To register all newly diagnosed patients with primary systemic vasculitides (PSV) in a large region in northern Germany.
To register all newly diagnosed patients with primary systemic vasculitides (PSV) in a large region in northern Germany.
Between January 1, 1998 and December 31, 2002 all newly diagnosed cases of PSV were identified in a large mixed rural/urban northern German region consisting of 2,777,275 habitants in a population-based prospective study. The following sources were used: departments of all hospitals, including their outpatient clinics; all departments of pathology; and the reference immunologic labs serving the catchment area.
During 5 years, 642 PSV patients were identified. The incidence rates for all PSV were between 40 and 54 cases per 1 million and per year. People at age 50 years and older had a 3–5-fold higher risk of developing PSV compared with those younger than 50 years. The incidence rates of antineutrophil cytoplasmic antibody (ANCA)–associated PSV (Wegener's granulomatosis [WG], microscopic polyangiitis [MPA], Churg-Strauss syndrome [CSS]) were between 9.5 and 16/million/year. WG occured 2–3 times more frequently than MPA or CSS.
Results of a population-based vasculitis register over 5 years for the incidence of PSV among 2.78 million habitants in northern Germany revealed a stable incidence for all PSV. Compared with other European studies coming from small regions or referral centers, the incidence rates for ANCA-associated PSV were the same as in Norway, lower than those in United Kingdom, but higher than those in Spain.
Little has been published on the epidemiology of the primary systemic vasculitides (PSV), particularly on their incidence and prevalence. Most published data derive from referral centers or very small regions, possibly leading to referral or selection bias (1–4). An increase in the incidence has been observed for most types of PSV, conspicuously for giant cell arteritis (GCA) (5). An ascending incidence of small-vessel vasculitides has been reported since the discovery in the mid-1980s of highly specific autoantibodies (antineutrophil cytoplasmic antibodies [ANCA]) that are associated with Wegener's granulomatosis (WG), microscopic polyangiitis (MPA), and Churg-Strauss syndrome (CSS) (6, 7). Recently published data on the epidemiology of ANCA-associated PSV and polyarteritis nodosa (PAN) from 1988 to 1998 from the Norwich Health Authority, UK, showed a slow increase over time as well as an increase with age (4). In contrast, the incidence rates for these 4 entities remained stable in Spain over an observed period of 14 years (1, 4). The overall yearly incidence of WG, MPA, CSS, and PAN was similar in the UK and Spain, with 18.9 and 18.3 cases per million inhabitants, respectively. Considerable differences were found for the single entities: the incidence of WG in Spain was only half that in the UK and also nearly half that in Norway (8). WG appears to be less frequent in southern than in northern Europe. Similarly, GCA shows a higher incidence in northern Europe than in the Mediterranean countries (5, 9–13). Cotch et al report marked regional differences of the prevalence of WG in New York, ranging from 0 to 170 cases per million (14). In a previous population-based study, we showed statistically significantly higher prevalence rates in the cities of 2 German regions (northern and southern Germany) compared with the rural areas, particularly for GCA (15).
These results prompted us to establish a vasculitis register in northern and southern Germany on January 1, 1998. In a population-based epidemiologic study covering an area of 5 million inhabitants, all patients newly diagnosed with PSV according to the Chapel Hill Consensus Conference (CHCC) were included in the register (16). No difference of the incidences and the entities between northern and southern Germany were observed within the first 2 years (17). Because of the relatively short observation period to obtain reliable population-based data, the vasculitis register was continued in Schleswig-Holstein (population 2.78 million) and the results from 1998 until 2002 are presented in this article.
The catchment area comprised 2,777,275 inhabitants (federal state of Schleswig-Holstein); more details are shown in Table 1. The calculations of incidence rates were based on the population statistics dated January 1, 2000 (obtained from the Department of Vital Statistics of the federal state of Schleswig-Holstein). In this area, the population was stable with an immigration/migration rate of 6–7% during 1995 and 1998; for the population aged ≥50 years, the rate was only 2–3%.
|Population as of January 1, 2000||2,777,275|
|Age ≥50 years, %||37|
|% of all men||34|
|% of all women||40|
|Medical university, n*||2|
|Hospital departments, n*||122|
|Internal medicine, n (%)†||62 (51)|
|ENT‡, n (%)||14 (11)|
|Ophthalmology, n (%)||2 (2)|
|Dermatology, n (%)||2 (2)|
|Pediatrics, n (%)||10 (8)|
|Orthopedics/surgery, n (%)||18 (15)|
|Gynecology, n (%)||3 (2)|
|Neurology, n (%)||6 (5)|
|Pathology, n (%)||5 (5)|
Between January 1, 1998 and December 31, 2002 all newly diagnosed patients with PSV were registered.
Prior to the start of the vasculitis register, detailed information about the vasculitis register was published in regional physician journals as well as university-based journals. All possible facilities were notified by mail and finally confirmed their participation with a responsible physician in the facility. Cases of PSV were reported regularly from the following sources: all hospital departments, including their outpatient clinics, and the hospital where the authors are employed; all departments of pathology; and all reference immunology laboratories of the described region. At 3-month intervals, all sources were asked by mail (up to 3 times if there was no initial response) to screen for PSV diagnosed in the previous 3 months. In case of no response, the responsible physician was contacted by phone. All reported patients were reevaluated by the authors regarding diagnosis, sex, date of birth, address, and initials. According to the guidelines of confidentiality, patients were first reported anonymously with postal code, second initials of first and last name, and diagnosis. With agreement of the patients, additional medical data were collected (first symptoms of the disease, organ manifestations at diagnosis, ANCA titer, histology, therapy).
All newly registered patients received a disease information pack containing written information on their disease, the location of patient education courses, and the addresses of support groups. The study was approved by the institutional review board and ethical committee of the University of Luebeck as well as the office for data protection in Schleswig-Holstein.
The following types of PSV, defined by the 1992 CHCC (16), were registered: GCA, Takayasu arteritis, WG, MPA, CSS, Henoch-Schönlein purpura (HSP), (isolated) cutaneous leukocytoclastic vasculitis (CLA), classical PAN, and Kawasaki syndrome. In case of definite diagnosis of primary systemic vasculitis with no certain assignment to any of the previously described definitions, the patient was registered as unclassified vasculitis, e.g., isolated cerebral vasculitis. The diagnosis had to be made within the study period.
All patients with secondary vasculitides were excluded, e.g., in association with other inflammatory rheumatic diseases, malignancies, or infectious diseases including hepatitis-associated vasculitides. Patients diagnosed before January 1, 1998 and patients who did not reside in the catchment area were excluded from the analysis.
The annual incidence rates for the years 1998–2002 were calculated as the number of newly diagnosed cases per 1 million inhabitants for all cases per year and for all types of PSV. We calculated 95% confidence intervals (95% CIs) for all incidence rates. Furthermore, we performed analyses on sex- and age-related rates for the population aged ≥50 years.
Overall, 833 cases were reported: 191 of them were excluded either because the disease was diagnosed before 1998, the patient did not live in the catchment area, or the diagnosis failed the inclusion criteria. Finally, 642 patients were included in the analysis: 152 cases in 1998; 138 cases in 1999; 126 cases in 2000; 111 cases in 2001; and 115 cases in 2002 (see Table 2). The 642 patients were reported, on average, from 2 different sources (range 1–4). Laboratory investigations of ANCA (indirect immunofluorescence and enzyme-linked immunosorbent assay for proteinase 3 and myeloperoxidase), antinuclear antibodies, double-stranded DNA, and extractable nuclear antigen were performed in 262 (41%) of 642 of the cases in the immunology reference laboratory Bad Bramstedt (Prof. Dr. W. L. Gross).
|n (m/f)||Age||n (m/f)||Age||n (m/f)||Age||n (m/f)||Age||n (m/f)||Age|
|GCA||45 (11/34)||71 (52–89)||28 (9/19)||72 (59–85)||27 (5/22)||71 (60–81)||27 (11/16)||76 (35–93)||37 (12/24)||74 (58–88)|
|CLA||26 (12/14)||66 (17–89)||23 (7/16)||51 (9–96)||24 (14/9)||67 (34–90)||17 (11/6)||67 (28–91)||10 (3/7)||72 (46–83)|
|WG||21 (9/12)||61 (25–79)||18 (11/7)||61 (23–87)||24 (14/10)||58 (24–85)||24 (12/12)||58 (22–77)||33 (18/14)||56 (12–79)|
|CSS||0||2 (0/2)||60 (59–62)||4 (2/2)||52 (36–68)||4 (3/1)||57 (46–76)||5 (3/1)||58 (29–60)|
|MPA||8 (5/3)||67 (57–80)||8 (4/4)||66 (55–79)||9 (6/3)||73 (48–81)||5 (0/5)||74 (67–76)||7 (3/4)||63 (36–77)|
|HSP||20 (17/3)||13 (4–80)||28 (11/15)||8 (3–82)||17 (9/7)||6 (2–60)||20 (11/9)||7 (4–85)||8 (5/3)||8 (3–60)|
|PAN||3 (0/3)||67 (66–80)||4 (3/1)||57 (30–80)||2 (0/2)||48 (45–51)||1 (0/1)||52||1 (0/1)||62|
|TA||2 (0/2)||44 (44–45)||2 (0/2)||n.i.||1 (0/1)||39||1 (0/1)||31||1 (1/0)||46|
|Kawasaki||1 (1/0)||6||2 (1/1)||4 (3–6)||2 (1/1)||3 (2–5)||3 (2/1)||6 (3–7)||0|
|UV||26 (12/14)||63 (27–96)||23 (10/12)||62 (31–80)||16 (7/8)||46 (31–87)||9 (4/5)||61 (27–78)||13 (6/7)||62 (39–81)|
|Overall||152 (67/85)||138 (56/77)||126 (48/65)||111 (54/57)||115 (51/61)|
Details of the 642 included patients are presented in Table 2. The most frequent type of vasculitis was GCA in 1998 (45 cases); in 1999 GCA and HSP were most frequent, with 28 cases each; from 2000 through 2002, GCA was most frequent, with 27–37 cases. During the complete observation period, a minimum of 28 patients (1999) to a maximum of 45 patients (2002) were diagnosed with ANCA-associated PSV, thereof two-thirds to three-quarters were patients with WG (18 of 28 in 1999 and 33 of 45 in 2002).
The overall incidence for all PSV in the federal state of Schleswig-Holstein in 1998 was 54 cases per million (95% CI 39–68) (see Table 3). In 1999 it was 48/million (95% CI 34–61); in 2000 it was 45/million (95% CI 34–61); in 2001 it was 40/million (95% CI 31–50); and in 2002 it was 42/million (95% CI 31–52). Yearly incidence rates of ANCA-associated PSV (WG, MPA, CSS) between 1998 and 2002 were 11, 9.5, 12, 12, and 16/million inhabitants, respectively. WG was the most common ANCA-associated PSV with incidence rates as follows: 8/million in 1998, 6/million in 1999, 8/million in 2000, 9/million in 2001, and 12/million in 2002. MPA was reported more frequently than CSS; details are presented in Table 3.
|All PSV||54 (39–68)||48 (34–61)||45 (34–61)||40 (31–50)||42 (31–52)|
|GCA||17 (8–24)||9 (3–15)||9 (3–15)||10 (4–16)||13 (6–20)|
|≥50 years†||46 (33–59)||25 (14–36)||31 (18–44)||27 (16–38)||37 (23–50)|
|ANCA-associated PSV||11 (5–18)||9.5 (4–16)||12 (5–19)||12 (5–19)||16 (8–24)|
|WG||8 (2–14)||6 (1–11)||8 (2–13)||9 (3–15)||12 (5–19)|
|MPA||3 (0–6)||2.5 (0–6)||3 (0–6)||2 (0–4)||3 (0–6)|
|CSS||0 (0)||1 (0–3)||1.5 (0–3)||1 (0–3)||2 (0–4)|
|PAN||1 (0–3)||2 (0–4)||1 (0–2)||0.4 (0–2)||0.4 (0–2)|
|HSP||7 (2–12)||10 (4–16)||6 (3–9)||7 (4–11)||3 (0–6)|
|CLA||9 (3–15)||8 (2–14)||9 (3–15)||6 (3–9)||4 (1–7)|
|TA||1 (0–3)||1 (0–3)||0.5 (0–2)||0.4 (0–2)||0.4 (0–2)|
|Kawasaki||0 (0)||1 (0–3)||1 (0–2)||1 (0–3)||0 (0)|
|UV||9 (3–15)||8 (2–14)||6 (3–9)||3 (0–6)||5 (2–8)|
PSV are 2–5 times more common in patients aged ≥50 years compared with the population younger than 50 years (see Table 4). The increase in the incidence of PSV with age was observed in both men and women (Table 4).
|All PSV in patients <50 years||20.3 (11–29)||24.8 (15–35)||24.5 (12–37)||20.4 (14–27)||12.4 (6–19)|
|All PSV in patients ≥50 years||114.3 (93–135)||87.0 (69–105)||78.4 (61–96)||74.0 (57–91)||90 (71–109)|
|All PSV in men||48.9 (35–63)||42.2 (30–55)||42.0 (27–57)||39.7 (30–40)||37 (25–49)|
|All PSV in women||59.4 (44–74)||53.7 (39–68)||45.1 (29–61)||39.3 (29–50)||42 (29–55)|
|All PSV in men <50 years||26.1 (16–36)||20.7 (12–30)||20.0 (12–28)||20.0 (11–29)||17.9 (10–26)|
|All PSV in men ≥50 years||97.7 (78–117)||88.4 (70–107)||83.0 (71–95)||77.4 (52–103)||72.8 (56–90)|
|All PSV in women <50 years||14.0 (7–21)||29.2 (19–40)||26.8 (17–36)||21.5 (12–32)||8.2 (3–14)|
|All PSV in women ≥50 years||127.2 (105–149)||86.0 (68–104)||72.8 (60–85)||67.0 (46–88)||93.3 (74–112)|
|GCA in men ≥50 years||25.6 (16–36)||18.6 (10–27)||10.9 (7–15)||21.5 (8–35)||13.4 (6–21)|
|GCA in women ≥50 years||60.9 (46–76)||30.5 (20–41)||37.3 (28–46)||28.2 (14–42)||42.3 (30–55)|
The overall incidence rates for PSV were comparable in men and women, only slightly higher frequencies without statistical relevance were calculated for women (see Table 4). Likewise for the sum of all PSV, no differences were found in the younger or older population, except that GCA was diagnosed in women aged ≥50 years twice as often as in men of the same age group (see Table 4).
Details regarding age and sex distribution of the different PSV entities are described in Table 2. The regional distribution of all 642 patients for the complete observation period is demonstrated in Figure 1.
The vasculitis register is the first prospective population-based study over a period of 5 years covering a large region of ∼2.78 million inhabitants. A previous investigation in cooperation with a similar structured region in Baden-Württemberg over a period of 2 years (University of Freiburg) revealed almost identical incidence rates for both regions in 1998 and 1999 (17). Because 2 years is too short a time period for epidemiologic investigations of rare diseases to obtain valid data, the vasculitis register in Schleswig-Holstein was continued for 5 years. The overall incidence rates of all PSV varied from 40/million in 2001 to a maximum of 54/million in 1998 (Table 3). Publications from other groups report data only from single entities of PSV. Compared with other rheumatic inflammatory systemic diseases, the incidence rate of PSV approximates that of systemic lupus erythematosus (SLE) (18–21). Interestingly, we observed similar prevalence rates of PSV and SLE in a previous study (15, 18). In relation to other autoimmune diseases, the incidence rate of PSV is approximately twice that of systemic sclerosis (22, 23), IgA nephritis, or multiple sclerosis (21). Other European centers published data on the epidemiology of ANCA-associated PSV (1, 2–4, 8) with distinct methodologic differences. The results were derived from referral centers or population-based studies covering very small regions (see Table 5). Although the observation period in those investigations comprised 11–15 years, our study recruited considerably more patients within 5 years. In a Norwegian population-based study analyzing the incidence of WG over 15 years, only 55 patients were identified; in a British study over 11 years, 40 patients were diagnosed with WG; and in a Spanish study over 15 years, 12 patients were diagnosed with WG (see Table 5). During the observation period of 5 years in our present study, 120 patients were diagnosed with WG in the catchment area. Within the past 10–15 years, a remarkable increase in interest and a change in the diagnostic procedures regarding ANCA-associated vasculitides have led to the development and establishment of American College of Rheumatology (ACR) classification criteria (24) and the CHCC definition (16) for the various forms of PSV. Those criteria and the implementation of ANCA testing in the mid-1980s allowed the diagnosis of mild and abortive forms of PSV with a subsequent increase in the incidence rates. Therefore, a comparison of our data with the results of the other European studies appears to be inappropriate. The present study revealed incidence rates for ANCA-associated PSV between 9.5 and 16 new diagnoses per million per year (see Table 5) compared with 18 in the UK and 13 in Spain; patients with (classical) PAN were included in the latter studies, however. The application of the CHCC definition and classification criteria influenced the epidemiologic data of PSV considerably, particularly the diagnosis of MPA and PAN. In the present study, we applied the CHCC definition because, unlike the ACR 1990 classification criteria, the CHCC definitions incorporate immunologic and immunohistochemical findings and distinguish clearly between the presumably very different pathogenic entities MPA and PAN. Furthermore, with regard to cutaneous vasculitides, the CHCC definitions make a clearer distinction between CLA and HSP than the ACR criteria. The ACR criteria do not include MPA and CLA as distinct entities. Patients with GCA, TA, WG, CSS, and HSP met the CHCC definitions and ACR classification criteria (24).
|UK (ref.4)†||Spain (ref.1)‡||Norway (ref.8)||Germany (this study)|
|Study region||NHA, Norfolk||Referral center, Lugo||3 northern counties||Schleswig-Holstein|
|Study type||Population based||Hospital based||Population based||Population based|
|Study population, n||413,000||208,271||464,000||2,777,275|
|Types of vasculitis||WG, CSS, MPA, PAN||WG, CSS, MPA, PAN||WG||All PSV|
|Incidence rates of ANCA-associated vasculitides¶|
The most frequently observed PSV was GCA with 164 cases, followed by WG with 120 cases, CLA with 100 cases, and HSP with 93 cases. All ANCA-associated PSV (n = 172) were reported more often than GCA. The complete assessment of all cases of GCA represents one major limitation of this study. Studies from small regions assessing only the incidence of GCA revealed incidence rates of ∼300 cases of GCA in the population group age ≥50 years. These incidence rates were remarkably higher than the incidence rates in our study for GCA (5, 12). GCA frequently is diagnosed and treated in outpatients, and is not regularly confirmed by histology or ANCA testing. Therefore a certain number of GCA cases are not identified by our sources.
Although in our present study and in the—also population-based—British study WG was the most frequent ANCA-associated PSV, in Spain the incidence of MPA was twice that of WG (1, 4). In our study, WG represents two-thirds to three-quarters of all ANCA-associated PSV (see Table 2). As opposed to our population-based study covering 2.78 million inhabitants, the data of the Spanish study were collected from a referral center comprising a region of only 208,271 inhabitants. The importance of methodologic aspects in epidemiologic investigations was demonstrated in an earlier Swedish study analyzing the incidence of small-vessel vasculitides (3). Tidman et al (3) considered patients from a department of nephrology and identified more patients with MPA (n = 70) than WG (n = 19). Our data, as well as the British data, revealed several times higher incidence rates of WG compared with MPA (see Tables 2 and 5). The incidence rates of WG in Schleswig-Holstein were slightly lower than in the UK but remarkably higher than in Spain (with 2.95) and comparable with Norway (8). This result could reflect a real difference between northern and southern Europe, similar to that found for GCA, under consideration of the methodologic differences.
The demographic data of all enrolled patients (sex, age at diagnosis) remained constant over the observation period (see Table 2). Interestingly, the median age at diagnosis of patients with WG was 60 years (range 22–87 years) whereas cohorts, recruited 10–20 years earlier, observed a median age at diagnosis between 40 and 50 years (25–28). Furthermore, the time between occurrence of the first symptom and diagnosis has shortened: in this study to a median time of 3 months (range 0–48 months) compared with previous cohorts with a median time of 9 months (0–288 months) (28) and 15 months (26). Compared with our previously published cohort of 155 WG patients, an increase of the median age at diagnosis of ∼15 years was found despite the shorter period between first symptom and diagnosis by two-thirds. The reason could be an increased awareness regarding small-vessel vasculitides in the older patient. In contrast, WG patients from the Norwegian epidemiology study were stable from 1984 to 1998 with a median age of 50 years at diagnosis (4). No data are available from the British study regarding the age of WG patients. In a previous study by the same British group, assessing the same region (6), 21 WG patients were diagnosed at a median age of 62 years (range 33–82 years). The time between first symptom and diagnosis was only 2 months (range 0.5–40 months).
The present study has certain limitations. The incidence rates for GCA and possibly also for CLA appear to be lower than expected in comparison with other epidemiologic studies. The main objectives of this study were the confirmation of the identical incidence rates in northern and southern Germany observed within the first 2 years of the vasculitis register and the almost complete registration of ANCA-associated PSV. Because demographic factors and the incidence rates of all PSV and the single entities remained stable over 5 years, we conclude that the data represent reliable epidemiologic data for PSV in Germany.
In summary, the results of the prospective population-based vasculitis register Schleswig-Holstein reveal an incidence rate for PSV between 40 and 54 per million per year for a large region of 2.78 inhabitants. GCA represents the most frequent PSV in this population, but the incidence rates are certainly underestimated. Following GCA in frequency is WG, which occurs almost as frequently as GCA. The incidence of ANCA-associated PSV is even higher than that of GCA, with WG always between two-thirds and three-quarters in this group. Compared with other European countries, the incidence of ANCA-associated PSV is lower than in the UK, and slightly higher than in Spain. The incidence rate of WG was lower than in the UK but remarkably higher than in Spain and comparable with Norway. Whether these differences reflect real regional distinctions in the incidence of WG possibly due to various exogenous expositions, or are related to methodologic aspects in the investigations, remains unclear.
The authors would like to thank Mrs. Claudia Möck for her excellent assistance.