Comment on: Comparable 12-month renal response outcomes with medium- vs high-dose starting prednisone in a real-world study of lupus nephritis: reply

Dear Editor, We thank Dr Duarte-Garcia for his interest in our manuscript and his sophisticated comments [1]. Previous work from his group on the dose-relationship between starting prednisone dose and response in lupus nephritis has stimulated our interest in this area [2]. Here, the author draws attention to an important methodological consideration regarding the dichotomization of prednisone dose, noting that treating glucocorticoid exposure as a continuous variable may better capture dose–response relationships. This comment is relevant to the interpretation of observational analyses’ data and aligns with ongoing efforts to refine glucocorticoid-sparing strategies, while maintaining efficacy. With respect to the suggestion to analyse prednisone dose as a continuous variable, this approach was not originally pursued in our analyses for two main reasons: First, our primary aim was to compare clinically relevant starting-dose thresholds, which we believe are more applicable to everyday practice and decision-making by treating physicians. We were concerned that modelling prednisone dose continuously could dilute this clinical message. Second, in our cohort, the majority of patients clustered around a limited number of commonly used starting doses (e.g. 30 or 40 mg), and given the relatively small sample size, we anticipated that the sparse representation of patients at other dose levels would limit the robustness and interpretability of continuous-dose modelling.