Response to Vera and Larrea: ‘Concerns regarding methodological and interpretive aspects in the evaluation of <scp>BTHC</scp> ‐ <scp>PAGGSM</scp> red cell storage’

We would like to thank the editor for the opportunity to respond to the concerns raised by Vera and Larrea [1] in their letter to the editor about the publication ‘Clinical evaluation of red cell concentrates in non-DEHP plasticized containers compared with standard Di-ethyl-hexyl-phthalate (DEHP) plasticized containers: A quasi-randomised observational multicentre study’ [2]. We would also like to thank the authors for their careful reflections on our work and for their interest in the clinical evaluation of non-DEHP blood storage systems. One of the points raised by Vera and Larrea relates to the significantly different adverse event rates between the study phases, raising concerns about the validity of pooling data. We would like to clarify that pooling of data from the two haemovigilance phases was not undertaken to mitigate observed discrepancy in outcomes but, rather, to increase statistical precision in the context of low transfusion reaction rates. For this reason, we pre-specified a strategy in which Phase 1 and Phase 2 data were pooled, while also performing analyses restricted to reactions with higher imputability (definite, probable or possible). This approach was designed to maximize the information on rare safety signals. Importantly, both haemovigilance phases followed the same predefined study design. A prospective statistical analysis plan was established before data collection and inspection, and we adhered to this plan rather than modifying it in response to observed results. We consider this particularly important to avoid data-driven analytical choices in safety evaluations. We further note that all results were also reported separately for each phase, enabling independent re-analysis by other investigators using alternative statistical methods if desired. With only two study phases, however, we see limited value in approaches such as random-effects or hierarchical models, since there is insufficient information to reliably estimate between-phase variance. Likewise, the added benefit of formal interaction testing is questionable, given that the two phases did not differ in study design. Additionally, Vera and Larrea correctly note the imbalance in group sizes between DEHP-saline-adenine-glucose-mannitol and N-butyryl-n-hexyl citrate/phosphate-adenine-glucose-guanosine-saline-mannitol. This reflects logistical availability of the non-DEHP product during the study period rather than differential allocation by risk. Important to note in this context is that adverse event collection and reporting were identical in both groups, minimizing the risk of systematic bias. Finally, we agree that greater details on the Bayesian design and sample size considerations are valuable. These details are provided in the study protocol and can be made available upon request to facilitate transparency and reproducibility. We appreciate the opportunity to clarify the concerns raised by Vera and Larrea and welcome continued discussion as haemovigilance data on alternative blood storage systems continue to accumulate. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.