Reply: Dual-Plane versus Prepectoral Breast Reconstruction in High–Body Mass Index Patients

Sir: We thank Zhang et al. for their interest in our article.1 Their comments and suggestions are addressed below. We are cognizant of the fact that very few events could potentially produce unreliable risk assessments. For this reason, we used "any complication" as the event and not individual complications in our stepwise regression model to increase the number of events per predictive variable. As for the number of predictive variables, we settled on eight variables—five categorical variables (i.e., dual-plane reconstruction, diabetes, neoadjuvant and adjuvant radiotherapy, and adjuvant chemotherapy) and three continuous variables (i.e., age, body mass index, and mastectomy specimen weight)—all of which have been shown in previous published studies to influence reconstructive outcomes. In univariate analysis, all five of the categorical variables were found to be significant predictors of any complication. Of note, only two of the five variables had fewer than 10 events and none had zero events; the latter would be more problematic than events fewer than 10. Furthermore, the 10-events-per-variable rule is considered to be too conservative, and others have found a range of circumstances in which coverage and bias are within acceptable levels despite fewer than 10 events per variable.2 Variables such as surgical duration of procedures; axillary management (e.g., axillary lymph node dissection, implant size); oncologic characteristics such as BRCA or PALB2 carrier; and postoperative breast tumor, axillary nodal, and metastasis stages verified by pathologic evaluation were not included in our analyses. Although we acknowledge that these could be potential confounding variables, the reality is that there could be any number of confounding variables, and it is not possible to include all variables and then have enough events per variable for logistic regression analysis. Thus, our approach was to start with variables known to be predictive of complications and include the plane of reconstruction, an unknown variable, and assess that as a predictor of complications after accounting for the known variables. Patients were selected for prepectoral reconstruction in our study, based on our published selection criteria,3 which were relaxed as we gained greater confidence with the technique.1 By the same token, patients for dual-plane reconstruction have also been selected. The dual-plane technique has been performed for the past decade or so, during which time the technique and patient selection have been optimized. Compared to the dual-plane technique, prepectoral reconstruction is a relatively newer technique that is still evolving. Patient selection and reconstruction in our study were performed according to standard protocols for each technique. Robust mastectomy flap perfusion is critical for immediate device-based breast reconstruction, regardless of the technique. Flap perfusion in our study was clinically or objectively assessed, and immediate reconstruction was performed only when perfusion was deemed adequate. When flap perfusion was compromised, reconstruction was delayed regardless of the reconstructive technique. Importantly, patients with poorly perfused flaps were not channeled into dual-plane reconstruction. Our study was a retrospective study. As such, there are inherent limitations, and we acknowledge that. Nonetheless, the result suggests that prepectoral reconstruction may be a better reconstructive option in patients with a high body mass index. This finding merits further study, as the optimal reconstructive approach in this population remains to be determined. We thank Zhang and colleagues for their insightful comments and suggestions. ACKNOWLEDGMENT A publication grant from Allergan, Madison, New Jersey, was used for writing, editorial, and data analyses assistance. DISCLOSURE Dr. Gabriel is a consultant for Allergan. Dr. Maxwell has no financial interest to report. No funds were received or utilized for the research reported in this communication. Allen Gabriel, M.D.G. Patrick Maxwell, M.D.Department of Plastic SurgeryLoma Linda University Medical CenterLoma Linda, Calif.