Reply to Katkuri et al

We thank Sharma et al1 for their reading of our article2 and for their constructive comments on study design and reporting. We welcome the opportunity to respond and to further contextualize our findings. First, we agree that our evaluation is an observational, single-centre, uncontrolled before–after comparison undertaken within a departmental service redesign. As stated in the Methods, Discussion, and Limitations, implementation of the On-Call Specialist Aortic Rota occurred alongside updates in operative strategy, perfusion management, and perioperative pathways. Accordingly, we framed our results as associative rather than causal, acknowledging risks of confounding, learning-curve effects, secular trends, and calendar-time bias. We, therefore, concluded that these findings require confirmation in multicentre cohorts using robust quasi-experimental designs. Nonetheless, reporting outcomes after service reconfiguration remains valuable in acute type A aortic dissection, where randomized evaluation of organizational change is rarely feasible. Second, regarding covariate adjustment, we chose a multivariable mortality model a priori to respect the limited number of outcome events and to minimize overfitting. Over-parameterization in logistic regression is known to produce unstable coefficients, exaggerated effect sizes, and inflated type I error, particularly when events are limited relative to the number of predictors.3 Age and previous cardiac surgery were chosen a priori based on established prognostic importance and supported by univariable screening, as detailed in the Statistical Analysis section. The inverse probability of treatment weighting (IPTW) analysis was performed as a complementary approach to address confounding from non-random group assignment, incorporating a broader set of pretreatment variables and achieving acceptable covariate balance (standardized mean differences