To the Editor:

We thank Dr. Christian for his stimulating comments on the classification and differential diagnosis of HBV-associated PAN. We fully agree that the fact that patients included in our trial fulfilled the ACR criteria for PAN (1) should be interpreted with caution. Because the ACR vasculitis criteria sets were established to distinguish one primary systemic vasculitis from others (2), these tools should be applied exclusively to patients with a definitive diagnosis of vasculitis. Because microscopic polyangiitis was separated from PAN (3), it has been demonstrated that it too could satisfy the PAN criteria (4). Conversely, in response to one of the points raised by Dr. Christian, our personal experience does not support that the presence of granulocytes in the vessel wall, a feature highlighted by the ACR PAN criteria, orients towards a diagnosis of PAN with adequate sensitivity or specificity.

Regarding the potential limitations of the ACR criteria for PAN and in light of the observation that some of our patients experienced rapid HBe seroconversion under the treatment protocol, Dr. Christian raised the important question as to whether our patients might not have had a “dermatitis–arthritis syndrome,” rather than genuine PAN. Like PAN and glomerulonephritis, the dermatitis–arthritis syndrome is a well-recognized HBV infection-associated disease. All 3 entities result from immune complex-mediated tissue damage, but occur at different periods during the course of viral infection. Indeed, the dermatitis–arthritis syndrome is considered to be a serum sickness-like disease that occurs during the prodromal phase of infection and resolves spontaneously by the time jaundice develops, whereas PAN and glomerulonephritis develop at later stages as a consequence of sustained HBV antigenemia (5–7).

Therefore, we persist in thinking that, on clinical grounds, the multisystem HBV-associated disorder of our 10 patients was unequivocally PAN. The median interval between the first systemic symptoms and therapy onset was 3.9 months (range 22 days–19.7 months). In addition, none of our patients had a transaminase peak or jaundice, findings that argue against the possibility that the extrahepatic manifestations pertained to prodromata of acute HBV infection. The fact that all patients initially had clinical manifestations involving one or more sites other than the joints or skin is also not suggestive of a dermatitis–arthritis syndrome. Furthermore, the vasculitis was biopsy-proven in 7 patients (2 of 5 positive neuromuscular samples with vessel inflammation in both tissues), and another patient had an abnormal angiogram. Because none of the patients had characteristic signs of microscopic polyangiitis (e.g., alveolar hemorrhage, glomerulonephritis, positive antineutrophil cytoplasmic antibodies), it can be considered that, at least for those 8 cases, the ACR criteria were accurately applied and would support their classification as PAN, with a reported specificity of 87%.

Dr. Christian's comments also give us the opportunity to briefly reemphasize the utility of the five-factor score (FFS). The FFS was derived from a survival analysis of 342 patients with PAN, microscopic polyangiitis or Churg–Strauss syndrome (8) that identified 5 variables predicting an increased risk of mortality. Therefore, the FFS assesses the prognoses of patients with 1 of these 3 vasculitides, but not disease activity. Consequently, the finding that 6 of our patients had an FFS of zero implies that PAN was morebenign in these individuals but not a “forme fruste” of the disease. However, it cannot be excluded nowadays that cli- nicians' potentially heightened awareness of HBV-associated PAN might prevent, as a result of an earlier diagnosis, the development of the full-blown clinical picture of this entity.

Alfred Mahr MD*, Loïc Guillevin MD*, Pascal Cohen MD*, For the French Vasculitis Study Group * Hôpital Cochin, Assistance Publique–Hôpitaux de Paris, Paris, France.