Authors' Reply: Utility of Urinary CD4⁺ T-Cell Count in Detecting ANCA-Associated Vasculitis Renal Relapse

We thank Brousse et al. for their comments¹ on our article “Urinary CD4⁺ T Cells Predict Renal Relapse in ANCA-Associated Vasculitis.”² In retrospect, we agree that instead of choosing the wording “outperforming,” it would have been wise to use a more neutral terminology, such as stating that urinary T cells yield numerically higher area under the curve values. In principle, the only statistically robust finding of our study is that urinary T cells can predict future kidney flares in ANCA-associated vasculitis. This was the predefined primary end point of our study, which we therefore were able to report with statistic precision. All other analyses are secondary end points, and it can be debated how helpful P values, multiple testing, and conventional significance thresholds are for interpreting these findings. DeLong test for comparison of urinary T cells with routine clinical biomarkers yielded P values of 0.026 for hematuria, 0.062 for creatinine, 0.076 for albuminuria, 0.072 for cANCA levels, and 0.003 for pANCA levels, which should be interpreted in context with the known limited ability of these single markers (when used in isolation) for flare prediction.

More importantly, we completely agree with the second aspect that Broussee and colleagues¹ raise in their comment: the aim should not be to establish one single biomarker “to rule them all” but instead aim for a meaningful combination of biomarkers for optimal prediction. Therefore, we agree that the category “outperformance of other biomarkers” may be misleading. Instead, as suggested, various biomarkers could be combined in a model. However, we believe that such a model should be derived from a larger dataset than our study. We envision that in the future, ANCA-associated vasculitis disease activity and risk of relapse will be determined by a combination of biomarkers that reflect different elements of the pathogenesis of ANCA-associated vasculitis: serologic activity (i.e., ANCA levels, complement products), damage (i.e., proteinuria, hematuria, creatinine), and active intrarenal inflammation (i.e., urinary T cells, urinary sCD163). These elements could be combined in a model as a score, or as reflection of individual aspects of pathology, which could mandate different treatments. Either way, we agree that a combination of biomarkers may pave the way to modern, personalized treatment of ANCA-associated vasculitis.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/E696.

Funding

None.

Author Contributions

Writing – original draft: Philipp Enghard.

Writing – review & editing: Luka Prskalo, Adrian Schreiber.