Short-term corticosteroids then lamivudine and plasma exchanges to treat hepatitis B virus–related polyarteritis nodosa


  • Lamivudine was provided by Glaxo Pharmaceuticals.



To assess the efficacy and safety of lamivudine, an antiviral agent that strongly inhibits hepatitis B virus (HBV) DNA replication, combined with plasma exchanges after short-term corticosteroids for HBV-related polyartertitis nodosa (PAN).


Ten patients (8 men, 2 women, mean ± SD age 50.4 ± 14.4 years) with previously untreated HBV-related PAN were included in a multicenter, prospective, observational trial. Oral prednisone (1 mg/kg/day) was given for 1 week, then tapered and withdrawn within 1 week. Then, lamivudine (100 mg/day or less in the case of renal insufficiency) was started for a maximum of 6 months. Plasma exchanges were performed simultaneously and scheduled as follows: 3/week for 3 weeks, 2/week for 2 weeks, then 1/week until hepatitis B e antigen (HBeAg) to anti-HBe antibody (HBeAb) seroconversion was obtained or until 2–3 months of clinical recovery was sustained. The primary trial endpoint was clinical recovery from HBV-PAN at 6 months. The secondary endpoint was loss of detectable serum HBeAg and HBV DNA, and HBeAg to HBeAb seroconversion at 9 months.


One death, attributed to catheter-related septicemia, was recorded. At 6 months, all 9 survivors had achieved clinical recovery and by 9 months, 6 of 9 (66%) had seroconverted.


The strategy of short-term steroids followed by lamivudine and plasma exchanges effectively led to recovery from HBV-PAN. Because of its oral administration and good safety profile, lamivudine should henceforth be considered the antiviral agent of choice to treat HBV-related PAN.


Polyarteritis nodosa (PAN) is a systemic necrotizing vasculitis affecting medium-sized arteries (1). Although its etiology remains elusive, the strong association between PAN and hepatitis B virus (HBV) infection (2–4) clearly suggested a causal link with this agent in many patients. Epidemiologic data suggest that the frequency of HBV-related PAN (HBV-PAN) might currently be declining in developed countries (5), most probably as a consequence of effective vaccination programs and mandatory screening of blood donors.

Conventional therapy of PAN relies on corticosteroids and, for severe cases, adjunctive cyclophosphamide. Although this regimen may also control HBV-PAN (3), prolonged use of corticosteroids (6, 7) or cytotoxic agents (8, 9) in HBV-infected patients raises some concern with regard to their stimulatory effect on virus replication and enhanced host immune response against the virus that may cause fulminant liver disease. Therefore, 15 years ago, our group devised an original therapeutic strategy based on short-term corticosteroids, followed by the combination of plasma exchanges and an antiviral agent that successfully treated HBV-PAN (10, 11). Since then, various antiviral agents, e.g., vidarabine (12) and interferon-2α (13), have had proven efficacy in this setting.

Lamivudine (also known as 3TC) is a nucleoside analog that potently inhibits HBV DNA replication. It has been approved as the treatment of choice for chronic HBV infection (14–16) and also has shown efficacy against acute HBV infection (17–19). The present trial was established to investigate the efficacy and tolerance of lamivudine combined with plasma exchange after short-term corticosteroids prescribed to 10 consecutive untreated patients with recent-onset HBV-PAN.



This multicenter, prospective, observational therapeutic trial included 10 consecutive patients with newly diagnosed and previously untreated HBV-PAN. To be eligible, patients had to be at least 18 years old and PAN had to fulfill the American College of Rheumatology (ACR) criteria (20); histologic proof of vasculitis was preferred but not required. Patients also had to have detectable hepatitis B surface antigen (HBsAg), hepatitis B e antigen (HBeAg), and detectable levels of serum HBV DNA determined using a hybridization assay or polymerase chain reaction (PCR). Abdominal angiography was recommended for patients with persistent abdominal pain. To avoid the risk of hemorrhage associated with the puncture of potentially present arterial hepatic microaneurysms, liver biopsy was not systematically required. Patients were excluded when any of the following were present: coinfection with hepatitis C virus (HCV) or human immunodeficiency virus (HIV); positive antineutrophil cytoplasm antibodies (ANCA); pancreatitis (because of its potential aggravation under lamivudine therapy); pregnancy or breastfeeding; absence of contraception for women of childbearing age.

All the patients underwent an in-depth investigation of their disease at study enrollment. Poor general condition included weight loss, defined as a loss of >5% of the usual body weight during 1 month, asthenia, or temperature >38°C without concurrent infection. Virologic investigations comprised complete screening for HBV markers (HBsAg, anti-HBc antibodies [Ab], HBeAg, anti-HBeAb, serum DNA], serologic screening for HCV and, when positive, PCR, and HIV serology. ANCA were sought by immunofluorescence and anti-proteinase 3 and anti-myeloperoxidase activities were systematically determined using an enzyme-linked immunosorbent assay. Disease severity was assessed with the five-factors score (FFS) (21). The FFS is defined by the following parameters: proteinuria > 1 gm/day, renal insufficiency (creatininemia > 1.58 mg/dl), cardiomyopathy, gastrointestinal manifestations, and central nervous system involvement. FFS was previously shown to have significant prognostic value with 5-year survival rates of 88% when FFS = 0, 74% when FFS = 1, and 54% when FFS ≥ 2 (21).


The therapeutic scheme used for the 10 patients is shown in Figure 1. Treatment was started with prednisone (1 mg/kg/day) for 1 week; for severely ill patients, corticosteroid therapy could be initiated by a methylprednisolone pulse (15 mg/kg) for 1–3 days. The prednisone dose was rapidly tapered by half every day until complete withdrawal by the end of week 2. Then, lamivudine and plasma exchanges were begun. Patients took 100 mg of oral lamivudine once a day; for patients with impaired renal function, the lamivudine dosage was adjusted to reflect creatinine clearance as calculated with the Cockcroft and Gault formula (Table 1) (22). Lamivudine was maintained for 6 months or stopped at the time of seroconversion to anti-HBsAb. Plasma exchange sessions were scheduled as follows: 3/week for 3 weeks, 2/week for 2 weeks, then 1/week for the following weeks; they were stopped at the time of HBeAg to HBeAb seroconversion (see below) or after stable clinical recovery had been obtained and maintained for ∼2–3 months. For each plasma exchange session, the circuit was primed with 1 liter of starch and 60 ml of plasma/kg were replaced with 4% albumin. Lamivudine was taken after each plasma exchange session to avoid drug removal during the procedure.

Figure 1.

Therapeutic scheme for hepatitis B virus-related polyarteritis nodosa. Plasma exchanges and lamivudine could be prescribed for no more than 2–3 months and 6 months, respectively. They were stopped as soon as HBe seroconversion was documented. MP = methylprednisolone; OCS = oral corticosteroids; * = optional.

Table 1. Lamivudine dose adaptation as a function of creatinine clearance (CrCl)
CrCl (ml/min)Daily dose (mg)
Day 1Thereafter

Outcome measures.

During 6 months of treatment and through months 9 of the trial, the main clinical and laboratory parameters of all the patients were closely monitored. Patients' outcomes were assessed with respect to response to therapy, relapses, and mortality. The primary trial endpoint was clinical recovery from PAN at 6 months, defined as no sign of active vasculitis, with either complete resolution or clear-cut stabilization of all signs of organ system disease activity. The secondary endpoint was the loss of detectable circulating levels of HBeAg and HBV DNA, and the appearance of anti-HBeAb (referred to as HBe seroconversion) at 9 months. Relapse was defined as the reappearance of clinical and/or laboratory signs of vasculitis.


This trial was approved by the Comité Consultatif des Personnes Participant à la Recherche Biomédicale (CCPPRB; reference 102-98).


Patient characteristics.

Ten HBV-PAN patients, 8 men and 2 women, mean ± SD age 50.4 ± 14.4 years, were included between January 1, 1999 and February 28, 2000. All the diagnoses met ≥3 of the 10 ACR criteria, with 4 items present in 3 patients, 5 items in 4 patients, 6 items in 1 patient, and 7 items in 1 patient. The patients' principal clinical manifestations and laboratory results at entry are summarized in Tables 2 and 3. Histologic proof of vasculitis was obtained in muscle, peripheral nerve, or skin biopsies for 7 patients. Three patients were intravenous drug abusers, who potentially contracted HBV infection through percutaneous exposure; for the remaining 7 patients, the route of HBV acquisition was unknown. Nine patients had anti-HBcAb of the IgM isotype, suggesting recent infection. Liver biopsy was performed in 3 patients and revealed active hepatitis with calculated histology activity index scores of 1, 1, and 10 (23). Microaneurysms or stenoses were observed in 1 of 3 patients who had abdominal angiograms. FFS were 5 for 1 patient, 2 for 2, 1 for 1, and 0 for the other 6.

Table 2. Demographics and principal clinical manifestations in 10 patients with HBV-related PAN*
  • *

    Values are numbers (%) unless otherwise specified. HBV = hepatitis B virus; PAN = polyarteritis nodosa.

  • Defined as creatininemia >1.58 mg/dl.

Age, mean ± SD (range) years50.4 ± 14.4 (29–67)
Poor general condition9 (90)
Weight loss, >5% body weight9 (90)
Fever >38°C7 (70)
Myalgias/arthralgias8 (80)
Mononeuritis multiplex9 (90)
 Superficial peroneal nerve, no.7
 Deep peroneal nerve, no.3
 Cubital nerve, no.3
 Bilateral involvement, no.4
Gastrointestinal tract involvement5 (50)
Cutaneous involvement2 (20)
Vascular nephropathy3 (30)
 Proteinuria >1 gm/day3 (30)
 Hematuria1 (10)
 Creatininemia, mean ± SD (range) mg/dl1.65 ± 1.96 (0.67–7.14)
 Renal insufficiency2 (20)
Orchitis, no./n (%)5/8 (63)
Hypertension3 (30)
Cardiomyopathy2 (20)
Central nervous system involvement1 (10)
Table 3. Detailed clinical and biologic manifestations in 10 patients with HBV-related PAN*
Patient no/sex/age, yearsClinical symptomsAngiography findingsVasculitis-proven biopsyCreatininemia (mg/dl)FFSAST/ALTHBeAg seroconversion/time
  • *

    Poor general condition was defined as asthenia, loss >5% of usual body weight, and/or fever >38°C without concurrent infection. HBV = hepatitis B virus; PAN = polyarteritis nodosa; FFS = five-factors score; ALT/AST = serum alanine/aspartate aminotransferase; HBeAg = hepatitis B e antigen; ESR = erythrocyte sedimentation rate; N = normal; mo = months.

1/M/55Poor condition, high ESR, fever, weight loss, myalgias, arthralgias, Raynaud's phenomenon, abdominal pain, proteinuria, hypertension, mononeuritis multiplexMicroaneurysmsNeuromuscular1.862N/NYes/6 mo
2/M/67Poor condition, high ESR, fever, weight loss, myalgias, arthralgias, mononeuritis multiplex Neuromuscular1.1101.5N/2NYes/6 mo
3/M/37Poor condition, high ESR, fever, weight loss, arthralgias, mononeuritis multiplex, orchitis  1.1306N/6NYes/2 mo
4/M/70Poor condition, high ESR, fever, weight loss, myalgias, arthralgias, purpura, abdominal pain, cholecystitis, proteinuria, hypertension, mononeuritis multiplex, orchitisNormalSkin0.872N/NNo (died)
5/M/50Poor condition, high ESR, fever, weight loss, myalgias, arthralgias, abdominal pain, mononeuritis multiplex, orchitis Neuromuscular0.6702N/2NYes/8 mo
6/F/61Poor condition, high ESR, fever, weight loss, myalgias, arthralgias, abdominal pain, hypertension, mononeuritis multiplex Neuromuscular0.8502N/2NNo
7/F/60Poor condition, high ESR, fever, weight loss, myalgias, arthralgias, mononeuritis multiplex  1.130N/NYes/0.5 mo
8/M/35Poor condition, weight loss, myalgias, arthralgias, gangrene, abdominal pain, mononeuritis multiplex, orchitis, cardiomyopathy Neuromuscular1.351N/1.5NYes/6 mo
9/M/29Poor condition, weight loss, cardiomyopathy, abdominal pain, cholecystitis, proteinuria, hematuria, anuria, central nervous system involvement, orchitisMicroaneurysms 7.1451.5N/1.5NNo
10/M/40Myalgias, arthralgias, high ESR, mononeuritis multiplex Muscular0.701.5N/1.5NNo

Outcome results.

At 6 months, 9 of 10 patients had achieved clinical recovery; the remaining patient died after an initial clinical response. None of the 9 survivors relapsed during the subsequent 3 months of followup. In a per-protocol analysis at 9 months, 6 (66%) of the 9 survivors had achieved HBe seroconversion with a mean ± SD treatment onset to seroconversion interval of 158 ± 85 days (range 13–252 days). HBsAg became undetectable in 3 of the 9 survivors concomitant with anti-HBsAb seroconversion. Among the 3 patients who had not seroconverted to anti-HBeAb, 2 had sustained suppression of serum HBV DNA to undetectable levels at several consecutive determinations. Alanine or aspartate aminotransferase were never elevated in any patient during therapy or after therapy withdrawal, particularly concurrently with the loss of markers of HBV replication. Treatment was well tolerated in 9 of 10 patients; the sole death was attributed to Staphylococcus epidermidis septicemia related to central catheter (inserted for plasma exchange) infection. Sequelae were end-stage renal failure in 1 patient requiring dialysis, and minor sensorimotor deficits following mononeuritis multiplex in 9 patients.


The results of this observational trial indicate that lamivudine combined with plasma exchange after short-term corticosteroids is an effective therapy for HBV-PAN. This strategy achieved clinical recovery in all 10 patients and a 66% HBe seroconversion rate, defined as suppression of serum HBeAg and HBV DNA to undetectable levels and the presence of anti-HBeAb. This seroconversion rate is higher than the 36% rate previously observed with vidarabine (12) and equal to that obtained with interferon-2α (13). Tolerance of this therapy was good, although 1 patient succumbed to septicemia that was attributed to S. epidermidis contamination of the central catheter used for plasma exchange. Therefore, we recommend that plasma exchange should be performed through a peripheral venous access, which is less susceptible to infectious complications.

The therapeutic modality used to treat HBV-PAN is specifically tailored to its pathogenesis (24). In this setting, corticosteroids are administered for only 2 weeks to rapidly contain 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 plasma exchange combined with an antiviral agent: plasma exchanges remove the immune complexes, which are the major determinant of vessel inflammation (24), and the antiviral agent is added to facilitate immune-complex clearance by diminishing the virus load and, eventually, to stop virus replication. Although this approach has never been evaluated in a randomized trial, indirect evidence of its effectiveness has been provided by long-term followup data of our former trials showing a 0% relapse rate among patients achieving HBe sercoconversion and a 7-year overall survival rate of 83% (25).

Data from patients whose duration of HBV carrier status could be defined in previous observations indicated that PAN occurs early during the course of HBV infection, most frequently within the 12 months following contamination (25). Because the natural history of acute HBV infection is characterized by a spontaneous anti-HBsAb seroconverion rate of >90% of adult cases (26), recovery might occur spontaneously. However, in light of the potential life-threatening organ damage caused by uncontrolled vasculitis, it would be unethical to withhold treatment waiting for a hypothetical self resolution or to evaluate the benefit of this therapeutic strategy in a randomized placebo-controlled trial.

Results of our study suggest that lamivudine should henceforth be prescribed as the antiviral agent of choice for HBV-PAN. This recommendation is in agreement with the results of large therapeutic trials on patients with chronic HBV infection that documented lamivudine as an effective and safe therapy for chronic hepatitis B (14–16). Compared with interferon-2α, lamivudine yields similar HBeAg seroconversion rates and offers additional advantages, i.e., oral administration, fewer side effects (27), and a potentially more rapid antiviral effect. However, the emergence of virus subpopulations bearing the YMDD mutation, which has been associated with reduced HBV susceptibility to lamivudine, was detected in 14–32% of lamivudine recipients after 1 year of therapy (14–16). Thus, the appearance of mutants increases with the time of exposure to lamivudine (28) and particularly concerns prolonged therapies of >6 months (29). Because our protocol prescribed lamivudine for 6 months, this problem should not be of major concern in the setting of HBV-PAN.

Recent therapeutic trials on chronic HBV infection suggested that the combination of the 2 major antiviral agents, lamivudine and interferon-2α, might be more effective than either agent alone (27, 30). Although those findings are still open to debate (31), they raise the question of whether combining lamivudine and interferon-2α might also be useful in HBV-PAN. Isolated case reports have indicated the efficacy of this antiviral combination against HBV-PAN, including 1 patient who had previously failed on interferon-2α alone (32, 33). Further investigation is nonetheless warranted to assess adequately the benefit of this combined therapy for HBV.

The lingering question is, what therapeutic option should be adopted for patients not clinically responding to short-term corticosteroids followed by lamivudine and plasma exchange? Although no such case was observed in this trial, the decision of how to manage such therapeutic failures should be made after careful consideration of the severity of the vasculitis. For patients with mild or moderate PAN, it might be reasonable to attempt to add interferon-2α, although the inadequacy of the data now available on the efficacy of combining lamivudine and interferon-2α has to be kept in mind. For critically ill patients, however, it would appear justified to switch to a conventional immunosuppressive strategy with prolonged corticosteroids and cyclophosphamide. Given their potential deleterious effects on the underlying hepatitis, treatment with these agents should be as short as possible, but no clear recommendation can be given concerning the optimal duration. Based on observations in chronic HBV carriers receiving cytotoxic agents, maintenance of lamivudine prophylaxis should then be considered to prevent HBV activation (34).


The authors are indebted to P. Marcellin, MD, PhD for valuable discussions, and to Ms J. Jacobson for editorial assistance.