Secondary Logo

Share this article on:

Hepatitis B Virus-Associated Polyarteritis Nodosa: Clinical Characteristics, Outcome, and Impact of Treatment in 115 Patients

Guillevin, Loïc MD; Mahr, Alfred MD; Callard, Patrice MD; Godmer, Pascal MD; Pagnoux, Christian MD; Leray, Emmanuelle MD; Cohen, Pascal MD

doi: 10.1097/01.md.0000180792.80212.5e
Article

Abstract: Hepatitis B virus-associated polyarteritis nodosa (HBV-PAN) is a typical form of classic PAN whose pathogenesis has been attributed to immune-complex deposition with antigen excess. We conducted the current study to 1) analyze the frequency of HBV infection in patients with PAN, in light of the classification systems described since 1990; 2) describe the clinical characteristics of HBV-PAN; 3) compare the evolution according to conventional or antiviral treatment; and 4) evaluate long-term outcome.

One hundred fifteen patients were included in therapeutic trials organized by the French Vasculitis Study Group and/or referred to our department for HBV-PAN between 1972 and 2002. To determine the frequency of HBV-PAN during the 30-year period, we analyzed a control group of patients with PAN without HBV infection, followed during the same period and diagnosed on the same bases.

Depending on the year of diagnosis, different treatments were prescribed. Before the antiviral strategy was established, some patients were given corticosteroids (CS) with or without cyclophosphamide (CY). Since 1983, treatment for patients with HBV markers has combined 2 weeks of CS followed by an antiviral agent (successively, vidarabine, interferon-α, and lamivudine) combined with plasma exchanges (PE).

Ninety-three (80.9%) patients entered remission during this period and 9 (9.7%) of them relapsed; 41 (35.7%) patients died. For the 80 patients given the antiviral strategy as intention-to-treat, 4 (5%) relapsed and 24 (30%) died vs 5 (14.3%) relapses (not significant [NS]) and 17 (48.6%) deaths (NS) among the 35 patients treated with CS alone or with CY or PE. HBe-anti-HBe seroconversion rates for the 2 groups, respectively, were: 49.3% vs 14.7% (p < 0.001). Patients who seroconverted obtained complete remission and did not relapse.

Thus, HBV-PAN, a typical form of classic PAN, can be characterized as follows: when renal involvement is present, so is renal vasculitis; glomerulonephritis due to vasculitis is never found; antineutrophil cytoplasmic antibodies (ANCA) are not detected; relapses are rare, and never occur once viral replication has stopped and seroconversion has been obtained. Combining an antiviral drug with PE facilitates seroconversion and prevents the development of long-term hepatic complications of HBV infection. The major cause of death is gastrointestinal tract involvement. Importantly, the frequency of HBV-PAN has decreased in relation to improved blood safety and vaccination campaigns.

Abbreviations: ab = antibodies, ACR = American College of Rheumatology, ag = antigen, ANCA = antineutrophil cytoplasmic antibodies, BVAS = Birmingham vasculitis activity score, CI = confidence intervals, CS = corticosteroids, CY = cyclophosphamide, FFS = five-factor score, GI = gastrointestinal, HBV = hepatitis B virus, HBV-PAN = hepatitis B virus-associated polyarteritis nodosa, HCV = hepatitis C virus, HIV = human immunodeficiency virus, NS = not significant, PAN = polyarteritis nodosa, PE = plasma exchanges.

From Department of Internal Medicine (LG, AM, PG, CP, P Cohen), Hôpital Cochin, Assistance Publique-Hôpitaux de Paris, Paris; and UPRES 3409 "Recherche Clinique et Thérapeutique" (LG, AM, PG, CP, P Cohen), Université Paris-Nord, Bobigny; Department of Pathology (P Callard), Hôpital Tenon, Paris; and Department of Public Health (EL), Faculté de Médecine, Rennes, France.

*See Appendix.

This work was supported by grants from the Ministère des Universités (UPRES 3409), the Ministère de la Santé (Programme Hospitalier de Recherche Clinique), and the Groupe d'Intérêt Scientifique-Maladies Rares (INSERM). Lamivudine treatment was conducted under the sponsorship of and with a grant from the Assistance Publique-Hôpitaux de Paris. Lamivudine was provided by GlaxoSmithKline.

Address reprint requests to: Loïc Guillevin, Department of Internal Medicine, Hôpital Cochin, 27, rue du Faubourg Saint-Jacques, 75679 Paris Cedex 14 Paris, France. Fax: 33.1.58.41.14.60; e-mail: loic.guillevin@cch.ap-hop-paris.fr.

Back to Top | Article Outline

INTRODUCTION

Hepatitis B virus-associated polyarteritis nodosa (HBV-PAN) is a typical form of classic PAN. PAN pathogenesis remains largely unknown and not extensively investigated since the 1970s, when authors19,29 suggested the responsibility of immune-complex deposition in antigen excess. The therapeutic strategy we developed was based on the necessity to clear immune complexes and to suppress viral replication so as to facilitate seroconversion and prevent the long-term consequences of HBV infection. HBV has been considered among the etiologic agents of PAN in 36% of cases5. This association, first described by Trépo and Thivolet26 and Gocke et al3, has been confirmed by many reports5,21. Estimates of the annual incidence of PAN in a general population range from 4.6 per 1,000,000 in England23 to 31 per 1,000,000 in France17 and 77 per 1,000,000 in a hepatitis B hyperendemic Alaskan Eskimo population18. In France, the frequency of HBV infection has progressively decreased as a consequence of the increased safety of blood transfusions and products, and of large vaccination campaigns aimed at teenagers and healthcare workers.

We previously established the specific clinical characteristics of this type of PAN5, which include high frequencies of orchitis, vascular nephropathy, and malignant hypertension. These findings, although based on 41 patients5, needed to be confirmed in a larger population, and HBV-PAN characteristics had to be analyzed according to new classification systems13,15.

We conducted the current study to 1) analyze the frequency of HBV infection in patients with PAN seen between 1972 and December 2002, using the classification systems devised since 1990; 2) describe the clinical characteristics of HBV-PAN; 3) compare the PAN evolution according to conventional or antiviral treatment; and 4) evaluate long-term outcome.

Back to Top | Article Outline

PATIENTS AND METHODS

Patients

One hundred fifteen patients were enrolled in the studies organized by the French Vasculitis Study Group (FVSG; see Appendix) since 1980, including those participating in prospective therapeutic trials and/or referred to our department for HBV-PAN. Patients followed between 1972 and December 2002 were eligible. Patients were selected for this retrospective study based on the following criteria: 1) PAN was biopsy proven or, when histologic proof could not be obtained, abdominal and/or renal angiography showing the presence of microaneurysms and/or multiple stenoses associated with clinical symptoms of vasculitis was considered sufficient for diagnosis; 2) HBs antigen (Ag) was present in the sera of all the patients and the demonstration of virus replication, assessed by HBeAg and viral DNA or DNA polymerase, was required for diagnosis. The 1990 American College of Rheumatology (ACR) classification criteria for PAN15 and the Chapel Hill nomenclature13,15 were retrospectively applied to these patients. Because the Chapel Hill nomenclature had not yet been adopted when we started to collect data concerning the patients included in this cohort, we carefully reviewed each patient's form and confirmed the diagnosis of PAN, excluding all other diagnoses, for example, microscopic polyangiitis. Four patients with cryoglobulinemia were included: 3 had histologically proven medium-sized artery involvement and clinical symptoms consistent with PAN; the fourth was coinfected with hepatitis C virus (HCV). To determine the frequency of HBV-PAN, we analyzed a control group of patients with PAN without HBV infection, observed during the same period and diagnosed based on the same criteria.

Back to Top | Article Outline

Serologic Assays

Patients' sera were tested (using a commercially available polyclonal radioimmunoassay) for the presence of the following HBV markers: HBsAg; anti-HBs antibodies (Ab), -HBcAb, -HBeAb, and -HBeAb. To search for HBV DNA in sera we used a spot-hybridization technique and HBV-DNA polymerase or viral DNA. HBV infection was usually demonstrated by the presence of HBsAg, HBeAg, and anti-HBcAb, and by the absence of anti-HBsAb and-HBeAb. Because of mutant viruses responsible for discordant serologic tests, the determination of the viral replication rate by viral DNA or DNA polymerase was mandatory for diagnosis.

Antineutrophil cytoplasmic antibodies (ANCA) had been sought in 72/115 patients. The technique used followed the recommendations of the European ANCA-Assay Standardization Group12.

Back to Top | Article Outline

Angiography

Abdominal angiography with renal, celiac, and mesenteric arteriography was performed only when abdominal manifestations or renal insufficiency were present, or a diagnosis was not obtained by biopsies. The arterial images were classified as normal, stenoses, occlusion, and/or microaneurysms. Both fusiform and circular dilatations were considered aneurysms. Renal infarctions and microaneurysms or stenoses of gastrointestinal (GI) tract or renal vessels seen on the angiograms were considered sufficient arguments for the diagnosis.

Back to Top | Article Outline

Liver Biopsies

Liver biopsies were not systematically performed and were usually taken during the acute phase of PAN when liver function was abnormal. Abdominal angiography was performed before biopsies in patients with abdominal pain or who had undergone angiography for a diagnostic purpose. In some cases, biopsies were taken during the course of the disease, for example, when abnormal liver function persisted, to evaluate the evolution toward chronic hepatitis or liver cirrhosis. When available, the biopsy samples were reviewed by each attending pathologist, who classified liver modifications according to METAVIR and ISHAQ classifications; biopsy samples were also subjected to central review. The 2 scores measure hepatic inflammation and fibrosis, and, because their results were highly correlated22,28, only those obtained with the METAVIR scoring system are given herein. We have divided the biopsy results into 2 groups, based on their fibrosis indexes: the first with an index of 2 and the other, >2.

Back to Top | Article Outline

Treatment

Depending on the time of diagnosis, different treatments were prescribed. Ninety-three patients were included in or treated according to the FVSG protocols. These trials have been described previously in detail5,7,9 and are summarized below.

Thirteen patients were included in a trial on PAN associated or not with HBV markers, microscopic polyangiitis, or Churg-Strauss syndrome (August 1980-December 1983)4, corticosteroids (CS) and plasma exchanges (PE) were administered to all patients, who were subsequently randomized to receive cyclophosphamide (CY; 2 mg/kg per day) for 1 year. Ten patients received CS+PE and 3, CS+PE+CY.

Since 1983, for patients with PAN and HBV markers, treatment has consisted of 2 weeks of CS followed by the combination of an antiviral agent, vidarabine until 1987 (49 patients)7, then interferon-α until the end of 1999 (15 patients)8, and PE. Seven patients were included in the antiviral protocol after failure of CS alone or combined with CY. CS were given as follows: a daily 15 mg/kg pulse of methylprednisolone, for 1-3 consecutive days, followed by 1 mg/kg per day of prednisone for 1 week. CS were then progressively stopped within the second week. In the case of relapse, treatment was adapted to the severity of the symptoms: CS, CY, or another immunosuppressive agent (3 patients). PE were scheduled 4 times/week for 3-4 weeks, then progressively tapered to 3 sessions/week for 2-3 weeks, 2 sessions/week for 2 weeks, then 1 session/week before being definitively stopped. During each session, 60 mL of plasma/kg were removed and replaced by 4% albumin or fresh-frozen plasma, the latter exclusively for patients with profoundly depleted clotting factor concentrations. The circuit was primed with dextran or starch. When vidarabine was the antiviral agent, 15 mg/kg per day were given for 1 week, then 7.5 mg/kg per day for 2 weeks. When interferon-α was prescribed, 3 MU 3 times/week were administered subcutaneously until seroconversion was obtained. In a limited number of patients, failure of the 3-MU regimen led to doubling of the interferon-α dose to 6 MU 3 times/week.

Since January 1999, lamivudine, a more recent antiviral agent, has been given orally to 16 patients at the initial dose of 100 mg/d (10 in a prospective trial9, 6 since that preliminary study). Otherwise, CS doses and PE scheduling were as described above.

Twenty-two patients were not included in a therapeutic protocol because they had been treated before we started the prospective FVSG trials. Nine patients had been treated with CS, 12 with a combination of CS+CY, and 1 with the combination of CS+CY+PE.

Back to Top | Article Outline

Evaluation Criteria

Two scores were applied retrospectively to evaluate disease activity. The five-factor score (FFS)6 that comprises the following items: classes of serum creatinine (≤ and > 1.58 mg/dL) and proteinuria (≤ and > 1 g/d), presence of severe GI-tract involvement (that is, small bowel perforation, pancreatitis, digestive bleeding), cardiomyopathy, and central nervous system involvement; the presence of each factor is accorded 1 point. Three scores are defined: 0, when no factor is noted; 1, when 1 factor is present; 2, when 2 or more factors are present.

We also applied the first version of the Birmingham vasculitis activity score (BVAS)16 at the time of diagnosis to test its ability to correlate with prognosis and to compare it with FFS. BVAS is a clinical index of disease activity based on symptoms and signs in 9 separate systems (systemic signs, skin, mucous membranes and eyes, ear-nose-throat, chest, heart and vessels, GI tract, kidney, nervous system). Maximum totals were retained for each system; the maximum score is 63. Disease features are considered and taken into account only when they are attributable to active vasculitis.

Uncontrolled vasculitis was defined as the appearance of new manifestations or aggravation of manifestations already present, despite administration of optimal treatment for it. Relapses were defined as the development of new manifestations or aggravation of those already present in a patient who had been considered to be in complete or partial remission. Remission was defined as the partial or complete disappearance or attenuation of symptoms of vasculitis. Sequelae can persist in patients in remission or who had recovered. Recovery is defined as patients in remission for at least 12 months after stopping treatment with no relapse. HBeAg to anti-HBeAb serococonversion was not obligatory for a patient to be considered in complete remission or recovered.

To gather information concerning the long-term survival of the patients, we sent a questionnaire to the hospital physicians who had enrolled them and/or to their treating practitioners. The following questions were asked: 1) Is the patient dead or alive? 2) If the patient is dead, what was the cause of death? 3) Had the patient suffered relapse(s) since inclusion in the study? If yes, what were the characteristics, date, and treatment of relapse(s)? 4) Is the patient still being treated for the vasculitis? and 5) Did the patient experience treatment side effects? When the information had not been obtained at the time of the inquiry or if the patient had not been seen, the physician of record was contacted by mail or phone. For patients lost to follow-up, demographic data were obtained from the French death registry, in which every French citizen's dates of birth and death are recorded.

Back to Top | Article Outline

Statistical Methods

The baseline characteristics of the study patients are expressed as percentages for categorical variables and as means ± SD for continuous variables. Quantitative variables were compared using the Student t-test or a nonparametric test, and qualitative variables with the chi-square test, or, when appropriate, the Fisher exact test. Patient survival was assessed by life-table analysis using the Kaplan-Meier method14 and compared using the log-rank test20. Standardized mortality ratios were calculated to compare the PAN patients' survival to that expected in the French general population over the period of observation. Estimation of the expected number of deaths was based on age- and sex-specific mortality rates of the French population using national life tables (1972-2003) provided by the Institut National des Études Démographiques (INSEE). All confidence intervals (CI) were calculated at the 95% level. Statistical analyses were computed using the SAS Statistical Package, version 8.12 (SAS Institute Inc., Cary, NC).

Back to Top | Article Outline

RESULTS

Classification

The patients included in this analysis had, in 114/115 cases, at least 3 of the ACR criteria for classification of PAN. One patient had 2 ACR criteria, 13 had 3, 26 had 4, 32 had 5, 23 had 6, 17 had 7, 2 had 8, and 1 had 9. The patient with 2 criteria had biopsy-proven PAN, in addition to clinical manifestations and angiographic signs highly suggestive of the diagnosis. When histology was available, all but 7 patients also satisfied the Chapel Hill nomenclature. Among these 7, 4 had the simultaneous association of medium- and small-sized vessel vasculitis, 3 had only small-sized vessel involvement. One of these latter had PAN and cryoglobulinemic vasculitis, and the 2 others had typical clinical symptoms and evolution of PAN.

Back to Top | Article Outline

Demography and Epidemiology

The 115 patients, 74 men and 41 women, had a mean age at disease onset of 51.1 ± 17.0 years (range, 20-80 yr) and were first seen between 1972 and December 2002. One hundred ten patients were white, 2 were Asian, and 3 were black (1 African and 2 Caribbean). Ninety-four patients had spent their entire lives in northern Europe, mainly France; 9 had lived for prolonged periods in southern European countries, 8 in North Africa, 2 in Southeast Asia, 1 in the French Antilles, and 1 in the Middle East.

HBV infection routes were clearly identified for 43 patients: sexual for 15 (10 heterosexual and 5 homosexual), intravenous drug abuse for 16, transfusions for 12 (during surgical procedures for 6). For 19 other patients, a possible source of infection was found: 15 patients had undergone surgery or various invasive investigations, including dialysis for 1 and acupuncture for another, interventions that could have been responsible for viral contamination. For these 15 patients, we had little or no information about the administration of blood or blood products that might have been prescribed. When a blood transfusion was incriminated, it had been given between 1972 and 1987. No new case of HBV-PAN attributable to blood products or transfusion has been observed since 1987. When it was possible to determine the interval between HBV infection, that is, blood transfusion and/or the precise date of sexual transmission (12 patients), and the first manifestations of PAN, it was 596 ± 628 days (range, 30-1695 d). Seven of the 12 patients had been infected during the year preceding the onset of PAN.

Hepatitis was diagnosed in 32 patients before PAN, and the interval between the 2 diseases was undetermined in only 1 patient. In 17 patients, PAN became manifest during the course of hepatitis, or immediately after its acute onset. One patient developed fatal fulminant hepatitis at the time of seroconversion under antiviral treatment. The mean time between the diagnoses of hepatitis and PAN was 217.8 ± 800.9 days.

Four patients were coinfected with HBV and HCV at the time of PAN diagnosis. No patient had a concomitant human immunodeficiency virus (HIV) infection at the time of diagnosis. In the 28 other patients, hepatitis was clinically asymptomatic or limited to asthenia. Only 1 patient had jaundice.

The frequency of HBV-PAN in our cohort of patients suffering from PAN has been calculated for 3-year (Figure 1) and 5-year periods (Table 1). During the entire study period, we observed 341 cases of PAN, 115 (33.7%) of which were due to HBV. The frequency of HBV-PAN among cases of PAN throughout the study period is detailed in Table 1 and Figure 1.

FIGURE 1

FIGURE 1

TABLE 1

TABLE 1

Back to Top | Article Outline

Clinical Symptoms

The initial clinical symptoms at PAN onset were available for 87 patients (Table 2). Among the 24 patients with arthralgia, 4 had polyarthritis at disease onset. Among the 6 patients with skin manifestations, 1 had urticarial vasculitis and the others had infiltrated purpura, with bullae in 1 patient. Leg edema was present in 3 patients, facial edema in 1 patient. One patient was vaccinated against the flu 3 months after hepatitis B onset and immediately thereafter developed mononeuritis multiplex due to the vasculitis.

TABLE 2

TABLE 2

The mean time between the first manifestations of PAN and its diagnosis was 84 ± 84 days (range, 9-335 d).

The clinical symptoms of vasculitis for all 115 HBV-PAN patients are detailed in Table 3. Disease severity, as assessed by FFS and BVAS, respectively, was 1.10 ± 1.16 (range, 0-5) and 17.8 ± 8.2 (range, 9-43). Seventy-one (61.7%) patients had an FFS ≥ 1 point, 36 (31.3%) had an FFS ≥ 2 points. Sixty-two patients (53.9%) had a BVAS ≥ 17.

TABLE 3

TABLE 3

Back to Top | Article Outline

Laboratory Investigations

Cryoglobulins were found in 4/72 (5.6%) patients, who were tested before treatment, and they disappeared in 1 patient, concomitantly with anti-HBeAb detection. They persisted in 2 other patients, who died in the weeks following diagnosis, and were not monitored in the fourth patient. The 4 patients had clinical symptoms consistent with PAN, 1 had renal infarcts, and 2 had microaneurysms. Medium-sized vessel involvement was histologically proven in 3 patients. One patient's outcome facilitated classification because she was coinfected with HBV and HCV; she had microaneurysms of abdominal arteries despite histologically proven small-sized vessel vasculitis. She recovered, HBeAb appeared but HCV and cryoglobulinemia persisted. Notably, this patient never relapsed. Monoclonal gammopathy was demonstrated in 2 patients (1 IgMκ and 1 IgGλ), 1 of whom had associated cryoglobulinemia, and no patient had symptoms of or developed myeloma or Waldenström macroglobulinemia. ANCA were not detected in the 72 patients who were tested.

Aspartate aminotransferase (ASAT) and alanine aminotransferase (ALT) levels were normal in 38 (33%) and 30 (26%) patients, respectively. Transaminase concentrations, 1.5-5.0 times the normal range, were observed in the other patients, except for 6 whose values ranged from 7 to 20 times the normal values.

Abdominal and renal angiographies were performed in 67 patients; microaneurysms and/or stenoses were found in 46 (68.2%). Thirty-three patients underwent angiography before liver biopsy; it was normal in 12, and revealed microaneurysms in 21 but not in liver vessels. Angiography was not performed before biopsy in 16 patients.

Viral replication was demonstrated in every patient. Since 1983, every patient has systematically been tested for HIV, and serology was negative in every case. HCV was sought in 40 patients and was positive in 3.

Liver biopsies were performed in 49 patients and a centralized review was obtained for 33. When the METAVIR classification was applied, 23 patients had a low fibrotic index (≤F2) and 10 had histologic signs of fibrosis (≥F3). For the 16 biopsies that could not be reviewed by our pathologist, the reports stated that the specimen was normal or contained moderate inflammation for 11 patients, and that fibrosis and chronic hepatitis were observed in 5. High-grade fibrosis was seen in 15 (30.6%) of the 49 liver biopsies.

Back to Top | Article Outline

Outcome

Mean follow-up from treatment onset was 69 ± 135.8 months (range, 9-7105 d [236.8 mo]). Ninety-three (80.9%) patients entered remission during follow-up. The 22 who did not achieve remission died after a mean time of 178 ± 199 days (range, 9-821 d); 15 of them died within the first 6 months after PAN diagnosis. The mean FFS of these 22 patients was 1.9 ± 1.3, and their mean BVAS was 18.3 ± 8.9.

Nine (9.7%) of the 93 patients who achieved remission subsequently relapsed; 1 of them died of stroke but the relationship between PAN and stroke could not be established. The mean time to relapse, measured from treatment onset, was 31.2 ± 46.9 months. Five of the 9 patients who relapsed did so during the first year. Forty-one (35.7%) patients died. The causes of death and the time of occurrence are detailed in Table 4. One patient died of fulminant hepatitis at the time of seroconversion, which coincided with remission of PAN. Thirteen patients died after 5 years of follow-up; none of these deaths was directly related to a disease flare. The 5-year survival rate was 72.5% (95% CI, 63.9-81.2) (Figure 2). The mortality of the patients significantly exceeded that of the French general population as reflected by a standardized mortality ratio of 4.04 (95% CI, 2.99-5.50). Standardized mortality ratios for men and women were 3.45 (95% CI, 2.11-5.33) and 4.88 (95% CI, 2.98-7.53), respectively.

TABLE 4

TABLE 4

FIGURE 2

FIGURE 2

When determined at the start of treatment, FFS and BVAS for survivors vs nonsurvivors, respectively, were as follows: 1.08 ± 1.16 and 17.82 ± 7.86 vs 1.17 ± 1.28 and 17.8 ± 8.88 (not significant [NS]).

We also assessed patient outcome according to the treatments assigned. For the 80 patients who received the antiviral strategy as intention-to-treat, 4 (5%) relapsed and 24 (30%) died compared with 5 (14.3%) relapses (NS) and 17 (48.6%) deaths (NS) among the 35 patients treated with CS alone or CS combined with CY or PE. Seven patients, who did not respond to antiviral drugs and PE, were given CS (n = 5) or CY+CS (n = 2); 4 of them died, 2 of liver cirrhosis, aggravated by alcohol abuse in 1; 1 succumbed to sudden death, and 1 died of stroke. One of them was a homosexual who subsequently became HIV positive. None of these 7 patients seroconverted.

Considering the outcome during the 18 first months of follow-up, we noted that, among the 35 patients receiving CS alone or combined with a cytotoxic agent as intention-to-treat, 9 (25.7%) died: 4 (11.4%) from uncontrolled vasculitis, 1 of lung cancer, 1 of liver cirrhosis, and the others of unknown causes. For the 80 patients treated according to the antiviral strategy, the 18-month outcome was that 14 (17.5%) patients died: 6 (7.5%) of uncontrolled vasculitis, 4 of treatment side effects (3 of septicemia via an indwelling PE catheter, 1 of fulminant hepatitis), 1 of liver cirrhosis, 1 relapsed and died of stroke, 1 succumbed to rhythm disturbances, and 1 died of gallstones. According to survival analysis, the risk of mortality did not differ significantly between the 2 groups (p = 0.46) (Figure 3A). Conversely, the event-free survival rate, defined as the absence of relapse or death, tended to be higher for the patients treated with a regimen containing an antiviral agent (p = 0.21) (Figure 3B).

FIGURE 3

FIGURE 3

Among the 35 patients treated with CS alone or combined with CY or PE, 5/34 (14.7%, data missing for 1 patient) seroconverted from HBeAg to anti-HBeAb and 3 of them from HBsAg to anti-HBsAb after a mean time of 909 ± 413 days (range, 583-1629 d). When the patients received the antiviral strategy as intention-to-treat, their HBeAg to anti-HBeAb seroconversion coincided with the cessation of viral replication, in 38/77 (49.4%; no data for 3 patients) after a mean of 645 ± 861 days (range, 13-3554 d) (p < 0.001). When vidarabine was used, 20/49 (40.8%) patients seroconverted in a mean time of 744 ± 955 days (range, 13-3237 d). Nine (64.3%) of the 14 (data missing for 1) patients receiving interferon-α seroconverted after a mean of 727 ± 1139 days. Among the 16 patients treated with lamivudine, 9/14 (64.3%; data missing for 2) seroconverted from HBeAg to anti-HBeAb after a mean interval from treatment onset of 198 ± 85 days (range, 13-724 d). When seroconversion was achieved, clinical remission was also obtained and no relapse occurred. Even when seroconversion did not occur, remission was also obtained in most patients, but subsequent relapse could occur. Indeed, all the relapses occurred in patients who did not seroconvert.

No relationship could be established among the intensity of hepatic cytolysis, liver function, and the seroconversion rate, or between the hepatitis incubation time, when known, and seroconversion. The occurrence of seroconversion and subsequent recovery was independent of liver function at the time of diagnosis, when it was known, and of the treatments for vasculitis preceding the introduction of the antiviral strategy.

Back to Top | Article Outline

DISCUSSION

The relationship between HBV and PAN is well known3,26, and this association demonstrated the potential role of infections as the cause or target for some systemic vasculitides. The main characteristics of HBV-PAN have been described previously5, and the results obtained with some treatments have been reported5,7,9. The aim of this study was to analyze the epidemiology, clinical characteristics, laboratory investigations, and outcomes, based on a large population followed over an extended period in internal medicine departments in France. This long-term follow-up of the patients and the different treatments prescribed, most of them prospectively, gave us the opportunity to compare a strategy combining antiviral agents and PE, to a conventional regimen with CS alone or with an immunosuppressant.

HBV-PAN represents the most typical form of classic PAN. Its pathogenesis is not well known, and old reports suggested that the deposition of circulating immune complexes in vessel walls was responsible for the disease11,27. We based our therapeutic strategy on the responsibility of immune complexes, and the empirical results obtained with our strategy seem to confirm this pathogenic hypothesis. This very pure group of PAN clearly shows the signs of the disease and could be, in association with exclusionary criteria-the presence of ANCA, lung capillaritis, and glomerulonephritis-a model for establishing diagnostic criteria of PAN.

With the results of the current study, we were able to clarify further the epidemiology of HBV-PAN. We observed that HBV was a major cause of PAN, being present in roughly half the cases, between 1972 and 1989. The decreased frequency of HBV-PAN observed since the early 1990s is probably the direct consequence of vaccination campaigns against HBV and the increased safety of blood transfusions and products with the systematic search for HBeAg. This evolution shows that HBV is a causal agent of PAN and not only a coincidence or triggering factor. Vaccination had a major impact in France. Between 1990 and 1997, the Ministry of Health promoted the systematic vaccination of teenagers and not only people at risk of contamination; 22,313,000 individuals were vaccinated (source: French Ministry of Health). Not only the number of cases of HBV-PAN, but also the number of cases of PAN of unknown origin observed during this period declined. However, that policy was stopped in 1998 because demyelinating neuropathies developed after immunization and could have been imputable to the vaccine. Our results also show that HBV-PAN has not disappeared, despite its dramatically decreased frequency.

No specific new clinical characteristics of the disease were identified. The vasculitis is responsible for renal insufficiency due to vascular infarcts, orchitis, GI involvement, and the presence on angiograms of microaneurysms, stenoses, and infarcts. When angiography was performed during complete remission, normalization was observed. We no longer require follow-up angiographies, because it is now clear that clinical recovery parallels angiographic normalization2.

Pertinently, the laboratory investigations showed that ANCA were absent. Therefore, it seems that ANCA can discriminate between small- and medium-sized vessel vasculitides and can be considered an exclusionary criterion for HBV-PAN. However, patients with HBV-PAN can have cryoglobulinemia. Usually, the presence of mixed cryoglobulins is sufficient to diagnose small-vessel vasculitis, called cryoglobulinemia, and to exclude other vasculitides. Nevertheless, we described herein few typical cases of cryoglobulinemia associated with HBV-PAN, a medium-sized vessel vasculitis, which is not surprising because cryoglobulinemia is frequently observed during the course of viral infections.

It is difficult to determine the exact time of HBV infection and only patients who received a documented HBV-contaminated blood transfusion or those who had sexual contact with an HBV-positive partner during a precise period of time can be analyzed for such a study.

In our population, infection occurred in most patients during the year preceding PAN. HBV-PAN is an early postinfection disease, and late occurrences are rare. In such cases, it could be advanced that PAN is the consequence of a new viral infection or a mutation of the virus, but neither has ever been demonstrated and this idea remains hypothetical. When PAN occurs years after HBV infection, a coincidence between the 2 events should also be considered, as the therapeutic approach depends on the etiology and should differ from the antiviral strategy.

Liver involvement is common in HBV-PAN. Hepatitis is frequent, but usually mild, sometimes asymptomatic and therefore misdiagnosed. Hepatitis was diagnosed before the occurrence of PAN in 32 patients. The clinical manifestations were moderate, with asthenia and moderately elevated transaminases, and jaundice was usually absent. The time of PAN occurrence demonstrates that it is a recent postinfection phenomenon. In 5 patients, hepatitis and PAN occurred simultaneously or in tandem. These cases support the intricate relationship between HBV and PAN: PAN is the exacerbation of the systemic manifestations occurring during the early phase of hepatitis. The usual response to HBV infection is the development of jaundice preceded by systemic manifestations, such as arthralgia, urticaria, migraine, and followed by dramatic increases of transaminases, which constitute the normal immune reaction to clear the virus. Systemic manifestations observed during the prodromal phase and PAN are associated with the deposition of immune complexes24,27. In PAN, owing to a nonelucidated immune response, the prodromal phase is exacerbated and transaminase levels of this hepatitis are only moderately elevated. This moderate liver response reflects the development of subclinical chronic hepatitis, which was frequently found in our patients. In contrast, patients mounting a normal response to the virus rarely develop chronic liver disease. In acute hepatitis, virus replication stops after a few weeks, when antibodies are detected, first against HBeAg then against HBsAg. In PAN, this seroconversion is low, especially when the treatment stimulates replication, as is the case when CS alone or combined with a cytotoxic drug is prescribed.

Based on this pathogenesis, an original therapeutic concept was elaborated10,25, combining the clearance of immune complexes and the suppression of HBV replication by an antiviral drug. In this setting, CS are administered for only 2 weeks, in order to contain rapidly the potentially life-threatening organ damage caused by the inflammatory phenomena of systemic vasculitis. The cornerstone of this protocol is based on the subsequent initiation of PE combined with an antiviral agent: PE remove the immune complexes, which are the major determinant of vessel inflammation, and the antiviral agent is added to facilitate immune-complex clearance by diminishing the virus load and, eventually, to stop virus replication. Although our group has been conducting prospective therapeutic trials on systemic vasculitides for more than 20 years, randomized trials are sometimes not feasible due to insufficient recruitment. Indeed, it is for this reason that our antiviral strategy has never been subjected to a randomized trial. Pertinently, we based our strategy on empirical evidence of its effectiveness provided by long-term follow-up data from our former trials showing a 0% relapse rate among patients achieving HBeAg to HBeAb serococonversion and a 7-year overall survival rate of 83%7. The first results obtained using this therapeutic strategy were excellent10,25, despite the partial efficacy and side effects of vidarabine, the first antiviral drug available7. Despite approximately 50% HBeAg to anti-HBeAb seroconversion, clinical recovery was obtained in the majority of the patients, even when they did not seroconvert.

The seroconversion rate was significantly higher in patients given the antiviral strategy than in those treated with immunosuppressants, and those receiving the latter regimens had higher probabilities of relapsing. Since we know that chronic HBV infection can be responsible for liver cirrhosis and carcinoma, the antiviral strategy is well adapted to the long term. According to the current analysis, survival rates did not differ between patients receiving CS alone or combined with an immunosuppressant and those receiving the antiviral regimen. However, the latter seems the most sensible approach to reduce the risk of developing long-term consequences of HBV infection and to suppress the side effects of prolonged therapies. The results of this study also confirm that the only effective action against the disease is stopping or containing virus replication. The improved efficacy of antiviral drugs facilitates our approach. Lamivudine and, should it fail, its combination with interferon-α, are easy to administer and usually well tolerated. In this strategy, PE remain a major component of the regimen because they clear the immune complexes or other pathogenic substances, which are potentially responsible for the clinical manifestations. PE effectiveness is also indirectly proven in patients who enter remission despite the failure to seroconvert and who did not receive any CS or cytotoxic agents. Treatment with antiviral drugs alone cannot be recommended because the seroconversion rate (only 50%-60%) was lower and often occurred several weeks after the end of the treatment administration. Because PAN is an acute, sometimes life-threatening disease, therapy should be initiated as quickly as possible after diagnosis; its effectiveness depends, in part, on this early introduction: at present, only PE can meet these requirements. A trend toward a lower mortality rate was observed when patients were treated with the antiviral strategy, but the difference did not reach statistical significance when compared to patients receiving CS and immunosuppressants. In another study1 in this same issue evaluating mortality of systemic necrotizing vasculitis patients during the first year after diagnosis, we showed that mortality due to GI manifestations was higher for HBV-PAN than non-HBV-PAN patients and lower for diffuse vasculitis. The results of the current study also show that there is no argument not to apply the PE + antiviral treatment to selected patients, rather than CS and immunosuppressants for GI involvement.

The results of this study, describing the very long-term follow-up of a homogeneous cohort of HBV-PAN patients, demonstrate the close relationship between HBV infection and the frequency of PAN in the general population, and also demonstrate the impact of public health initiatives on its prevalence. The goal of total disappearance of HBV-PAN can be achieved in the near future only if the vaccination campaigns, safety of blood transfusions and products, and prophylactic methods against HBV dissemination by intravenous drug addicts are assured. Our findings also confirm those previously published that demonstrated the effectiveness of the antiviral agent and PE strategy, and its superiority to treatment by CS alone or in combination with cytotoxic agents.

Back to Top | Article Outline

APPENDIX

French Vasculitis Study Group members, in alphabetical order, who contributed patients to this study: Philippe Arlet (Toulouse), Jacques Barrier (Nantes), Eric Bercoff (Rouen), Philippe Bielefeld (Dijon), Philippe Bironne (Gonesse), Olivier Blétry (Suresnes), Jean-Pierre Bouchon (Ivry-sur-Seine), Philippe Cassan (Vichy), Elisabeth Cassuto-Viguier (Nice), Boyan Christoforov (Paris), Pierre Deblois (Gonesse), Albert Faradji (Strasbourg), Jacques Glowinski (Gonesse), Pierre Godeau (Paris), Pierre Hausfater (Paris), Bernard Hurault de Ligny (Caen), Jean-Marie Idatte (Paris), Bernard Imbert (Grenoble), Jean Jouquan (Brest), Marcel-Francis Kahn (Paris), Claire Larroche (Bobigny), Frédéric Liozon (Limoges), Véronique Loustaud-Ratti (Limoges), Jean-Claude Marcheix (Clermont-Ferrand), Philippe Petitjean (Strasbourg), Jacques Pourrat (Toulouse), Olivier Pourrat (Poitiers), Viviane Queyrel (Lille), François Rossi (Vendôme), Jean Roudier (Marseille), Michel Ruel (Nanterre), Didier Smadja (Fort-de-France), Alain Sobel (Créteil), Bertrand de Toffol (Tours), Bertrand Wechsler (Paris).

Back to Top | Article Outline

REFERENCES

1. Bourgarit A, Le Toumelin P, Pagnoux C, Cohen P, Mahr A, Le Guern V, Mouthon L, Guillevin L. Deaths occurring during the first year after treatment onset for polyarteritis nodosa, microscopic polyangiitis, and Churg-Strauss syndrome. A retrospective analysis of causes and factors predictive of mortality based on 595 patients. Medicine (Baltimore). 2005;84:323-330.
2. Darras-Joly C, Lortholary O, Cohen P, Brauner M, Guillevin L. Regressing microaneurysms in 5 cases of hepatitis B virus-related polyarteritis nodosa. J Rheumatol. 1995;22:876-880.
3. Gocke DJ, Hsu K, Morgan C, Bombardieri S, Lockshin M, Christian CL. Association between polyarteritis and Australia antigen. Lancet. 1970;2:1149-1153.
4. Guillevin L, Jarrousse B, Lok C, Lhote F, Jais JP, Le Thi Huong D, Bussel A. Longterm followup after treatment of polyarteritis nodosa and Churg-Strauss angiitis with comparison of steroids, plasma exchange and cyclophosphamide to steroids and plasma exchange. A prospective randomized trial of 71 patients. The Cooperative Study Group for Polyarteritis Nodosa. J Rheumatol. 1991;18:567-574.
5. 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. Medicine (Baltimore). 1995;74:238-253.
6. 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. Medicine (Baltimore). 1996;75:17-28.
7. 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. J Rheumatol. 1993;20:289-298.
8. 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. 1994;53:334-337.
9. Guillevin L, Mahr A, Cohen P, Larroche C, Queyrel V, Loustaud-Ratti V, Imbert B, Hausfater P, Roudier J, Bielefeld P, Petitjean P, Smadja D. Short-term corticosteroids then lamivudine and plasma exchanges to treat hepatitis B virus-related polyarteritis nodosa. Arthritis Rheum. 2004;51:482-487.
10. Guillevin L, Merrouche Y, Gayraud M, Jarrousse B, Royer I, Leon A, Baudelot J. Periarterite noueuse due au virus de l'hepatite B. Determination d'une nouvelle strategie therapeutique chez 13 patients. Presse Med. 1988;17:1522-1526.
11. Guillevin L, Ronco P, Verroust P. Circulating immune complexes in systemic necrotizing vasculitis of the polyarteritis nodosa group. Comparison of HBV-related polyarteritis nodosa and Churg-Strauss angiitis. J Autoimmun. 1990;3:789-792.
12. Hagen E, Daha M, 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. Kidney Int. 1998;53:743-753.
13. Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross WL, Hagen EC, Hoffman GS, Hunder GG, Kallenberg CG, McCluskey RT, Sinico A, Rees A, van Es LA, Waldherr R, Wiik A. Nomenclature of systemic vasculitides. Proposal of an international consensus conference. Arthritis Rheum. 1994;37:187-192.
14. Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481.
15. Lightfoot RW Jr, Michel BA, Bloch DA, Hunder GG, Zvaifler NJ, McShane DJ, Arend WP, Calabrese LH, Leavitt R, Lie JT, Masi AT, Mills JA, Stevens JC, Wallace SL. The American College of Rheumatology 1990 criteria for the classification of polyarteritis nodosa. Arthritis Rheum. 1990;33:1088-1093.
16. Luqmani R, Bacon P, Moots R, Janssen B, Pall A, Emery P, Savage C, Adu D. Birmingham Vasculitis Activity Score (BVAS) in systemic necrotizing vasculitis. Q J Med. 1994;87:671-678.
17. Mahr A, Guillevin L, Poissonnet M, Ayme S. Prevalences of polyarteritis nodosa, microscopic polyangiitis, Wegener's granulomatosis, and Churg-Strauss syndrome in a French urban multiethnic population in 2000: a capture-recapture estimate. Arthritis Rheum. 2004;51:92-99.
18. McMahon BJ, Heyward WL, Templin DW, Clement D, Lanier AP. Hepatitis B-associated polyarteritis nodosa in Alaskan Eskimos: clinical and epidemiologic features and long-term follow-up. Hepatology. 1989;9:97-101.
19. Pernice W, Sodomann CP, Luben G, Seiler FR, Sedlacek HH. Antigen-specific detection of HBsAG-containing immune complexes in the course of hepatitis B virus infection. Clin Exp Immunol. 1979;37:376-380.
20. Peto R, Pike M. Conservation of the approximation (O - E)/E in the logrank test for survival data or tumor incidence data. Biometrics. 1973;29:579-584.
21. Prince AM, Trepo C. Role of immune complexes involving SH antigen in pathogenesis of chronic active hepatitis and polyarteritis nodosa. Lancet. 1971;1:1309-1312.
22. Rozario R, Ramakrishna B. Histopathological study of chronic hepatitis B and C: a comparison of two scoring systems. J Hepatol. 2003;38:223-229.
23. Scott DG, Bacon PA, Elliott PJ, Tribe CR, Wallington TB. Systemic vasculitis in a district general hospital 1972-1980: clinical and laboratory features, classification and prognosis of 80 cases. Q J Med. 1982;51:292-311.
24. Trepo C, Guillevin L. Polyarteritis nodosa and extrahepatic manifestations of HBV infection: the case against autoimmune intervention in pathogenesis. J Autoimmun. 2001;16:269-274.
25. Trepo C, Ouzan D, Delmont J, Tremisi J. Superiorite d'un nouveau traitement etiopathogenique guerissant la periarterite noueuse due au virus de l'hepatite B par la combinaison d'une breve corticotherapie, de vidarabine et d'echanges plasmatiques. Presse Med. 1988;17:1527-1531.
26. Trepo C, Thivolet J. Hepatitis associated antigen and periarteritis nodosa (PAN). Vox Sang. 1970;19:410-411.
27. Trepo C, Zuckerman A, Bird R, Prince A. The role of circulating hepatitis B antigen/antibody immune complexes in the pathogenesis of vascular and hepatic manifestations in polyarteritis nodosa. J Clin Pathol. 1974;27:863-868.
28. Wright M, Goldin R, Fabre A, Lloyd J, Thomas H, Trepo C, Prada, M. HENCORE collaboration. Measurement and determinants of the natural history of liver in hepatitis C virus infection: a cross sectional and longitudinal study. Gut. 2003;52:574-579.
29. Zuckerman AJ. Proceedings: hepatitis B, immune complexes, and the pathogenesis of polyarteritis nodosa. J Clin Pathol. 1976;29:84-85.

Cited By:

This article has been cited 1 time(s).

Medicine
Deaths Occurring During the First Year After Treatment Onset for Polyarteritis Nodosa, Microscopic Polyangiitis, and Churg-Strauss Syndrome: A Retrospective Analysis of Causes and Factors Predictive of Mortality Based on 595 Patients
Bourgarit, A; Toumelin, PL; Pagnoux, C; Cohen, P; Mahr, A; Guern, VL; Mouthon, L; Guillevin, L
Medicine, 84(5): 323-330.
10.1097/01.md.0000180793.80212.17
PDF (108) | CrossRef
Back to Top | Article Outline
© 2005 Lippincott Williams & Wilkins, Inc.