If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of General Practice and Orthopedics & Sports Medicine, Erasmus MC University Medical Center Rotterdam, PO-box 2040, Rotterdam 3000 CA, the Netherlands
Many experts call for a focus on the primary prevention of osteoarthritis.
•
Preventing the development of osteoarthritis is not as straightforward as it may seem.
•
Osteoarthritis prevention research should be deemed highly important to oppose the predicted increase in osteoarthritis development and associated costs to health care and society in the near future.
Introduction
The Need for Primary Osteoarthritis Prevention
For years, osteoarthritis (OA) has been ranked among the diseases with the largest impact on patients and society.
Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019.
The theoretic basis of primary prevention of OA might sound simple, appealing, and feasible: preventing the onset of OA among subjects without but at high risk for OA. Nevertheless, there are many pitfalls in different aspects related to the initiation of preventive measures in both clinical and research settings. To name a few: how and where can we identify the appropriate target groups? What interventions have an acceptable risk-benefit profile in the absence of symptoms in the ones we treat? How do we establish adherence to the intervention when patients cannot experience any benefits of the intervention? How do we evaluate the preventive effectiveness of the intervention in a slowly developing disease such as OA? The current review addresses some of these key challenges in preventing OA, provides insights into our current knowledge, and highlights some essential knowledge gaps in OA prevention. For this, we use a slightly modified version of the well-known PICO approach; in the absence of OA at the initiation of any primary preventive therapy, the P for “patient” in the traditional PICO should be replaced by either “population” or “target group.” As trivial as this might seem, this has important implications, mainly for the choice and uptake of a preventive intervention. Given the high prevalence, we will focus on knee, hip, and hand OA.
Identifying the target group
Risk Factors for Osteoarthritis Illness and Disease
The traditional approach to identify a target population for preventive measures for any condition is through identifying its risk factors. For OA, many risk factors have been identified over the years. Some of the most studied and well-known risk factors for knee OA development include older age, female sex, overweight/obesity, joint trauma, genetic predisposition, and occupational loading.
Female sex, older age, overweight/obesity, occupational loading, local muscle weakness, and genetic predisposition are all known to increase the risk for incident hand OA.
A very important point to consider here is the fact that most studies on risk factors for OA development have focused on the development of OA disease, that is, pathologic changes in joint tissues. In the available systematic reviews on risk factors for OA development, very little evidence is available for risk factors for OA illness, that is, symptoms and complaints of OA.
Nevertheless, it is the actual OA illness that drives the large burden for patients and causes major direct and indirect costs for society and health care. Moreover, radiographic OA (OA disease) in the absence of pain was not associated with mortality, whereas OA pain (illness) in the presence and absence of radiographic OA was associated with a 35% to 37% increased risk for mortality in the general population.
Knee osteoarthritis and time-to all-cause mortality in six community-based cohorts: an international meta-analysis of individual participant-level data.
Given the general importance of OA illness over OA disease, the focus of OA prevention should be on OA illness, and better insights into risk factors for the onset of OA illness are required.
After selecting established risk factors for OA development in the joint of interest, taking the OA illness versus disease concept into account, the identification of a target group for the prevention of OA could focus either on those at risk for a certain risk factor (eg, those at risk for knee joint trauma or at risk for overweight/obesity) or on those with a certain risk factor (eg, those with a recent knee joint trauma or those with overweight/obesity). The list of risk factors described here is far from exhaustive but is rather meant to illustrate that a strong risk factor for OA development does not always easily translate into the identification of a feasible target population. Both modifiable and nonmodifiable risk factors will be addressed; modifiable risk factors (eg, lifestyle, body weight, occupational loading) are amenable to interventions and therefore help to shape the preventive interventions, whereas nonmodifiable risk factors (eg, age, sex, genetic predisposition) are not amenable to interventions but can be used to identify the right target population.
Identification of Individuals at Risk for Osteoarthritis Risk Factors
For one of the strongest risk factors for knee OA development, joint trauma,
identification of individuals at risk seems relatively straight forward; sports that put players at increased risk for joint trauma include soccer and rugby for meniscal injuries and American football, soccer, and gymnastics for anterior cruciate ligament (ACL) injuries.
Selecting individuals participating in these “high-risk sports” could be a very feasible approach to identify a potential target population for preventive measures. Nevertheless, the incidence of ACL injuries per 1000 hours of athlete exposure ranges between 0.10 and 0.17 only, which leads to very high “numbers needed to treat.”
Also, for other risk factors for OA development in the knee, hip, or hand, for example, overweight/obesity or high occupational loading, different groups of individuals at risk for these risk factors can fairly easily be identified, for instance, low socioeconomic status girls/women who have an increased risk for overweight/obesity
For other OA risk factors, the identification of the right target group to prevent these risk factors might be more challenging. Local muscle weakness is a known risk factor for both knee and hand OA development.
Being highly modifiable, local muscle weakness could be seen as a great target for preventive interventions. However, with an average of only 1% decline in muscle mass (a proxy for muscle strength) per year from a peak between the age of 20 and 30 years,
identifying those at risk for local muscle weakness is challenging, as there is no threshold to define the presence of muscle weakness. Indications for a stronger decline after the age of 50 years and the strong link to physical inactivity could help to identify those at risk for local muscle weakness. One has to keep in mind that local muscle weakness could also be an early sign of OA disease and therefore not causally related to OA development. If so, targeting local muscle weakness in order to prevent OA development will be ineffective.
Identification of Individuals with Osteoarthritis Risk Factors
Focusing on risk factors for OA illness, identifying individuals with overweight/obesity (for knee and hip OA illness), participating in high-impact sports (for hip OA illness), or having physically demanding jobs (for hip OA illness) seems doable.
Although the presence of other OA illness risk factors such as hip shape morphology, (mild) hip dysplasia, and local muscle strength can be determined accurately and reliably, the feasibility of screening for the presence of these risk factors among subjects free of OA symptoms can be questioned. Besides the low prevalence of these risk factors in the open population, for example, 0% to 13% for cam impingement among nonselective populations,
From a prevention perspective, the importance of the association between aging and OA incidence is somewhat questionable. Not only is age itself nonmodifiable, the impact on the number of years lived in good health will be less substantial when preventing OA among elderly individuals (eg, 80+ years) than among middle-aged individuals. Next to that, the exposure to/development of many risk factors for OA incidence occurs during adolescence and early adulthood (eg, joint injuries, overweight/obesity, and occupational overload). When using these risk factors to define potential target populations for preventive interventions, the optimal “window of opportunity” likely has passed at an older age. That is why experts in the field call for a lifespan approach to OA prevention (Fig. 1): prevention and treatment of risk factors for OA development at those stages of life where these risk factors are developing or amendable to treatment.
it is the actual combination of the prevalence of the risk factor and the strength of its association to OA development that will determine the importance of that risk factor, in the light of OA prevention. For example, knee joint injuries are one of the strongest risk factors for future knee OA development.
By calculating the population attributable fraction, an estimation of the proportion of new cases in the population that could be avoided if the risk factor was removed, Silverwood and colleagues showed that the number of new cases of knee OA/pain in a 3-year follow-up study of 3907 middle-aged men and women (aged 50 years and older) that could be attributed to knee injuries was only 5.1%.
Hence, preventing all knee injuries among these 3907 individuals would only have led to a 5.1% lower incidence of knee OA. Despite the fact that the association between overweight/obesity and knee OA development is less strong than that for knee injuries, the higher prevalence of overweight/obesity resulted in a population attributable fraction of 24.6%.
These results might look very promising for a preventive trial on weight loss among overweight/obese subjects free of knee OA/pain. Nevertheless, with a total incidence of knee pain of only 24% in 3 years,
the actual effect of preventing all cases of overweight/obesity in the given population would result in a reduction of the total incidence of 24.6% × 24% = 6% in the first 3 years. A quick sample size calculation for a trial with 24.6% incidence in the control group and 18.6% in the intervention group would require 2 groups of 735 individuals followed-up for 3 years (α = 95%, β = 80%).
The aforementioned examples illustrate the importance of the selection of target populations based on the interaction of multiple risk factors for future preventive interventions. Risk factors such as female sex/gender or genetic predisposition might not be relevant on their own for the selection of a target population for OA prevention, as these factors are nonmodifiable. However, given the known interaction between female sex/gender, knee joint injuries, and adiposity on knee OA development,
focusing on the prevention of weight gain among female athletes who suffered from a knee injury might provide a more feasible treatment target. Unfortunately, little is known about the interaction of risk factors for OA development and should therefore be a focus of future research initiatives.
To better communicate individuals’ risk for OA development and to visualize the potential of certain interventions, risk stratification tools, such as those for cardiovascular events in which multiple risk factors are incorporated, might improve the understanding among high-risk individuals and could also help to motivate them to adhere to any preventive intervention.
Designing the intervention
So, for the selection of OA risk factors, either to prevent the occurrence of the risk factor or to intervene once the risk factor is present, both the prevalence of the risk factor and the strength of its association with OA development are of great importance for the relevance of the selected target population. Obviously, the next step to consider is if and when the selected risk factor is modifiable (ie, amenable to any intervention).
Challenges in Preventing Osteoarthritis Risk Factors
As indicated earlier, overweight/obesity is one of the major drivers for the development of OA.
Given these complex interactions, designing interventions that prevent overweight/obesity is very challenging. Despite the multidisciplinary nature of primary care, which is generally seen as the optimal setting for targeting the prevention of overweight/obesity, there is little to no evidence for effective interventions.
already in 2007, there was a call from the World Health Organization for more upstream interventions by countries to “develop its own needs-driven portfolio of appropriate and realistic interventions, and involve many stakeholders from all relevant sectors in a transparent and explicit process.”
Organisation, W.H., The challenge of obesity in the WHO European Region and the strategies for response, F. Branca, H. Nikogosian, and T. Lobstein, Editors. 2007.
The initiatives for the prevention of joint injuries illustrate another major challenge when designing a preventive intervention, namely adherence/uptake. There is sufficient high-quality evidence available showing that injury prevention programs (ie, plyometrics, strengthening, and agility exercises) are effective in the prevention of anterior cruciate ligament injuries.
A Majority of anterior cruciate ligament injuries can Be prevented by injury prevention programs: a systematic review of randomized controlled trials and Cluster-randomized controlled trials with meta-analysis.
Nevertheless, implementation of these programs into real-world settings is a major challenge, as adherence to these programs in the real world is generally very low.
This observation closely relates to a broader challenge for preventive therapies, as “there is no glory in prevention.” To further OA prevention, we need to align our efforts and knowledge with experts from behavioral research. How can we motivate individuals at risk for OA risk factors to modify their lifestyle, if there is no positive feedback from that actual intervention? After all, these individuals do not have any joint symptoms, so they cannot experience any relief of symptoms to keep them motivated to adhere to an intervention. Are individuals at risk for OA risk factors willing to consider medical interventions or should we aim for nonpharmacologic interventions only? How should we educate these individuals regarding their future risk of OA in order to motivate them to get into action, rather than demoralizing them by presenting a future with a chronic condition? These questions highlight some of the current knowledge gaps in OA prevention.
Challenges in Treating Osteoarthritis Risk Factors
When designing a preventive intervention, the presence of a (modifiable) risk factor for OA might help to overcome the lack of motivation for preventive measures among the target population. For instance, among those with overweight/obesity, tracking the body weight over time to illustrate their weight loss will provide participants with a measure of their progress during the intervention. Still, we know that achieving sustainable lifestyle changes among subjects with overweight/obesity is very hard, and adherence to these interventions remains a challenge.
Similarly, targeting individuals with a history of joint trauma might facilitate the willingness for preventive interventions, as these individuals have a strong preference to keep physically active but also often fear for a reinjury.
Activity preferences, lifestyle modifications and re-injury fears influence longer-term quality of life in people with knee symptoms following anterior cruciate ligament reconstruction: a qualitative study.
A recently developed framework for an intervention for managing knee OA risk factors after ACL injuries does show that the first steps toward OA prevention after knee joint injuries are currently being taken.
Davies AM, Wong R, Steinhart K, et al., Development of an Intervention to Manage Knee Osteoarthritis Risk and Symptoms Following Anterior Cruciate Ligament Injury. Osteoarthritis Cartilage, 2021. (in press).
The presence of an OA risk factor does not always facilitate the process of OA prevention. Participating in high-impact sports puts individuals at risk for hip OA development, and high occupational loading is known to increase the risk for hand and hip OA. Nevertheless, targeting these risk factors to prevent the subsequent onset of OA will be challenging. How to motivate athletes to change the sport they participate in toward one with less/no risk for OA development? Also, most employees with high occupational loadings will require retraining in order to take on new jobs, are likely to have a preference for their current job, and might lack a track-record to be competitive when applying for another and less burdensome job. Given the lack of preventive trials in OA research, the selection of effective interventions for the prevention of OA is very hard.
Next to that, there are other aspects that challenge the design of future preventive interventions, as described in the earlier section. Again, the list of examples is far from exhaustive, but rather an impetus for an in-depth discussion on what interventions will be effective and feasible in a preventive setting.
Measuring the effect
Every trial or intervention requires predefined primary outcome and/or treatment goal. In a preventive intervention aimed to prevent the onset of OA, obviously, the development of OA should be evaluated primarily. As easy as this may seem, defining the actual outcome for a preventive trial in OA has some challenges and implications.
Measuring the Incidence of Osteoarthritis Illness
As indicated earlier, it is the illness of OA (ie, symptoms and complaints of OA) that seems the most important outcome for OA prevention. Pain is often defined as the most important complaint by patients with OA.
Nevertheless, how we measure pain or define the presence of pain in OA is very variable. Current clinical guidelines use the presence of “activity-related joint pain” (irrespective of frequency or severity) to define the presence of pain in hip and knee OA.
Scientific reports using the American College of Rheumatology (ACR) classification criteria to evaluate clinical hip/knee OA define the presence of pain as “pain on most days of the month,” without specifying whether this is during any specific activity.
Comparison of three sets of clinical classification criteria for knee osteoarthritis: a cross-sectional study of 13,459 patients treated in primary care.
Others specifically measure pain during an activity, such as walking, stair descending/ascending, or rising from a chair, using specific questions from validated OA questionnaires such as WOMAC and KOOS/HOOS
Recent injury, severe radiographic change, and lower Quadriceps strength increase risk of knee pain Exacerbation during walking: a within-person knee-matched study.
Co-existing patterns of MRI lesions were differentially associated with knee pain at rest and on joint loading: a within-person knee-matched case-controls study.
Patient Acceptable symptom state in knee osteoarthritis patients Succeeds across different patient-reported outcome measures assessing physical function, but Fails across other Dimensions and Rheumatic diseases.
Minimum clinically important improvement and patient acceptable symptom state in pain and function in rheumatoid arthritis, ankylosing spondylitis, chronic back pain, hand osteoarthritis, and hip and knee osteoarthritis: results from a prospective multinational study.
; before an intervention, each patient defines his/her “acceptable state,” and the percentage of patients reaching that state is used in the comparison of the intervention arm. Up to now, the clinical relevance and feasibility of PASS have never been tested in a preventive setting.
An important aspect to consider when choosing a potential outcome of symptomatic OA is the fact that in early stage OA, pain can be absent for prolonged periods of time, with short and very sudden short flare-ups.
When pain is highly fluctuating, one can question the clinical relevance of a single measure in time of something like “pain during walking” or the PASS. To illustrate this fluctuation problem, although 52% of first-time presenters in primary care with knee and/or hip pain fulfilled the ACR-criteria for clinical knee/hip OA at baseline, only 17.5% fulfilled these criteria at each of the 5 follow-up measures over a 10-year period and only 14% never fulfilled these criteria during follow-up.
Most likely, measuring symptomatic outcomes more often (eg, daily) over a longer period (eg, several weeks or months) will provide a better insight into the symptoms of patients with early stage OA. With the use of mobile applications for data capture, this will be feasible
but will require advanced statistical methods to properly evaluate intervention effects in these highly correlated repeated measures. Measuring multiple domains of pain is highly recommended in the field of chronic pain research/management.
Given the challenges in measuring OA illness in a preventive setting, state-of-the-art knowledge from related fields (eg, chronic pain and pain assessment) needs to find its way into OA prevention.
Measuring the Incidence of Osteoarthritis Disease
Whether defined using traditional radiography and Kellgren and Lawrence criteria,
Structural features of OA do matter when considering total joint arthroplasty, as patients with only mild structural OA are significantly more dissatisfied after an arthroplasty than those with more severe structural OA.
Mild knee osteoarthritis predicts dissatisfaction after total knee arthroplasty: a prospective study of 186 patients aged 65 years or less with 2-year follow-up.
Moreover, in the light of cost savings, preventing arthroplasties due to OA will save on health care costs, as these are primarily attributable to these joint arthroplasties. However, when targeting subjects without disease for preventive intervention and evaluating the effects on joint arthroplasties, one will need to have multiple decades of follow-up to reach sufficient numbers of cases that underwent surgery; this scenario seems very unrealistic.
The Need for Surrogate Outcomes
Both OA illness and OA disease develop very slowly. Therefore, when evaluating potential outcomes for preventive trials in OA, the annual incidence rate of the outcome is an important feature for the feasibility of such a trial. Annual incidence rates of OA (either hands, knees or hips) within subjects at high risk are only ±2% to 5% for OA illness and disease.
Population trends in the incidence and initial management of osteoarthritis: age-period-cohort analysis of the Clinical Practice Research Datalink, 1992-2013.
To overcome this challenge, OA prevention urgently needs validated surrogate outcomes. These surrogate outcomes can lead to a “reduction in sample size and trial duration when a [...] distant disease is replaced by a more frequent or proximal endpoint,” and this could “reduce costs and enhance the feasibility” of preventive clinical trials.
Because of the success of surrogate outcomes in many other chronic diseases, treatment effects of certain medications are no longer evaluated using primary outcomes that have direct impact on patients (eg, cardiovascular events or mortality). For instance, in 436 registered randomized controlled trials on diabetes drugs, most trials (82%) used glycosylated hemoglobin as the primary outcome rather than patient-oriented outcomes.
One of the few existing examples of the evaluation of surrogate outcomes for OA development showed that a measure of cartilage integrity on MRI was significantly affected by weight loss and that its change was related to long-term knee OA development.
Unfortunately, in another study, none of the changes in structural MRI features were related to long-term pain progression among patients with knee pain and a meniscal tear that were randomized over after meniscal surgery and physical therapy.
Exploring and validating more potential surrogate outcomes for different target groups and different potential preventive interventions is deemed a top priority for the field of OA prevention research.
Conclusions and recommendations
With the aforementioned examples, we have tried to illustrate some of the key challenges in OA prevention (Table 1). Despite the call for shifting the focus of OA research toward the early disease stage and OA prevention, there is a clear lack of knowledge for many aspects of the early phase of OA.
Given the association with OA incidence in several joints, combined with the increasing prevalence of overweight/obesity worldwide, lifestyle interventions to prevent weight gain or to treat the presence of overweight are of great importance for the prevention of OA. Of course, OA prevention is not the only reason to strife for a healthy lifestyle in the population; positive health effects of a healthy lifestyle and a healthy body weight are known for, for example, diabetes, coronary heart disease, hypertension, cancer, and early death.
From a research perspective, it would be naïve to not align initiatives to gain knowledge on OA prevention with ongoing and future efforts with these other medical fields. From a societal perspective, perhaps focusing efforts on preventing overweight/obesity would be more appealing and therefore more feasible, than narrowing the scope down to OA prevention only. Given the important effect of OA pain on mortality
Knee osteoarthritis and time-to all-cause mortality in six community-based cohorts: an international meta-analysis of individual participant-level data.
OA outcomes should then be considered a major concern for the overall population that should be taken into account when evaluating the societal effects of interventions focusing on preventing overweight/obesity.
Two essential aspects of OA prevention clearly lack solid high-quality scientific evidence: risk factors for OA illness and surrogate outcomes. Inevitably, the signs of OA illness (ie, joint pain, functional limitations, health-related quality of life) form the burden to patients, drive patients to seek health care, and lead to early retirement.
Therefore, the selection of target populations and potential interventions should be informed by knowledge on risk factors for the development of OA illness. To facilitate future studies on OA prevention, also the evaluation and validation of surrogate outcomes for long-term OA development are urgently needed. Surrogate outcomes have the potential to shorten future trials on OA prevention and with that reduce costs and enhance the feasibility. As many studies in the field of OA have measured both short-term (ie, 1–2 years) changes in OA-related features (eg, patient-reported outcomes, objective physical functional test, and imaging) and have evaluated the long-term (ie, >5 years) development of OA, there is a huge potential for the identification and validation of surrogate markers already. Based on a stepwise approach, not only data from RCTs can be used for the validation of surrogate outcomes but also data from observational studies can help to build the evidence around potential surrogate outcomes for OA development.
In the absence of validated surrogate outcomes, and in line with the narrative of the societal approach to OA prevention, an alternative approach to enhance the feasibility of preventive trials could be to set strong OA risk factors as primary outcome for preventive trials (eg, ACL injuries) and estimate the subsequent preventive effect on OA development, rather than following all patients long enough for OA to develop.
In conclusion, despite the urge for a more proactive rather than a reactive approach to OA treatment, there are still many challenges in OA prevention. All aspects of OA prevention, selecting the proper target population, designing an optimal intervention, and measuring the intervention effects, still hold many challenges and have important knowledge gaps. Nevertheless, given the enormous burden of OA for patients and society, and the expected increase in OA prevalence, OA prevention is a very important field that deserves our attention and dedication.
Clinics care points
•
To prevent osteoarthritis development, the onset and interaction of multiple osteoarthritis risk factors should direct preventive strategies, for example, weight gain after menopause and muscle strength loss after joint trauma.
•
Behavioral strategies, such as Motivational Interviewing, should be applied to explore the internal motivation for lifestyle changes in individuals at risk for osteoarthritis development.
Disclosure
All authors declare no conflicts of interest related to the content of this article.
References
Diseases G.B.D.
Injuries C.
Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019.
Knee osteoarthritis and time-to all-cause mortality in six community-based cohorts: an international meta-analysis of individual participant-level data.
Organisation, W.H., The challenge of obesity in the WHO European Region and the strategies for response, F. Branca, H. Nikogosian, and T. Lobstein, Editors. 2007.
A Majority of anterior cruciate ligament injuries can Be prevented by injury prevention programs: a systematic review of randomized controlled trials and Cluster-randomized controlled trials with meta-analysis.
Activity preferences, lifestyle modifications and re-injury fears influence longer-term quality of life in people with knee symptoms following anterior cruciate ligament reconstruction: a qualitative study.
Davies AM, Wong R, Steinhart K, et al., Development of an Intervention to Manage Knee Osteoarthritis Risk and Symptoms Following Anterior Cruciate Ligament Injury. Osteoarthritis Cartilage, 2021. (in press).
Comparison of three sets of clinical classification criteria for knee osteoarthritis: a cross-sectional study of 13,459 patients treated in primary care.
Recent injury, severe radiographic change, and lower Quadriceps strength increase risk of knee pain Exacerbation during walking: a within-person knee-matched study.
Co-existing patterns of MRI lesions were differentially associated with knee pain at rest and on joint loading: a within-person knee-matched case-controls study.
Patient Acceptable symptom state in knee osteoarthritis patients Succeeds across different patient-reported outcome measures assessing physical function, but Fails across other Dimensions and Rheumatic diseases.
Minimum clinically important improvement and patient acceptable symptom state in pain and function in rheumatoid arthritis, ankylosing spondylitis, chronic back pain, hand osteoarthritis, and hip and knee osteoarthritis: results from a prospective multinational study.
Mild knee osteoarthritis predicts dissatisfaction after total knee arthroplasty: a prospective study of 186 patients aged 65 years or less with 2-year follow-up.
Population trends in the incidence and initial management of osteoarthritis: age-period-cohort analysis of the Clinical Practice Research Datalink, 1992-2013.
00111-7&id=gr1.jpg){kind=link}