The largest multicentre data collection on prepectoral breast reconstruction: The iBAG study

Abstract Background and Objectives In the last years, prepectoral breast reconstruction has increased its popularity, becoming a standard reconstructive technique by preserving pectoralis major anatomy and functionality. Nevertheless, the lack of solid and extensive data negatively impacts on surgeons’ correct information about postoperative complication rates and proper patient selection. This study aims to collect the largest evidence on this procedure. Methods A multicentre retrospective audit, promoted by the Barcelona Hospital, collected the experience of 30 centers on prepectoral breast reconstruction with Braxon ADM. The study had the scientific support of INPECS and IIB societies which provided the online database Clinapsis. Results A total of 1450 procedures were retrospectively collected in a 6‐year period. Mean age 52.4 years, BMI 23.9, follow‐up 22.7 months. Reconstruction was carried out after a tumor in 77.1% of the cases, 20.1% had prophylactic surgery, 2.8% had revisions. Diabetes, smoke, and immunosuppression had an influence on complications occurrence, as well as implant weight. Capsular contracture was associated with postoperative radiotherapy, but the overall rate was low (2.1%). Complications led to implant loss in 6.5% of the cases. Conclusions The international Braxon Audit Group multicentre data collection represents a milestone in the field of breast reconstruction, extensively improving the knowledge on this procedure.


| INTRODUCTION
Breast cancer nowadays represents the most common tumor among women, with over 2 million new cases in 2018. Currently, its surgical treatment is based on oncoplastic surgery, a synthesis between those two disciplines with the final goal of an outcome balanced between the best oncologic and cosmetic results. Indeed, since its conception, breast reconstruction helped patients look "normal" when dressed; more recently, advancements in surgical techniques and medical technologies have raised the bar so that patients can feel esthetically pleasing also unclothed. 1 During the past decades, the surgical approach to breast cancer has evolved from radical mastectomy to the development of breast-conserving surgery and reconstructive techniques. In fact, after a nipple or skin-sparing mastectomy, the opportunity of having performed an immediate reconstruction represents a great advantage for patients, owing to its significant psychosocial benefits. Currently, immediate implant-based breast reconstruction (IBR) represents 81.9% of all breast reconstructive procedures. 2 However, only in the last decade, surgeons approached the prepectoral technique considering the importance of the role of the pectoralis major muscle and the possibility to avoid its recruitment for implant coverage: starting with the first cases described by Berna et al, the subcutaneous breast implant positioning becomes a concrete option, along with a complete breast implant wrapping with a large sheet of ADM to prevent the direct contact of the silicone prosthesis with the surrounding tissues. 3,4 Prepectoral breast reconstruction (PPBR) has quickly become more and more popular due to its satisfying esthetic outcomes, low functional detriment, and low complication rate. In fact, whereas the traditional submuscular reconstructive technique involves several drawbacks, such as unpleasant esthetic results and postoperative discomfort due to implant displacement, animation deformity, and capsular contracture, PPBR has overcome most of them. 5,6 Unfortunately, despite the current multiplicity of different devices committed to PPBR, there is still a lack of high-quality evidence to support the safety of these new materials. In fact, most of the published studies are limited to small single-center experiences with no homogeneous outcomes. [7][8][9] As mentioned above, the first biologic membrane able to completely cover a silicone implant to position it subcutaneously was a 0.6 mm-thick preshaped porcine ADM (Braxon; DECOmed Srl, Venice, Italy). 3 To date, this device represents the most widely used one for PPBR in European and UK breast centers and the only ADM with a specific patented design that allows a standardized wrapping technique, reducing the number of variables so that data can be more easily compared among different centers. 10 After a preliminary multicentre study on 100 PPBR cases carried out with Braxon ADM, 11 the first author encouraged a larger EU-UK database on PPBR performed with the same device to obtain valuable evidence about outcomes and risk factors potentially linked to postoperative complications: for this reason, the group of study was defined as international Braxon Audit Group (iBAG). Our study wants to represent an extensive and homogeneous retrospective multicentre data collection that sets the basis for future randomized prospective studies. indeed, some general guidelines were already published about patient's selection for PPBR procedure and they include women with no history of radiation, non-or ex-smokers, a good subcutaneous layer, a well-perfused skin flap, an estimated mastectomy weight of less than 500 g and tumors not invading skin or chest wall, or those with inflammatory nature. Considering the retrospective nature of this study, the authors do not set any exclusion criteria with the aim to verify any association between patient characteristics and postoperative complications. Thus, if some of the 30 centers involved in the iBAG study enlarged the indication, therefore offering PPBR to thin or obese women, as well as diabetics or smokers, data coming from these cases were analyzed as well. The sole patients excluded from the analysis were those with less than 3 months of follow-up.

| Ethical standards
The research protocol of the study was approved by the Ethics Committee of the Santa Creu i Sant Pau Hospital; each participating center was required to obtain local committee approval. In accordance with the Declaration of Helsinki of 1975, as revised in 1983, all patients were asked to provide written consent for the use of their personal data, as part of the standard surgical consent of their institutions. Patients' data were visible only to the supervisors of each center where the reconstructive procedure was carried out; no one else had access to those records. At the end of the period of data collection, the coordinating center had the authorization to manage all the anonymized data and to analyze them through statistical protocols.

| Material and surgical procedure
After mastectomy or surgery performed for revision purposes, the reconstructive phase involves the preparation of the ADM, which requires rehydration of 5 minutes in room-temperature saline solution to facilitate its handling. Then, the ADM is wrapped around the breast implant, its excess is cut away and it is closed by mean of absorbable 3/ 0 suture ( Figure 1). Then, the completely wrapped implant is placed inside the breast pocket and fixed with cardinal stitches to the pectoralis major muscle fascia and to the subcutaneous layer through quilting sutures in the anterior part of the Braxon ADM, thus eliminating dead spaces and supporting ADM integration.
Surgeons involved in the study performed mastectomies according to specific clinical scenarios and their routine practice with the only common characteristic that the reconstructive phase did not entail the pectoralis major; implant selection (round or anatomical shape, fixedvolume or adjustable implant, as well as temporary tissue expander) depended on surgeons' preference. Breast reconstruction was then classified as two-stage when performed with an expander or one-stage when a definitive implant was used. In some cases, the ADM was used for patients requiring revision surgery due to implant-related complications or undesired surgical outcomes.
In all cases, perioperative and postoperative antibiotics and drains were used according to local policy.

| Statistical analysis
Considering the relevant sample size, χ² and the Student t test tests were firstly used to study all independent variables potentially influencing the studied outcome, particularly postoperative complications. A model was then designed for all variables to explain the evidence derived from the data set; it was based on hierarchical and generalized additive models (GAM), mainly dictated by the poor plausibility of a linear link between regressors and the response variable, which need to exploit the flexibility of the GAM models. Moreover, for each variable there were differences between several sample units dictated by some specific characteristics of the data collection, thus motivating the use of hierarchical models. Variables with more than 20% of missing data were excluded from the statistical analysis.   Table 1.
F I G U R E 1 Standardized wrapping technique of Braxon ADM around the breast implant. The patented shaped allows a complete cover acting as an internal "implant bag" which distributes the weight of the silicone prosthesis around all the breast pocket and hides it from the direct contact with the organism. ADM acellular dermis matrix [Color figure can be viewed at wileyonlinelibrary.com] 850 | A t-test analysis revealed a significant difference in hospitalization time mean for patients with hypothyroidism (P-value = .016), with preoperative chemotherapy (P-value = .004), patients who underwent axillary surgery (P-value = .008), with postoperative chemotherapy (Pvalue = .003) or radiotherapy (P-value = .018). All these patients were more likely to have longer hospital stay; also, the type of reconstruction (one-or two-stage) influenced hospital stay (P-value = .001).     Patients were further divided into two subgroups: those undergone radiation therapies (pre-or postoperatively) and those not irradiated, to evaluate with a univariate analysis the potential impact of this treatment on postoperative complications (Table 3).
Through univariate analysis, it was investigated also if the administration time of radiation therapy, pre-or postoperatively, could represent a significant risk factor for complication development.
Radiation therapy appeared to be a statistically significant risk factor for the development of postoperative seroma, capsular contracture, rippling, and implant loss. Interestingly, at multivariate analysis, the negative impact of the postoperative treatment was confirmed only on capsular contracture, having no statistically significant association with the occurrence of all other complications (Table 3).   Overseas, studies have been published by American authors which have performed PPBR on larger cohorts, yet in a two-stage procedure with the use of tissue expander covered with ADM positioned on the frontal part only. 16 Moreover, the use of porcine ADM is still uncommon among the main American scientific papers, so that most evidence derives from human dermis such as AlloDerm. 17 The analysis of these studies should bring up concern on more aspects, the fact that two-stage techniques require a second surgery to exchange the device with a definitive implant and the relative increased risk of tumor recurrence. 18 In European daily practice, a direct-to-implant single-stage PPBR is preferably offered, since in most women a second procedure can be easily spared. Moreover, multiple surgeries are associated with an increased infection risk which could find confirmation in higher infection rates reported in the main American PPBR experiences. 19 In fact, Nahabedian reported very high infection rates (12.8% by patient; 8.1% by breast) not in line with iBAG study outcomes, while implant loss rate was similar (6.5%) and slightly different necrosis and seroma rates (9.7% and 4.8%). 17 High implant and expander loss rate, as well as infection one, were also found in Bettinger and Schnarrs publications about prepectoral expander/implant reconstructions, with overall infection rates of 9.1% for the first one and 11.8% for the second, reinforcing again our hypothesis. 20,21 Furthermore, patient's risk factors revealed by our statistical analysis find confirmations in other publications: Schnarrs highlighted higher postoperative complications risk in smokers and patients with >500 g breasts, 20 while Bettinger observed increased complication rates in previously irradiated breasts and no significant association between chemotherapy and postoperative adverse events. 21 In iBAG analysis a secondary interest was to investigate RT impact on postoperative outcomes; the multivariate analysis did not reveal any statistically significant correlation between RT and postoperative complications, except for capsular contracture, that even so did not reach a high percentage (about 5%) ( Table 3). Very few publications until now have focused their studies on PPBR and RT and moreover their populations were limited to small cohorts with a maximum of 54 patients undergone unilateral RT. [22][23][24] Even in these series the authors had experienced only with the two-stage reconstructive procedure and not with a direct-to-implant (DTI) PPBR. With regard to the impact of radiotherapy on PPBR outcomes, the iBAG study collected 190 irradiated patients for a total of 198 DTI-PPBR, which to our knowledge represent the largest evidence on this issue. The outcomes in Table 3  The role of ADM in the prevention of periprosthetic fibrosis was evaluated in several papers, also before the conception of PPBR, and its positive effect was confirmed in mutual agreement. 27 In conclusion, the iBAG data collection represents the largest study on PPBR performed using a complete implant wrapping with a preshaped porcine ADM. Data from this retrospective multicentre audit confirm the effectiveness of the technique and the low postoperative complication rate, especially capsular contracture; moreover, an ideal patient profile has been drawn by assessing the risk factors involved in postoperative complications' onset.
The authors agree to consider this study limited by its retrospective nature, especially for those patient's characteristics and intraoperative details that were collected in a percentage too small to reach the value sufficient for the statistical analysis and which in the future should be further investigated to better describe the correct patient selection. For this reason, a prospective study is planned: starting from the outcomes of this large data collection on PPBR we will analyze the impact of a more careful patient selection on the postoperative complication as well as implant loss rate. A prospective evaluation of the effect of radiotherapy is planned too. Other limits of this study were that esthetic and patient-reported outcomes were not reported and that postoperative pain scores were available only for a minority of patients so that could not be included as an outcome measure. This leaves room for further discussion of results, in particular in a longer follow-up and a greater number of irradiated patients so that to draw more stable conclusions in this interesting subgroup of patients.

| CONCLUSIONS
The iBAG study brings the largest evidence on PPBR, collecting the