Treatment of polyarteritis nodosa: Comment on the article by Guillevin et al


To the Editor:

I read with interest the recent article Guillevin et al (1). The authors are members of a cooperative study group in France that has reported (over the past 2 decades) several studies of systemic vasculitis syndromes, in particular evaluation of regimens that include antiviral drugs in the treatment of hepatitis B virus (HBV)–associated polyarteritis nodosa (PAN). As a preface to my comments, I submit the following points as background.

First, all patients with acute HBV infection probably experience an early phase that is the equivalent of one-shot acute serum sickness. From the classic studies of Krugman et al (2), it is known that the first HBV reactant to appear in the circulation is HBV surface antigen (HBsAg), which percolates thoughout vascular and extravascular compartments for a period of weeks before the appearance of viral DNA or anti-HBV antibodies and before the onset of clinical illness. When the host immune complex is expressed, the distribution of HBsAg (as in acute serum sickness) determines organ system involvement; this may be subclinical, mild, moderately severe (arthritis/cutaneous vasculitis syndrome), or severe, including multisystemic patterns of involvement consistent with PAN. In addition to organ-directed inflammation, anti-HBsAg antibody, reacting with circulating antigen, can mediate diffuse vascular injury via immune complex deposition. The majority of patients with PAN (whether HbsAg positive or not) have an acute illness; if they escape irreversible injury to vital organ systems during that phase, they survive (with or without chronic residua) and infrequently manifest recurrent active vasculitis. The keys to success are diagnosis early in the course of the illness and prompt institution of treatment.

Second, there are 2 objectives in the management of HBV-associated PAN: 1) the first priority is suppression of inflammation with corticosteroids, sometimes combined with cytotoxic drugs and plasma exchange (the efficacy of the latter is difficult to judge because, when used, it is almost always combined with other treatments), and 2) institution of antiviral therapy. Two agents (lamivudine and interferon-α) have proven efficacy in the suppression of HBV viremia and promotion of seroconversion. Because of the relative rarity of HBV-associated vasculitis, there are no prospective controlled therapeutic trials. The tendency to standardize management by protocols, i.e.,“one plan fits all,” is not appropriate, considering the clinical and pathologic heterogeneity described above.

Third, the American College of Rheumatology (ACR) criteria for classification of PAN were designed to assure some semblance of uniformity in reporting cases, but they are not accurate in either establishing or excluding diagnosis. Three of 10 criteria are required to meet the ACR classification for polyarteritis. The 10 criteria are as follows: 1) weight loss (at least 4 kg), 2) livedo reticularis, 3) testicular pain/tenderness, 4) myalgia, weakness, or tenderness, 5) mononeuropathy or polyneuropathy, 6) diastolic blood pressure >90 mm Hg, 7) elevated serum blood urea nitrogen or creatinine, 8) hepatitis B infection, 9) abnormal arteriography, and 10) biopsy specimen showing a small or medium-sized artery containing granulocytes with/or without mononuclear leukocytes. There are endless scenarios illustrating how bedside application of these criteria for diagnosis would lead one astray.

Finally, sometimes sequential observations of an individual patient over time can be particularly instructive; an example is the followup over more than 3 decades of the first patient in whom Koch's postulates were fulfilled regarding HBV etiology of PAN. In this subject, patient # 5 in 2 publications (3, 4), PAN developed 3 weeks after known infusion of a single unit of HbsAg-positive blood; more than a decade later she had seroconverted to all HBV reactants except HbsAg, and that antigen was (at least in part) in immune complex form. After another decade, the patient was still alive and fairly well. HbsAg antigen had disappeared, and anti-HBsAg antibodies were measurable. Beyond the first several months of her illness there were no clinical manifestations of active vasculitis.

Reviewing the study of Guillevin et al and in light of comments above, I accept the fact that their patients did have HBV-associated syndromes, but I am not convinced that all 10 patients had PAN. Seven subjects had relatively mild disease (zero on the “5 factor score”), and 6 patients seroconverted between 0.5 and 6 months, periods not inconsistent with natural seroconversion following HBV infection. It is stated that 7 patients had “histologic proof of vasculitis,” and the site of biopsy in 5 subjects (see Table 3) was described as “neuromuscular.” Does that mean both nerve and muscle sites were positive? It would be useful to know if all of the “positive” biopsy specimens demonstrated leukocytes in artery walls (ACR classification cri- teria) because perivascular leukocyte exudation can be a feature of almost any inflammatory process.

Why am I raising these questions? I think there is a wide spectrum of HBV-associated vascular syndromes. All of them warrant prompt treatment, but there is a strong case for individualizing treatment regimens according to anatomic patterns and severity, adjusting the dose and duration of corticosteroid, and for patients with very severe disease, employment of cyclophosphamide (maybe even plasma exchange). I strongly suspect that patients with the relatively common HBV-associated arthritis and dermalvasculitis syndrome have vasculitis in other sites. However, in the absence of evidence of visceral involvement and considering the overall benign character of that problem, aggressive management is not warranted. Finally, I believe the authors have made a good case for antiviral therapy of patients with HBV-associated PAN, given the relative safety of lamivudine therapy and the high frequency of persistent HBsAg antigenemia in such subjects.

Charles L. Christian MD*, * University of Florida, Jacksonville.