Internationally, prevalence estimates for osteoarthritis show wide variability depending on the age and sex of the studied population, the method of case identification used, and the specificity of joint sites included. Currently, there is no generally agreed “gold standard” for identifying cases of osteoarthritis in epidemiologic studies. Despite this lack of standardisation, it is consistently demonstrated in population-based studies, worldwide, that osteoarthritis prevalence is positively associated with increasing age and that the greatest disease burden is attributable to involvement of the hip or knee joints. To estimate the true burden of osteoarthritis involving the hips or knees, comprehensive accounting of all associated morbidity is required. The identification of modifiable risk factors for disease incidence and progression is needed.
Descriptive epidemiology
Case definitions
The major purposes of case definitions are to distinguish between individuals who have the disease from those who do not and to promote comparability between different research studies. Case-definition methods need to be distinguished from diagnostic criteria, which are used in clinical practice to differentiate between different types of illness to determine the appropriate management strategy and to enhance communication among the professionals. In research settings, medical diagnosis can also be used as a method of case identification .
In epidemiological studies, osteoarthritis cases can also be identified on the basis of radiographic examination (e.g., the Framingham Study) or self-report (e.g., the Norwegian Study of Musculoskeletal Disorders) , although variations exist within each of these broad categories. A recent systematic review of the literature identified 25 different criteria for identifying cases of knee osteoarthritis alone , while, in an earlier study , seven different criteria were identified for the classification of hip osteoarthritis. Currently, there is no generally agreed-upon ‘gold standard’ for identifying cases of osteoarthritis in epidemiologic studies.
Medically diagnosed osteoarthritis
In population-based morbidity surveys, cases of medically diagnosed osteoarthritis can be identified through a clinical examination of survey participants or by using medical administrative databases . The major advantages of osteoarthritis case-identification methods based on medical examinations are that these methods tend to have high inter-rater reliability and high validity. The sensitivity of clinical diagnosis for differentiating cases of symptomatic osteoarthritis from other types of joint problems varies between 89% and 94%, while the specificity varies between 86% and 98% .
Case-identification methods based on diagnostic criteria are especially useful in clinical studies and in therapeutic trials where it is important to distinguish individuals with symptomatic osteoarthritis from those who have other forms of joint conditions . However, the use of such case-identifying methods in population-based studies of morbidity is subject to limitations, as the diagnostic criteria tend to be biased towards more severe, established disease . While information on the clinical cases of osteoarthritis provides important insights into the burden of this condition on the health-care system, from the public policy standpoint, it would also be important to identify individuals with early and/or mild forms of osteoarthritis. This information would be useful for estimating the extent of future health-care needs and may provide opportunities for the prevention of osteoarthritis progression. The case-identification methods based on medical examination also tend to have low feasibility in population-based morbidity studies due to high costs, time needed to assess a large number of people and the need for highly trained, experienced clinicians to perform medical examinations of study participants during the data-collection phase .
Radiographic methods
One of the major advantages of radiographic case definitions is that radiographs form part of the diagnostic criteria for osteoarthritis and, as such, have high internal validity. In addition, at least three systematic reviews reported that radiographic methods of case identification have high inter- and intra-rater reliability . However, findings for the sensitivity and specificity of radiographic case-definition methods are less encouraging. In a systematic review of the validity of various radiographic case-definition methods, the sensitivity of radiographic osteoarthritis compared with clinical examination ranged between 54% and 78% and the specificity ranged between 33% and 78% . Higher values of sensitivity and specificity were generally obtained in the studies where more radiographs of the same joint were taken for each participant. Increased number of radiographs, however, would tend to add to the costs of what is already recognised as an expensive method of data collection. However, radiological surveys also tend to identify many cases of asymptomatic osteoarthritis . As these individuals are not experiencing pain or functional limitations related to their osteoarthritis, such cases would have low relevance for public-health policy, leading to inflated prevalence estimates of this condition. Although this problem can be overcome by taking into account the presence of clinical symptoms, radiographic methods of osteoarthritis case identification tend to be biased towards structurally advanced osteoarthritis. This means that symptomatic individuals with early osteoarthritis who have not yet shown structural changes would not be identified, which could lead to missed opportunities for prevention and early intervention .
Self-report
Major advantages of osteoarthritis case identification based on self-report are that it is inexpensive, easy to implement, does not require specialist training for individuals collecting data and poses no additional health risks to the participants . One of the frequently used self-report methods is based on the self-report of being previously diagnosed with osteoarthritis by a health professional . A problem with this method is that individuals are frequently unable to identify the specific rheumatic condition that affects them. In a survey of 472 rheumatology outpatients, only 65% correctly self-reported the specific type of arthritis diagnosis they had . Similar findings were obtained in a more recent study, where almost a quarter of individuals with severe osteoarthritis, who were scheduled to undergo a joint replacement surgery, did not correctly report the type of joint condition they had. Self-reported doctor-diagnosed (SRDD) osteoarthritis had moderate sensitivity (77%) but very low specificity (46%) . Both the sensitivity and the specificity of SRDD osteoarthritis among individuals aged ≥50 years were reportedly improved by an addition of a question regarding the presence of joint pain, stiffness or swelling . However, this method of osteoarthritis case identification had not yet been tested in the general population. One additional problem with the use of SRDD osteoarthritis is that it substantially underestimates the frequency of osteoarthritis among the study participants, with the prevalence estimates for SRDD osteoarthritis almost 20% lower than the prevalence of medically verified osteoarthritis (75% vs. 92%) .
Disease burden
Mortality
Osteoarthritis is not regarded as a fatal disease and only a handful of studies have examined mortality among individuals with osteoarthritis. In 2007, diseases of the musculoskeletal system were reported as the underlying cause of death in less than 1% of all registered deaths in Australia. A further 3% of registered deaths had diseases of the musculoskeletal system as a contributing or associated cause of death . The standardised death rate for arthritis and musculoskeletal diseases has increased from 4.3 per 100 000 population in 1998 to 4.6 per 100 000 population in 2007. The standardised death rate for males in 2007 was 3.6 per 100 000, and 5.3 per 100 000 for females. Of all deaths due to arthritis and musculoskeletal disease in 2007, 751 (69%) were females, predominantly in the 75–94 year-old age group.
Low overall mortality rates for osteoarthritis notwithstanding, at least two population-based studies have reported decreased survival for people with radiographic osteoarthritis. In a US-based First National Health and Nutrition Examination Survey (NHANES) , women with radiographic evidence of knee osteoarthritis (aged 55–59 years at baseline) had a higher risk of death than the general population, even after adjusting for age and co-morbidities. The second population-based survey, conducted in Finland, reported that adjusting for age, education, physical stress at work, smoking and body mass index (BMI), there was an increase in all-cause mortality for women who had osteoarthritis in at least two distal interphalangeal joints of each hand. For men, the presence of osteoarthritis in any finger joint was associated with increased cardiac mortality . One additional US study, where participants were 269 women with radiographically defined generalised osteoarthritis, reported that the probability of survival decreased by 1.5 years for each additional three joints affected by osteoarthritis . The strength of the association was not affected by BMI, smoking or co-morbidities.
Given the more recent findings of an increased risk of death due to gastrointestinal and cardiac events associated with non-steroidal anti-inflammatory drugs (NSAIDs), information on mortality among individuals with osteoarthritis might provide insights into the iatrogenic burden of this condition. However, the role of medication side effects in the increased mortality in osteoarthritis has not yet been fully investigated. Deaths that occur due to medication side effects may not be attributed to osteoarthritis in the official Deaths Registries. Hence, taking into account only those deaths that list osteoarthritis as the immediate, underlying or contributory cause could underestimate the mortality burden due to the treatment of this condition. In addition, while osteoarthritis does not cause death directly, it limits mobility and physical activity, thus increasing the risk of obesity and cardiovascular diseases . The effect on the mental health of patients with osteoarthritis has not previously been considered as a possible cause of increased mortality in osteoarthritis. This is despite evidence that chronic pain increases all-cause mortality . Clarification of the relationship between mortality and osteoarthritis would require a prospective cohort study carefully designed to control for several potential confounders, such as age, BMI, medication use and co-morbid conditions.
There is also a small but real short-term increase in the perioperative risk of death associated with all-joint replacement surgery for OA including anaesthetic risk, complications during surgery and complications during recovery. This is recorded as 30-day or 60-day mortality and is reported to be approximately 0.75% but does increase with age. The improvements in quality of life and mobility must be weighed up against this small increase risk of death . Favourable patient selection means that patients undergoing total joint replacement for osteoarthritis tend to have better general health than their age- and sex-matched peers, which makes it difficult to interpret the full impact of joint arthroplasty on mortality .
Morbidity
Although mortality has limited usefulness in quantifying the burden of osteoarthritis, the non-fatal nature of this condition means that individuals will live for many years in a lasting state of disability, thus contributing to morbidity estimates. The major indicators of morbidity at population level are prevalence (number of individuals with existing disease at a given point in time) and the joints involved. The disability morbidity associated with osteoarthritis is mostly attributable to the large weight-bearing joints, the hips and knees. An important aspect of morbidity studies is the method used to identify individuals with hip and knee osteoarthritis in the general population, as different methods are likely to yield a different group of individuals .
Disease burden
Mortality
Osteoarthritis is not regarded as a fatal disease and only a handful of studies have examined mortality among individuals with osteoarthritis. In 2007, diseases of the musculoskeletal system were reported as the underlying cause of death in less than 1% of all registered deaths in Australia. A further 3% of registered deaths had diseases of the musculoskeletal system as a contributing or associated cause of death . The standardised death rate for arthritis and musculoskeletal diseases has increased from 4.3 per 100 000 population in 1998 to 4.6 per 100 000 population in 2007. The standardised death rate for males in 2007 was 3.6 per 100 000, and 5.3 per 100 000 for females. Of all deaths due to arthritis and musculoskeletal disease in 2007, 751 (69%) were females, predominantly in the 75–94 year-old age group.
Low overall mortality rates for osteoarthritis notwithstanding, at least two population-based studies have reported decreased survival for people with radiographic osteoarthritis. In a US-based First National Health and Nutrition Examination Survey (NHANES) , women with radiographic evidence of knee osteoarthritis (aged 55–59 years at baseline) had a higher risk of death than the general population, even after adjusting for age and co-morbidities. The second population-based survey, conducted in Finland, reported that adjusting for age, education, physical stress at work, smoking and body mass index (BMI), there was an increase in all-cause mortality for women who had osteoarthritis in at least two distal interphalangeal joints of each hand. For men, the presence of osteoarthritis in any finger joint was associated with increased cardiac mortality . One additional US study, where participants were 269 women with radiographically defined generalised osteoarthritis, reported that the probability of survival decreased by 1.5 years for each additional three joints affected by osteoarthritis . The strength of the association was not affected by BMI, smoking or co-morbidities.
Given the more recent findings of an increased risk of death due to gastrointestinal and cardiac events associated with non-steroidal anti-inflammatory drugs (NSAIDs), information on mortality among individuals with osteoarthritis might provide insights into the iatrogenic burden of this condition. However, the role of medication side effects in the increased mortality in osteoarthritis has not yet been fully investigated. Deaths that occur due to medication side effects may not be attributed to osteoarthritis in the official Deaths Registries. Hence, taking into account only those deaths that list osteoarthritis as the immediate, underlying or contributory cause could underestimate the mortality burden due to the treatment of this condition. In addition, while osteoarthritis does not cause death directly, it limits mobility and physical activity, thus increasing the risk of obesity and cardiovascular diseases . The effect on the mental health of patients with osteoarthritis has not previously been considered as a possible cause of increased mortality in osteoarthritis. This is despite evidence that chronic pain increases all-cause mortality . Clarification of the relationship between mortality and osteoarthritis would require a prospective cohort study carefully designed to control for several potential confounders, such as age, BMI, medication use and co-morbid conditions.
There is also a small but real short-term increase in the perioperative risk of death associated with all-joint replacement surgery for OA including anaesthetic risk, complications during surgery and complications during recovery. This is recorded as 30-day or 60-day mortality and is reported to be approximately 0.75% but does increase with age. The improvements in quality of life and mobility must be weighed up against this small increase risk of death . Favourable patient selection means that patients undergoing total joint replacement for osteoarthritis tend to have better general health than their age- and sex-matched peers, which makes it difficult to interpret the full impact of joint arthroplasty on mortality .
Morbidity
Although mortality has limited usefulness in quantifying the burden of osteoarthritis, the non-fatal nature of this condition means that individuals will live for many years in a lasting state of disability, thus contributing to morbidity estimates. The major indicators of morbidity at population level are prevalence (number of individuals with existing disease at a given point in time) and the joints involved. The disability morbidity associated with osteoarthritis is mostly attributable to the large weight-bearing joints, the hips and knees. An important aspect of morbidity studies is the method used to identify individuals with hip and knee osteoarthritis in the general population, as different methods are likely to yield a different group of individuals .
Prevalence
Internationally, prevalence estimates for osteoarthritis show wide variability depending on the age and sex of the studied population, the method of case identification used and the specificity of joint sites included. There have only been a few, large, population-based surveys that have included a radiographic examination of the hip or knee allowing a case definition of symptomatic knee or hip osteoarthritis, if pain is present in a joint with radiographically evident osteoarthritis. As outlined above, a case definition requiring attainment of structural changes visible on radiographs is likely to underestimate the true disease prevalence as early osteoarthritis will be undetected. Furthermore, usually only one view of the joint is acquired in large, population-based surveys. On the other hand, prevalence estimates based only on symptomatic criteria will include many cases where most of the ‘pain, aching or stiffness’ in the region of the joint may not be attributable to osteoarthritis.
Symptomatic knee osteoarthritis and knee pain ( Table 1 )
Symptomatic knee osteoarthritis and knee pain are both highly prevalent among people aged 50 years and over worldwide, with a consistently higher prevalence among women compared with men . From these large population-based surveys that have included knee radiographs and restricting to within-country comparisons of similar gender and age cohorts, it appears that the prevalence of symptomatic knee osteoarthritis is higher in rural regions compared with urban regions. The prevalence of symptomatic knee osteoarthritis among people aged 60 years and over was almost double in Wuchuan county, a rural region in Northern China, compared with the prevalence found in Beijing ( Table 1 ). A similar difference is seen between the higher prevalence found in the Johnston County cohort, compared with that demonstrated in the NHANES-3 and the Framingham Osteoarthritis Study. However, there were major differences in the ethnic and socioeconomic mixes and case definitions used by these cohorts from the US, which may explain some of inconsistent prevalence findings. Within Europe, the prevalence of symptomatic knee osteoarthritis was markedly higher in the national or mixed urban–rural cohorts of Greece and Spain, compared with the purely urban cohort from Rotterdam.
| Country | Source (ref.) | Region | Cohort (N) | Age (years) | Symptomatic OA male/female | Knee pain male/female |
|---|---|---|---|---|---|---|
| China | Beijing OA | Urban | 1781 | 60+ | 7/18% | 17/30% |
| Wuchuan OA | Rural | 1027 | 50–59 60+ | 3/9% 13/27% | 44/56% | |
| Shanghai COPCORD | Urban | 2093 | 40+ | 4/10% | 6/13% | |
| Greece | ESORDIG | National | 8740 | 50–59 60–69 70–79 80+ | 3/11% 7/21% 15/30% 16/27% | Not available |
| Netherlands | Rotterdam | Urban | 2895 | 55–64 65–74 75–84 | <1/7% 5/11% 6/11% | 13/21% 10/24% 16/21% |
| Spain | Basque Country | Mixed | 7577 | 60–69 70–79 80–89 | 8/15% 10/17% 7/11% | 22/36% 27/46% 25/44% |
| USA | NHANES-3 | National | 3128 | 60–69 70–79 80+ | 10/14% 14/19% 15/20% | 16/21% 20/26% 23/28% |
| Framingham OA | Mixed | 1418 | 60–69 70–79 80+ | 6/8% 8/13% 5/16% | 16% (vs 10% knee OA) | |
| Johnston County | Rural | 3068 | 45–54 55–64 65–74 75+ | 6/7% 14/18% 21/20% 19/30% | 29/37% 38/45% 47/49% 38/64% |
As a larger proportion of participants in rural cohorts are required to have occupations with a heavy physical workload, often up to an advanced age, the higher prevalence of knee pain in the rural cohorts was expected; that almost 50% of people aged 60 years and over in these rural cohorts have knee pain “on most days” is surprising. While most of these large population-based studies required “pain on most days for a month or longer” for a participant to be identified as having knee pain, the Johnston County study and the Basque Country study also included the terms ‘aching’ and ‘stiffness’ to this case definition, a strategy likely to inflate prevalence estimates of knee (or hip) pain. However, the term ‘pain’ is a very individual assessment. It could also be argued that chronic aching and stiffness will inevitably lead to pain.
Symptomatic hip osteoarthritis and hip pain ( Table 2 )
The prevalence of symptomatic hip osteoarthritis was lower compared with symptomatic knee osteoarthritis within each of the cohorts with radiographs on both joints . In fact, symptomatic hip osteoarthritis was only found occasionally in three large urban cohorts in China and Vietnam . Among the European and US cohorts , the highest prevalence of symptomatic hip osteoarthritis was demonstrated in the Johnston County cohort. This higher prevalence may be partly explained by the broader case definition of pain to include ‘aching or stiffness’; however, the case definition used in the Basque Country study also included ‘aching or stiffness’. Some of the higher prevalence may be explained by the rural setting of the Johnston County study as farming has frequently been associated with a two- to threefold increased risk of radiographic hip osteoarthritis among men .
| Country | Source (ref) | Region | Cohort (N) | Age (years) | Symptomatic OA male/female | Hip pain male/female |
|---|---|---|---|---|---|---|
| China | Beijing OA | Urban | 1506 | 60+ | 5 cases only | 6/11% |
| Shantou COPCORD | Urban | 2040 | 16+ | 3 cases only | 13% | |
| Vietnam | Hanoi COPCORD | Urban | 2119 | 55+ | 3 cases only | 13/15% |
| Greece | ESORDIG | National | 8740 | 50–59 60–69 70–79 80+ | <1/2% 1/4% 1/4% 1/4% | Not available |
| Netherlands | Rotterdam | Urban | 2895 | 55–64 65–74 75–84 | 2/3% 2/6% 2/9% | 8/15% 8/17% 9/18% |
| Spain | Basque Country | Mixed | 7577 | 60–69 70–79 80–89 | 5/7% 8/9% 9/9% | 19/31% 21/32% 16/27% |
| USA | Johnston County | Rural | 3068 | 45–54 55–64 65–74 75+ | 7/4% 6/10% 8/11% 16/18% | 30/31% 30/40% 34/45% 40/45% |
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
