A retrospective, Italian multicenter study of complex abdominal wall defect repair with a Permacol biological mesh

Complex abdominal wall defects (CAWDs) can be difficult to repair and using a conventional synthetic mesh is often unsuitable. A biological mesh might offer a solution for CAWD repair, but the clinical outcomes are unclear. Here, we evaluated the efficacy of a cross-linked, acellular porcine dermal collagen matrix implant (Permacol) for CAWD repair in a cohort of 60 patients. Here, 58.3% patients presented with a grade 3 hernia (according to the Ventral Hernia Working Group grading system) and a contaminated surgical field. Permacol was implanted as a bridge in 46.7%, as an underlay (intraperitoneal position) in 38.3% and as a sublay (retromuscolar position) in 15% of patients. Fascia closure was achieved in 53.3% of patients. The surgical site occurrence rate was 35% and the defect size significantly influenced the probability of post-operative complications. The long-term (2 year) hernia recurrence rate was 36.2%. This study represents the first large multi-centre Italian case series on Permacol implants in patients with a CAWD. Our data suggest that Permacol is a feasible strategy to repair a CAWD, with acceptable early complications and long-term (2 year) recurrence rates.

Pre-operative, operative and post-operative procedures. All patients underwent a pre-operative clinical assessment and the CAWD was evaluated by computed tomography (CT). The CT images were used to identify the CAWD site, the width and number of the defects, the anatomical position and trophic status of the rectus abdominis muscles, loss of domain, fistula and the presence of any other wounds or complications. Loss of domain was defined when the herniated organs lost their right of domain inside the abdominal cavity and so the volume of the hernia could no longer be reduced to the abdominal cavity.
All patients were operated on under general anesthesia and were given perioperative antibiotics according to hospital protocol. In each of the four hospitals, the same surgeon performed the CAWD repair for all patients. The surgeons across all four hospitals used one of three main surgical techniques to close the defect with the Permacol implant: (1) sublay, placement of the biologic mesh in the retromuscolar space in an extra-peritoneal position after the closure of the rectus posterior fascia; (2) underlay, placement of the mesh in an intraperitoneal position deep to anterior abdominal wall defect and then closure of the fascia over the mesh; or (3) bridge, placement of the mesh in an intraperitoneal position, with the center part of the mesh forming a bridge between the edges of the rectus sheath when fascial closure was not feasible. All patients attended 3, 6, 12 and 24-month post-operative follow-up examinations with the original surgeon. All patients were assessed by abdominal examination, and patients with clinical suspicion of recurrence were also assessed by CT to rule out recurrence. Study outcomes. The study outcomes were as follows: (1) surgical site occurrence (SSO) at 30 postoperative days; and (2) recurrence rate at 1-year and 2-years follow-up. SSO was defined as the presence of one or more of the following features: wound dehiscence, skin/fat necrosis, cellulitis/abscess, hematoma, seroma and/or enterocutaneous fistula. Recurrence was defined as a fascial defect at the surgical site. Bulging was defined as a relaxing of the abdominal wall at the mesh site due to stretching of the mesh, without any interruption between the mesh and the abdominal wall layer 21,25 . Bulging was considered in the SSO group and was not considered as a recurrence. The complications were defined according to the Clavien-Dindo classification system 26 .

Statistical analyses.
All collected data were reported in a specific study database and statistical analyses were performed using SPSS software (IBM SPSS Statistics for Windows, version 23.0). Descriptive statistics were used to characterize the study population by using frequencies, mean and median values. A two-tailed Chi-square test was used to statistically compare proportions. Continuous variables were tested for normality using the Kolmogorov-Smirnov test and compared by Student's t-test or Mann-Whitney test for independent samples, as appropriate. A p-value < 0.05 was considered statistically significant. informed consent. Informed consent was obtained from all individual participants included in the study.

Results
cohort characteristics. We recruited 60 eligible patients from four Italian hospitals who underwent wall reconstruction with Permacol implantation for a CAWD between January 2010 and May 2016 (Table 1). Most patients underwent Permacol implantation at the "Cannizzaro" (n = 21) or "Policlinico-Vittorio Emanuele" (n = 24) hospitals. The remaining patients were treated at the Civil Hospital of Lentini (n = 5) or the "San Vincenzo" hospital (n = 10). The mean age was 62 years (range 41-88 years) with 39 (65%) females and 21 (35.5%) males. The ASA score was high (III-IV) in 63.3% of patients. Sixteen patients (26.6%) had a BMI >35, such as 16 (26.6%) patients had diabetes and 16 (26.6%) patients had pulmonary disease; 22 (36.6%) patients were smokers. Abdominal surgery for neoplastic disease had been previously performed in 43.3% of the patients (Table 1) www.nature.com/scientificreports www.nature.com/scientificreports/ Defect characteristics. The abdominal wall defect was categorized as a recurrent incisional hernia in 12 (20%) patients: six cases were the first recurrence, two cases were the second recurrence, three cases were the third recurrence and one case was the fourth recurrence. Most (96.7%) defects were located along the midline, with only two cases (3.3%) being lateral. According to the Chevrel classification, the midline defects were predominantly M3-M4 and W3-W4 (Table 1).
Surgical procedure. Surgery was performed as an emergency in 20% of cases. Permacol was implanted in the sublay position in nine (15%) patients, in the underlay position in 23 (38.3%) patients and as a bridge in 28 (46.7%) patients. Fascial closure was achieved in 53.3% of the surgical procedures. The median mesh size was 540 cm 2 (range 120-1,400 cm 2 ), and in three patients two Permacol meshes were implanted to ensure comprehensive coverage of the defect. In three other patients, a partially absorbable synthetic lightweight multifilament mesh was implanted onlay (placement of the mesh over the anterior abdominal wall into the subcutaneous tissue) to enforce the midline fascial closure. These three patients had a grade 2 hernia according the modified VHWG system; none of these three patients required ostomy closure. In these cases, the surgeon decided to enforce the midline because the incisional hernia was at least a second recurrence and the abdominal wall muscles were of poor quality. In 12 (20%) cases, a previously implanted and infected synthetic mesh was removed during the Permacol implant surgery. No patients required component separation. The mean operating time was 192 min (range 100-360 min). In all patients, two drains were inserted into the subcutaneous tissue and these were removed when the output was <50 mL/day.  (Table 2). These complications included: 14 (66.6%) patients with wound infections, five (23.8%) with seroma, one (4.7%) with skin necrosis and one (4.7%) with bulging. Most SSOs (57.1%) were grade 1-2 according to the Clavien-Dindo classification system. One patient with a wound infection died from pneumonia with respiratory failure. In one patient with minor wound dehiscence, the Permacol mesh was removed after transcutaneous migration of the implant. Here, the patient's body did not incorporate the mesh, but rather created a thick fibrotic layer without a hernia defect; the mesh was rejected a little at time, until the surgeon decided to remove it. All wound infections were managed with advanced wound dressings and in four patients VAC (vacuum assisted closure) therapy was applied with good outcomes. VAC therapy was also used in patients with skin necrosis. Early post-operative complications occurred in 50% of patients with a BMI > 35, and 37.1% of patients with modified VHWG grade 3. The complication rate was significantly lower in patients classified as Chevrel W1-2 (7.6%) compared to those classified as Chevrel W3-4 (44.4%) (p = 0.011) ( Table 3).
Post-surgical follow-up. The mean hospital length-of-stay was 12.75 days (range 3-65 days); the mean hospital length-of-stay was significantly longer for patients with early complications (18 days vs. 10 days) (p < 0.05; Mann-Whitney test). The 30-day mortality rate was 6.7% and was unrelated to the mesh implant. Cardiovascular complications were the cause of death in two patients, and pneumonia with respiratory failure was the cause of death in another two patients. The mean follow-up was 21.5 ± 17.11 months, but 10 patients (16.7%) were lost to follow-up. Excluding these 14 patients, the hernia recurrence rate in the remaining 46 patients was 32.6% (Table 4), with the majority (93.3%) recurrences occurring within 12 months. Recurrence was confirmed radiologically by CT scan. There were no statistically significant differences in the recurrence rate according to the surgical repair technique. Patients classified as Chevrel W3-4 were more likely to develop a recurrent hernia than patients classified as Chevrel W1-2 (40.0% versus 9.1%); however, this difference between the two groups was not statistically significant (p = 0.07). Finally, 53.3% of patients who developed complications presented with a recurrent hernia within 24 months of surgery.

Discussion
Biological meshes have been used for abdominal surgery for ~20 years, and their use is mainly dictated by the type of surgical wound. Some studies have suggested that a biological mesh can better resist infections than a synthetic mesh, and are thus recommended for use in contaminated fields or in patients with VHWG grade 3 23 . Data from a recent meta-analysis that compared the outcomes from biological versus synthetic meshes in contaminated wounds, however, showed that a biological mesh is not superior to a synthetic mesh in terms of SSO or recurrences 27 . In our opinion, the evidence from this meta-analysis is weak due to the poor quality of the studies included. Patients with CAWD typically present with associated problems due to comorbidities, contaminated wounds, fistulas or synthetic mesh infection. In these cases, the surgical repair procedure is extremely difficult and complex. For this reason, we consider that using a biological mesh for CAWD repair at least avoids the problems and risks directly incurred by a synthetic mesh.
A literature review on biological meshes found that Permacol is one of the most studied mesh types in the published literature 28 , especially in the context of CAWD repair. These studies, however, are mainly retrospective, single institution case series 5,9,[12][13][14][28][29][30][31][32][33][34] with only three multicenter retrospective studies 3,15,22 and one cross-sectional study 21 published to date. Our series of 60 patients represents the first large, multi-centre Italian study on Permacol mesh implants in patients with CAWD. We found that most patients had more than one comorbidity that was not well compensated, which explains why 63.3% of our patients had a high ASA score. Consistent with a previous study 3 , the ASA score did not significantly impact on the incidence of early complications or hernia recurrence.
In 2010, the VHWG proposed a grading system for SSO risk and concluded that a biological implant should be considered in patients with a potential risk of infection (grade 2-3), and preferred in high risk patients (grade 4) 6   www.nature.com/scientificreports www.nature.com/scientificreports/ indicated due to the presence of a defect >10 cm, multiple previous repair attempts or multiple defects, as indicated by CAWD characteristics 5 .
The surgical technique can have an impact on clinical outcomes. Giordano et al. found that independent of the mesh position, fascial closure is the only significant factor that affects hernia recurrence, with a higher recurrence rate when closure is not achieved (18.2% versus 5.4%) 5,15 . In our cohort, we achieved fascial closure in 53.3% patients, but this did not significantly impact the complication or recurrence rate (despite the fascia not being closed in 52.4% patients with complications and 60% of patients with recurrences). In our case series, no patients required component separation. This finding might be because the technique has only recently been implemented in our practice for CAWD repair, and the technique extends the complexity of an already challenging surgical procedure. Over recent years, a biosynthetic mesh to repair a CAWD is often considered in alternative to biological mesh, when a complete fascial closure is achieved with anterior or posterior component separation with or without transversus abdominis release (TAR) technique.
During our patient enrolment period (2010-2016) most of the participating surgeons widely used the bridging technique and the component separation technique was rarely applied in combination with a biological mesh. Our practice has changed over recent years, with a preference for underlay and sublay mesh placement even in combination with the component separation technique. The bridge technique is associated with an increased recurrence rate. Bridging defects with a biological mesh leads to stretching and laxity and can, therefore, result in failure of the repair, or at least in a bulging 5,29 . A previous study found that when Permacol was used to bridge a fascial defect, the hernia recurrence rates were >80% 15 . Here, we found that only 32.1% patients encountered a recurrence 24 months after bridging repair with Permacol. Only one patient developed a bulging. We did not consider a bulge as a recurrence because, even if the midline is not closed over the biological mesh, there is no interruption between the mesh and the abdominal wall layer 21,25 . Rather, we consider that for patients requiring CAWD repair, the onset of a postoperative bulging or abdominal wall weakness is an acceptable result and does not represent complete failure of the procedure.
The Permacol collagen matrix promotes new collagen deposition and thus confers biocompatibility and immunogenic ability that promotes infection resistance 6,7,10 . With regards to the risk of infection and the complexity of the cases in our series, we found that 35% of patients developed a complication, mainly either a wound infection (23.3%) or seroma (8.3%). The onset of complications significantly prolonged the hospital length-of-stay. A previous large series reported a high complication rate of 40.5%, with 19% seromas, 15.2% wound infections and 3.2% hematomas occurring after Permacol implantation 3 . These rates have been confirmed by a recent multicenter audit that reported an SSO rate of 48.8% consisting mainly of abscesses (66.4%) 22 . Only 29.6% cases in this series required a repeat intervention; these cases were mostly due to intraperitoneal complications (31.1%), wound abscesses (36.5%), seroma (10.8%), hematoma (12.2%) or wound dehiscence (9.5%) 22 . Such complications, however, can be readily managed by bedside incision and drainage and oral antibiotics 20,21 . Similar to previously published studies 22 , we found that the defect size, expressed by the Chevrel classification in width (W), influenced the rate of post-operative complications. Defects >10 cm typically require greater mobilization of the muscolofascial units; here, tissue microvascularization must be sacrificed and a third space must be created. This process tends to be associated with a higher rate of post-operative infections. Finally, biological mesh removal has been reported in case studies where an infection with mesh disruption has occurred 21,22 ; only rarely has a biological mesh been removed because of rejection 29,35,36 . In our series, the mesh was removed in only one case.
Using Permacol as a crosslinker increases graft stability and durability, even in contaminated fields. Here, the hernia recurrence rate at 12 months is lower (6.6%) than the recurrence rate when using a non-cross-linked porcine mesh (21.2%) 28 . In our study, the overall hernia recurrence rate was 25% at 24 months follow-up. The majority of these affected patients (93.3%) had already presented with a hernia recurrence at 12 months. Although the wound class has been suggested to increase the risk of hernia recurrence following CAWD repair 5 , we found no such effect on the recurrence rate in our series. Our data are consistent with the published literature 5,[13][14][15]33 , especially when comparing our findings to studies where a long-term follow-up of at least 24 months was achieved. Further studies are now needed to identify the risk factors for hernia recurrence and complications, using multivariate analyses.

conclusions
The high cost of Permacol means that it cannot be recommend for routine abdominal wall reconstruction 37 . As such, some clinicians only support this type of mesh for use in CAWD repair 5,7,15 . Unfortunately, the data regarding the superiority of using a biological mesh versus a synthetic mesh for CAWD repair are poor. Randomized studies in the context of CAWDs are difficult to perform due to a lack of uniformity in CAWD classification, the various possible surgical techniques that can be performed and high patient heterogeneity. For this reason, some clinicians do not support the use of a biological mesh for CAWD repair until more conclusive data are available 27,38,39 .
Our study represents the first large multi-centre Italian study on Permacol implantation in patients with CAWD. The use of biological meshes has increased rapidly over the past 10 years, but high-level evidence is lacking to demonstrate superiority. Our study, although retrospective, supports the use of Permacol in CAWD repair. Consistent with previous reports 5,15,16,21,22 , we propose that Permacol is safe and feasible, and show that it elicits an acceptable early SSO and long-term recurrence rate. This multicenter experience using Permacol should encourage surgeons who encounter challenging surgical situations like CAWDs to use a biological implant such as Permacol; continued use of Permacol will help grow the evidence base. Future multicenter studies that compare the different types of biological and biosynthetic mesh are now required. In addition, a long-term follow-up is now warranted to better evaluate the utility of biological mesh implants for abdominal wall repair in high-risk patients.