We thank Dr. Puéchal and Drs. Watts, Lane, and Scott for their interest in our work and their comments. As pointed out by Dr. Puéchal, our results differ not only from those of the hospital-based study by Watts et al (1), but also from those of a study of rates of hospitalization due to RA-related complications in California (2), in which a decline over time in the rate of hospitalization due to vasculitis was observed. As discussed in our article, we believe that a thorough survey of a community-based sample provides a better estimate of the burden of severe extraarticular disease in patients with RA than data based on hospital registers. In addition, differences in case definitions may explain part of these discrepancies.
Dr. Puéchal suggests that our restricting the analysis to the first decade of illness might have affected the results, since a substantial number of patients present with vasculitis >10 years after disease onset. We agree that long-term followup would yield more information on possible changes over time. There is an inherent problem, however; there is no way we could have long-term data for patients with recently diagnosed RA. In order to avoid bias due to differences in length of followup, we have grouped the patients by decade of RA diagnosis and, for the purpose of the analysis, calculated cumulative incidence rates with truncated followup. As suggested by Dr. Puéchal, we have also analyzed the incidence of rheumatoid vasculitis beyond the first decade of illness. Since our survey was performed in 2001, long-term data for patients with RA diagnosed in 1985–1994 are not available. We analyzed the 20-year cumulative incidence of severe extraarticular disease by decade of diagnosis for patients with RA onset in 1955–1964, 1964–1975, and 1975–1984. The 20-year cumulative incidence rates for severe vasculitis were 3.6%, 4.3%, and 6.2%, respectively (P = 0.88). Thus, in our sample there is no evidence for a decline in the incidence of rheumatoid vasculitis, but we cannot exclude the possibility of a change in the 10-year cumulative incidence in patients diagnosed as having RA after 1994, nor can we exclude the possibility of a change in the incidence beyond the first decade of illness in patients diagnosed after 1984. Data on this can only be provided by future studies.
Drs. Watts, Lane, and Scott comment on some relevant differences between our study and theirs (1). However, their statement indicating that our report “does not include any data from after 1994” is not correct. Our followup was through December 31, 2000, although only patients with RA diagnosed up to 1995 were included. We did not see any decrease in the incidence of vasculitis in patients with onset of RA in 1985–1994 compared with patients with onset of RA before 1985, indicating that there was no major fall in the incidence of rheumatoid vasculitis in the 1990s corresponding to that observed by Watts et al (1).
We agree with Drs. Watts, Lane, and Scott that the differing patterns for the incidence of primary systemic vasculitis and systemic rheumatoid vasculitis observed in the Norfolk study are of major interest. However, we do believe that the best estimate of the incidence of extraarticular manifestations, including vasculitis, comes from thorough studies of defined populations of patients with RA.