Authors' Reply: Clarifying Crossover Interpretation and Volume-Signal Assessment in ASSIST

We thank Cheng et al. for their Letter to the Editor1 regarding our JASN article2 and agree that crossover trials necessitate close attention to the sequence of randomization, as there may be carryover effects from the first period. Only the second treatment period was extended to 24 weeks, as the premise was to determine whether atrasentan afforded further proteinuria reduction beyond 12 weeks when combined with a sodium-glucose cotransporter 2 (SGLT2) inhibitor. Thus, we are unable to present a sequence-specific result for this secondary end point. As stated in the article, however, there was no clear indication of a carryover effect at week 12. Furthermore, absolute values of urinary protein-to-creatinine ratio are clearly depicted at each time point in Figure 3B.2 A mixed-model analysis is requested for volume-related safety end points, body weight, B-type natriuretic peptide, hemoglobin, and BP, as was done for the primary end point. The Consolidated Standards of Reporting Trials 2010 Statement3 is cited as justification; however, the Consolidated Standards of Reporting Trials guideline does not state this is necessary for all potential effects. Indeed, while hypothesis testing using P values is extremely important for primary, secondary, and even tertiary end points in clinical trials, they must be used appropriately and judiciously.4 We clearly illustrate the sequence- and period-specific results of all volume-related parameters throughout the trial in Figures 5A and 6, A, C, and E, of our publication.2 Moreover, since all participants in Atrasentan and Sodium-Glucose Cotransporter 2 Inhibitor Efficacy and Safety (ASSIST) were taking an SGLT2 inhibitor, these additional analyses would not address the posed question of whether SGLT2 inhibitors blunt the potential for atrasentan to cause fluid retention.