Vasculitis is defined as an inflammation of vessel walls; the different types are classified according to the size and type of vessels involved and the presence or absence of associated fibrinoid necrosis and/or granulomas
. Patients with polyarteritis nodosa (PAN), the predominant medium-sized vessel vasculitis, commonly present with general symptoms, peripheral neuropathy, and cutaneous involvement. In some instances, hepatitis B virus can be the etiologic agent of PAN (HBV-PAN) 34 . Microscopic polyangiitis (MPA) involves smaller vessels than PAN, and can be responsible for alveolar hemorrhage and rapidly progressive necrotizing glomerulonephritis 45,56 . Churg-Strauss syndrome (CSS) and Wegener granulomatosis (WG) are also small vessel necrotizing vasculitides. CSS is characterized by asthma, hypereosinophilia, and frequent cardiac involvement 24,34,45 . WG predominantly affects the upper and lower respiratory tract and the kidneys 19,22 . Characteristically, these small vessel vasculitides are associated with the presence of antineutrophil cytoplasmic antibodies (ANCA): P-ANCA, which refers to the perinuclear immunofluorescence labeling they generate, are found in 50% of CSS patients 1,32,72 and 80% of MPA patients 22 , whereas C-ANCA, which yield a diffuse cytoplasmic immunofluorescence, can be detected in 60%-90% of patients with WG 16,31 . Approximately 1% of patients with rheumatoid arthritis develop vasculitis (rheumatoid arthritis-associated vasculitis, RAAV), which clinically and histologically resembles classic PAN 35,63 . 3,14
The gastrointestinal (GI) tract also may be involved in these vasculitides, with reported rates of 40%-60% for PAN
, 30%-56% for MPA 20,41,44,55 , 20%-50% for CSS 19,45 , 5%-11% for WG 18,27 , and 1%-10% for RAAV 57 . Clinically, the spectrum of GI symptoms is wide, ranging from mild transient abdominal pain to life-threatening complications requiring emergency surgery, for example, peritonitis, bowel infarction, or hemorrhage. Studies before 1970 reported almost 100% mortality for the patients with these severe GI manifestations, whereas in 1990 the 10-year mortality was 56% 3,49 . Indeed, at least for PAN, CSS, and MPA, severe GI involvement has been identified as an independent factor associated with poor outcome 27 . 19,28
In the current study, we analyzed the clinical and radiologic parameters of 62 patients who had GI involvement of PAN, MPA, WG, CSS, or RAAV, followed at our institution over 22 years (1981-2002). Because this group is larger than previously published series
, we attempted to obtain an exhaustive, in-depth, up-to-date description of GI manifestations occurring in patients with systemic necrotizing vasculitides, and to identify precisely those manifestations that might predict a poor prognosis. 8,25,41,44 PATIENTS AND METHODS
Patients and Inclusion Criteria
Data from 344 patients with systemic necrotizing vasculitides and followed in the Department of Internal Medicine at Hôpital Avicenne (Bobigny, France) have been compiled since 1980 in a computerized bank, in a standardized form. We focused on patients with systemic small or medium-sized vessel vasculitis (PAN, MPA, WG, CSS, and RAAV) fulfilling the 1990 American College of Rheumatology classification criteria (for PAN, WG, and CSS)
or satisfying the Chapel Hill consensus conference nomenclature definitions (for MPA) 42,46,51 . Patients with RAAV were enrolled provided they had rheumatoid arthritis meeting the American College of Rheumatology criteria 33,34 and histologically proven vasculitis. Patient inclusion also required the presence of GI involvement considered to be associated with the vasculitis, defined as: 1) GI symptoms that were present at the time of vasculitis diagnosis (or within the next 3 months) and responded to specific therapy for vasculitis; 2) GI symptoms that occurred during a relapse, diagnosed on the basis of extraintestinal features of the vasculitis and/or responded to specific therapy for vasculitis; and/or 3) GI tract vasculitis that was histologically proven on biopsy or at autopsy. In March 2002, 62 patients in the database fulfilled these criteria; their diagnoses had been made between June 1980 and August 2001; 18 of them were included in the 1995 study by Guillevin et al 2 . 26 Data Collection
We retrospectively reviewed the medical charts of these 62 patients with respect to demographics, clinical presentation, laboratory, radiologic and histologic findings, therapy, and outcome. We categorized abdominal and GI manifestations in the following 12 items: abdominal pain, nausea or vomiting, hematemesis, diarrhea, hematochezia or melena, acute pancreatitis, acute cholecystitis, appendicitis, intestinal occlusion, peritonitis, bowel perforation, and GI ischemia or infarction. In agreement with former studies
, we further differentiated these patients as having or not having a surgical abdomen, defined as severe GI manifestations that required surgery, or at least a surgical consultation or opinion. The following clinical extraintestinal manifestations of vasculitides were also recorded: fever > 38.5 °C; weight loss (>3 kg in the 3 months before diagnosis); myalgia; arthralgia; cutaneous symptoms (purpura, livedo, nodules); digital ischemia or Raynaud phenomenon; peripheral neuropathy (mononeuritis multiplex); central nervous system involvement (seizure, stroke, confusion); ophthalmologic involvement; ear, nose and throat involvement (sinusitis, laryngitis, otitis media, nasal crusting, bleeding, or obstruction); asthma; lung involvement (cough, dyspnea, hemoptysis, pulmonary nodules, fibrosis); cardiac involvement (cardiomyopathy, pericarditis); hypertension (diastolic blood pressure >90 mm Hg); renal involvement (proteinuria >0.4 g/24 hr and/or hematuria >10 red blood cells/mL and/or renal insufficiency with creatininemia >140 μmol/L and/or biopsy-proven glomerulonephritis); and orchitis. Results of gastroscopy, colonoscopy, abdominal computerized tomography (CT) scan, and celiomesenteric and renal angiographies (interpreted as normal or revealing irregular stenoses and/or microaneurysms) were also noted, as were results of GI histologic analyses when available. Disease activity at the time of GI symptoms was ascertained with the Birmingham vasculitis activity score (BVAS) 27,28,44 , which scores 9 organ/system symptoms with a final score ranging from 0 to a maximum of 63. Likewise, for patients with PAN, MPA, or CSS, we calculated the prognostic five-factor score (FFS) that correlates with mortality 48 . Briefly, FFS includes the following 5 parameters: proteinuria ≥1 g/24 hr, creatininemia ≥140 μmol/L, cardiomyopathy, central nervous system involvement, and severe GI involvement (excluding, as defined by FFS, mild abdominal pain and isolated appendicitis or cholecystitis). 19,28 Statistical Analyses
Continuous variables are reported as mean values ± standard deviations (SD) with ranges, and were compared using the Student t-test or Kruskal-Wallis test when appropriate. Categorical variables were compared between groups with the chi-square or, when appropriate, Fisher exact test. A 2-sided p value < 0.05 was considered statistically significant. Survival curves were plotted using the Kaplan-Meier method
. Survival time was calculated from the onset of the first GI manifestations to either death or the date of the patient's last status assessment. Uni- and multivariate Cox proportional hazard models 36 were constructed to examine the relative risk of death as a function of each of the recorded categories of GI involvement and the results of endoscopic explorations, celiomesenteric angiographies, demographics, and extraintestinal symptoms. The multivariate linear regression model included all the parameters that had a p value ≤ 0.20 in the univariate analysis and used a backward selection procedure. Statistical analyses were conducted with SAS Statistical Package, version 8.12 (SAS Institute, Cary, NC). 12 RESULTS
The 62 patients identified were distributed as follows: 38 patients with PAN (including 21 HBV-PAN), 11 CSS, 6 WG, 4 MPA, and 3 RAAV. The male:female ratio was 2.9. Their mean age at the time the vasculitis was diagnosed was 48 ± 18 years (range, 10-86 yr) (
Table 1). TABLE 1:
Demographics, Gastrointestinal Manifestations and Radiologic Findings in 62 Patients With Small and Medium-Sized Vessel Vasculitides*
GI manifestations were present at diagnosis for 50 (81%) patients, but occurred during a subsequent relapse for the remaining 12 (19%), with a mean interval from diagnosis of 30 ± 29 months (range, 5-68 mo). Abdominal pain was the most frequent and almost constant finding (60 patients, 97%). The pain varied widely in intensity and location, and no preferential site could be identified. For 11 (18%) patients, mild abdominal pain was the sole manifestation of GI involvement. Eleven (18%) others presented with abdominal tenderness due to peritonitis. However, a surgical abdomen was not systematically associated with intense abdominal pain or tenderness.
The whole spectrum of GI symptoms is summarized in
Table 1. Nausea and vomiting were present in one-third of the patients; 27% had diarrhea, which was attributed to an exudative enteropathy in 1 CSS patient. Endoscopy or surgery revealed ulcers or ulcerations of the upper intestinal tract in the 4 (6%) patients with hematemesis. Signs of mucosal ulceration and/or inflammation were seen during colorectal endoscopy in 6 of the 10 (16%) patients with hematochezia or melena. Surgical Abdomen
Twenty-one patients (34%) had a surgical abdomen as defined above, and their individual characteristics are listed in
Table 2. Etiologies of surgical abdomen included appendicitis, cholecystitis, intestinal occlusion, acute pancreatitis, vascular rupture (1 patient, 2%), and/or bowel ischemia. Two additional patients underwent surgery because of massive GI bleeding, attributed to specific gastric ulcerations, at the onset of a PAN for 1, and attributed to colon inflammation, without histologic features of vasculitis, for another patient 10 months after he had been diagnosed with CSS. None of the cases of appendicitis or cholecystitis was isolated or revealed the vasculitis. The 3 (5%) patients who had clinical and laboratory features of acute pancreatitis also had bowel perforations. Among the 10 patients with bowel ischemia, 9 had surgical or radiologic signs of intestinal perforations (small intestine, 2 patients; large bowel, 3 patients; cecum, 1 patient; multiple localizations, 3 patients). Peritonitis was diagnosed in 11 patients, attributed to intestinal perforation in 8, cholecystitis in 1, and bowel ischemia without evidence of perforation on intestinal macroscopic examination in 1. The remaining patient had hemoperitonitis, resulting from erosive rupture of the hepatic vascular pedicle, which revealed PAN. TABLE 2:
Clinical and Epidemiologic Findings in the 21 Patients With GI Tract Manifestations and Surgical Abdomen
Continued) Associated Extraintestinal Clinical Manifestations of Vasculitis
The extraintestinal manifestations associated with these systemic necrotizing vasculitides as a function of the presence or absence of a surgical abdomen are reported in
Table 3. Comparing the frequencies of these manifestations between the 2 subgroups revealed no statistically significant differences. TABLE 3:
Clinical Extraintestinal Manifestations in Patients With Vasculitis and GI Tract Involvement, With Respect to the Presence or Absence of a Surgical Abdomen*
Activity (BVAS) and Prognostic (FFS) Scores
BVAS and FFS are given in
Table 3. Mean BVAS did not differ significantly according to the presence or absence of surgical abdomen: 24 ± 8 (range, 6-41) and 18 ± 8 (range, 6-39), respectively (p = 0.51). Among the 53 patients with PAN, MPA, or CSS, FFS was 0 for 10 (19%) patients who were considered to have no severe GI involvement. FFS was ≥2 for 9 of 35 (26%) patients without surgical abdomen, and for 10 of 18 (56%) patients with surgical abdomen (p = 0.02, comparing patients with FFS ≤ 1 and those with FFS ≥ 2). Angiographic Findings
Thirty-nine patients (31 PAN, 5 CSS, 2 WG, and 1 RAAV) underwent angiography, with signs of vasculitis detected in 26 (67%) (23 PAN, 3 CSS) (see
Table 1). Microaneurysms and/or stenoses ( Figure 1) were found in 22 (56%) and 7 (18%), respectively, and affected the renal (17 patients), hepatic (15 patients), mesenteric (11 patients), and/or splenic (6 patients) arteries. The distal mesenteric artery was thrombosed in 1 patient. Parenchymal infarctions were seen in 9 (23%) patients, affecting the kidneys (8 patients) or the liver (1 patient). No differences were observed for patients with or without surgical abdomen, respectively: 7/11 (64%) and 19/28 (68%) patients (p = 0.92). Serial abdominal angiographies were available for 5 PAN patients (4 HBV-PAN and 1 non-HBV-related PAN). Angiographic abnormalities had totally disappeared in 2 patients, 30 and 7 months, respectively, after the first angiographies, and had partially and progressively regressed in 2 others, respectively 28 months and 8-20-32 months after the first angiographies, while all of them were in clinically stable remission. Conversely, the fifth patient, with non-HBV PAN, suffered from bowel ischemia and perforations 23 months after the first angiographic exploration, while microaneurysms remained unchanged. FIGURE 1:
Celiomesenteric angiography in a 68-year-old man with PAN. Multiple hepatic and celiomesenteric microaneurysms and arterial narrowing (arrows).
Abdominal CT Findings
Abdominal CT scans were available for 24 patients and revealed abnormalities in 18 (75%) (see
Table 1). CT findings included enlarged pancreas (3 patients, all of whom had been diagnosed with acute pancreatitis) ( Figure 2), and epiploic infiltrate (3 patients, all with a definite diagnosis of peritonitis). Parenchymal infarctions were observed in 3 (13%) PAN patients, affecting kidneys (2 patients) or spleen (1 patient). An additional finding was thickening of the intestinal wall, observed in 6 patients who presented with various features of GI involvement. Abdominal CT also disclosed hepatomegaly in 8 (33%) patients, among whom 3 had HBV-PAN, associated with splenomegaly in 5 (21%). FIGURE 2:
Acute pancreatitis with pseudocystic formation (large arrow) and epiploic infiltrates (thin arrow) in a 41-year-old woman, 43 months after the diagnosis of PAN. Several days after this abdominal scan was obtained she underwent surgery, during which multiple bowel perforations were observed. She died the day after surgery.
Upper and/or lower GI endoscopies were performed in 28 and 9 patients, respectively, who had pyrosis, persistent abdominal cramps, or GI hemorrhages. Among the former, 24 (86%) patients had upper GI ulcerations, including gastroduodenal ulcerations in 17 (61%) and/or esophageal ulcerations in 7 (25%), whereas 4 (11%) had isolated gastritis. Six (67%) patients had colorectal ulcerations and 1 had rectal nonulcerative mucosal inflammation. All of the ulcerations observed were diagnosed within the first 3 months following the diagnosis of systemic necrotizing vasculitides in 18 of these patients.
Samples were taken from 36 patients during endoscopy (upper, 17 patients; lower, 6 patients) and/or surgery (GI tract, 8 patients; appendix, 6 patients; gallbladder, 5 patients) or at autopsy (2 patients) and subjected to histologic examination. None of the upper GI tract biopsies showed signs of vasculitis. Colorectal mucosal biopsies harbored histologic signs of vasculitis in 3 patients: 1 with PAN and 2 with WG. In 1 of these 2 WG patients, histologic analysis also revealed signs of granulomatous colitis.
Among the 8 surgically removed GI tract samples, vasculitis was seen in the gastric walls of 1 patient who underwent partial gastrectomy for massive hematemesis due to gastric ulcers, and in another patient at autopsy. Examination of small intestine or colon specimens removed from 7 patients during surgery (for bowel ischemia with or without perforation, or hematochezia) and obtained during 2 autopsies (with vasculitis-related deaths) detected signs of ischemic and/or thrombotic necrosis due unequivocally to vasculitis in 8 of 9 cases. Features of vasculitis were present in 4 of the 6 patients with appendicitis. One patient's appendix appeared histologically normal, but vasculitis was present in the adjacent cecal tissue, whereas the remaining patient was considered to have banal suppurative appendicitis. Vasculitis was found in the gallbladder walls in 4 of the 5 patients with cholecystitis and was associated with lithiasis in 2. In the remaining patient, cholecystitis was suppurative. Autopsy of 1 patient with pancreatitis revealed vasculitis of the pancreatic vessels, associated with multiple bowel perforations.
Finally, vasculitis of the GI tract was histologically proven in 18 (50%) patients.
Among the 21 patients with surgical abdomen, 16 (76%) underwent surgery. One CSS patient had a cecal perforation diagnosed with barium enema and abdominal CT, which resolved under medical treatment. Four patients died before surgery could be performed, despite concurrently administered medical treatment.
All but 1 patient with HBV-PAN received steroids, which were combined with an immunosuppressant for 32 (52%), mainly cyclophosphamide (30 patients, 48%), and/or underwent plasma exchanges (37 patients, 60%). HBV-PAN patients received steroids only at the initial phase of systemic necrotizing vasculitides, and more frequently underwent plasma exchanges (18/21 patients, 86%), combined with antiviral drugs (vidarabine for 13 patients, interferon-α for 9, and lamivudine for 3).
Over a mean follow-up of 74 ± 69 months (range, 1-241 mo), 16 (26%) patients died, with respective 6-month and 5-year survival rates of 86% (95% CI, 78-94) and 76% (95% CI, 65-87). Causes of death were directly related to GI involvement for 8 (50%) patients, all of whom presented with a surgical abdomen. Death occurred at the time of the first GI event in 4 of them, and during a GI relapse in the others. The 8 vasculitis-unrelated deaths were due to cardiovascular diseases (5 patients: 2 with refractory congestive heart failure, 1 hemorrhagic stroke after internal carotid endarterectomy, 1 myocardial infarction, and 1 ischemic stroke associated with atrial fibrillation), and infectious causes (3 patients: 1 with nosocomial sepsis that occurred 1 month after bowel perforation, 1 pneumocytosis associated with invasive pulmonary aspergillosis in the context of myelodysplasia, 1 cerebral toxoplasmosis in a patient who had become infected with human immunodeficiency virus long after having been diagnosed with systemic necrotizing vasculitis). Among the 46 survivors, 15 (33%) subsequently suffered ≥1 relapse involving the GI tract.
Kaplan-Meier plots of survival are shown in
Figure 3. Nine surgical abdomen patients died, compared with 7 patients without surgical abdomen; their respective 6-month and 5-year survival rates were 71% (95% CI, 52-90) and 56% (95% CI, 35-77) for the patients with surgical abdomen, and 94% (95% CI, 87-101) and 82% (95% CI, 70-94) (p = 0.08) for those without. According to multivariate analysis, only peritonitis (hazard ratio [HR] = 4.3, p < 0.01), bowel perforations (HR = 5.7, p < 0.01), intestinal occlusion (HR = 5.5, p < 0.01), and GI infarction/ischemia (HR = 4.1, p < 0.01) were significantly associated with higher mortality (see Figure 3). Pancreatitis considered alone only tended to have a poor prognostic value (HR = 4.0, p = 0.07), probably due to the small number of patients affected (n = 3). The respective 6-month and 5-year survival rates were 60% (95% CI, 35-85) and 46% (95% CI, 19-73) for the 15 patients with these 4 very severe GI manifestations, versus 96% (95% CI, 90-102) and 87% (95% CI, 74-100) for the 47 patients without them (HR = 4.4, p < 0.005; and HR = 5.7, p < 0.005 after adjusting for age and the presence or not of renal involvement). Univariate analysis did not identify any clinical extraintestinal manifestations of vasculitis or angiographic findings as predictors of poor outcome. Compared to patients with FFS = 0, FFS ≥ 1 significantly predicted higher mortality (p = 0.02). The respective 6-month and 5-year survival rates were 81% (95% CI, 70-92) and 71% (95% CI, 60-82) for the 31 patients diagnosed before January 1990, and 93% (95% CI, 85-101) and 84% (95% CI, 71-97) for the 31 patients diagnosed after January 1990 (HR = 1.6, p = 0.42) ( Figure 4). FIGURE 3:
Kaplan-Meier survival curves for the 62 patients with GI manifestations (A), and according to the presence or absence of surgical abdomen (B), peritonitis (C), intestinal occlusion (D), bowel perforation (E) or GI infarction/ischemia (F). Analyzed using a Cox proportional regression model.
Kaplan-Meier survival curves for the 62 patients with GI manifestations according to the date of diagnosis (before or after 1990). Analyzed using a Cox proportional regression model.
Previous observations have suggested that GI signs of systemic necrotizing vasculitides are isolated and are the first signs to reveal the disease in up to 13%-16% of cases
. Conversely, in the current series, only 1 patient presented with GI manifestations exclusively, indicating that isolated vasculitis of the GI tract is a rather uncommon scenario. Clinically, abdominal pain was observed in 97% of patients and was therefore an almost constant sign, a finding that is in agreement with previous studies focused on GI involvement 27,41 . In other terms, the absence of abdominal pain would theoretically preclude vasculitis of the GI tract, although it cannot be ruled out that systematic abdominal investigations, such as abdominal CT or angiography, would have detected subclinical GI involvement. 4,9,27,55
Twenty-one percent of patients presented with GI hemorrhage, either hematemesis (6%) and/or rectal bleeding or melena (16%). A high proportion of patients had endoscopically proven gastroduodenal (27%) and/or colorectal ulcers or ulcerations (10%). Former studies of patients with PAN reported specific gastric or intestinal ulcerations in only 6%
. In this context, it has been emphasized that vasculitis-related gastroduodenal ulcerations typically had an extensive and profound pattern 4,69 , but we could not confirm this finding. The high frequency of mucosal ulcerations observed in the current series is not easily explained, but could be due to the wider use of endoscopy, which might have increased their detection. This elevated rate also raises the question of whether some ulcerations might be iatrogenic and, in particular, induced by corticosteroids; however, the GI toxicity of corticosteroids remains unproven 71 . Furthermore, as emphasized by others 17,41,52,69 , endoscopic biopsies seem to have low sensitivity to diagnose vasculitis, even though the endoscopic appearance is suggestive, but might be explained by their superficial character. Whereas upper GI tract biopsies are rarely contributive 69 , colon biopsies may, in some instances, contribute to establishing a specific diagnosis of systemic necrotizing vasculitides 9 . In our patients, 3 colonic ulcerations were unequivocally attributed to systemic necrotizing vasculitides based on histologic examination. It should be kept in mind, however, that colon endoscopy can cause iatrogenic perforations during or after the procedure 43,71 , although no such complication was observed in our series. 44
GI involvement in systemic necrotizing vasculitides has been associated with high mortality
, and previous investigations have suggested that this high lethality would concern, in particular, the subgroup of patients presenting with severe GI symptoms 14,17,65 . Defined as those potentially requiring surgery, these severe abdominal manifestations have been reported to occur in 22%-54% of patients with GI tract involvement 8,27,44 , compared with 34% of patients in the current study. Peritonitis occurred in 18% of our patients, while 16% were diagnosed with bowel ischemia or infarction, 15% with GI perforation(s), 5% with pancreatitis, 8% with acute specific cholecystitis, and 10% with appendicitis. 9,14,27,44
Assessment of our patients' outcomes showed that the 5-year mortality rate was 44% for patients with surgical abdomen versus 18% for those without. A previous analysis
, based on 53 patients with PAN or CSS (18 of whom had GI manifestations and were also included in this study, compared with 35 patients without GI manifestations), reported a 10-year mortality rate of 56% for patients with surgical abdomen, versus 18% for patients with nonsurgical GI involvement and 20% for those without any GI manifestations. Based on 24 PAN patients with GI involvement, in 2002 Levine et al 26 reported a lower (23%) overall mortality for surgical patients and 9% for those with less severe GI involvement, although in that study the duration of patient follow-up was not stated. Because our analysis indicated that surgical abdomen only tended to carry a poor outcome, it appeared pertinent to examine more closely the surgical manifestations associated with poor prognosis, and to determine their predictive potential. According to uni- and multivariate survival analyses, only peritonitis, bowel perforation, GI ischemia or infarction, and intestinal occlusion independently predicted increased mortality, with 6-month and 5-year survival rates of 60% and 46%, respectively, for the 15 patients with these severe surgical manifestations. None of the other GI symptoms was significantly associated with poor outcome. Regarding mortality, all 3 RAAV patients died, emphasizing the poor prognosis of RAAV in general, and especially for those patients with GI involvement 44 , who are often older than patients with primary systemic small and medium-sized vessel vasculitis. 3,49,66
These findings are in accordance with a former study of patients with PAN, MPA, or CSS whose authors concluded that cholecystitis, appendicitis, and mild transient isolated abdominal pain had no prognostic impact
. Specific cholecystitis was clinically observed in 2%-17% of patients with PAN 28 and up to 40% in autopsy studies 4,6,58 . Cholecystitis 47 or appendicitis 47,59 has been reported to occur alone, and even to be an unexpected finding on histologic examination of presumed banal appendicitis 20,50 , but none of our patients had this clinical picture. It has been postulated that cholecystitis related to systemic necrotizing vasculitides could be cured by medical therapy alone, provided that the absence of calculi and its nonsuppurative nature are definitively confirmed, but in routine practice, this approach appears to be overconfident 7,55 . Indeed, all of our patients with cholecystitis or appendicitis promptly underwent surgery, in combination with medical treatment. 6,7,47
Acute pancreatitis also did not appear to be associated with mortality in our multivariate analysis. However, for the 3 patients with pancreatitis in this study, the intestines, mostly the small intestine, were also severely involved, as in most earlier reports
. Previous studies described pancreatitis as having been diagnosed clinically in less than 3% of PAN patients 14,21 and in 9%-50% of autopsy studies 27 . Pancreatic involvement in systemic necrotizing vasculitides may reflect severe GI ischemia or may even result from one or more GI perforations close to the pancreas. Thus, we think that medical therapy alone may be insufficient to control the vasculitic process in these patients, and that they should undergo exploratory laparotomy. 14,68
In our patients, GI symptoms were associated with a wide range of other organ/system involvements including signs of peripheral neuropathy (61%), cardiac (10%), renal (37%), cutaneous (44%), or pulmonary involvement (19%). It has been emphasized that PAN patients with GI tract involvement have less frequent cardiac involvement
, but this finding was not confirmed in our series. Peripheral and central neurologic involvement was slightly more frequent in the patients with surgical abdomen, as in another series of PAN patients 41 , but this difference did not reach statistical significance. These findings would seem to indicate that the severities of GI and extraintestinal involvement are not necessarily associated. 73
With regard to the poor specificity of findings, a definite diagnosis of systemic necrotizing vasculitides of the GI tract appears to be difficult to establish based on clinical symptoms or endoscopies. In this respect, radiologic explorations might be a useful tool to diagnose vasculitis or its complications. Angiography may show signs of vascular involvement with microaneurysms and/or stenoses that are more characteristic of medium-sized vessel vasculitis, like PAN
, but have been reported occasionally in MPA, CSS, or WG 10,15,67 . Because these angiographic abnormalities regress after clinical remission is achieved under therapy 15,54 , serial angiographies are usually neither necessary nor useful. It remains to be shown to what extent angio-magnetic resonance imaging, a less invasive procedure, might be able to detect microaneurysms. The potentially most specific abdominal CT finding was parietal thickening of some part of the GI tract, as observed in one-quarter of our patients who had CT scans. Abdominal CT frequently visualized nonspecific abnormalities, for example, hepatomegaly and splenomegaly in 33% and 21% of our patients, respectively, mostly in HBV-PAN patients but also in others, in agreement with previous reports showing hepatic enlargement in 20%-63% of patients with PAN 13 and 9%-20% of patients with MPA 39,61 , and splenomegaly in 23% of PAN patients 45 . Our calculations indicate that the sensitivities of abdominal CT and angiography do not exceed 75%, and that they do not appear to have prognostic value. 65
A potential drawback of the current study could be that it included patients with different forms of systemic necrotizing vasculitides. However, pooling these patients remains relevant since they share similar physiopathogenic and/or histologic characteristics. Hence, combining them into a single group allowed us to increase the statistical power of our analyses. Pertinently, the current series included a high number of HBV-PAN patients, although this form of PAN now accounts for less than 10% of all PAN patients
. This recruitment could be interpreted as a selection bias, since our institution was particularly involved in the management of and some therapeutic trials involving HBV-PAN patients 26,38 . The similar frequencies of the various GI symptoms in these systemic necrotizing vasculitides would seem to preclude a clear-cut difference based on this parameter. Although GI involvement in HBV-PAN patients was reported to be more frequent than in non-HBV-related PAN patients 26,29,30 , the clinical characteristics of these 2 subgroups of our patients were statistically similar (data not shown). Some differences were, however, directly related to the particular histologic characteristics of each entity. As clearly seen in 1 patient's case, endoscopy may reveal granulomatous colitis in WG, which can be misdiagnosed easily as inflammatory bowel disease 27 . Eosinophilic colitis has been held responsible for diarrhea or ulcerations in CSS 11,70 , but no such conclusion could be drawn based on 2 colon biopsies and 1 large bowel surgical specimen taken from our CSS patients. 53,60
As suggested by the literature and our data, it appears that survival rates have improved over the past few decades, probably reflecting better combined medical and surgical therapeutic strategies. HBV-PAN may be successfully treated with a specific combination of an antiviral agent and plasma exchanges
. Severe systemic necrotizing vasculitides-related GI involvement should be considered a clear indication to initiate corticosteroids in combination with a potent immunosuppressive agent, primarily cyclophosphamide 23,30 . Furthermore, the high early mortality observed in our patients suffering from severe GI symptoms might indicate that alternative therapeutics should be considered to control the disease more rapidly. In this context, anti-TNF alpha 19,40 or anti-CD20 antibodies 5,37 might constitute a valuable option in WG. 62 REFERENCES
1. Aasarod K, Iversen BM, Hammerstrom J, Bostad L, Vatten L, Jorstad S. Wegener's granulomatosis: clinical course in 108 patients with renal involvement.
Nephrol Dial Transplant
2. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis.
3. Babian M, Nasef S, Soloway G. Gastrointestinal infarction as a manifestation of rheumatoid vasculitis.
Am J Gastroenterol
4. Bailey M, Chapin W, Licht H, Reynolds JC. The effects of vasculitis on the gastrointestinal tract and liver.
Gastroenterol Clin North Am
5. Bartolucci P, Ramanoelina J, Cohen P, Mahr A, Godmer P, Le Hello C, Guillevin L. Efficacy of the anti-TNF-alpha antibody infliximab against refractory systemic vasculitides: an open pilot study on 10 patients.
6. Blidi M, Quang Tri N, Cassan P, Guillevin L. [Acute cholecystitis in periarteritis nodosa. 8 cases].
Ann Med Interne (Paris)
7. Burke AP, Sobin LH, Virmani R. Localized vasculitis of the gastrointestinal tract.
Am J Surg Pathol
8. Cacoub P, Le Thi Huong D, Guillevin L, Godeau P. [Causes of death in systemic vasculitis of polyarteritis nodosa. Analysis of a series of 165 patients].
Ann Med Interne (Paris)
9. Camilleri M, Pusey CD, Chadwick VS, Rees AJ. Gastrointestinal manifestations of systemic vasculitis.
Q J Med
10. Chatel A, Garnier T, Bigot JM, Toueg C, Helenon C. [Arteriography in polyarteritis nodosa. Diagnostic findings and value of repeated examinations].
11. Corne P, Grosleron S, Cayrol B, Reynaud D, Yeche S, Capron F, Dubois A. [Wegener's disease: a rare cause of granulomatous colitis].
Gastroenterol Clin Biol
12. Cox DR. Regression models and life tables.
J R Stat Soc
13. Darras-Joly C, Lortholary O, Cohen P, Brauner M, Guillevin L. Regressing microaneurysms in 5 cases of hepatitis B virus related polyarteritis nodosa.
14. Desbazeille F, Soule JC. [Digestive manifestations of vasculitis].
Gastroenterol Clin Biol
15. D'Izarn JJ, Boulet CP, Convard JP, Bonnin A, Ledoux-Lebard G. [Arteriography in polyarteritis nodosa. 15 cases].
J Radiol Electrol Med Nucl
16. Ewert BH, Jennette JC, Falk RJ. The pathogenic role of antineutrophil cytoplasmic autoantibodies.
Am J Kidney Dis
17. Finkbiner RB, Decker JP. Ulceration and perforation of the intestine due to necrotizing arteriolitis.
N Engl J Med
18. Fraioli P, Barberis M, Rizzato G. Gastrointestinal presentation of Churg-Strauss syndrome.
19. Gayraud M, Guillevin L, Le Toumelin P, Cohen P, Lhote F, Casassus P, Jarrousse B. Long-term followup of polyarteritis nodosa, microscopic polyangiitis, and Churg-Strauss syndrome: analysis of four prospective trials including 278 patients.
20. Genereau T, Lortholary O, Royer I, Lhote F, Darras-Joly C, Guillevin L. [Digestive manifestations of periarteritis nodosa].
Gastroenterol Clin Biol
21. Godeau P, Guillevin L, Tucat G, Wechsler B, Bletry O, Herreman G, Lenriot JP. Periarterite noueuse avec pancreatite aigue. Deux nouvelles observations.
22. Guillevin L, Le Thi Huong D, Godeau P, Jais P, Wechsler B. Clinical findings and prognosis of polyarteritis nodosa and Churg-Strauss angiitis: a study in 165 patients.
Br J Rheumatol
23. Guillevin L, Lhote F, Leon A, Fauvelle F, Vivitski L, Trepo C. Treatment of polyarteritis nodosa related to hepatitis B virus with short-term steroid therapy associated with antiviral agents and plasma exchanges. A prospective trial in 33 patients.
24. Guillevin L, Lhote F, Sauvaget F, Deblois P, Rossi F, Levallois D, Pourrat J, Christoforov B, Trepo C. Treatment of polyarteritis nodosa related to hepatitis B virus with interferon-alpha and plasma exchanges.
Ann Rheum Dis
25. Guillevin L, Lhote F, Cohen P, Sauvaget F, Jarrousse B, Lortholary O, Noel LH, Trepo C. Polyarteritis nodosa related to hepatitis B virus. A prospective study with long-term observation of 41 patients.
26. Guillevin L, Lhote F, Gallais V, Jarrousse B, Royer I, Gayraud M, Benichou J. Gastrointestinal tract involvement in polyarteritis nodosa and Churg-Strauss syndrome.
Ann Med Interne (Paris)
27. Guillevin L, Lhote F, Gayraud M, Cohen P, Jarrousse B, Lortholary O, Thibult N, Casassus P. Prognostic factors in polyarteritis nodosa and Churg-Strauss syndrome. A prospective study in 342 patients.
28. Guillevin L, Cohen P, Gayraud M, Lhote F, Jarrousse B, Casassus P. Churg-Strauss syndrome. Clinical study and long-term follow-up of 96 patients.
29. Guillevin L, Durand-Gasselin B, Cevallos R, Gayraud M, Lhote F, Callard P, Amouroux J, Casassus P, Jarrousse B. Microscopic polyangiitis: clinical and laboratory findings in eighty-five patients.
30. Guillevin L, Cohen P, Larroche C, Queyrel V, Loustaud-Ratti V, Imbert B, Hausfater P, Ramanoelina J, Lacombe F, Roudier J, Bielefeld P, Petitjean P, Smadja D. Treatment of HBV-related polyarteritis nodosa (PAN) with lamivudine and plasma exchanges: a prospective, multicenter, pilot trial in 10 patients.
31. Hagen EC, Daha MR, Hermans J, Andrassy K, Csernok E, Gaskin G, Lesavre P, Ludemann J, Rasmussen N, Sinico RA, Wiik A, van der Woude FJ. Diagnostic value of standardized assays for anti-neutrophil cytoplasmic antibodies in idiopathic systemic vasculitis. EC/BCR Project for ANCA Assay Standardization.
32. Hoffman GS, Kerr GS, Leavitt RY, Hallahan CW, Lebovics RS, Travis WD, Rottem M, Fauci AS. Wegener granulomatosis: an analysis of 158 patients.
Ann Intern Med
33. Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross WL, Hagen EC, Hoffman GS, Hunder GG, Kallenberg CG, et al. Nomenclature of systemic vasculitides. Proposal of an international consensus conference.
34. Jennette JC, Falk RJ. Small-vessel vasculitis.
N Engl J Med
35. Kallenberg CG, Mulder AH, Tervaert JW. Antineutrophil cytoplasmic antibodies: a still-growing class of autoantibodies in inflammatory disorders.
Am J Med
36. Kaplan E, Meier P. Non parametric estimation from incomplete observations.
Am Stat Assoc
37. Lamprecht P, Voswinkel J, Lilienthal T, Nolle B, Heller M, Gross WL, Gause A. Effectiveness of TNF-alpha blockade with infliximab in refractory Wegener's granulomatosis.
38. Lau CF, Hui PK, Chan WM, Fung TT, Tung YM, Loo CK, Yip WC, Lam KM. Hepatitis B associated fulminant polyarteritis nodosa: successful treatment with pulse cyclophosphamide, prednisolone and lamivudine following emergency surgery.
Eur J Gastroenterol Hepatol
39. Le Thi Huong D, Wechsler B, Bletry O, Guillevin L, Langlois P, Tran Ba Loc P, Godeau P. [Necrotizing angiitis and abdominal manifestations].
Ann Gastroenterol Hepatol (Paris)
40. Le Thi Huong D, Wechsler B, Guillevin L, Bletry O, Langlois P, Baumer R, Godeau P. [Digestive manifestations of periarteritis nodosa in a series of 120 cases].
Gastroenterol Clin Biol
41. Le Thi Huong D, Wechsler B, Grand d'Esnon A, Vignes B, Baculard A, Montagne JP, Chigot JP, Guesquiere F, Polonovski C, Godeau P. [Severe digestive vasculitis: value of bolus of cyclophosphamide].
Gastroenterol Clin Biol
42. Leavitt RY, Fauci AS, Bloch DA, Michel BA, Hunder GG, Arend WP, Calabrese LH, Fries JF, Lie JT, Lightfoot RW Jr, et al. The American College of Rheumatology 1990 criteria for the classification of Wegener's granulomatosis.
43. Leen EJ, Rees PJ, Sanderson JD, Wilkinson ML, Filipe MI. A case of Churg-Strauss syndrome: tissue diagnosis established by sigmoidoscopic rectal biopsy.
44. Levine SM, Hellmann DB, Stone JH. Gastrointestinal involvement in polyarteritis nodosa (1986-2000): presentation and outcomes in 24 patients.
Am J Med
45. Lhote F, Cohen P, Guillevin L. Polyarteritis nodosa, microscopic polyangiitis and Churg-Strauss syndrome.
46. Lightfoot RW Jr, Michel BA, Bloch DA, Hunder GG, Zvaifler NJ, McShane DJ, Arend WP, Calabrese LH, Leavitt RY, Lie JT, et al. The American College of Rheumatology 1990 criteria for the classification of polyarteritis nodosa.
47. LiVolsi VA, Perzin KH, Porter M. Polyarteritis nodosa of the gallbladder, presenting as acute cholecystitis.
48. Luqmani RA, Bacon PA, Moots RJ, Janssen BA, Pall A, Emery P, Savage C, Adu D. Birmingham Vasculitis Activity Score (BVAS) in systemic necrotizing vasculitis.
49. Marcolongo R, Bayeli PF, Montagnani M. Gastrointestinal involvement in rheumatoid arthritis: a biopsy study.
50. Marlier R, Candaele N. Periarterite noueuse a symptomatologie abdominale aigue.
Acta Gastro Enterol Belg
51. Masi AT, Hunder GG, Lie JT, Michel BA, Bloch DA, Arend WP, Calabrese LH, Edworthy SM, Fauci AS, Leavitt RY, et al. The American College of Rheumatology 1990 criteria for the classification of Churg-Strauss syndrome (allergic granulomatosis and angiitis).
52. Matolo NM, Garfinkle SE, Wolfman EF Jr. Intestinal necrosis and perforation in patients receiving immunosuppressive drugs.
Am J Surg
53. Memain N, De Bandt M, Guillevin L, Wechsler B, Meyer O. Delayed relapse of Churg-Strauss syndrome manifesting as colon ulcers with mucosal granulomas: 3 cases.
54. Miller DL. Angiography in polyarteritis nodosa.
AJR Am J Roentgenol
55. Muller-Ladner U. Vasculitides of the gastrointestinal tract.
Best Pract Res Clin Gastroenterol
56. Munster T, Weis NM, Krogsgaard K. [Polyarteritis nodosa and hepatitis B infection treated with lamivudine].
57. Pinkney JH, Clarke G, Fairclough PD. Gastrointestinal involvement in Wegener's granulomatosis.
58. Sautereau D, Cessot F, Gainant A, Andrieux V, Letard JC, Devalois B, Catanzano G, Claude R, Pillegand B. [Alithiasic cholecystitis and viral hepatitis B disclosing periarteritis nodosa].
Gastroenterol Clin Biol
59. Schwartz IS, Mendelow H, Winkler L. Polyarteritis nodosa presenting as acute cholecystitis.
Am J Clin Pathol
60. Shimamoto C, Hirata I, Ohshiba S, Fujiwara S, Nishio M. Churg-Strauss syndrome (allergic granulomatous angiitis) with peculiar multiple colonic ulcers.
Am J Gastroenterol
61. Solis-Herruzo JA, Munoz-Yague T, Ruizdelgado C. Laparoscopic findings in polyarteritis nodosa.
62. Specks U, Homburger HA, De Remee RA. Implications of c-ANCA testing for the classification of Wegener's granulomatosis: performance of different detection systems.
Adv Exp Med Biol
63. Specks U, Fervenza FC, McDonald TJ, Hogan MC. Response of Wegener's granulomatosis to anti-CD20 chimeric monoclonal antibody therapy.
64. Stone JH, Hoffman GS, Merkel PA, Min YI, Uhlfelder ML, Hellmann DB, Specks U, Allen NB, Davis JC, Spiera RF, Calabrese LH, Wigley FM, Maiden N, Valente RM, Niles JL, Fye KH, McCune JW, St Clair EW, Luqmani RA. A disease-specific activity index for Wegener's granulomatosis: modification of the Birmingham Vasculitis Activity Score. International Network for the Study of the Systemic Vasculitides (INSSYS).
65. Travers RL, Allison DJ, Brettle RP, Hughes GR. Polyarteritis nodosa: a clinical and angiographic analysis of 17 cases.
Semin Arthritis Rheum
66. Tsai JT. Perforation of the small bowel with rheumatoid arthritis.
South Med J
67. Vallat M, Vallat JM, Julien J, Leleu JP, Pesme D, Leger H. [Periarteritis nodosa with fatal liver and gallbladder involvement. Arteriographic aspects].
68. Walton EW. Giant-cell granuloma of the respiratory tract (Wegener's granulomatosis).
Br Med J
69. Williams DH, Kratka CD, Bonafede JP, Katon RM. Polyarteritis nodosa of the gastrointestinal tract with endoscopically documented duodenal and jejunal ulceration.
70. Wilson RH, Kerr PP, McLoughlin J, Gormley M. Symptomatic colitis as the initial presentation of Wegener's granulomatosis.
Br J Clin Pract
71. Wilson RT, Dean PJ, Upshaw JD, Wruble LD. Endoscopic appearance of Wegener's granulomatosis involving the colon.
72. Yi ES, Colby TV. Wegener's granulomatosis.
Semin Diagn Pathol
73. Zizic TM, Classen JN, Stevens MB. Acute abdominal complications of systemic lupus erythematosus and polyarteritis nodosa.
Am J Med