Ann Rheum Dis 68:564-567 doi:10.1136/ard.2008.099051
  • Clinical and epidemiological research

High frequency of venous thromboembolic events in Churg–Strauss syndrome, Wegener’s granulomatosis and microscopic polyangiitis but not polyarteritis nodosa: a systematic retrospective study on 1130 patients

  1. Y Allenbach,
  2. R Seror,
  3. C Pagnoux,
  4. L Teixeira,
  5. P Guilpain,
  6. L Guillevin,
  7. for the French Vasculitis Study Group
  1. The National Referral Center for Necrotizing Vasculitides and Systemic Sclerosis, Department of Internal Medicine, Université Paris Descartes, Hôpital Cochin, Assistance Publique–Hôpitaux de Paris, Paris, France
  1. Dr Christian Pagnoux, Department of Internal Medicine, Hôpital Cochin, 27, rue du faubourg Saint-Jacques, 75679 Paris Cedex 14, France; christian.pagnoux{at}
  • Accepted 29 October 2008
  • Published Online First 17 November 2008


Objective: To determine the frequency and risk factors of venous thromboembolic events (VTE) in Wegener’s granulomatosis (WG), microscopic polyangiitis (MPA) and, the so far unstudied, Churg–Strauss syndrome (CSS) and polyarteritis nodosa (PAN).

Methods: Retrospective, systematic analysis and comparisons were made between the characteristics of patients in the VTE group and non-VTE group. 1130 patients with WG, MPA, CSS or PAN were identified from the French Vasculitis Study Group cohort.

Results: During a mean follow-up of 58.4 (45.8) months, 83 VTE occurred in 74 (6.5%) patients, with a median vasculitis–VTE diagnosis interval of 5.8 months (−3 to +156). VTE occurred in seven of 285 (2.5%) patients with PAN, 19 of 232 (8.2%) with CSS, 30 of 377 (8%) with WG and 18 of 236 (7.6%) with MPA. Multivariate analysis retained age, male sex or previous VTE or stroke with motor deficit as being associated with a higher VTE risk. The adjusted odds ratio (95% confidence interval) for VTE was 2.88 (1.27 to 6.50) for patients with WG, MPA or CSS compared with PAN (p = 0.01).

Conclusions: Our results suggest that, like WG and MPA, patients with CSS are at a greater risk of VTE, than those with PAN. The reasons for this difference remain to be elucidated.

Results from the WeCLOT (Wegener’s Clinical Occurrence of Thrombosis) study1 showed a higher incidence of venous thromboembolic events (VTE) in patients with Wegener’s granulomatosis (WG) than the general population.2 That was later confirmed in two retrospective studies,3 4 which also included microscopic polyangiitis (MPA) and renal limited vasculitis. However, only anecdotal case reports of VTE associated with polyarteritis nodosa (PAN)5 or Churg–Strauss syndrome (CSS)6 have been published to date.

We conducted a systematic retrospective study to determine VTE frequency in patients with WG, MPA, CSS or PAN, and to identify factors associated with its occurrence.


Patient selection

Data from patients enrolled in the French Vasculitis Study Group (FVSG) cohort between 1985 and 2006 have been systematically entered in a standardised form into a computerised database at diagnosis and updated at regular intervals during follow-up. All adverse events have also been recorded. For this study, we conducted a systematic search through the FVSG computerised database to select the patients who satisfied the American College of Rheumatology79 and/or Chapel Hill Nomenclature classification10 criteria for WG, MPA, CSS or PAN (hepatitis B virus infection-related or not).

Patients with VTE were defined as having ⩾1 VTE, confirmed by vascular ultrasonography, ventilation–perfusion lung scan and/or computed tomography angiography, that had occurred within the 6 months before or after systemic necrotising vasculitis (SNV) was diagnosed.

Data collection

For all patients, the following parameters were assessed at the time of SNV diagnosis: age; sex; medical history, including previous cancer or VTE; type of SNV; date of SNV diagnosis; clinical manifestations, especially lower limb peripheral motor neuropathy and stroke with motor deficit; biological parameters, including nephrotic range proteinuria ⩾3 g/24 h, C-reactive protein (CRP) level, and antineutrophil cytoplasm antibody (ANCA) test results by indirect immunofluorescence and enzyme-linked immunosorbent assay and Birmingham vasculitis activity score (BVAS).11 12 For patients who had VTE, date, type of VTE (deep venous thrombosis and/or pulmonary embolism) and the patient’s characteristics at the time of VTE occurrence were also recorded.

Statistical analysis

Categorical variables are reported as numbers and/or percentages and were compared using a χ2 or, when appropriate, Fisher’s exact test. Quantitative variables are reported as mean (SD) or median (range) and compared using Student’s t-test. For all statistical analyses, p⩽0.05 was considered significant. For univariate analysis, characteristics of patients in the VTE group were compared with those of non-VTE group. Variables significantly associated with VTE occurrence in that analysis were entered into a multivariate logistic-regression model. Statistical analyses were conducted using SAS software, version 9.1 (SAS Institute Inc, Cary, North Carolina, USA).


Patient characteristics

The systematic search of the FVSG database identified 1130 patients diagnosed with WG (n = 377), MPA (n = 236), CSS (n = 232) or PAN (n = 285) (fig 1), among whom 807 (71.4%) have been included in prospective clinical trials.13 14 Their M:F sex ratio is 1.28. Mean age, at SNV diagnosis, was 52.5 (16.8) years. Mean follow-up from first SNV symptoms was 58.4 (45.8) months. Eighty-three VTE occurred in 74 (6.5%) patients (fig 1); seven patients suffered recurrent thromboses. The first VTE was deep venous thrombosis in 46 (62.2%) patients, pulmonary embolism in 16 (21.6%), or both of these in 12 (16.2%). The median SNV diagnosis/first VTE interval was 5.8 (−2.4 to +156) months. Fifty-seven (68.7%) of the VTE occurred within 3 months before the SNV diagnosis or 6 months after (n = 42), or a relapse (n = 15), with rates for PAN and small-sized vessel vasculitides of 3.27 and 7.26 per 100 person-years, respectively, during these periods, as compared with 0.58 and 1.84 per 100 person-years during the entire follow-up period. When the first VTE was diagnosed, mean BVAS was 12.6 (10.5) and mean CRP was 67.0 (80.9) mg/l.

Figure 1

Patient flowchart. Identification of patients with VTEs in the French Vasculitis Study Group database. CSS, Churg–Strauss syndrome; FVSG, French Vasculitis Study Group; HCV, hepatitis C virus; HIV, human immunodeficiency virus; MPA, microscopic polyangiitis; PAN, polyarteritis nodosa; SNV, systemic necrotising vasculitis; VTE, venous thromboembolic event; WG, Wegener’s granulomatosis. *Other vasculitides included giant cell arteritis, Takayasu’s arteritis, Kawasaki disease, HCV-related cryoglobulinaemia with vasculitis, thromboangiitis obliterans, Henoch–Schönlein purpura, vasculitis complicating rheumatoid arthritis, primary angiitis of the central nervous system and unclassifiable vasculitis.

Differences between Wegener’s granulomatosis, microscopic polyangiitis, Churg–Strauss syndrome and polyarteritis nodosa

VTE occurred in 30 (8%) patients with WG, 18 (7.6%) with MPA and 19 (8.2%) with CSS (p = 0.97). VTE were significantly less frequent in patients with PAN (2.5%) compared with other vasculitides (7.9%; p = 0.001). Hepatitis B virus status did not affect the frequency of VTE in patients with PAN (3% in PAN vs 1.2% in hepatitis B virus–PAN; p = 0.38). Throughout follow-up, the VTE rate for patients with WG, MPA or CSS was 1.84 per 100 person-years, compared with 0.58 for patients with PAN.

For WG, MPA or CCS, the VTE frequency did not differ significantly between patients who were ANCA positive (46 of 521; 8.8%) and ANCA negative (13 of 197; 6.6%) (p = 0.33). For the former, enzyme-linked immunosorbent assay specificity did not influence VTE frequency (p = 0.50).

Compared with patients with PAN, patients with WG, MPA or CSS had comparable frequencies of previous cancer (p = 0.94) or VTE (p = 0.98), nephrotic range proteinuria (p = 0.98) and/or stroke with motor deficit (p = 0.74). Conversely, peripheral motor neuropathy occurred more frequently in patients with PAN (61.1%) than those with other SNV (35.9%; p<0.0001).

Factors associated with the occurrence of venous thromboembolic events

According to our univariate analysis (table 1), older age at SNV diagnosis (p = 0.002), male sex (p = 0.04), previous VTE (p = 0.02), nephrotic range proteinuria ⩾3 g/24 h (p = 0.03) and/or stroke with motor deficit (p = 0.03) were significantly associated with VTE. Conversely, limb motor neuropathy occurred less frequently in patients in the VTE group than the non-VTE one (p = 0.001). Mean BVAS and CRP levels at SNV diagnosis did not significantly differ between groups and a history of cancer was not associated with a higher VTE risk (p = 0.16).

Table 1 Characteristics* of patients in the VTE and non-VTE groups and factors associated with VTE occurrence

Our multivariate analysis significantly associated VTE with WG or MPA or CCS versus PAN (adjusted odds ratio (aOR) = 2.88; p = 0.01), older age at SNV diagnosis (aOR = 1.29 by 10-year increments; p = 0.002), male sex (aOR = 1.79; p = 0.03), previous VTE (aOR = 3.48; p = 0.02), and stroke with motor deficit (aOR = 5.82; p = 0.02). Lower limb motor neuropathy appeared to be associated with a lower risk of VTE (aOR = 0.44; p = 0.004). Analyses for these VTE risk factors conducted for the SNV subgroup, excluding PAN, provided similar results.


This is the first study to determine VTE frequency and study its risk factors in patients with CSS or PAN along with those patients with WG or MPA. It was based on a large cohort of 1130 patients with SNV with long-term follow-up. We found that patients with CSS, like patients with WG or MPA, had a high VTE rate, especially during the active phases of their disease. Notably, the VTE frequency was lower in patients with PAN than those with these small-sized vessel vasculitides. In addition, age, male sex, previous VTE and stroke with motor deficit at SNV diagnosis were associated with a higher risk of VTE, while peripheral motor neuropathy carried a lower risk.

The retrospective design of this study may have incurred an underestimation of the VTE rate. However, the use of a standardised form to record patients’ data in the FVSG database, the regular updating of patient outcomes and the fact that >70% of the patients had been enrolled in prospective clinical trials should have limited these potential biases.

Indeed, VTE frequencies for our patients with WG and MPA were comparable with those previously reported.1 3 4 Our results also confirmed that VTE occurred mainly during active phases of WG and MPA,1 3 4 but also of CSS and PAN. Even though the VTE frequency was significantly lower in patients with PAN than patients with WG, MPA or CSS (0.58 versus 1.84 per 100 person-years, respectively), supporting the fact that VTE could be a more specific complication of these latter SNV, it remains higher in this patient group than the general population (0.3 per 100 person-years in the Swedish population).2

The reasons and underlying mechanisms for the higher VTE rate in WG, MPA and CSS, ie, small-sized vessel SNV, compared with PAN, remain unclear. Importantly, all diagnoses were reassessed to ensure that they satisfied American College of Rheumatology and Chapel Hill criteria and, in particular, to ascertain that patients with PAN or MPA had not been misclassified. We found no differences between patients with PAN and those with small-sized vessel vasculitides concerning most of the VTE risk factors, like immobilisation due to motor deficit, nephrotic range proteinuria and/or systemic inflammation. Peripheral motor neuropathy, which was significantly more frequent in PAN, was even associated with a lower risk of VTE. One hypothesis could be that ANCA, detectable in many patients with small-sized vessel SNV, but not PAN, play a role in VTE. However, in contrast to some previously published results,3 4 we found no relationship between VTE frequency and ANCA positivity and/or specificity for patients with WG or MPA, but also patients with CSS. Our patients were not systematically tested for hypercoagulability abnormalities. However, Sebastian et al15 showed that the rate(s) of anticardiolipin and/or anti-β2-glycoprotein antibodies was not higher in patients with WG.

The results of this study confirm a higher and somewhat similar risk of VTE for WG and MPA, but also patients with CSS, as compared with patients with PAN. The reasons for the higher VTE rate in SNV and this difference between PAN and small-sized vessel vasculitides remain unclear and speculative. Physicians should definitely be aware of this VTE risk during active phases of vasculitides, especially for their patients with small-sized vessel vasculitides.


  • Competing interests: None.