In the subset of patients who entered sustained remission during the study, increases in serum PR3-ANCA levels were also not associated with subsequent clinical relapses. Forty percent of patients in clinical remission who had an increase in serum PR3-ANCA levels relapsed within one year, but 60% did not.
This study is important because the sample was large, the follow-up was long, and clinical assessments were standardized and used a validated measure of vasculitis activity. The antibody tests had large inter-assay variation, but sera from all time points of each patient were tested together in the same assay. The data were carefully analyzed and considered not only correlations over time, but whether antibody concentrations decreased with clinical remission or increased with relapse.
But does this study close the book completely on this topic? Likely not. While the definition of sustained remission was credible (a score of zero on the Birmingham Vasculitis Activity Score for Wegener granulomatosis for six months), relapse was defined as any score above zero on this scale. We do not know if minor symptoms such as arthralgias or fever constituted the majority of relapses, or if glomerulonephritis or pulmonary involvement represented the majority of relapses. Maybe the lack of association between antibody levels and relapse was a consequence of many “relapses” being due to mild symptoms alone. The study did not grade the severity of relapse, leaving one to wonder if PR3-ANCA levels might be associated with severe relapses.
The study also does not address whether serum PR3-ANCA levels correlated with clinical disease activity in subsets of patients with particular disease manifestations. For example, at study entry, patients with renal involvement were more likely to have PR3-ANCA antibodies than those without renal involvement. It would be interesting to know if antibody levels correlated with the activity of glomerulonephritis or risk of relapse among these patients. PR3-ANCA levels were found to be decreased among patients in sustained remission, and the time to achieve remission was shorter among those with a decrease in antibody levels. Although the latter association was not statistically significant, the power of the study to detect a 60% difference in time to remission as significant might have been limited. These findings suggest the possibility that the lack of association in the overall cohort may be hiding associations in a clinically relevant subset.1
Diagnostic tests are used to separate diseased from nondiseased persons. Good diagnostic tests are able to detect affected individuals regardless of whether the disease is active or not. In contrast, good evaluative tests accurately reflect disease activity. Therefore, it should not be surprising that good diagnostic tests are rarely good evaluative tests. This study suggests that serum PR3-ANCA in patients with Wegener granulomatosis follow this rule. The authors conclude that ANCA levels should not be used to guide treatment. Whether this apples to all subsets of patients and different levels of disease severity is still uncertain. While I would not use PR3-ANCA levels to guide treatment, I might monitor patients more closely if I knew their antibody levels were rising.