Rheumatoid arthritis (RA) is a complex disease that affects approximately 0.5% of the adult population worldwide, and occurs in 20–50 cases per 100 000 annually, mainly in women after their 40s. The onset of the disease has important diagnostic, prognostic and therapeutic implications and is yet to be defined. The distribution of the disease, in terms of both occurrence and clinical expression, has unclear geographical borders that may reflect differences in genetic admixture, environmental factors and socio-demographic determinants. Some diseases co-occur more frequently than expected with RA, as it is the case of cardiovascular disease, infections or lymphoma, but others in lower frequency than expected, such as cancer or schizophrenia. RA is associated with increased mortality rates compared with the general population in the majority of cohorts published, and the expected survival of RA patients is likely to decrease 3–10 years. As in the general population, the leading cause of death among patients with RA is cardiovascular disease, and deaths due to malignancy occur at a comparable incidence; however, patients with RA are at greater risk of mortality due to infection. Many genes have been implicated in the susceptibility of RA, all of which with a modest effect on isolation. Gene-environment interactions appear as the most plausible underlying cause of RA. Age, sex, smoking, shared epitope and others correlate with its RA. The most important determinants of prognosis in RA are the severity at presentation and the management of the disease, both of which are subject to inequalities.
Rheumatoid arthritis (RA) is not an uncommon disease that continues to present challenges to modern medicine; however, recent advances have contributed positively to its course. The epidemiology of RA is a fascinating one: a learning tool for epidemiology and for rheumatology.
Descriptive epidemiology
The challenges of the descriptive epidemiology of RA are common to chronic and complex diseases. Among other challenges, it is unclear when the disease actually starts; there is no consensus on how early we should be calling it RA. Further, many patients may actually be in remission at the time a prevalence study takes place, thus making the epidemiological count difficult.
Case definitions: early arthritis versus early RA
Therapeutic strategies for the treatment of RA have changed significantly since the turn of the century. The emphasis is now on early intervention with the aim of preventing disability and irreversible damage, resulting in a race to find the earliest point in time to treat the disease safely. Erosions, which lead to subsequent joint damage, occur early in the course of the disease and progress rapidly ; hence, a case definition of early RA should target RA before the development of erosions. Until now, the 1987 American College of Rheumatology (ACR) criteria have been used to define RA, but have been criticised for having a low ability to discriminate between patients who will develop persistent, disabling or erosive disease and those who will not be in early RA or early arthritis . Recently, the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) Rheumatoid Arthritis Classification Criteria, which focus on the early features of the disease, have been published . Clinical synovitis is central in the new criteria, which include the number of involved joints, serologic abnormality, elevated acute-phase and symptom duration.
The working definition for RA at its onset varies between studies. Even the name varies: ‘pre-RA’, ‘early RA’, ‘undifferentiated RA’, ‘early arthritis’, ‘undifferentiated arthritis’, etc. Definitions that do not include the term ‘RA’ are more sensitive but less specific, as they include those who as of yet have not developed any clear characteristics of RA, but may go on to. The problem arises when comparing studies, as suspected, early, undifferentiated arthritis may have a different definition in each one. Further, the terms ‘early’ and ‘recent onset’ have also been changing over the years, from 2 years to 4 weeks; thus, ‘early’ is becoming earlier. The new classification criteria, which focus on earlier stages of the disease, are welcomed. Given the methodologically rigorous development of the latest ACR/EULAR criteria, it would seem prudent to recommend these for any future population-based epidemiological surveys of RA incidence and/or prevalence.
Incidence
According to a 2005 systematic review and more recently published reports, the annual incidence rate of RA varies between 20 and 50 cases per 100 000 in North American and Northern European countries, and it may be lower in Southern European countries . Unfortunately, incidence studies are not available from developing countries.
Measuring disease incidence generally implies the establishment of disease registries. Most RA registries have been established in highly developed areas or are nested within epidemiological cohorts with well-established limits . Further, estimating the incidence of RA confronts us with two major problems; one is the delay between symptom onset and the medical consultation, and the second is that the case definition used most commonly in reported studies, the 1987 ACR criteria, depends on the time elapsed between symptom onset and assessment of RA criteria, and on how the criteria are applied.
The incidence of early arthritis is also being studied. The use of different case definitions makes the estimates vary as widely as 25–115 per 100 000 .
Prevalence
A systematic review undertaken in 2005 of the prevalence of RA found an estimated prevalence between 0.2% and 1.2% . With the exception of some African studies , RA has been found in all areas of the world in which it has been studied. The prevalence rates for females tend to be considerably higher than the rate for males.
While many textbooks report a worldwide prevalence of approximately 1%, this rate may be based on the prevalence found in US and UK studies. Higher prevalences of up to 7.1% have been reported in some settlements, such as closed American Indian tribes or Eskimos . Lower prevalence rates have been reported in countries around the Mediterranean , Central Europe , Asia and South America . These regional differences in the occurrence of RA and the changes in incidence over time are discussed in more detail below.
Interestingly, a study measuring the prevalence of RA in people migrating to England found that the observed prevalence of RA in these migrants was within the range of the country of origin rather than that of the English population .
Morbidity
Morbidity covers the spectrum of what can be related and not related to the expression of the disease. In general, the line between related and not related is becoming hazy. On one side, it is clear that the so-called extra-articular features of RA, such as rheumatoid nodules, and, perhaps, interstitial lung disease, are so named because of the high frequency of occurrence and the clear relation to RA activity and severity . However, recently, as inflammation has become a clear trigger for cardiovascular (CV) disease , some authors would include CV events among the extra-articular features of RA. In addition, treatment may cause adverse events that become another form of co-morbidity. In any case, it appears that patients with RA commonly have higher numbers of co-morbidities than patients with non-inflammatory diseases , even at presentation . Among other co-morbidities, RA patients show higher rates of infections, pulmonary and renal disease and CV events .
It should be taken into account that co-morbidity studies are far more difficult to compare and to pool than studies on the occurrence of RA, as the study subjects are not usually population- or community-based but hospital samples and clinical cohorts that may have variable degrees of baseline severity, activity and treatment. Further, it is important to note, the case definition for specific co-morbidities varies among studies. Therefore, conclusions from co-morbidity studies should be judged cautiously.
CV events
In the vast majority of studies, patients with RA show a higher risk of major adverse CV events than controls, and RA activity appears as a predictor of major adverse CV events independent of traditional CV risk factors . Patients with very early disease, or even pre-RA, may also be at risk . To put it in perspective, however, these risks are threefold smaller than those of systemic lupus erythematosus (SLE) . There is an ongoing debate on the confounding effect of adverse CV risk profile on the independent association of disease activity score and major adverse CV events in RA. Despite a large body of evidence in favour of a relation between inflammation and CV disease, there may be other factors, not only the classical ones. For instance, the higher occurrence of thyroid disease in RA may amplify RA CV risk . There are also emerging CV risk factors related to the inflammatory process. C-reactive protein (CRP) is a clear independent prognostic factor of CV mortality, even in subjects without RA . Rheumatoid factor (RF) is associated with a sixfold risk of CV death in RA, but also in non-RA populations . Homocysteine levels seem to play a leading role in patients treated with methotrexate, especially in those not receiving folic acid supplements . Other treatments may confer an additional risk, as is the case for non-steroidal anti-inflammatory drugs (NSAIDs) or corticosteroids.
Infections
The rate of infections, in general, is increased in RA compared with other diseases, but more important is the fact that specific opportunistic infections are increased, such as tuberculosis, around fourfold or herpes zoster , around twice the risk. Undoubtedly, the rates are clearly related to both the immunosuppressant drugs used in the treatment of RA and to the level of systemic inflammation.
Cancer
The overall incidence and mortality of cancer in RA is not greater than the expected, although there is an increased risk of haematopoietic and lung cancers in RA patients compared with the general population . The risk of lymphoma is clearly increased in RA, and is closely related to the degree of inflammation, even in early RA .
Depression and schizophrenia
Although most studies show an increase of depression in RA, the impact of the instrument used for the case definition may be too high . Schizophrenia shows the opposite pattern: its occurrence is reduced in RA .
Mortality
In the majority of cohorts published, RA has been shown to be associated with increased mortality rates compared with the general population. The results of studies differ significantly, with a range of Standardized Mortality Ratios (SMRs) from 0.87 to 3.0 being reported. Depending on the exclusion criteria used, the median and mean SMR associated with RA fall within a range of 1.5–1.8 . While treatment for RA has improved dramatically in recent years, this is yet to be consistently accompanied by a significant reduction in mortality rates, although it is possible that mortality is a lagging indicator . A recent study from the Netherlands compared the mortality rates in three cohorts of RA patients with the general Dutch population in the periods 1993–1995 (mainly NSAID use), 1996–1998 (mainly hydroxychloroquine and salazopyrin use) and 1999–2006 (methotrexate and biologic use). They found increased SMRs of 1.35 and 1.23 in the first two periods and a decreased SMR of 0.49 in the more recent period when treatment regimens have been more aggressive . In a prospective RA cohort from Australia, life tables were used to calculate median life years lost in patients with RA compared with population-based age and gender norms. Life expectancy was reduced by 8 years in males and by 9 years in females. The SMR for males in this population was 1.66 (95% confidence interval (CI) 1.3,2.02) and 1.81 (95% CI 1.53, 2.08) in females. In the 1980s disease-onset cohort, median life years lost was 8 years in males and 10 years in females, whereas in the 1990s disease-onset cohort median life years lost had dropped marginally to 6 and 7 years, respectively .
CV disease accounts for around 40% of all deaths in RA , with cancer (17%), infection (14%), musculoskeletal disease (9%), respiratory disease (9%) and renal disease (6%) causing the majority of other deaths . Despite total CV deaths being similar to that in the general population, they occur at an earlier age, meaning that RA is associated with a 60% increased risk of CV death compared with the general population. Notably, this trend has not changed over time . Risk is also increased for infection, pulmonary, gastrointestinal and renal diseases, with the risk of death from infection being particularly marked at greater than 10 times the rate in the general population . Most studies do not find excess cancer deaths though risk is increased for lymphoma and lung cancer .
The most significant predictors of mortality in RA are functional status (as measured by the Health Assessment Questionnaire (HAQ)), age, male sex and co-morbidity . Other predictors of mortality include: disease severity (joint count, erythrocyte sedimentation rate (ESR), RF-positive), socioeconomic factors and extra-articular disease . Although the absolute risk of mortality amongst those with RA is higher amongst those of male sex and older age, the relative risk of mortality are higher amongst women and young people . In addition to age and gender, the more recent studies suggest that mortality risk is associated with higher baseline HAQ scores and high CRP levels, supporting the argument for earlier treatment that treats to a target to reduce inflammation .
There are several limitations to mortality studies. Many studies have inconsistent case definitions, or are based on registries or clinicians notes, which rely on clinician diagnosis as opposed to a systematic case definition. Studies are often drawn from clinics or discharge registries, meaning they may not reflect all those with RA but only those who have RA severe enough to come in contact with the medical system. Studies also vary substantially in years of follow-up and in whether they are prospective or retrospective. Finally, RA mortality studies are drawn overwhelmingly from Western European countries and the USA, meaning that a global picture of RA mortality is still unclear. These factors help to explain both the variability associated with mortality studies as well as their limitations. Yet, further variability is likely to be seen in future estimates associated with variations in availability and uptake of new treatments.
Geographical variation
Geographical difference in occurrence or outcome
Studies of the incidence and prevalence of RA suggest variations between different populations even within the same country . Potential explanations include regional variation in behavioural factors, climate, environmental exposures, RA diagnosis and genetic factors . Traditionally, it has been stated that RA is less frequent in some countries than others, and the commonly reported prevalence is the rate found in US and UK studies. However, it may be that the estimates in the US, and perhaps in the UK, may be higher than expected, and the world prevalence is not as high as the 1% fixed in most textbooks. It has been reported that the prevalence in Northern European countries is also high, but again, this is debatable, as studies in Norway , Sweden and Lithuania , for instance, do not show such a pattern. A study from Greece has not shown the low prevalence that is commonly attributed to Mediterranean countries . The prevalence rates of RA in the different regions of the world are shown in Table 1 .
| Region & Country | Prevalence (%) | ||
|---|---|---|---|
| Total | Males | Females | |
| Asia | |||
| Japan | 0.20 | 0.14 | 0.25 |
| China | 0.20–0.34 | 0.0–0.76 | 0.0–2.23 |
| Taiwan | 0.16–0.65 | 0.36–1.20 | |
| India | 0.19–2.50 | 0.19 | 1.24 |
| Pakistan | 0.14 | 0.0–1.08 | 0.0–2.86 |
| Bangladesh | 0.0–0.20 | 0.56–1.20 | |
| Philippines | 0.17–0.24 | ||
| Thailand | 0.0 | 0.23 | |
| Indonesia | 0.20–0.30 | ||
| Malaysia | 0.09 | 0.21 | |
| Vietnam | 0.10 | 0.44 | |
| Europe | |||
| CzechRepublic | 0.0–0.77 | 0.11–2.88 | |
| Hungary | 0.23 | 0.48 | |
| Bosnia & Herzegovina | 0.46 | ||
| The Former Yugoslav Republic of Macedonia | 0.09 | 0.29 | |
| Russian Federation | 0.0 | 0.0–10.50 | |
| Lithuania | 0.0–1.56 | ||
| Spain | 0.20–2.70 | 0.20 | 0.80 |
| Sweden | 0.50–0.66 | 0.37–0.50 | 0.90–0.94 |
| Norway | 0.44 | 0.19 | 0.67 |
| France | 0.0–0.79 | 0.09 | 0.51 |
| Netherlands | 1.60 | 4.60 | |
| Greece | 0.57 | 0.30 | 1.00 |
| Ireland | 5.00 | ||
| United Kingdom | 0.30 | 0.02–2.18 | 0.12–2.99 |
| Middle East | |||
| Turkey | 0.38 | 0.15 | 0.77 |
| Saudi Arabia | 0.19 | 0.25 | |
| Iran | 0.10 | 0.55 | |
| Kuwait | 1.26 | ||
| Oman | 0.36 | ||
| North America | |||
| USA(incl. Alaskans, Pima Indians) | 0.73 | 0.43–1.52 | 0.00–3.79 |
| South America | |||
| Mexico | 0.30–2.46 | 0.20 | 0.40 |
| Chile | 1.33 | ||
| Argentina | 0.20 | 0.06 | 0.32 |
| Cuba | 0.48 | 1.20–1.72 | |
| Brazil | 0.13–2.09 | 0.09 | 0.68 |
| Australasia & Oceania | |||
| Australian Aborigines | 0.00 | ||
| Tokelau | 0.0 | 0.0–2.12 | |
| Africa | |||
| Lesotho | 0.0–13.0 | F: 0.0 – 6.90 | |
| Nigeria | 0.00 | ||
Studies focussed on the clinical expression of RA in populations also show variations. Studies in closed or isolated populations, such as those of American Indians and Alaska Natives, show more severe patterns . A high frequency of specific high-risk human leucocyte antigen (HLA) alleles in these populations may account for some of the increased risk, but other genetic factors are also likely to contribute. On the contrary, studies in Asia generally show less severe patterns, while African studies show conflicting results . Considerable variability has been found between studies, even between the same populations, with differences seen in clinical features, mainly in extra-articular disease, and considerable differences also evident in disease management and access to specialised health care . Variability in access to early diagnosis and the latest treatments for RA is likely to have a significant impact on the disability and mortality burden from RA and will influence regional variations in epidemiology estimates into the future.
What are the reasons?
A large part of the variability between studies, both of occurrence or of clinical expression, may be related to methodological differences. Some authors have highlighted the importance of data sources in the ascertainment of cases, and, thus, in the estimate of occurrence . Further, the age and sex distribution of the background population may be at play, as in the young and in males it is a rare disease . In addition, part of the variability may be due to the timing of the study with more recent studies showing lower prevalence rates in general. Changes in prevalence over time may not completely be related to time trends, but to a shift in criteria or to the methodology used in the reported studies.
The differences in the expression of the disease between populations may be a problem when applying RA criteria. This is the case for Asian populations, for instance, where the prevalence of nodules is low , or in relationship with RF positivity background rates in regions where RF may be associated with other chronic disease and infection. The use of the more specific anti-cyclic citrullinated peptide (CCP) antibody in the criteria may help to overcome this. The prevalence rates probably have a genetic basis, especially in endogamic societies, such as those of circumpolar natives or North-American Indians, although an environmental influence cannot be excluded.
In addition, several studies have highlighted the importance of the rural-urban gradient that has not been clearly understood . Some studies that show very low national prevalence estimates were actually conducted only in cities , sometimes even within a single city, thus showing the prevalence of urban RA, not national RA.
While some factors have been shown to be related to the severity of disease in some regions, these factors do not consistently affect occurrence or incidence of RA . Environmental factors have been studied in some detail, but are yet to yield any significant or consistent explanation for the onset or variations in RA. While spatial clusters of inflammatory polyarthritis have been seen, they have not provided an explanation for disease onset . A study on the effect of latitude and sun exposure also failed to show any significant association .
Geographical variation
Geographical difference in occurrence or outcome
Studies of the incidence and prevalence of RA suggest variations between different populations even within the same country . Potential explanations include regional variation in behavioural factors, climate, environmental exposures, RA diagnosis and genetic factors . Traditionally, it has been stated that RA is less frequent in some countries than others, and the commonly reported prevalence is the rate found in US and UK studies. However, it may be that the estimates in the US, and perhaps in the UK, may be higher than expected, and the world prevalence is not as high as the 1% fixed in most textbooks. It has been reported that the prevalence in Northern European countries is also high, but again, this is debatable, as studies in Norway , Sweden and Lithuania , for instance, do not show such a pattern. A study from Greece has not shown the low prevalence that is commonly attributed to Mediterranean countries . The prevalence rates of RA in the different regions of the world are shown in Table 1 .
| Region & Country | Prevalence (%) | ||
|---|---|---|---|
| Total | Males | Females | |
| Asia | |||
| Japan | 0.20 | 0.14 | 0.25 |
| China | 0.20–0.34 | 0.0–0.76 | 0.0–2.23 |
| Taiwan | 0.16–0.65 | 0.36–1.20 | |
| India | 0.19–2.50 | 0.19 | 1.24 |
| Pakistan | 0.14 | 0.0–1.08 | 0.0–2.86 |
| Bangladesh | 0.0–0.20 | 0.56–1.20 | |
| Philippines | 0.17–0.24 | ||
| Thailand | 0.0 | 0.23 | |
| Indonesia | 0.20–0.30 | ||
| Malaysia | 0.09 | 0.21 | |
| Vietnam | 0.10 | 0.44 | |
| Europe | |||
| CzechRepublic | 0.0–0.77 | 0.11–2.88 | |
| Hungary | 0.23 | 0.48 | |
| Bosnia & Herzegovina | 0.46 | ||
| The Former Yugoslav Republic of Macedonia | 0.09 | 0.29 | |
| Russian Federation | 0.0 | 0.0–10.50 | |
| Lithuania | 0.0–1.56 | ||
| Spain | 0.20–2.70 | 0.20 | 0.80 |
| Sweden | 0.50–0.66 | 0.37–0.50 | 0.90–0.94 |
| Norway | 0.44 | 0.19 | 0.67 |
| France | 0.0–0.79 | 0.09 | 0.51 |
| Netherlands | 1.60 | 4.60 | |
| Greece | 0.57 | 0.30 | 1.00 |
| Ireland | 5.00 | ||
| United Kingdom | 0.30 | 0.02–2.18 | 0.12–2.99 |
| Middle East | |||
| Turkey | 0.38 | 0.15 | 0.77 |
| Saudi Arabia | 0.19 | 0.25 | |
| Iran | 0.10 | 0.55 | |
| Kuwait | 1.26 | ||
| Oman | 0.36 | ||
| North America | |||
| USA(incl. Alaskans, Pima Indians) | 0.73 | 0.43–1.52 | 0.00–3.79 |
| South America | |||
| Mexico | 0.30–2.46 | 0.20 | 0.40 |
| Chile | 1.33 | ||
| Argentina | 0.20 | 0.06 | 0.32 |
| Cuba | 0.48 | 1.20–1.72 | |
| Brazil | 0.13–2.09 | 0.09 | 0.68 |
| Australasia & Oceania | |||
| Australian Aborigines | 0.00 | ||
| Tokelau | 0.0 | 0.0–2.12 | |
| Africa | |||
| Lesotho | 0.0–13.0 | F: 0.0 – 6.90 | |
| Nigeria | 0.00 | ||
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
