Amongst older adults without knee pain, those who were obese were nearly three times more likely than those of normal weight to develop severe knee pain in a subsequent three-year period. Including progression from non-severe pain, almost one-fifth of all new cases of severe knee pain in people aged 50 years and over in a three-year period could be avoided if excess weight was prevented.
The implications of our findings for public health, clinicians and researchers are clear as they illustrate how the population burden of knee pain could diminish if the proportion of overweight and obese people (i.e. those with a BMI ≥ 25) in the total population of older people were to decline. Being obese or overweight were predictors of onset of severe knee pain at three years in responders who were free of knee pain at baseline. Avoidance of excess weight, in this group in particular, is likely, therefore, to assist with primary prevention of knee pain in the older population. As knee pain is a common and disabling condition in older adults living in the community, the potential for health gain in the population is large.
Our findings add to previous longitudinal research on the link between obesity and knee OA [8
]. Gelber and colleagues studied male medical students (median follow-up 36 years) and found that a greater BMI in men aged between 20 to 29 was associated with an increased risk of subsequent knee OA [8
]. Our population-based findings reinforce the need for early primary prevention. In our study, obesity (defined as BMI >30) was also a strong predictor of progression of non-severe knee pain to severe knee pain at three years. Public health interventions targeted at avoiding excess weight in those with non-severe knee pain are likely to assist with secondary prevention.
Our study was not an intervention study. Our analyses have not taken into account the effect of any weight change. It cannot be concluded from our study that weight loss, rather than avoiding excess weight in the first place, would reduce the amount of severe and disabling knee pain in older people. However, weight loss is recommended as treatment for knee osteoarthritis [27
]. Given that osteoarthritis is a chronic but generally non-fatal disease, people with established severe knee pain and disability might usefully use weight reduction to manage and alleviate these symptoms. So it is likely that our estimates of the impact of avoiding excess weight, which are based on the incidence of new pain or the onset of severe or disabling knee pain in the general population, are underestimates of the total impact which interventions aimed at both excess weight avoidance and weight reduction in the population might have on incident and prevalent problems combined. Our results suggest, however, that the main advantage in population terms is in preventing obesity; the excess risk is small in the overweight group, and so, despite the fact that the overweight category contains more people overall, the number of potentially preventable cases in this group is small compared to those generated by the obese group.
In our study, responders in the obese category would have to reduce their weight by a median of 13.5 pounds (6.1 kg) to move into the overweight category. Those overweight would have to reduce their weight by a median of 12.4 pounds (5.6 kg) to move into the normal weight category. In a study exploring the effects of actual weight loss, a loss of approximately 5.1 kg decreased the odds of developing symptomatic knee OA by 50% [15
Identifying and implementing effective interventions for avoiding weight gain or achieving weight loss in the older population remains a challenge. There is evidence that long-term weight loss needs to be supplemented with exercise [28
]. Barriers to physical activities that are experienced by older adults (for example, fear of pain, misconceptions about benefits of exercise, environmental factors) [29
] must also be addressed. There is evidence of beneficial effects of exercise per se in knee OA patients [30
] so providing information about the benefits of both weight loss and exercise to the community is a realistic start. Mehrortra and colleagues studied the prevalence of professional advice to lose weight among obese adults with arthritis who had a recent physician visit [31
]. Less than half of obese arthritis sufferers recalled receiving advice on weight loss at their last visit. Providing information is a short-term strategy but in the long-term a wider public health approach will be required.
Approximately 98% of the British population is registered with a GP [32
] and the register provides a convenient sampling frame of a local population. The prevalence of knee pain at baseline was 46.8% [14
]. To take account of non-response bias at baseline we standardised this to the age and gender distribution of the entire older population of the participating general practice registers. The population estimate of prevalence was unchanged by this. As responders were slightly older than non-responders at baseline, but slightly younger at follow-up, the age of our sample was similar at time of follow-up to the original study population.
Based on mid-2000 population estimates, [33
] the demographic profile of the study sample is similar to that of North Staffordshire as a whole (the region in which the 3 GP practices were based) and of England and Wales. In our study, 56% of baseline responders were female, compared with 54% of the over 50 population in North Staffordshire and England and Wales. The proportion of people aged over 75 years in the baseline survey sample was 21%, compared to 22% in North Staffordshire and 23% in England and Wales. The study sample was slightly different to the UK as a whole, however, in terms of ethnicity as 99% of responders were white UK/European origin, a figure that does reflect the make up of the North Staffordshire population. The potential for prevention of knee pain may be different in populations with a different ethnic mixture. However, we estimate that 19% of new cases of severe knee pain could be avoided by a shift downwards in BMI category and this is similar to that derived in a US study by Leveille et al [16
] who reported 18% of arthritis cases were attributable to obesity.
This is a population-based study and our aim was to report the impact of excess weight avoidance on knee pain in the general population, regardless of consultation. This approach required the use of self-reported knee pain data instead of, for example, radiographic evidence of changes within the knee joint. The Knee Pain Screening Tool has been previously reported as valid and reliable [17
]. However, individuals with musculoskeletal pain often experience episodes of pain separated by pain free periods. The two surveys used in this study may have failed to capture some such recurrently painful episodes, thereby introducing some underestimation of episodic knee pain.
We also used self-reported data to calculate BMI and this may underestimate true population BMI [34
]. However, the mean BMI score in our population was is in line with national trends [35
]. The test-retest reliability of our BMI data was very good. Any misclassification would mean that our risk ratios are underestimates of the true associations between knee pain and obesity and overweight.
Another limitation is that there may be confounding factors for which our analysis has been unable to account. Such influences might include, for example, the use of medications, occupation, physical activities and other health problems like cardiovascular disease.