Results

Table 3 shows the number of patients in each diagnostic category. Ninety seven patients had primary vasculitis and 21 patients had secondary vasculitis.

All patients with suspected vasculitis were included prospectively. After inclusion of 14 patients with giant cell arteritis that uniformly fulfilled the nomenclature it was decided to stop inclusion of more patients with giant cell arteritis.

The diagnosis of giant cell arteritis were in all cases based on characteristic biopsy findings. It is well known that in approximately 10–30 per cent of patients with symptoms and signs of temporal arteritis cannot be demonstrated by temporal artery biopsy.8 In this study no attempts were made to assess the number of patients with negative biopsy.

The organ involvements, the results of biopsies, and the presence of surrogate parameters in 27 patients with WG are shown in table 4. According to the Chapel Hill nomenclature the diagnosis of WG requires the demonstration of granulomatous inflammation of the respiratory tract, and necrotising vasculitis affecting small to medium sized vessels. Granulomatous inflammation of the respiratory system was demonstrated by surrogate parameter in all cases, but only confirmed by biopsy in 10 cases. Necrotising vasculitis was only biopsy verified in eight cases (patient numbers 2, 11, 22, 98, 99, 106, 120, 123). However, biopsy confirmed necrotising glomerulonephritis was demonstrated in 11 patients, and PR3-ANCA was positive in all but two patients.

According to the Chapel Hill nomenclature, the diagnosis of MPA, requires demonstration of necrotising small vessel vasculitis, not restricted to the skin, and without granulomatous inflammation in the respiratory system. Table 5 shows the organ involvement, the results of biopsies, and presence of surrogate parameters in 12 patients who were considered to have MPA. Only three patients (patient numbers 34, 85, 95) had small vessel vasculitis demonstrated extracutaneously, thus fulfilling the Chapel Hill nomenclature. However, in nine patients glomerulonephritis was demonstrated by biopsy or surrogate parameter. In patients numbers 89 and 94 the occurrence of haemoptysis and transient lung infiltrates indicated pulmonary capillaritis.

Table 6 shows the clinical findings in five patients with cryoglobulinaemia in the blood. All patients had involvement of the skin, and in all cases leucocytoclastic vasculitis was demonstrated by biopsy. Patient number 115 had biopsy confirmed crescentic glomerulonephritis, but in no other patients could vasculitis be demonstrated in any other organs. In all cases analysis of C3d showed increased values indicating complement activation. Further analysis of these patients showed that patient number 86 suffered from Waldenström's macroglobulinaemia, and patient numbers 17 and 56 had hepatitis C viraemia, as demonstrated by the presence of hepatitis C RNA by polymerase chain reaction in the blood. Patient number 29 fulfilled all diagnostic criteria for systemic lupus erythematosus complicated with secondary Sjögren's syndrome. Accordingly, only patient number 115 seemed to have essential cryoglobulinaemic vasculitis. However, repeated skin biopsies and the kidney biopsy failed to show immune deposits in vessel walls. As the diagnosis of essential cryoglobulinaemic vasculitis requires the presence of cryoglobulin deposits, it was concluded that, according to the Chapel Hill nomenclature, this patient should be diagnosed as MPA.

The greatest group of 37 patients had cutaneous leucocytoclastic angiitis, as demonstrated by biopsy. Twenty nine patients had palpable purpura, whereas skin manifestations dominated by ulcers were seen in only eight patients. Seven patients presented with rashes that suggested cutaneous vasculitis, which, however, could not be confirmed by skin biopsy. The number of patients in the remaining diagnostic categories was considered too few for further evaluation.

Discussion

The great number of publications on classification of primary and secondary vasculitis demonstrate the inherent problems in this field.9-15 Furthermore, diagnostic criteria have not been worked out, and major problems regularly arise in the clinic when a diagnostic decision must be made with respect to the individual patient.

The advantage of the Chapel Hill nomenclature is its clarity and simplicity. Ten primary vasculitis syndromes are defined according to vessel size. The small vessel vasculitides are subdivided according to the presence or absence of granulomatous inflammation in the respiratory system, eosinophilia, and immune deposits in the vessel walls.

The Chapel Hill nomenclature was not intended to function as classification or diagnostic criteria. However, the suggestion that certain clinical findings could act as surrogate parameters implied to us that the Chapel Hill nomenclature could be used as diagnostic criteria in the daily clinic. The main purpose of this study was to test this hypothesis.

In the clinical evaluation of patients with suspected vasculitis it is essential to obtain representative biopsy specimens to (1) confirm the vasculitis diagnosis, (2) reveal the presence of granulomatous inflammation, and (3) assess the presence of immune deposits in the vessel walls. However, because of the segmentary and dynamic nature of the vasculitis process, it is often difficult in practice to obtain representative biopsies. Furthermore, in patients with lung involvement transbronchial or percutaneous biopsies often are not informative regarding the presence of chronic granuloma, although such biopsies may be essential for excluding neoplasia and chronic infection. In such cases open lung biopsy is clearly indicated, but regularly not feasible because of the inherent risk of this procedure in vulnerable patients with debilitating disease. Therefore, it was decided to consider longstanding lung infiltrates or cavitation as surrogate parameters for granulomatous inflammation in the lungs, if malignancy and infection were excluded.

Similarly, biopsies in the ENT area regularly fail to establish the presence of granulomatous inflammation, but helped ruling out malignancies and chronic infection.

Accordingly, bloody nasal discharge or crusting, chronic sinusitis, otitis and/or mastoiditis, bone and/or cartilage destruction or acute hearing loss were considered manifestations of granulomatous inflammation, if malignancy and infection could be excluded. Angiographic or ultrasonic evidence of vasculitis was accepted in patients with other signs of vasculitis in the form of general symptoms and/or organ symptoms. Finally the presence of haematuria and proteinuria and red blood cell casts was considered a sign of glomerulonephritis.

In this study the diagnosis of giant cell arteritis was accepted only if biopsy of the temporal artery had demonstrated characteristic findings. Similarly, only cases with biopsy confirmed vasculitis in the skin were included in the study. In both diseases it must be accepted that approximately 10–30 per cent of biopsies fail to confirm the clinical diagnosis.

With respect to the diagnosis essential cryoglobulinaemic vasculitis no patient fulfilled the criteria suggested in the Chapel Hill nomenclature. Most of the patients with vasculitis and cryoglobulinaemia in the blood were secondary vasculitides, indicating the need in such patients to perform an intensive investigation to exclude the possibility of secondary vasculitis.

Application of the Chapel Hill nomenclature supplemented with the surrogate parameters described above proved insufficient in establishing the diagnoses of WG and MPA. It appears from table 4 that the simultaneous demonstration of granulomatous inflammation that involved the respiratory tract and necrotising inflammation affecting small to medium sized vessels, was only possible in 8 of 27 patients believed to have WG.

Table 5 shows similar data for the MPA group. In only 3 of 12 cases could extracutaneous small vessel vasculitis be demonstrated in patients with no signs of granulomatous inflammation in the respiratory tract, thus fulfilling the Chapel Hill inspired diagnostic criteria.

Eleven WG patients and nine MPA patients had glomerulonephritis shown by biopsy or surrogate parameter, which in the clinical decision process was of decisive importance for the diagnosis. Furthermore, all but two of the 27 WG patients were PR3-ANCA positive, which was also of importance for the diagnostic decisions.

Our conclusion is that the Chapel Hill nomenclature, even supplemented with surrogate parameters, fails to act as diagnostic criteria in the diagnosis of WG and MPA.

We propose the following diagnostic criteria for WG:

(1) Biopsy or surrogate parameter for granulomatous inflammation in the respiratory system and

(2) Biopsy verified necrotising vasculitis in small to medium sized vessels or biopsy/surrogate parameter for glomerulonephritis or positive PR3-ANCA test and

(3) Lack of eosinophilia in blood and biopsy samples.

It should be noted that these criteria were fulfilled by all of the patients (27 of 27) considered to have WG. At the same time, none of the patients with diagnoses other than WG fulfilled these criteria.

We propose the following diagnostic criteria for MPA:

(1) Biopsy verified necrotising vasculitis in small vessels and/or glomerulonephritis with few or no immune deposits and

(2) Involvement of more than one organ system as indicated by biopsy verified vasculitis in small to medium sized vessels or surrogate parameter for glomerulonephritis and

(3) Lack of biopsy and surrogate parameter for granulomatous inflammation in the respiratory system.

These criteria were fulfilled by 9 of 12 patients diagnosed with MPA, while none of the patients with diagnoses other than MPA fulfilled these criteria.

We are currently planning to test the diagnostic criteria proposed for WG and MPA in a prospective study where patients are collected from all major university hospitals in Denmark.