Background Mechanical joint stress imposed by high body mass index (BMI) is associated with increased risk of knee and hip osteoarthritis. This prospective study investigated the independent and joint association of BMI and physical exercise on risk of knee and hip osteoarthritis.
Methods The study includes 15 191 women and 14 766 men in the Norwegian HUNT Study without pain or physical impairment at baseline. Occurrence of self-reported physician-diagnosed osteoarthritis was assessed at 11 years of follow-up.
Results BMI was positively related to risk of knee osteoarthritis (Ptrend<0.001), with an RR of 4.37 (95% CI 3.01 to 6.33) in women and 2.78 (95% CI 1.59 to 4.84) in men, comparing obese and normal weight persons. No clear association was observed for hip osteoarthritis. Obesity increased the risk of severe activity-limiting osteoarthritis, with an RR of 2.30 (95% CI 1.68 to 3.15) and 2.50 (95% CI 1.56 to 4.03) in women and men, respectively. Physical exercise did not modify the above associations (Pinteraction>0.34). Exercise intensity was not associated with risk of osteoarthritis in any BMI category; that is, obese persons reporting high-intensity exercise had an RR of 1.28 (95% CI 0.59 to 2.79) for severe osteoarthritis compared with inactive persons.
Conclusion High BMI increases the risk of knee osteoarthritis and severe osteoarthritis. Physical exercise does not increase the risk of osteoarthritis at any level of BMI, suggesting that exercise could be encouraged also among individuals with excessive body mass, without concern for an increased risk of osteoarthritis.
- prospective study
- knee pain
- hip pain
Statistics from Altmetric.com
Osteoarthritis is a chronic degenerative joint disease that contributes substantially to lower limb disability, especially among older adults.1–3 Body mass has been positively associated with hip and knee osteoarthritis, indicating that mechanical stress is a contributing factor to development of osteoarthritis in weight-bearing joints.4–8 It may therefore be hypothesised that physical exercise, particularly among individuals with high body mass, increases the risk of knee and hip osteoarthritis by imposing additional joint stress.
Physical exercise has been associated with a higher risk of knee and/or hip osteoarthritis in some studies,9–11 while others have reported inverse or no association.12–14 Few prospective studies have investigated if physical exercise amplifies the risk of osteoarthritis among individuals with high body mass. However, longitudinal data from the Framingham Study have shown an increased risk of knee osteoarthritis among older people with high body mass index (BMI) who performed heavy physical activity at least 3 h/day.15 In contrast, physical exercise was not associated with higher risk of knee osteoarthritis among younger Framingham residents with BMI above median level of the study population.16
We have used longitudinal data from a large population-based health survey of Norwegian women and men to prospectively examine the independent association between BMI and risk of self-reported physician-diagnosed knee and hip osteoarthritis and also to investigate if exercise increases the risk of knee or hip osteoarthritis among people with high BMI.
Subjects and methods
In Nord-Trøndelag County in Norway, all inhabitants aged 20 years and older were invited to participate in two waves of a large health survey (the HUNT Study), the first in 1984–1986 (HUNT 1) and the second in 1995–1997 (HUNT 2).
Among 87 285 eligible persons, 77 216 (89%) accepted the invitation to HUNT 1, filled in a questionnaire and attended a clinical examination. Body mass and height were measured at the examination, and the participants were given a second questionnaire to complete at home and return in a pre-stamped envelope. At HUNT 2, 94 187 persons were invited to participate, and 66 215 (70%) accepted the invitation. The procedures were similar as those described for HUNT 1, although the questionnaires and clinical examination were more comprehensive. Detailed information about the HUNT Study can be found at http://www.ntnu.edu/hunt website.
For the purpose of the present study, we selected all 24 357 women and 21 568 men who had participated in both surveys. Of these, we excluded 4085 women and 3446 men without baseline information on physical exercise, 167 women and 113 men without data on musculoskeletal pain, 13 women without information on BMI and 59 men and 360 women classified as underweight (BMI <18.5 kg/m2). Moreover, we excluded 1511 women and 1521 men who reported to be physically impaired at baseline or who had no information on this variable. To obtain a study sample of persons without pain in knees or hips at baseline (ie, at the time of HUNT 1), participants who at HUNT 2 reported pain in lower limbs that had lasted for 10 years or more were excluded (3030 women and 1663 men), leaving a total of 15 191 women and 14 766 men for statistical analysis.
All participants in the HUNT Study gave a written informed consent upon participation, and the study was carried out according to the Declaration of Helsinki. The protocol for the current study was approved by the Regional Committee for Ethics in Medical Research (project no. 4.2008.206, REK midt, Norway).
Body mass index
Standardised measurements of body height (to the nearest centimetre) and body mass (to the nearest half kilogram) obtained at the clinical examination at baseline (HUNT 1) were used to calculate BMI as body mass divided by the squared value of height (kg/m2). Participants were then classified into three BMI categories according to cut points suggested by the WHO,17 that is, normal weight (BMI 18.5–24.9 kg/m2), overweight (BMI 25.0–29.9 kg/m2) or obese (BMI ≥30.0 kg/m2).
Chronic pain and osteoarthritis
Questions about musculoskeletal symptoms were adopted from the Standardised Nordic Questionnaire.18 At follow-up (HUNT 2), the participants were asked: “During the last year, have you had pain and/or stiffness in your muscles and limbs that lasted for at least three consecutive months?” with response options yes/no. If ‘yes’, the participants were required to specify the pain-afflicted body area(s) and to indicate whether these complaints had caused reduced leisure-time activity (yes/no). Response options for the lower limbs were ‘hips’, ‘knees’ and ‘ankle/feet’. Only knees and hips were included in the analyses. In addition, participants filled in information about duration of symptoms, that is, number of months if symptoms had lasted <1 year and number of years if symptoms had lasted ≥1 year. The latter was used to exclude those who reported pain duration ≥10 years (ie, at baseline). The diagnosis of osteoarthritis was confirmed (yes/no) by the question: “Has a doctor ever said that you have a degenerative joint disease (osteoarthritis)?” However, no information about the time of diagnosis or the joint(s) affected by osteoarthritis was collected. Joint location of chronic pain was therefore used to indicate location of osteoarthritis.
Based on the information about chronic pain and osteoarthritis, we constructed two mutually exclusive outcomes: (1) knee osteoarthritis, that is, only chronic knee pain along with self-reported physician-diagnosed osteoarthritis and (2) hip osteoarthritis, that is, only chronic hip pain along with self-reported physician-diagnosed osteoarthritis. An additional outcome variable was ‘severe osteoarthritis’, which included those who reported chronic knee and/or hip pain along with osteoarthritis and reduced ability to carry out desired leisure-time activity due to knee and/or hip symptoms.
At baseline (HUNT 1), the participants were asked to complete a questionnaire that included questions on frequency, duration and intensity of leisure-time physical exercise per week (ie, walking, swimming or other sports). The frequency question had five response options (0, <1, 1, 2–3, ≥4 times per week; coded 1–5). Individuals who reported less than one exercise session per week were classified as inactive. Individuals who reported exercising once a week or more were asked about the average duration (<15, 15–30, 31–60, >60 min; coded 1–4) and intensity (no sweat or heavy breathing, sweat and/or heavy breathing and nearly exhausted; coded 1–3). In the analysis, the response option ‘no sweat or heavy breathing’ was categorised as low intensity, while ‘sweat and/or heavy breathing’ and ‘nearly exhausted’ were collapsed into high intensity. Based on the information on frequency and duration, we calculated the average number of hours spent on physical exercise per week; the frequency response option two to three times per week was counted as 2.5 times and ≥4 times per week counted as five times, whereas the duration response options <15, 15–30, 31–60 and >60 min were counted as 10, 25, 45 and 75 min, respectively. Individuals who accumulated ≥1 h exercise per week were classified as physically active.
A generalised linear model for the binomial family (log link) was used to estimate RRs for knee osteoarthritis, hip osteoarthritis and severe osteoarthritis in different categories of BMI compared with the reference group with normal weight. A trend test was calculated by treating the BMI categories as an ordinal variable in the regression model. BMI was also analysed as a continuous variable, estimating the RR for one unit increase. Precision of the estimated RRs was assessed by 95% CIs. All analyses were adjusted for the potentially confounding effect of age (continuous), occupation (manual, non-manual, farmer/fisher, non-worker and unknown) and smoking (never, former, current and unknown).
In corresponding analysis, we estimated the combined effect of BMI (ie, normal weight, overweight, obese) and physical exercise (inactive vs active as defined above) on risk of knee and hip osteoarthritis, using inactive persons with normal weight as the reference category. Statistical interaction between BMI and exercise was evaluated in a likelihood ratio test of a product term of these factors in the regression model. In a supplementary analysis, we increased the amount of exercise required to be classified as active by restricting the active category to persons who reported at least two exercise sessions per week.
Finally, we conducted two sets of analysis stratified by BMI categories; first, to assess the independent effect of exercise (categorised as inactive or active) on risk of knee and hip osteoarthritis, and second, to examine if exercise intensity (categorised as inactive, low intensity or high intensity) was differentially associated with the risk of knee and hip osteoarthritis in normal weight, overweight and obese individuals.
All statistical tests were two-sided, and all statistical analyses were performed using Stata for Windows (V.10.0 © StataCorp LP, 1985–2007).
Table 1 presents baseline characteristics of the study population stratified by BMI and physical exercise. At follow-up, 219 (1.4%) women and 132 (0.9%) men reported knee osteoarthritis, while 220 (1.4%) women and 102 (0.7%) men reported hip osteoarthritis. Of those confirming a diagnosis of osteoarthritis, 323 (2.1%) women and 175 (1.2%) men reported severe osteoarthritis (ie, reduced leisure-time activity due to osteoarthritis-related symptoms in knees and/or hips).
As shown in table 2, BMI was positively associated with both knee osteoarthritis and severe osteoarthritis (all Ptrend<0.001) but not with hip osteoarthritis (Ptrend 0.21 in women and 0.29 in men). Compared with normal weight persons, women and men who were classified as obese had an adjusted RR of 4.37 (95% CI 3.01 to 6.33) and 2.78 (95% CI 1.59 to 4.84) for knee osteoarthritis, respectively. The corresponding RR for severe osteoarthritis was 2.30 (95% CI 1.68 to 3.15) in obese women and 2.50 (95% CI 1.56 to 4.03) in obese men, whereas the RR for hip osteoarthritis was 1.15 (95% CI 0.74 to 1.78) in women and 1.44 (95% CI 0.72 to 2.88) in men. The adjusted RR for knee osteoarthritis per unit increase in BMI was 1.12 (95% CI 1.09 to 1.15) in women and 1.09 (95% CI 1.04 to 1.15) in men, whereas for severe osteoarthritis, the corresponding RR was 1.07 (95% CI 1.05 to 1.10) and 1.09 (95% CI 1.04 to 1.14).
In analysis of the combined effect of BMI and exercise, the RRs for the various outcome measures were largely similar among inactive and active participants (table 3). There was no indication of an interaction between BMI and exercise (p from likelihood ratio test >0.34 for all associations). Moreover, exercise was not associated with the risk of osteoarthritis within any of the BMI categories (all ps>0.38). Increasing the amount of exercise required to be classified as active to at least two exercise sessions per week did not change the results (p from likelihood ratio test >0.81 for all associations). The risk of knee osteoarthritis was 3.06 (95% CI 1.69 to 5.53) among obese people who were physically active at least two times per week for a total of at least 1 h (data not shown), whereas the corresponding risk among inactive people was 3.34 (95% CI 2.01 to 5.54) (table 3).
Table 4 presents results from additional analyses of exercise intensity. Overall, there was no difference in risk between inactive individuals and those who usually exercised with low or high intensity for any of the outcome measures within each category of BMI.
This study shows that BMI has a strong positive association with risk of knee osteoarthritis and also with osteoarthritis-related symptoms that impair the ability to carry out desired leisure-time activity (ie, severe osteoarthritis). No clear association was observed between BMI and risk of hip osteoarthritis. Analysis of the combined effect of BMI and physical exercise showed no indication of an interaction, suggesting that exercise imposes no different effect on risk of knee or hip osteoarthritis across categories of BMI. Results from more detailed analysis of exercise intensity (inactive vs low or high intensity) stratified by categories of BMI were also in line with this. The results support the notion that high BMI is a strong risk factor for knee osteoarthritis and that regular exercise, irrespective of intensity, does not enhance this risk.
The strong dose–response relation between body mass and risk of knee osteoarthritis is in accordance with similar findings in other studies.8 Few prospective studies have investigated the association of BMI with both the risk of knee and hip osteoarthritis within the same population. However, a recent study by Lohmander and colleagues6 is in agreement with the results from the present study, reporting a much stronger association for risk of severe knee osteoarthritis than for severe hip osteoarthritis. Interpreting our results, it should be noted that the data only allowed for specification of the joint location of chronic pain but not the joint location of osteoarthritis. Thus, the location of chronic pain was used as a proxy for the joint location of osteoarthritis and the construction of the mutually exclusive outcome measures of knee and hip osteoarthritis.
A striking finding in our study is the high risk of severe activity-limiting osteoarthritis both among overweight and obese individuals compared with those who were normal weight. In Western societies, more than half of the population can be classified as overweight or obese,19 and a substantial proportion of adult women and men are therefore at risk of developing osteoarthritis-related symptoms in knees and hips that interfere with the ability to carry out desired leisure-time activity. In particular, this may have severe impact on health and quality of life by reducing the possibility to perform physical activity and thereby maintenance of a healthy lifestyle. Accordingly, recent studies have shown a substantial impact of osteoarthritis on mortality,20 particularly among those with walking disability.21
It is well established that a sedentary lifestyle increases the risk of serious medical conditions, such as type 2 diabetes mellitus,22 ,23 hypertension24 and cardiovascular diseases.25 Lower limb osteoarthritis are therefore of particular concern due to the impact on ambulatory activity level and the possibility of being physically active. Importantly, the current study indicates that regular exercise of 1 h or more per week among overweight and obese individuals does not enhance the risk of knee or hip osteoarthritis. Moreover, this association also holds true for high-intensity exercise, that is, there was no increased risk associated with high-intensity exercise within categories of BMI compared with low-intensity exercise or inactivity. Other studies investigating the effect of recreational exercise on risk of osteoarthritis have provided mixed results.26 Some studies indicate that participation in certain high impact sports may impose an increased risk of osteoarthritis,2 ,8 while other studies have found no effect of long-distance running13 ,27 ,28 or walking/sports participation.29 In the HUNT Study, there is no information available about type of exercise or sports participation. Thus, although there was no evidence of increased risk with high-intensity exercise, we cannot rule out the possibility that participation in high impact sports can increase the risk of knee and hip osteoarthritis disproportionally within categories of BMI.
Notably, we are only aware of two studies, both from the Framingham Study, that have used a prospective design to investigate the effect of exercise on risk of knee osteoarthritis among overweight and obese individuals.15 ,16 Felson and colleagues16 found no effect of exercise among middle-aged women and men with BMI above median for the study population, while McAlindon and colleagues15 reported an increased risk of knee osteoarthritis from more than 3 h of heavy physical activity per day among older people in the highest third of BMI. These contrasting findings may result from different categorisations of BMI and physical activity and also from focus on different age groups of the population. Neither of these two studies was of sufficient size to enable BMI categorisation according to cut-off points recommended by the WHO,17 and they did not investigate whether exercise increases the risk of hip osteoarthritis among overweight and obese individuals. Thus, the current study adds additional and important evidence in refuting the hypothesis that exercise increases the risk of osteoarthritis in knees and hips among individuals with excess body mass.
The strengths of the current study include its prospective design, the large population-based sample, exclusion of individuals with pain and physical impairment at baseline and the long follow-up period. However, there are also some limitations that should be considered in interpretation of the results. First, the confirmation of physician-diagnosed osteoarthritis was based on self-reports, and accuracy could therefore be influenced by poor recall. Moreover, it is generally accepted that complete accuracy in diagnosing of osteoarthritis requires a thorough clinical examination and radiographic testing. However, this is not easily implemented in large observational studies, and the issue is further complicated since radiographic findings are not always in agreement with clinical symptoms.30 Some studies have indicated that self-reports of osteoarthritis have acceptable reliability to be used in epidemiological studies.30 ,31 Although less than ideal, self-reported physician-diagnosed osteoarthritis may therefore be appropriate for case ascertainment in large community studies. Second, self-reports about chronic knee and hip pain were used as proxy to indicate joint location of osteoarthritis. Although misclassification cannot be ruled out, we consider it unlikely that individuals with, for example, knee osteoarthritis will report chronic hip pain only and not chronic knee pain (and vice versa). Third, information on physical exercise and BMI was obtained only at baseline and changes during the follow-up period could not be taken into account. This also applies to injuries that may have occurred during the follow-up period and that increased the risk of osteoarthritis, for example, anterior cruciate ligament injury.32 Finally, it cannot be ruled out that people with high BMI and low exercise level to a lesser degree participated in the second survey. If persons who were lost to follow-up also had a higher risk of osteoarthritis, the results of the present study could be underestimated.
The results from the present study support the notion that BMI is an independent and important modifiable risk factor for osteoarthritis in knees and that high BMI increase the risk of severe activity-limiting osteoarthritis. Moreover, there was no evidence that regular physical exercise, irrespective of intensity, increases the risk of osteoarthritis among overweight and obese individuals. These findings indicate that exercise should not be discouraged among overweight and obese individuals out of concern for increased risk of knee and hip osteoarthritis.
What is already known on this subject
Excess body mass is an independent risk factor for hip and knee osteoarthritis, indicating that mechanical stress is a contributing factor to the development of osteoarthritis in major weight-bearing joints.
However, whether physical exercise in combination with excess body mass amplifies the risk of knee and hip osteoarthritis in the general adult population remains unresolved.
What this study adds
This study shows that regular physical exercise, irrespective of exercise intensity, does not increase the risk of knee or hip osteoarthritis among overweight and obese women and men in the general adult population.
Thus, physical exercise should not be discouraged among individuals with excess body mass out of concern for increased risk of knee and hip osteoarthritis.
The Nord-Trøndelag Health Study (The HUNT Study) is a collaboration between HUNT Research Centre (Faculty of Medicine, Norwegian University of Science and Technology NTNU), Nord-Trøndelag County Council and the Norwegian Institute of Public Health. The authors thank Professor Brian Hopkins (Lancaster University) for his advice and comments during the preparation of this paper.
Competing interests None declared.
Ethics approval The Regional Committee for Ethics in Medical Research (project no. 4.2008.206, REK midt, Norway).
Provenance and peer review Not commissioned; externally peer reviewed.
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.