Contribution from the Field
Thromboembolic events as a complication of antineutrophil cytoplasmic antibody–associated vasculitis
Antineutrophil cytoplasmic antibody (ANCA)–associated vasculitis is dominated by inflammatory occlusion of small vessels causing tissue ischemia in a variety of organs. The clinical picture is characterized by the involvement of multiple vital organs with damage in the kidneys; ear, nose, and throat (ENT) region; airways; central nervous system; and rarely the digestive tract. Prerequisites in the pathogenesis of ANCA-associated vasculitis are the expression of ANCA antigens on the surface of tumor necrosis factor α (TNFα)–primed neutrophils and monocytes, the ligation of ANCA to their antigenic targets, and the induction of cellular activation via Fcγ receptor cross linking. The consecutive generation of reactive oxygen species, cytokine production, and chemokine release is believed to induce the fibrinoid necrosis of small vessels (1), the hallmark of ANCA-associated vasculitis, and vessel wall infiltration with monocytes and lymphocytes (2). Typically, this process is not linked to significant immune complex deposition (so-called pauci immune), which distinguishes this disease from other types of small-vessel vasculitides and inevitably sets focus on the infiltrating cells as initiators of the inflammatory processes.
The introduction of immunosuppressive protocols with cytotoxic agents and glucocorticoids has substantially improved survival, with remission rates >90% and 5-year survival rates >80% (3). However, patients now face the potential of substantial treatment-related morbidity such as infections, bladder carcinoma, and myelosuppression. In addition to the typical clinical constellation and common complications, physicians are confronted with the problem of venous thromboembolism (VTE) in some cases of ANCA-associated vasculitis (4, 5). During routine clinical treatment of patients with ANCA-associated vasculitis and while conducting a retrospective outcome study (3), we noticed several patients with thromboembolic events. This observation led to a more specific investigation of VTE in ANCA-associated vasculitis. In this report, we present a series of patients with thromboembolic events in ANCA-associated vasculitis.
Patients and Methods
After we had become aware of the potential clinical significance of VTE, we retrospectively reviewed all medical records of patients with ANCA-associated vasculitis who had been treated at the Department of Nephrology and Hypertension, University of Erlangen-Nürnberg, Germany, between 1986 and 2001 for the presence of this complication at the initial presentation.
Patients were entered into the study at the time point when a new diagnosis of ANCA-associated vasculitis was established. All data were retrospectively registered at diagnosis and during followup by systematically reviewing medical records on patients' history and laboratory analysis. The Disease Extent Index was used as a parameter to assess disease activity (6).
Patients were classified as having Wegener's granulomatosis by either the presence of granulomatous inflammation in a biopsy specimen or the presence of clinical signs strongly suggestive of granulomatous disease (involvement of the upper respiratory tract with nasal inflammation [purulent/bloody nasal discharge], sinusitis, or otitis media or lower respiratory tract manifestation with pulmonary nodules or fixed infiltrates). Microscopic polyangiitis was ascribed upon both the absence of granuloma formation in a biopsy specimen and the absence of clinical signs strongly suggestive of granulomatous disease. Renal limited vasculitis was diagnosed in patients with biopsy-proven pauci-immune necrotizing glomerulonephritis without symptoms of systemic vasculitis. All patients had renal involvement with histologically proven focal necrotizing glomerulonephritis with few or no immunoglobulin deposits (pauci immune).
All patients were tested for ANCA by indirect immunofluorescence and enzyme-linked immunosorbent assay (ELISA). The indirect immunofluorescence tests and ELISA systems used for ANCA detection were manufactured by Euroimmun (Lübeck, Germany). The ELISA system used a positive cutoff <7 units/ml.
Duplex sonography was performed in all patients with a VTE. Diagnoses for pulmonary embolism were established by radionuclide studies with ventilation-perfusion lung scans or by computed tomography. Because this was a retrospective analysis, the individual method that was used was based solely on the treating physician's choice. Prior VTE was not assessed in a standardized fashion and is only reported if data were available in the medical records.
A total of 105 patients were identified who fulfilled the criteria of small-vessel vasculitis according to the Chapel Hill consensus conference (7) and could be ascribed a diagnosis of Wegener's granulomatosis, microscopic polyangiitis, or renal limited vasculitis. All 105 patients had no previous diagnosis and had a new diagnosis of ANCA-associated vasculitis that received initial treatment at our center. None of the patients had received any immunosuppressive medication before the diagnosis was made. All patients had renal involvement; further prominent manifestations were ENT and pulmonary disease. ANCA testing by indirect immunofluorescence showed a positive test result in 98 (93%) patients. By ELISA, proteinase 3 ANCA (PR3-ANCA) or myeloperoxidase ANCA (MPO-ANCA) were found in 97 (92%) patients. Fifty-nine (56%) patients had PR3-ANCA and 38 (36%) patients had MPO-ANCA.
A VTE event was discovered in 13 patients at the active phase of ANCA-associated vasculitis. One of these patients, an MPO-ANCA–positive man (Table 1, number 13), had experienced venous thrombosis at relapse. The event was found in the remaining 12 patients when they presented for the diagnosis of vasculitis. The clinical features of all patients presenting with VTE are shown in Table 1. There were 9 males and 4 females. The mean ± SD age was 53.6 ± 16.9 years (range 16–74 years). In patients with VTE, PR3-ANCA were identified in 10 (76%) patients and MPO-ANCA were identified in 3 (24%) patients. Wegener's granulomatosis was diagnosed in 8 patients and microscopic polyangiitis in 3 patients. The remaining 2 patients had renal limited vasculitis.
Table 1. Summary of clinical features in 13 patients with antineutrophil cytoplasmic antibody (ANCA)–associated vasculitis presenting with venous thromboembolism*
|1||F||16||WG: kidney, lung||Iliac/caval vein thrombosis, PE||PR3-ANCA (67 units/ml)||Death with right-sided heart failure due to fulminant PE|
|2||M||30||WG: kidney, lung, ENT||VT||PR3-ANCA (43 units/ml)|| |
|3||M||39||WG: kidney, lung, ENT||VT||PR3-ANCA (89 units/ml)|| |
|4||M||51||WG: kidney, ENT||VT and PE detected on autopsy||PR3-ANCA (80 units/ml)||Death due to pneumocystis pneumonia|
|5||F||53||MPA: kidney, skin||VT||PR3-ANCA (16 units/ml)|| |
|6||M||54||RLV: kidney||VT with iliac/caval vein thrombosis, PE||PR3-ANCA (140 units/ml)||Death with right-sided heart failure due to fulminant PE|
|7||M||54||RLV: kidney (RPGN)||VT||PR3-ANCA (95 units/ml)|| |
|8||M||58||WG: kidney, ENT, heart||VT (caval vein thrombosis), PE||PR3-ANCA (219 units/ml)||Death due to coronary arteritis (autopsy)|
|9||F||63||WG: kidney, lung, ENT||PE||PR3-ANCA (136 units/ml)|| |
|10||M||64||WG: kidney, ENT, eye||VT||MPO-ANCA (168 units/ml)|| |
|11||F||67||WG: kidney, ENT||VT, PE||PR3-ANCA (121 units/ml)|| |
|12||M||74||MPA: kidney, Lung||VT, PE||MPO-ANCA (271 units/ml)||PE 1 year after diagnosis|
|13||M||74||MPA: kidney, lung, skin||VT at relapse||MPO-ANCA (116 units/ml)|| |
Of the 13 patients with VTE, a deep venous thrombosis was identified in 12 patients and an additional pulmonary embolism was identified in 6 patients. A 67-year-old woman (patient number 11) experienced a pulmonary embolism without the objective evidence of a deep venous thrombosis. Death due to fulminant pulmonary embolism occurred in 2 patients (patients 1 and 6) who had extensive thrombosis of both iliac veins ascending into the inferior vena cava. Caval vein thrombosis was also detected in a 58-year-old man who experienced multiple pulmonary embolisms and died of myocardial infarction as a consequence of coronary arteritis, confirmed on autopsy, due to active vasculitis.
In addition, pulmonary embolism was detected in an MPO-ANCA–positive, 74-year-old woman (patient number 12) during stable remission 1 year after the initial diagnosis. During followup, pulmonary embolism occurred in 2 other patients, a 58-year-old MPO-ANCA–positive woman and a 52-year-old PR3-ANCA–positive man (data not shown). These 2 patients had not had a prior VTE at the diagnosis of active vasculitis and the event occurred during quiescent disease. One patient (number 10) had a previous thromboembolism before diagnosis. In summary, during retrospective followup of 105 patients, 16 VTEs occurred over 367.5 person-years of observation, corresponding to an incidence rate of 4.3 per 100 person-years (95% confidence interval 1.6–10.2).
Neither typical risk factors for thromboembolic events, such as malignancy, surgery or trauma, pregnancy, oral contraceptives, hormone replacement therapy, immobilization, obesity, and smoking, nor thrombophilic defects could be identified in these patients. Patients with established thromboembolic events were evaluated for thrombophilic defects as follows: all patients were tested for deficiencies of proteins C and S, for antithrombin deficiency, and for antiphospholipid antibodies. Nine patients were also tested for activated protein C resistance (factor V Leiden). The remaining 4 patients were diagnosed before this test became available for clinical routine. The patients without thromboembolism were not evaluated for coagulation defects. Mean ± SD platelet counts were 315 ± 168 × 103/mm3. None of the patients had nephrotic range proteinuria. Compared with the total cohort, patients with VTE had slightly lower levels of proteinuria (mean ± SD 1.4 ± 0.6 gm/dl versus 1.8 ± 2.0 gm/dl).
In this case series, we report 13 patients who presented with a VTE at the time of diagnosis of active ANCA-associated vasculitis. The most striking observations were that these events all occurred during the active phase of the vasculitis, and that these patients were predominantly positive for PR3-ANCA or had Wegener's granulomatosis. We observed 3 patients who had pulmonary embolism at a time when the disease had reached remission.
Venous thrombosis in ANCA-associated vasculitis has also been reported by other investigators. Kiykým et al described a case of an extensive iliac and inferior vena cava thrombosis in a female patient with ANCA-associated vasculitis (4). In the same way, 2 of our patients died from massive pulmonary embolism in the course of caval vein thrombosis. A third patient also had caval vein thrombosis but died from coronary arteritis.
Merkel et al analyzed data from the Wegener's Granulomatosis Etanercept Trial, a randomized study of etanercept for the induction and maintenance of remission in Wegener's granulomatosis (5). During the recruitment phase, 13 (7.2%) patients were found to have had a prior VTE. Another 16 patients had a first-time VTE during the trial, and 10 (63%) of these 16 patients had active Wegener's granulomatosis at the time of the event. The incidence of VTE was 7.0 per 100 person-years, which was markedly increased over a risk of 0.3 and 0.98 per 100 person-years compared with a normal population and a cohort of SLE, respectively (5).
Our data are in line with the findings by Merkel et al. Although the incidence rate of 4.3 per 100 person-years in our study is lower than that in the study by Merkel at al, it is still conspicuously higher than the incidence rates in their control groups. The cohort in the Merkel et al study consisted exclusively of patients with Wegener's granulomatosis, mostly PR3-ANCA–positive patients, whereas our group comprised patients with Wegener's granulomatosis, microscopic polyangiitis, and renal limited vasculitis. If PR3-ANCA positivity has an influence on the rate of VTE, then this might be one reason why the incidence rates differ.
The reasons for these events are not obvious at first but suggest disease-specific changes in the venous system, such as venulitis. However, because ANCA-associated vasculitis is a systemic disease, it seems more likely that the thromboembolic disorders are initiated by the systemic inflammatory condition. In this context, the secretion of the proinflammatory cytokines TNFα and interleukin-1 may trigger thrombotic processes by increased endothelial expression of tissue factor (2). Furthermore, thromboembolic events may be caused indirectly by PR3, which has been detected in elevated levels in sera from patients with Wegener's granulomatosis (8). In vitro, tissue factor expression and activity was increased in human umbilical vein endothelial cells upon stimulation by PR3 and elastase (9). In addition, PR3 and other serine proteinases are known to cause apoptosis of endothelial cells in vitro (10), and apoptotic endothelial cells have been shown to become procoagulant and proadhesive for platelets (11). In this context, it is important to mention that circulating endothelial cells showing characteristics of necrosis and apoptosis were isolated from patients with active ANCA-associated vasculitis (12).
Thrombosis is not an unusual feature in systemic vasculitis. Behçet's disease is commonly complicated by venous and also arterial thrombosis. Although a number of thrombophilic factors may contribute to thrombosis in individual cases (13), the primary reason for the disposition to clot seems to lie in the inflammatory process in the vessel wall in which endothelial injury and dysfunction are thought to be the main factors (14). It is unclear if vasculitic venous involvement accounts for the thrombotic processes observed in ANCA-associated vasculitis.
In summary, these data suggest that thromboembolic events may be seen as a typical complication of ANCA-associated vasculitis. Of course, this study has some limitations, one being that it is retrospective in nature. Therefore, there was no formal method of VTE data ascertainment, and it is unclear if all VTE was actually recorded. Furthermore, no standardized disease assessment tool such as the validated Birmingham Vasculitis Activity Score was used. Despite the awareness of the statistical difficulties associated with a case series, we found the accumulation of thromboembolic events to be impressive and noteworthy. To further assess the relevance of thromboembolic events in ANCA-associated vasculitis, the prospective evaluation of a large cohort would be valuable, such as the multicenter activities of the European Vasculitis Study Group (15). According to our series, however, and the data of other investigators (5), the diagnosis of VTE should be suspected and pursued routinely in patients with active ANCA-associated vasculitis to prevent fatal outcomes.