|Home | About | Journals | Submit | Contact Us | Français|
This article reviews the published risk factors associated with incident osteoarthritis of the lower extremity weight-bearing joints. Systemic risk factors include factors such as age, ethnicity, gender and genetic variables. Local risk factors are variables such as obesity, previous knee injury and occupational activities. Challenges in the study of incident osteoarthritis, and promising potential future study directions are also reviewed.
Osteoarthritis is a significant public health problem due to its major impact on disability and associated morbidities in the elderly. The incidence of osteoarthritis is expected to increase as the population ages to include a larger number of the elderly and with growing incidence of obesity throughout the world. The economic impact of osteoarthritis has also been estimated to be as high as 3% of the gross domestic product because of work-days missed due to joint pain. 
Despite the large number of people affected with OA, there are few known modifiable risk factors for incident OA of the hip or knee. While age is clearly a significant risk factor for the development of the disease, the association of incident osteoarthritis of the knee or hip with other systemic and local risk factors is not clear. This review will focus on published risk factors for incident disease in the knee and hip that may be relevant in clinical practice.
The definition of osteoarthritis varies in reported studies and can include self-reported osteoarthritis obtained from a questionnaire, radiographic definitions of osteoarthritis, and symptomatic osteoarthritis as defined by self-reported joint pain and radiographic evidence of osteoarthritis . In many of the larger cohort studies studies, radiographic osteoarthritis has been the preferred definition of incident osteoarthritis. Radiographic definitions are based largely on the Kellgren-Lawrence radiographic classification which grades the extent of radiographic osteoarthritis from 0 to 4, based on the presence and severity of individual radiographic features such as osteophytes and joint space narrowing.[4•] However, radiographic findings in osteoarthritis do not always correlate with clinical symptoms, and for that reason, studies which use the radiographic definition of OA only may not include patients with clinical disease. Studies may include different groups of patients depending on which definition of osteoarthritis is used.
The pathogenesis of osteoarthritis is likely multifactorial, and an interplay of systemic risk factors such as obesity or older age and local risk factors such as mechanical load both contribute to creating disease in any one joint. However, while systemic factors may predispose to an increased risk of osteoarthritis, local abnormal joint biomechanics, whether from injury or other cause, may be the initiator of a cascade of changes in the local joint environment that can, if unchecked result in osteoarthritis. While systemic factors such as age and sex are not modifiable risk factors, an understanding of their potential contribution to incident disease development can be helpful information for patients who want to know why they have the condition. Local factors, on the other hand, may have the potential for intervention and modification of disease risk, especially as more is understood about these risks.
Aging is thought to have an adverse effect on the ability of the joint to protect itself from biomechanical stress, perhaps because of changes in the articular cartilage (such as thinning of the non-calcified cartilage) or in increased joint laxity, which can predispose to injury through aberrant joint loading, past injury with resultant biomechanics, or other age-related factors. Increased age has been shown to be associated with an increased incidence of osteoarthritis in several studies.[5, 6] In an HMO-based survey of subjects a 10-fold increase in incident and prevalent osteoarthritic disease of the hand, hip and knee was found in subjects from ages 30 to 65 , with a steeper increase seen up to and around age 50, and a decline after age 70. The association of age with increased incident hip osteoarthritis has been shown to be modified by increased BMI, as in a recent study which modeled a lifetime risk of symptomatic knee osteoarthritis in a Johnston County cohort. Murphy et al. modeled the likelihood of developing symptomatic hip osteoarthritis in at least one hip by age 85, using longitudinal data from the Johnston County Osteoarthritis Project, and found that the lifetime risk of developing hip osteoarthritis was 25.3% (95%CI: 21.3–29.3), even after adjusting for race, BMI, sex or hip injury history.
Many of the better characterized studies of racial and ethnic differences in OA patterns come from large database studies. Reports on differences in knee osteoarthritis between African-Americans vs. Caucasians report a higher prevalence knee osteoarthritis in African-Americans. The NHANES-III data analysis reported a higher incidence of knee osteoarthritis in African Americans than Cacausians [OR = 1.65, 95%CI: 1.17–2.37], as well as more symptomatic knee osteoarthritis in African-Americans than Caucasians: [OR = 1.52, 95%CI: 1.06–2.19]. Results obtained from the Johnson County Osteoarthritis Project also support a 6% higher prevalence of radiographic knee osteoarthritis among African-Americans than Caucasians [32.4%(95%CI:29.8–35.1) v. 26.8%(25.3–28.4)]. Gait patterns may also differ between ethnic or racial differences in OA prevalence; a recent study by Chang et al., reported that Caucasians had a higher risk of valgus thrust with walking than African-Americans, with African-Americans possibly more prone to lateral compartment knee osteoarthritis than Caucasians. .
Differences in patterns of osteoarthritis between Caucasians and Han Chinese have also been characterized well by the Beijing Osteoarthritis Study and the Framingham Study. While the overall prevalence of radiographic knee osteoarthritis was found to be similar between Chinese and Caucasian men, Chinese subjects were found to have a higher prevalence of lateral compartment knee osteoarthritis compared to Caucasians, who had predominantly medial knee osteoarthritis. In addition to race, differences in OA pattern by gender also were observed to be different between Chinese women and Caucasian women, with a higher prevalence of radiographic knee OA in Chinese women compared to Caucasians (prevalence ratio = 1.45; 95%CI: 1.31–1.60), despite a lower body mass index in Chinese women.
There is some evidence that women have higher rates of incident knee osteoarthritis than men as they age. Data from the Framingham Knee Osteoarthritis Study, a population-based study of osteoarthritis, reported a 1.7 times higher incidence of osteoarthritis of the knee in women than in men (95%CI: 1.5–2.7).[11•] Similarly, women were estimated to have a lifetime increased risk of symptomatic knee osteoarthritis of 46.8% (95%CI: 41.2–52.5) The reasons for this difference are not completely clear, but an increase in osteoarthritis observed in women at the time of menopause has lead to the hypothesis that hormonal issues may play a role in the development of osteoarthritis. However, results from both observational studies and larger randomized trials on the association of estrogen and osteoarthritis have been mixed.
Observational studies have suggested a decreased risk of knee osteoarthritis in subjects who self-report taking estrogen. A study of the association of postmenopausal estrogen intake and radiographic hip OA suggested a protective effect of estrogen, with women currently using oral estrogen having a significantly reduced risk of any OA of the hip (OR = 0.62; 95% CI: 0.49–0.86). Studies of the potential association of estrogen and symptomatic or radiographic knee osteoarthritis have been conflicting. While some studies have not shown a significant protective effect of estrogen and incident radiographic knee osteoarthritis, others have found more evidence for a protective effect of estrogen. In the Women’s Health Initiative, a randomized double-blind, placebo-controlled trial of estrogen replacement, estrogen supplementation was found to be associated with a slightly significant decreased rate of either hip or knee arthroplasty, (HR = 0.84, 95% CI: 0.7–1.00), but the hazard ratios were insignificant in individual analysis of hip and knee arthroplasty. While these results are of interest, the association of estrogen and incident knee and hip osteoarthritis is still not clearly defined.
Several studies have suggested a protective effect of antioxidant vitamins such as Vitamin C and D in preventing progression of existing hip and knee OA, but protection against incident OA has not been clearly shown in large observational studies.[15, 16] As of this writing there is little evidence to support the routine use of antioxidants in the prevention of osteoarthritis of the knee or hip. Selenium is another dietary factor that has been shown to be associated with an increased risk of Kashin-Beck disease, a degenerative osteoarticular disease, and supplementation of selenium with iodine may have an improved effect on this condition.
Twin studies have shown an association between genetics and the risk of radiographic hand and knee OA in women. A genetic study of the Framingham cohort study showed a good fit for a Mendelian genetics model of genetic influence on the presence of radiographic hand and knee osteoarthritis in 337 nuclear families. Currently genome-wide association studies (GWAS) using the Rotterdam cohort have identified the 7q22 locus as a potential locus for polymorphisms associated with generalized osteoarthritis . Specific gene polymorphisms associated with proteins that regulate Wnt pathway signaling such as ANP32A , and polymorphisms of genes that encode proteins in the Transforming growth Factor-Beta (TGF-Beta) pathway have also been targeted for susceptibility gene studies, with promising results in the GDF5, ASPN and SMAD3 gene regions .These results bear close watching and may help to provide further insight into disease mechanism.
Local factors that affect the risk of incident hip or knee OA are thought to work through abnormal joint loading that alters local biomechanical forces across the joint.
While malalignment (hip-knee-ankle angle) has been shown to be a risk factor for the worsening of existing knee osteoarthritis the association of malalignment and incident knee has been conflicting. While a study of Rotterdam et al. found that baseline valgus malalignment was associated with a slightly increased risk of incident knee osteoarthritis, a similar analysis found that malalignment did not influence incident OA knee in the Framingham cohort, suggesting that malalignment is less important for incident disease.
Obesity has been shown to be associated with an increased risk of incident knee osteoarthritis in several studies [27–30]. A study from Finland reported on 823 subjects without baseline knee osteoarthritis, and a strong association with incident knee osteoarthritis and BMI was found, adjusting for age and gender (OR = 1.75; 95%CI: 1.0, 2.8), and a higher odds ratio (OR = 7.0; 95% CI: 3.5, 14.1) for subjects in the highest categories of BMI (BMI = 25–29.9 or >30.0) compared to subjects with BMI <25.0 . Similar results were found in the Framingham study, where 598 subjects without knee osteoarthritis were found to have an increased risk of incident knee osteoarthritis if they had a higher baseline BMI (OR = 1.6 per 5-unit increase; 95%CI: 1.2–2.2). These results are important, as they are consistent in their findings that heavier subjects have a high risk of incident knee osteoarthritis and point to a potentially modifiable risk factor when discussing this disease with patients.
Hip osteoarthritis, more than knee osteoarthritis, can be associated with local anatomic abnormalities. Subclinical anatomic abnormalities, i.e. abnormalities apart from the well-recognized diagnoses such as slipped capital femoral epiphyses (SCFE) have been shown to increase the risk of hip osteoarthritis. Lane et al. showed that a decrease in the center-edge angle or acetabular dysplasia, defined as an acetabular depth of <9 mm in older white women was associated with an increased risk of hip osteoarthritis (OR = 3.3; 95%CI: 1.1–10.1 and OR = 2.8; 95%CI: 1.0–7.9, respectively). Similar results have also been seen in other studies of women with hip osteoarthritis . While an earlier study of male subjects did not find an association of acetabular dysplasia with radiographic hip osteoarthritis in 1516 pelvic radiographs, a subsequent analysis using the Rotterdam cohort data found acetabular dysplasia to be an important risk factor in incident radiographic hip disease in both sexes . Variations in the shape of the femoral head have also been shown to be associated with an increased risk of incident hip osteoarthritis.[35, 36] Lynch et al. showed that subtle variations in an active-shape model of the femoral head accounted for up to 81% of the variance in femoral head shape in hips and that certain shapes had a higher association with incident hip osteoarthritis. These results are also important as they may explain early onset of osteoarthritis in the young patient without other clear risk factors for the disease.
For knee osteoarthritis, local knee injury has been shown to be a risk factor for incident knee osteoarthritis in several studies. Specifically, anterior cruciate ligament injury up to a decade previously in healthy soccer players was found to increase the prevalence of radiographic knee osteoarthritis. Meniscal resection has also been found to independently increase the risk of incident knee osteoarthritis (RR = 2.6, 95% CI: 1.3–6.1), with total menisectomy having a stronger risk of incident knee osteoarthritis up to 15–22 years after the resection . This risk factor can subsequently result in earlier-age onset of osteoarthritis, since menisectomies (either partial or total) are often done in the setting of acute injury in the younger, more active patient.
These results are important in informing our approach to a potential modifiable risk factor for incident knee disease. In younger athletes, efforts to prevent acute injury should be optimized, and if meniscal injury occurs, every possible effort should be made to preserve meniscal tissue to prevent accelerated disease.
Results of activities and association with incident knee or hip osteoarthritis are conflicting. High-level, intense participation in sports, such as that done by elite athletes has been associated with an increased risk of both hip and knee osteoarthritis in adults in some studies, [39, 40] such as a retrospective study of ex-athlete females in England who were found to have a higher incidence of osteophytes than their age-matched population controls in the tibio-femoral joint (OR = 3.57, 95% CI: 1.89–6.71). However, other studies of recreational runners have not found this consistent association.[41, 42] In a self-selected group of middle aged runners from a very active community in northern California with an average age of 63 years, participants were followed over a five-year period and did not develop accelerated clinical or radiographic osteoarthritis of the knees at a higher rate than controls . Similarly, observational studies using the Framingham cohort data did not show any additional risk or benefit associated with incident radiographic knee osteoarthritis and participation in activities such as recreational walking or jogging, even in subjects with a BMI >30 kg/m2 
Repetitive bending required by certain occupations has been shown to be associated with an increased risk of radiographic knee osteoarthritis [45, 46]. In one study, men whose jobs required knee bending had higher rates of subsequent radiographic knee OA than men whose jobs required neither (43.4 vs. 26.8%; OR = 2.22, 95% CI 1.38, 3.58). While occupation may not be a modifiable risk factor for all patients, knowledge of these patients’ increased risk of knee osteoarthritis could lead to a strategy to minimize other potentially modifiable risk factors, such as avoiding injury as much as possible or attempting more aggressive weight loss in these patients.
Osteoarthritis is a complex disease that is the result likely of a combination of systemic factors such as aging and local factors such as biomechanical injury from previous trauma resulting in abnormal load-bearing in the joint. Identifying consistent risk factors for incident disease of the large weight-bearing joints is challenging because of several factors such as varying definitions of disease and the long delay from the time of the exposure until the disease becomes clinically manifest . The more advanced imaging techniques for the joints may provide the ability to better correlate the structural changes in the joint to the clinical symptoms. Also, the increasing knowledge of genetic associations with the disease of OA may provide a more complete understanding of epidemiology of this disease.
Despite these limitations, there are several modifiable risk factors that have been consistently identified as associated with incident radiographic knee osteoarthritis, in particular: obesity, prior knee injury and repetitive bending. While these are not easily changed, targeting these known risk factors for incident disease may be a successful strategy for optimizing future disease outcomes.
Disclosure No conflicts of interest relevant to this article were reported.