Most of the subjects were moderate exercisers. There were only a few subjects who practised a sport at competitive level. The effect of physical exercise on the risk of knee arthroplasty was considered for different life periods separately: 
29 yr, 30–49 yr, 49 yr and during the subject's lifetime (including the preceding categories and exposure after the age of 50 yr). For the physically active subjects, total exposure was divided into two classes according to the mean cumulative number of hours of physical exercise. In those aged 49 yr or under, a low level of cumulative exercise was defined as fewer than 8654 exercise hours in the men and fewer than 6862 exercise hours in the women. A high level of cumulative exercise was defined as at least 8654 exercise hours in the men and at least 6862 exercise hours in the women. One extra class was established for those who reported having practised some sport but whose cumulative number of exercise hours could not be calculated. There was also a reference category, consisting of subjects who stated that they had not taken part in any regular physical exercise for at least 2 yr.
One of the authors, who had long experience in the occupational health service, reviewed the occupational history of each subject according to job title, and classified them into three groups according to the amount of physical stress involved: low (office work, teaching, management etc.), medium (factory work, service etc.) and high (farming, forestry, mining, etc.). A subject was considered to have a history of knee injury if the injury had led to consultation with a physician.
Statistical analysis
The data were analysed by logistic regression [9] and the results were expressed as odds ratios with 95% confidence intervals (CI). Models were constructed separately for men and women. Variables that showed statistical significance (P<0.05) in the univariate analyses were included in the multivariate analyses. Age and body mass index were included as continuous variables. The cumulative number of hours of physical exercise was entered both as a continuous and as a categorical variable. The statistical significance of physical exercise in the models was tested with the likelihood ratio test [9].
|
Results |
|---|
|
TopAbstractIntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
Of the 805 subjects studied, only 300 (37.2%) had participated in some physical exercise regularly by the age of 49 yr (Table 1
). Among the cases, there were more subjects who reported that they had not practised any regular physical exercise by the age of 29 yr (P=0.01), by the age of 49 yr (P=0.004) and during their lifetime (P=0.001) compared with the controls.
|
Of the potential confounders in women, body mass index was associated with knee arthroplasty in the multivariate analysis (Table 2
). In the men, knee injury increased the risk of knee osteoarthritis requiring arthroplasty threefold in comparison with subjects having no history of knee injury (Table 2). In the multivariate analysis for men, heavy physical work (as indicated by expert assessment of physical work stress according to job title) showed a protective effect on knee arthroplasty. In the women, heavy physical work stress seemed to increase the risk of knee arthroplasty in the univariate analysis.
|
Physical exercise was related to the risk of knee arthroplasty (Table 2
). Both in the men and in the women, the risk was less in those with a high number of cumulative hours of exercise by the age of 49 yr compared with those who had no regular physical exercise. The association was significant (P<0.05) in the univariate analysis in the women and in the multivariate analysis in the men.
Of the specific types of sports cross-country skiing, walking and swimming were associated with a statistically significant decrease in the risk of knee osteoarthritis in the women (Table 3). In the men, despite the trend of inverse association between all types of physical exercise except swimming and motor sport and the risk of knee arthroplasty, no significant association was found for specific types of sports.
|
The relationship between total physical exercise and the risk of knee arthroplasty seemed consistent across different age-periods in both sexes (Table 4
). In the models with cumulative hours of physical exercise as a continuous variable, the adjusted odds ratio was 0.91 (95% CI 0.80–1.01) per 1000 h of physical exercise in the men and 0.96 (95% CI 0.92–1.01) in the women. The association between cumulative hours of physical exercise and the risk of knee arthroplasty was non-linear, as inclusion of the quadratic term of the hours of physical exercise improved the models both in the men (P=0.001) and in the women (P=0.001). No significant interaction could be observed in either men or women between knee injury, physical work stress and physical exercise with respect to the risk of knee osteoarthritis.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
In this study, the subjects represented a mainly rural population who had practised either no regular physical exercise or whose exercise level had been, in most subjects, moderate. In both sexes, moderate physical exercise during different periods of life seemed to be inversely related to the risk of knee osteoarthritis requiring knee arthroplasty. The tendency was constant in both sexes across different age-periods of physical exercise.
The cases in this study were subjects with severe primary knee osteoarthritis requiring arthroplasty, and were identified from the Finnish Registry of Arthroplasty [8], where all joint prostheses are reported. In Finland most knee replacement operations are performed within the public health-care sector, which means that all citizens have practically the same opportunity to receive a prosthesis. Controls were drawn from the general population and there was no control with respect to whether they had osteoarthritis. This may have attenuated the associations detected in this study.
A potential source of error in this study is that joint problems restrict the practice of physical exercise. To minimize selection resulting from this, subjects with a disabling chronic disease in their lower extremities were excluded. The causation of other lower extremity disease was investigated by means of an open question in the interview. We assume that hardly any of the excluded subjects had exercise-induced disease. Subjects with earlier injury in the lower extremities were included and controlled for in the analyses. We also considered the associations between physical exercise and knee osteoarthritis during different age-periods, and from this point of view the results seem to be consistent. At least for the subjects who had practised regular physical exercise before the age of 29 yr, there should not have been any major selection bias.
In case–control studies, recall error with respect to physical activity may introduce bias [10]. The cases in our study may have been more aware of their physical activity than the controls. The magnitude of bias in this kind of setting is not known. In general, retrospective collection of data is subject to recall error. The repeatability of data on physical exercise depends on the test–retest interval, the type and intensity of physical activity, the method used to collect the data, and the study population [11]. The use of questionnaires has been found to have fairly high short-term repeatability, but long-term repeatability has been lower. In a comparison of several surveys, the test–retest correlation varied between 0.67 and 0.95 for vigorous physical activity and between 0.24 and 0.60 for low-intensity activity [11]. Bias may also occur when interviewers do not treat cases and the controls equally, but in the present study the interviewers did not know who was a case and who was a control.
Assessment of the effect of physical exercise on the risk of a slowly progressing disease such as osteoarthritis is difficult in epidemiological studies. In this study, physical exercise was assessed cumulatively and the effect of exposure on the risk was considered for different age-periods. Unfortunately, there was a considerable number of subjects in this study who reported that they had practised some exercise but for whom the cumulative number of exercise hours could not be calculated. This may have been a result of our method of calculation. The cumulative numbers of hours spent practising specific types of sports were calculated with the use of several variables. Missing data in one of these variables led to failure to calculate the number of exercise hours of the subject concerned.
What was the continuity of specific types of sport in this questionnaire? In Finland, the types of sports that are being practised at a given time depend to a great extent on the season. We assume that those who do practise physical activity are active throughout the year. They change the type of sport they practise according to weather and season. In studying the causal factors of knee osteoarthritis, the level of exposure for a number of years before diagnosis needs to be considered. In fact, the duration of the latency period is unknown. Vingård et al. [12] estimated exposure up to the age of 49 yr in their study of hip osteoarthritis. We examined exposure over the entire lifetime of our subjects.
In our study, physical work stress was not associated with an increased risk of knee osteoarthritis, which is in discordance with a recently published review [13]. The analyses in the present study were based on expert assessment of physical work stress according to job titles. However, assessment of physical work stress is difficult in epidemiological studies. Vingård and her colleagues [14] postulated that the assessment of physical work stress according to job title is only a rough guide to the level of exposure [14], which may explain our result.
Of the specific types of sport, those giving the greatest protection against knee osteoarthritis seemed to be cross-country skiing, walking and swimming in women. Only rough analyses could be carried out because the number of subjects was too small to test the associations in more detail. In the present analyses, it was not possible to take into account the intensity of physical exercise and the age-periods in which it was practised.
In two earlier studies, physical exercise showed some protective effect on the risk of knee osteoarthritis. Kohatsu and Schurman [15] reported that a higher proportion of controls (57%, n=46) than of cases with knee osteoarthritis (36%, n=46) had walked regularly for exercise. In a study comparing knee radiographs between middle-aged female teachers of physical education and an age-matched sample from the general population, the teachers had a lower prevalence of knee osteoarthritis. The teachers had undertaken habitual physical activity over many years [16]. In two studies from the Framingham cohort, the risk of knee osteoarthritis was not increased in those who practised physical activity habitually [6–7].
The results of this study seem to disagree with those of several previous studies [1–2, 17–18]. One explanation may be that in all these studies the exposed subjects had a high level of physical activity. Roos and colleagues studied knee osteoarthritis in former soccer players [1]. Of the potential confounders, only age was considered. In the case–control study of Kujala and colleagues there were more hospitalizations due to knee osteoarthritis in former Finnish elite athletes than in controls [17]. Another study using the same database revealed an increased risk of knee osteoarthritis in soccer players and weight-lifters [2]. The increased risk was attributed to joint injuries in the soccer players and to being overweight in the weight-lifters. However, these studies cannot demonstrate whether the practice of sport predicts osteoarthritis per se or the likelihood of hospitalization once joint problems have occurred. The latter seems possible, because former sportsmen, even in old age, may be more likely than others to seek care in order to maintain their physical activity. However, in a retrospective cohort study [18] the risk of knee osteoarthritis was increased not only in female ex-athletes but also in women from the general population who reported long-term sport activity. The authors speculated that the duration rather than the frequency of sport activity is important. Recreational physical activity has also been observed to increase the risk of osteoarthritis of the hip [19]. In population studies, it is not easy to obtain even a rough estimate of the amount of physical activity.
Three recent reviews have dealt with the association of physical activity with the risk of osteoarthritis [3, 20–21]. It was concluded that physical activity increased the risk of osteoarthritis if it was a part of sport activity that continually exposed normal joints to high levels of impact or torsional loading, and also if the activity consisted of a sport that could lead to injury of supporting structures such as ligaments, tendons and meniscuses [3]. Moderate habitual exercise does not seem to increase the risk of osteoarthritis [3, 20–21]. Recent findings also suggest that quadriceps weakness can be a contributory risk factor for the progression of joint damage in knee osteoarthritis [22], which provides indirect evidence that physical exercise in general may not be detrimental to the knee joint.
In humans, there are no previous studies in which the risk of knee osteoarthritis has been compared between physically inactive subjects and those practising physical exercise. Animal studies have indicated that both immobilization and high-intensity physical exercise may contribute to articular degeneration, but this appears not to be the case for physical activity at moderate intensity [23] and regular loading [24]. Perhaps the optimal effect of physical exercise in the prevention of knee osteoarthritis is achieved by regular and moderate exercise.
Recreational physical exercise may decrease rather than increase the risk of severe knee osteoarthritis requiring knee arthroplasty. This could be an important message for health promotion in general and for the prevention of knee osteoarthritis in particular.
|
Notes |
|---|
Correspondence to: P. Manninen, Department of Public Health and General Practice, University of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland.
|
References |
|---|
|
TopAbstractIntroductionSubjects and methodsMethodsResultsDiscussionReferences |
|---|
- Roos H, Lindberg H, Gardsell P, Löhmander LS, Wingstrand H. The prevalence of gonarthrosis in former soccer players and its relation to meniscectomy. Am J Sports Med1994;22:219–22.
[Abstract/Free Full Text] - Kujala UM, Kettunen J, Paananen H, Aalto T, Battie MC, Impivaara O et al. Knee osteoarthritis in former runners, soccer players, weight lifters and shooters. Arthritis Rheum1995;38:539–46.[Web of Science][Medline]
- Lane NE. Physical activity at leisure and risk of osteoarthritis. Ann Rheum Dis1996;55:682–4.
[Abstract/Free Full Text] - Lane NE, Bloch DA, Hubert HB, Jones H, Simpson U, Fries JF. Running, osteoarthritis and bone density: Initial 2-year longitudinal study. Am J Med1990;88:452–9.[Web of Science][Medline]
- Lane NE, Michel B, Bjorkengren A, Oehlert J, Shi H, Bloch DA et al. The risk of osteoarthritis with running and aging. A 5-year longitudinal study. J Rheumatol1993; 20:461–8.[Web of Science][Medline]
- Hannan MT, Felson DT, Andersson JJ, Naimark A. Habitual physical activity is not associated with knee osteoarthritis: the Framingham study. J Rheumatol1993; 20:704–9.[Web of Science][Medline]
- McAlindon TE, Wilson PW, Aliabadi P, Weissman B, Felson DT. Level of physical activity and the risk of radiographic and symptomatic knee osteoarthritis in the elderly: the Framingham study. Am J Med1999;106: 151–7.[Web of Science][Medline]
- Paavolainen P, Hämäläinen M, Mustonen H, Slätis P. Registration of arthroplasties in Finland. Acta Orthop Scand1991;62(Suppl. 241):27–30.
- Breslow NE, Day NE. Statistical methods in cancer research. Vol. I. The analysis of case–control studies. Lyon: International Agency for Research on Cancer, 1980.
- Hernberg S. Introduction to occupational epidemiology. Chelsea, MI: Lewis Publishers, 1992.
- Jacobs DR, Ainsworth BE, Hartman TJ et al. A simultaneous evaluation of 10 commonly used physical activity questionnaires. Med Sci Sports1993;25:81–91.
- Vingård E, Alfredsson L, Goldie I, Hogstedt C. Sport and osteoarthritis of the hip. An epidemiologic study. Am J Sports Med1993;21:195–200.
[Abstract/Free Full Text] - Jensen LK, Eenberg W. Occupation as a risk factor for knee disorders. Scand J Work Environ Health1996; 22:165–75.[Web of Science][Medline]
- Vingård E, Alfredsson L, Goldie I, Hogstedt C. Occupation and osteoarthrosis of the hip and knee: A register-based cohort study. Int J Epidemiol1991; 20:1025–31.
[Abstract/Free Full Text] - Kohatsu ND, Schurman DJ. Risk factors for the development of osteoarthritis of the knee. Clin Orthop1990; 261:242–6.
- White JA, Wright V, Hudson AM, He D. Relationships between habitual physical activity and osteoarthritis in ageing women. Public Health1993;107:459–70.[Medline]
- Kujala UM, Kaprio J, Sarna S. Osteoarthritis of weight bearing joints of lower limbs in former elite male athletes. Br Med J1994;308:231–4.
[Abstract/Free Full Text] - Spector TD, Harris PA, Hart DJ, Cicuttini FM, Nandra D et al. Risk of osteoarthritis associated with long-term weight-bearing sports—a radiologic survey of the hips and knees in female ex-athletes and population controls. Arthritis Rheum1996;39:988–95.[Web of Science][Medline]
- Lane NE, Hochberg MC, Pressman A, Scott JC, Nevitt MC. Recreational physical activity and the risk of osteoarthritis of the hip in elderly women. J Rheumatol1999;26:849–54.[Web of Science][Medline]
- Buckwalter JA, Lane NE. Athletics and osteoarthritis. Am J Sports Med1997;25:873–81.
[Abstract/Free Full Text] - Lequesne MG, Dang N, Lane NE. Sport practice and osteoarthritis of the limbs. Osteoarthritis Cartil1997; 5:75–86.
- Slemenda C, Brandt KD, Heilman DK, Mazzuca S, Braunstein EM, Katz BP et al. Quadriceps weakness and osteoarthritis of the knee. Ann Intern Med1997; 127:97–104.
[Abstract/Free Full Text] - Newton PM, Mow VC, Gardner TR, Buckwalter JA, Albright JP. The effect of lifelong exercise on canine articular cartilage. Am J Sports Med1997;25:282–7.
[Abstract/Free Full Text] - Arokoski JAP, Jurvelin JS, Väätäinen U, Helminen HJ. Normal and pathological adaptations of articular cartilage to joint loading. Scand J Med Sci Sports2000; 10:186–98.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. F. Zoeller Jr Physical Activity: Physical Activity in the Management of Osteoarthritis of the Knee and Hip American Journal of Lifestyle Medicine, August 1, 2007; 1(4): 264 - 266. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Hendry, N. H Williams, D. Markland, C. Wilkinson, and P. Maddison Why should we exercise when our knees hurt? A qualitative study of primary care patients with osteoarthritis of the knee Fam. Pract., October 1, 2006; 23(5): 558 - 567. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Karpansalo, T A Lakka, P Manninen, J Kauhanen, R Rauramaa, and J T Salonen Cardiorespiratory fitness and risk of disability pension: a prospective population based study in Finnish men Occup. Environ. Med., October 1, 2003; 60(10): 765 - 769. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Galois, S. Etienne, L. Grossin, C. Cournil, A. Pinzano, P. Netter, D. Mainard, and P. Gillet Moderate-impact exercise is associated with decreased severity of experimental osteoarthritis in rats Rheumatology, May 1, 2003; 42(5): 692 - 693. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||