Epidemiology of ANCA-associated Vasculitis

The epidemiology of the antineutrophil cytoplasm antibody (ANCA)-associated vasculitides (AAV), comprising Wegener’s granulomatosis, microscopic polyangiitis, and Churg-Strauss syndrome, poses considerable challenges to epidemiologists. These challenges include the difficulty of defining a case with a lack of clear distinction between the different disorders, case capture, and case ascertainment. The AAV are rare and therefore a large population is required to determine the incidence and prevalence, and this poses questions of feasibility. Despite these difficulties a considerable body of data on the epidemiology of the AAV has been built in the past 20 years with an interesting age, geographic, and ethnic tropism gradually being revealed. Most of the data come from White populations of European descent, and the overall annual incidence is estimated at approximately 10–20/million with a peak age of onset in those aged 65 to 74 years.

The antineutrophil cytoplasm antibody (ANCA)-associated vasculitides (AAV) comprise Wegener’s granulomatosis (WG), microscopic polyangiitis (MPA), and Churg-Strauss syndrome (CSS). For historical reasons polyarteritis nodosa (PAN) is often considered with the AAV although the presence of ANCA is now not considered to be a feature of PAN.

The epidemiology of the AAV poses considerable challenges to epidemiologists. The first is the difficulty of defining a case with a lack of clear distinction between the different disorders. There are 2 main systems of case definition or classification in current use: the American College of Rheumatology (ACR) (1990) classification criteria and the Chapel Hill Consensus Definitions (CHCC). There are several problems with these when used for epidemiology purposes. MPA does not feature in the ACR system but does in the CHCC and neither system use ANCA as a criterion. The CHCC were intended as definitions only and not classification criteria. Hence there are no validated classification criteria for MPA. To overcome this many studies have used both in parallel, but this leads to considerable overlap between categories. To improve the situation, an algorithm was devised by international consensus to incorporate both systems and this has been validated in 2 separate populations and shown to reliably classify patients with AAV into WG, MPA, CSS, and PAN with a minimum of unclassified patients.

The second difficulty is case capture. The AAV are rare and therefore a large population is required to determine the incidence and prevalence, and this poses questions of feasibility. A large population increases the risk of incomplete case detection but permits a reasonable number of cases to be collected in a practicable time frame; whereas a smaller population requires a much longer time frame to collect the necessary cases, which also may not be feasible. Statistical methods of capture-recapture analysis enable estimates to be made of the number of missing cases.

The third difficulty is case ascertainment. The AAV are rare potentially life-threatening conditions and therefore usually come to the attention of physicians. Ascertainment of cases can therefore be achieved by monitoring clinical facilities and hospital activity statistics. The AAV are multisystem and therefore surveillance of many different specialties is necessary. However, patients with fulminating disease may die before diagnosis and not be ascertained.

The rarity of the conditions makes prospective case-control studies difficult to conduct because the population size required to achieve statistical confidence is in excess of that readily available. Thus, much of the data on risk factors are derived from retrospective studies with inherent potential bias.

Despite these difficulties, a considerable body of data on the epidemiology of the AAV has been built in the past 20 years. However, much of the data comes from White populations of European descent. There are relatively few studies from non-White populations and none from Africa or the Indian Subcontinent.


There is a broad consensus that for primary, systemic, medium- and small-vessel vasculitis (including WG, CSS, PAN, and MPA) the overall annual incidence is approximately 10 to 20/million and the peak age of onset is 65 to 74 years ( Table 1 ).

Table 1

Incidence of AAV

Place Period Incidence (Per Million) References
Australian Capital Territory, Australia 1995–1999 17.0
2000–2004 16.2
Lugo, Spain 1988–1994 13.0
Norwich, UK 1988–2008 20.1
Crete, Greece a 1995–2003 19.5
Sweden 1997–2006 21.8

a Primary systemic vasculitides including Henoch-Schönlein purpura.


In 1936, Wegener first described a disease characterized by necrotizing granulomata of the upper and lower respiratory tract, focal glomerulonephritis, and necrotizing systemic vasculitis. The annual incidence of WG in the past decade has been estimated to be 8 to 10/million. WG is slightly more common in men than women.


WG is generally considered to be rare in childhood with an incidence of 0.3/million. However, a recent Canadian study in the Southern Alberta childhood population reported that the average incidence of childhood WG during 1993 to 2008 was 2.75/million/y, which is comparable with the incidence observed in adults. This was driven primarily by a high incidence in the last 5 years of the study of 6.39/million/y. This could be a regional phenomenon and further studies in pediatric populations are needed to see if this is indeed a global trend.

Many series report a peak age of onset of 64 to 75 years. A recent Swedish study reported the peak age to be more than 75 years; this is probably because of increased recognition of vasculitis in the very elderly.

Time Trends

The incidence of WG seems to have increased from the 1980s to 1990s. An increase in the annual incidence of WG from 0.7/million (1980–1986) to 2.8/million (1987–1989) was reported by Andrews and colleagues in Leicester, United Kingdom. Studies conducted during the late 1990s and 2000s in the United Kingdom suggest a relatively stable incidence. No significant change was noted during the 20-year period of a study conducted in secondary care by the Norwich group in the United Kingdom. There was an increase in the annual prevalence from 28.8/million in 1990 to 64.8/million in 2005 in a primary care population. This was felt to be to the result of improved outcome because of improved treatment regimens.

Some Scandinavian studies suggest that the incidence of WG has increased in the last 2 decades with little change in clinical symptoms at presentation. In particular, a recent study from Finland reported an increase from 1.9/million in 1981 to 1985 to 9.3/million in 1996 to 2000 with a similar increase up to 1998 in Tromsø in northern Norway reported by Koldingsnes and Nossent. In contrast, Swedish and German studies did not report such an increase.

There are several possible explanations for these conflicting data. The earlier increase may be related to better case recognition after the introduction of ANCA testing in the 1980s and early 1990s, but this effect should have diminished by now. Since the 1990s the ACR classification criteria or CHCC definition have been used for most studies. The use of different criteria will result in different estimates of incidence. Some studies have used International Classification of Diseases (ICD) codes for case identification, assuming diagnostic validity based on previous studies. In other studies, when assigned ICD codes for AAV in clinical registries were used to capture cases, only about 60% of identified cases fulfilled ACR criteria or were validated in accordance with the consensus algorithm. These methodological differences may have resulted in a variation in the reported incidence rates. Overall it is likely that there has not been a significant increase in the 1980s and early 1990s.

Geographic Factors

Most of the studies have been conducted in Europe, United Sates, Australia, and Japan in a secondary care setting ( Tables 2 and 3 ). The annual incidence of WG since 1986 in Europe is in the range of 2 to 10/million. The only study from a primary care setting suggests that the incidence is 8.4/million, indicating that the secondary care figures accurately reflect the occurrence in the community at large.

Table 2

Annual incidence of WG

Year Place Criteria Incidence (Per Million) References
1980–1986 Leicester, UK Fauci a 0.7
1987–1989 Leicester, UK Fauci 2.8
1988–2008 Norwich, UK EMEA 10.8
1990–2005 UKGPRD, UK ACR 8.4
1992–1996 Kristiansand, Norway ACR 6.6
1984–1998 Tromso, Norway ACR 8.0
1971–1993 Lund, Sweden ACR 2.1
1975–2001 Sweden ICD 0.78
1997–2006 South Sweden ACR, CHCC 9.8
1980–1985 Finland ACR 1.9
1996–2000 Finland ACR 9.3
1998–2002 Schleswig-Holstein, Germany CHCC 6–12
1995–2003 Crete, Greece ACR 6.6
1988–1997 Lugo, Spain ACR 4.8
1988–2001 Lugo, Spain CHCC 2.95
1990–1999 Vilnius, Lithuania ACR 2.1
1993–2004 Western Montana, USA ACR 8.6
1985–2004 South Australia ACR, ICD 11.2
1995–1999 Australian Capital Territory, Australia ACR 8.8
2000–2004 ACR 8.4
1990–2004 Lima, Peru CHCC 0.5

Abbreviations: ACR, American College of Rheumatology criteria (1990); CHCC, Chapel Hill Consensus Conference definition; EMEA, European Medicines Agency Algorithm; ICD, International Classification of Diseases; UKGPRD, United Kingdom General Practice Research Database.

a Fauci et al 1983.

Table 3

Prevalence of WG

Year Classification Place Prevalence (Per Million) References
2008 a EMEA Norwich, UK 130
1990 ACR UKGPRD, UK 28.8
2005 ACR UKGPRD, UK 64.8
2000 ACR Paris, France 23.7
1994 CHCC Germany (north) 58
1994 CHCC Germany (south) 42
1977–2001 ICD Denmark 100
2003 b ACR/CHCC Southern Sweden 160
1992–96 ACR Norway 53.0
1999–2003 ACR d Canterbury,
New Zealand
2003 c ACR d 93.5
1986–90 ACR New York, USA 32.0
1986–90 ACR USA 26.0
2004 ACR Western Montana, USA 90

Abbreviations: ACR, American College of Rheumatology criteria (1990); CHCC, Chapel Hill Consensus Conference definition; EMEA, European Medicines Agency Algorithm; ICD, International Classification of Diseases; UKGPRD, United Kingdom General Practice Research Database.

a Point prevalence on 31 December, 2008.

b Point prevalence on 1 January, 2003.

c Point prevalence on 31 December, 2003.

d Using unmodified ACR criteria.

There are several studies on the incidence of WG from Europe. In Finland, the annual incidence of the population increased from 1.9/million in 1981 to 1985 to 9.3/million in 1996 to 2000. Only minor changes in the signs and symptoms at diagnosis were observed in this 20-year period. In the population-based Swedish Inpatient Register during the period 1975 to 2001, the incidence of WG increased from 0.33/million in the period 1975 to 1985 to 0.77/million in 1986 to 1990, to 1.19/million in 1991 to 2001, resulting in a mean incidence of 0.78/million. In southern Europe (Greece) the overall annual incidence of primary systemic vasculitides (PSV) was 19.5/million. The incidence of WG was 6.6/million and was more prevalent in younger patients (<65 years old). In Spain the annual incidence of WG was estimated at 3/million.

The prevalence of WG has now been estimated in several populations. In the United Kingdom, the point prevalence on 31 December 2008 was 130/million. In Germany in 1994 the prevalence of WG in the north and south of the country was reported to be 58/million and 42/million, respectively. The estimated prevalence of WG in the United States was 26/million in 1986 to 1990. WG seems to be much less common in Japan than the United Kingdom. Although the overall occurrence of renal vasculitis is similar in Japan and the United Kingdom, the clinical phenotype is very different, with MPA predominating in Japan and WG being more common in the United Kingdom.

Epidemiologic studies in the Southern Hemisphere regions, using modified ACR criteria allowing for ANCA positivity in the absence of granulomatous vasculitis, have shown a 5-year period prevalence for WG at 152/million. The 5-year incidence of WG in South Australia is higher than that in the same latitudinal region in New Zealand. This geographic variation might be relevant to exposure to different environmental factors. O’ Donnell and colleagues in New Zealand described a positive north-south gradient in the incidence of WG. This supports a hypothesis of a latitude-dependent risk factor that can affect both global hemispheres. WG is more common in southeastern Australia than in southern Europe, whereas the opposite occurs in MPA. There is a trend for higher incidence of WG in rural compared with urban areas.

There are limited data from the Indian subcontinent; however, case series suggest that there is increasing recognition of WG.

Ethnic Factors

Most studies have been based on White population data. Data from the French prevalence study suggested that WG is less common than MPA in non-Europeans. The rate among Europeans was double that of New Zealander Maoris or Asians in New Zealand. In Japan, WG is not as common as MPA although the incidence of overall primary vasculitides does not differ from that in Europe. pANCA-associated disease is more frequent than cANCA-associated disease in the Japanese population, which is not the case in European or US patients. In China, WG also seems to be less common than MPA.


There are familial cases in WG, but they are rare. Familial heritability is similar to that seen in rheumatoid arthritis and it has been estimated at a relative risk of 1.56. Clusters of WG occurring in families usually involve no more than 2 affected members and those most commonly affected are first-degree relatives. Distant relatives are less affected, which is suggestive of a more dominant role for environmental factors.

HLA associations have been reported; the strongest for the AAV is with HLA DPB10401. Associations with polymorphisms in genes encoding key regulators of the immune response such as CTL4A and PTPN22 have been identified. Alpha 1-antitrypsine deficiency has been repeatedly reported as a genetic susceptibility factor.

Environmental Factors

The notion of seasonality has been suggested by reports of a higher incidence rate of onset in winter (29.8%) than in summer (14.3%). This has been variable and not consistent with studies in the United States. Difficulty in establishing an accurate date of onset of the disease process, methodological discrepancies, and variation of triggering factors in different geographic areas may explain this inconsistency. Lane and collegaues were unable to confirm the occurrence of seasonal onset for WG or any other AAV.

A north-south declining gradient in disease risk in the Northern Hemisphere and the opposite in the Southern Hemisphere suggest there is indeed a latitudinal variation in occurrence of WG and the other AAVs. This could be explained by ambient ultraviolet radiation, which was shown to have a protective immunomodulatory effect on the onset of WG. Possible mechanisms consider the effect of vitamin D on the immune system. However, ultraviolet radiation is known to be associated with autoimmunity especially in conditions mediated by TH1 cells, such as multiple sclerosis, type 1 diabetes mellitus, and Crohn disease.


Infection is closely related to vasculitis not only as a triggering factor but also as a consequence of intensive immunosuppression used for the treatment of primary vasculitides. De novo expression and relapse of vasculitis could be related to infection. The hypothesis that infection acts as a trigger for autoimmunity is not new. That could occur via direct microbial toxicity affecting the endothelium, either by invasion (eg, Rickettsiae, Bartonella, or cytomegalovirus) or the effect of microbial toxins. Humoral or cellular immune response to infections can lead to vasculitis via immune complex pathways (such as in hepatitis C and cryoglobulinemia).

For WG, the only clear microbial association is that with Staphylococcus aureus infection. S aureus is frequently isolated in cultures from the upper airways of patients with WG and nasal carriage has been linked with higher risk of relapse in WG. The relative risk for relapse is modulated by the presence and type of S aureus and was calculated at 3.2. The presence of toxic-shock-syndrome-toxin-1 increases the relative risk for relapse to 13.3.


No specific drug has been linked with WG. There have been several clinical observations of cocaine abuse followed by WG suggestive of an active induction of a proteinase 3 (PR3)-ANCA–positive vasculitis by cocaine. Detection of human leukocyte elastase (HLE)-ANCA in a patient with cocaine-induced midline destructive lesions mimicking the midfacial osteocartilagenous changes of WG may be a valuable diagnostic marker. It is rarely detected in primary WG or MPA, thus is suggestive of a drug-induced vasculitis. Drug-induced PR3-ANCA and myeloperoxidase (MPO)-ANCA may be associated with vasculitis mimicking AAV. The mechanisms involved in the break of tolerance and induction of ANCA (eg, superantigens, neutrophil extracellular traps, Th17 cells, protease-activated receptor-2) are likely to be shared in a degree between drug-induced or infection-induced and primary AAV.

Occupational Factors

Several studies have investigated possible links between occupation and development of vasculitis. A hospital discharge case-control study in Sweden with 2288 cases and 10 controls per case did not find any association with 32 occupations. Systemic vasculitis has been associated with exposure to particulate silica (eg, quartz, granite, sandstone, and grain dust). Pulmonary silicosis in individuals exposed to high levels of silica (eg, miners and quarrymen) has been shown to associate with systemic vasculitis.

Nuyts and colleagues found an odds ratio (OR) of 5.0 for silica exposure in 16 cases of WG compared with community controls. Lane and colleagues, in a case-control study in Norfolk, UK, reported an OR of 3.0 in 75 cases of AAV (47 cases of WG). There was no significant association between MPA or WG and occupational silica exposure. Hogan and colleagues found an OR of 4.6 for reported silica exposure in ANCA-positive patients (36 MPA, 21 WG, 8 necrotizing glomerulonephritis) compared with renal controls. In a population-based case-control study by Hogan and colleagues, silica exposure was found in 78 (60%) of 129 case patients and in 49 (45%) of 109 control subjects. The results showed no increased risk for disease from low/medium exposure relative to no exposure (OR 1.0) but identified increased risk with high exposure (OR 1.9).

A case-control study performed at the US National Institutes of Health (NIH) reported an association with inhaled fumes and particulates and pesticides with WG compared with healthy and rheumatic disease controls but not compared with respiratory disease controls. Nuyts and colleagues reported significantly raised OR for exposure to various metals and welding fumes (OR 2.0) in a group of renal patients including WG and glomerulonephritis. The Norfolk case-control study did not observe an association with occupational metal exposure and primary systemic vasculitis.

Links between occupational exposure to hydrocarbons such as paints and glues and systemic vasculitis have been studied by different groups with varying outcomes. Pai and colleagues reported significantly higher hydrocarbon exposure in male MPA and WG patients compared with matched blood donors and nonsignificantly greater exposures in female patients with pulmonary hemorrhage. Lane and colleagues estimated OR of 4.8 for exposure to occupational solvents compared with matched controls for PSV overall. Heavy metal exposure (mercury, lead) has been associated with WG. However, the large Swedish study did not find any association with relevant occupations.

Farming in the year before the onset of vasculitis has been associated with primary systemic vasculitis with an OR of 2.3 (WG 2.7 and MPA 6.3). The association appeared stronger in livestock than crops. This was not demonstrated in studies by the NIH and Knight and colleagues, which did not find a significant association between WG and farming.


The original description of periarteritis nodosa by Kussmaul and Maier was of a patient with inflammation and necrosis of medium-sized arteries leading to aneurysm formation and organ infarction. Davson and colleagues described patients with segmental necrotizing glomerulonephritis who also had features of PAN with extrarenal small and medium artery involvement. The term microscopic polyarteritis was used to describe these patients in whom the dominant feature was rapidly progressive renal failure. The current term for these patients is MPA. The dominant feature of PAN is organ infarction (intestine, nerve) as a result of involvement of medium-sized arteries. This illness is now termed classic PAN. The literature on the epidemiology of PAN and MPA has to be carefully interpreted, because many older studies used the term polyarteritis nodosa as a generic term for any form of necrotizing vasculitis. The accurate classification of patients as MPA is also difficult as there are no validated classification criteria for MPA. The introduction of the European Medicines Agency Algorithm (EMEA) has helped to ensure consistency of classification across studies.

Geographic/Ethnic Factors

The etiopathogenesis of the AAV is unknown and it is therefore difficult to differentiate the effects of ethnicity from environmental factors when studying the occurrence of the AAV in different populations.

Most studies on the epidemiology of vasculitis have been conducted in White populations, and there is relatively little data from outside Europe, United States, Australia, and Japan. There are very few studies looking at the occurrence of MPA in different ethnic groups in the same geographic region. From the data that are available it is clear that there are major differences between populations. The only study of a multiethnic population was conducted in Paris and showed that the prevalence of AAV in persons of European ancestry was twice (104.7/million) the rate in non-Europeans (52.5/million). MPA was more frequent than WG in non-Europeans. The non-European population was derived from the Maghreb, sub-Saharan Africa, Asia, and the Caribbean and comprised 28% of the study population (1.09/million).

The occurrence of MPA shows some striking variations in different populations. In Europe, the overall occurrence of AAV is roughly stable across populations; however, there is a north-south variation in the ratio of WG to MPA. In the northern European populations WG is more common than MPA; in southern European populations MPA is more common. There is an even more striking difference in the occurrence of renal AAV between the United Kingdom and Japan, together with a variation in the clinical phenotype of AAV between European populations and Japan. The authors have recently shown that ANCA-associated renal vasculitis in Japan is almost entirely caused by MPA, whereas in Europe only 50% is associated with MPA, the remainder being associated with WG or CSS. Renal vasculitis in Japan was exclusively caused by pANCA-MPO, whereas in Europe 33.3% is associated with the presence of cANCA-PR3 ( Table 4 ). The clinical features of the patients with renal vasculitis also support the rarity of renal vasculitis in WG in Japan, as upper respiratory tract manifestations and ear, nose, and throat features were much less common.

Oct 1, 2017 | Posted by in RHEUMATOLOGY | Comments Off on Epidemiology of ANCA-associated Vasculitis

Full access? Get Clinical Tree

Get Clinical Tree app for offline access