An incisional hernia (IH) is a type of hernia caused by an incompletely healed surgical wound [1,2,3]. IH after midline laparotomy and other anterior abdominal wall operations is a common, challenging problem for surgeons [1]. Accordingly, various methods of suture closure and mesh reinforcement have been used to treat IH and restore abdominal wall integrity [4,5,6,7].

Biological mesh can be used in septic contexts in which synthetic materials are unsuitable [8]. These biological mesh, which are made of collagen matrix, allow cellular colonization by the host tissue, neovascularisation, and progressive replacement by the host tissue. This avoids the occurrence of chronic infections on the mesh, which would require removing the mesh.

To date, no randomized studies have evaluated the efficacy of biological mesh compared to that of herniorrhaphies in contaminated/infected fields, and very few studies have considered the differences among various types of biological mesh available for clinical use; moreover, all available data are from retrospective or prospective non-randomized studies. As biological mesh is widely used, it is necessary to evaluate its efficacy. This study investigated the risk factors for the recurrence of incisional hernias after using biological mesh.

Materials and methods

All patient files at Limoges University Hospital from February 2009 to February 2015 that mentioned Permacol™ biological mesh placement were retrospectively compiled. We included patients with incisional hernias only treated with biological mesh, regardless of the reason mentioned by the surgeon.

All patients were contacted by phone and underwent new clinical examinations in September 2016. Clinical recurrence was determined by physical examination and was defined as a detectable gap in the abdominal wall with or without visceral bulging (clinical recurrence). The computed tomography (CT) findings were also analysed (radiologic recurrence). The data extracted from the registry included patient age, sex and other characteristics, such as Body Mass Index (BMI), smoking habits, diabetes, previous other abdominal wall hernias, American Society of Anesthesiologists (ASA) score, hernia characteristics (location, width, length, primary and recurrent hernia), surgical characteristics (open or laparoscopic, operating time, emergency surgery, oncological context, intestinal resection), mesh placement technique (onlay, inlay, sublay, underlay, or intraperitoneal placement), closure of the fascia (yes or bridge), Altemeier wound classification (clean/clean-contaminated/contaminated/dirty) [9] and Ventral Hernia Working Group (VHWG) septic status [10]. The sizes of the defects were measured by CT, if performed preoperatively. In the absence of CT, the measurements were based on clinical information. Postoperative morbidity according to the Clavien–Dindo classification and infectious prosthetic complications were assessed.

The primary evaluation criterion was the recurrence of incisional hernia after a minimum of 18 months. This recurrence was further assessed by CT with a minimum follow-up of 18 months up to 6 years. In cases that lacked morphological data, a clinical evaluation was performed to look for cases of recurrence.

Statistics

Qualitative variables are described as the frequency distribution percentage. Quantitative variables are described as the median and mean. Comparisons of categorical variables between two groups of subjects were made using chi-square or Fisher’s exact tests. The distributions of quantitative variables were compared using Student’s t tests or non-parametric Mann–Whitney U tests for unpaired series or Wilcoxon’s test for paired series for variables that did not follow a normal distribution. Univariate analyses were followed by multivariate analyses to determine the profiles of ‘recurrent’ patients. To be part of the first multivariate model, the variables had to present a significance of ≤ 0.20 in the univariate analyses. The multivariate model was simplified using the step-down method to obtain the final model. The relevant interactions between the variables that were present in the final model were investigated. In all analyses, p < 0.05 was taken to indicate statistical significance. Statistical analyses were performed using SAS version 9.1.3 (SAS Institute, Cary, NC, USA).

Results

Baseline patient characteristics

The retrospective database included 105 patients treated with Permacol™ biological mesh from February 2009 to February 2015. Thirteen patients were excluded because they did not have IH (fistula repair with biological mesh intervention, n = 9; prophylactic reinforcement, n = 2; prosthetic reinforcement during hiatus hernia repair, n = 1; strangulated umbilical hernia repair, n = 1). No data were available regarding the surgical procedure of five patients. Postoperative assessments were not possible in 19 patients (early death, n = 6; lost to follow-up, n = 13). Thus, 68 patients with IH who were treated using biological mesh were included in this study. Patient consent to use their medical data for research purposes was obtained before surgery. These patients consisted of 38 women (55.9%) and 30 men (44.1%) with an average age of 65.5 years (28–89 years) and an average BMI of 30.6 kg/m2 (Table 1). The rates of diabetes, regular smoking, and previous abdominal surgeries (> 2) in the study population were 25%, 32.4%, and 67.6%, respectively. In preoperative evaluations, 3 patients were classified as ASA I (4.4%), 35 patients as ASA II (51.5%), 26 patients as ASA III (38.2%) and 4 patients as ASA IV (5.9%) (Table 2).

Table 1 Demographic data
Full size table
Table 2 Comorbid conditions
Full size table

Incisional hernia characteristics and surgical characteristics (Tables 3, 4 and 5)

Table 3 Incisional hernia characteristics
Full size table
Table 4 Surgical context
Full size table
Table 5 Situation of the biological mesh placements
Full size table

All patients underwent general anaesthesia and were given broad-spectrum antibiotics at the induction of anaesthesia. According to the Altemeier classification, 24 patients (35.3%) were classified as grade III, and 11 patients (16.2%) were classified as grade IV. According to the Ventral Hernia Working Group grading system, 27 patients were classified as grade III (39.7%), and 18 patients were classified as grade IV (26.5%). All operations were laparotomies. Among them, 15 were performed in an emergency setting for acute symptoms (22.1%) and intestinal resection was performed in 32 patients (47.1%). The biological mesh was placed intraperitoneally with closure of the anterior fascia in 27 cases (“fascia approximated” group, 39.7%). The meshes were placed in the retromuscular pre-fascial plane in 1 case (1.5%) and the pre-aponeurotic plane in 1 case (1.5%). Closure of the anterior fascia in front of the mesh was not achieved in 39 cases (57.3%), including 20 cases (“intraperitoneal bridging” group, 29.4%) in which the mesh was placed intraperitoneally and 19 cases in which the mesh was placed between the edges of the fascia (inlay bridging group, 27.9%).

Postoperative complications

Of the 68 patients included in the analyses, 43 (63.2%) developed one or more postoperative complications within 90 days, and most patients had wound complications. The rate of complications classified as Clavien–Dindo grade I–II was 19.1% and those as grade II–III was 44.1%. The mortality rate was 0%. Overall, the number of postoperative surgical site infections (pSSIs) was 37 (54.4%). All 90-day postoperative outcomes are presented in Tables 6 and 7.

Table 6 Postoperative complications
Full size table
Table 7 Patient demographics and comparison between two groups
Full size table

There were 42 recurrences, corresponding to a rate of 61.8%. The rate of recurrence was significantly lower in the fascia approximated group (37%) than in the «inlay bridging» group (73.7%, p = 0.02) and «intraperitoneal bridging» group (85%, p = 0.001) (Fig. 1). The median time from hernia mesh repair to recurrence was 18 months (Fig. 2A). Healing was achieved in 26 patients (38.2%). Furthermore, hernia recurrence was clinically obvious in 34 of these patients (81%). For 10 patients, hernia recurrence was confirmed by radiology reports (23.8%). Among the 42 patients, 7 patients required the placement of new biological mesh (10.3%).

Fig. 1
figure 1

The rate of recurrence in each treatment group

Full size image
Fig. 2
figure 2

Kaplan–Meier curve for hernia recurrence over time, for A all patients, B whether fascia was approximated or bridged (intraperitoneal onlay mesh without pre-fasciotomy closure + bridging repair) and C degree of wound contamination [with or without postoperative surgical site infection (pSSI)]

Full size image

The median time from hernia mesh repair to recurrence according to fascia closure was 8 months in the “fascia bridged” group. This time was not reached in the “fascia approximated” group, p = 0.001. The median time from hernia mesh repair to recurrence was 5 months in the group with pSSIs versus 72 months in the group without pSSIs, p < 0.0001 (Fig. 2B, C).

Univariate analyses of risk factors for procedural failure indicated an increased risk of recurrence in cases of surgical site infection (OR = 9.39, [3–29.46]), complications according to Clavien–Dindo grade III and higher (OR = 22.73, [2.81–183.6]), absence of fascial closure in front of the mesh (OR = 8.69, [2.83–26.64]), operating time > 180 min (OR = 4.85, [1.23–19.16]), and VHWG score > 2 (OR = 3.62, [1.27–10.34]). (Table 8).

Table 8 Univariate analysis of risk factors for recurrence
Full size table

In logistic regression analyses, the risk factors for recurrence were postoperative infections (OR = 6.2, [1.6–36.34]), placement of bridged biological mesh (OR = 22.3, [4.08–121.62]), and postoperative morbidity grade III and above (OR = 16.7, [1.33–208.61]) (Table 9).

Table 9 Multivariate analysis of risk factors for recurrence
Full size table

Discussion

To the best of our knowledge, our study is one of the largest series to evaluate the long-term outcomes after IH treated with biological mesh. Building on a 6-year experiment on the use of biological mesh in the department of digestive surgery, this study showed a significant recurrence rate of 61.8%. Notably, this recurrence rate was more noticeable with bridged biological mesh placement and was higher in patients with surgical site infections.

In view of the results of this study, the use of biological prostheses must be questioned. Our study confirms the conclusion reached in the paper written by Majumder et al. that suggested that biological mesh is associated with a higher recurrence rate [11]. In their study, which included 126 patients undergoing major ventral hernia repair in clean-contaminated/contaminated fields, the synthetic mesh improved wound morbidity rates over that of biological mesh [11]. The important recurrence rate of this expensive mesh needs to be a key consideration when each medical file is discussed by surgeons before hernia surgery. The optimal operative solutions for ventral hernia repair in clean or contaminated fields have been discussed for many years. This is due to the fear of infective complications affecting permanent synthetic implants, which has led to a trend towards the overuse of biological mesh. Thus, recommendations have been established to guide surgeons in the choice of mesh [12, 13].

The rate of recurrence in this study was high (61.8%) and can be explained by the rate of Altemeier grades of III–IV (51.5%) and the rate of bridging technique used by the surgeons (57.7%). This bridging technique is used in specific situations when closure of the fascia cannot be achieved and when the surgeon cannot perform the component separation technique. Although this technique is useful in difficult situations, it is associated with a high recurrence rate in the literature.

Among the patients with recurrence in our study, 16.6% had a second surgery with biological mesh. In septic contexts in which synthetic materials are unsuitable, the surgeon preferred used a biological meshes on these select patients, despite the first recurrence. IH has often adverse effects on quality of life (QoL), as disability, pain, long-term dysfunction and dissatisfaction [2, 3, 14].

However, the rate of recurrence in our study is similar to that reported by Abdelfatah et al. (66%) [15] but is higher than that reported in other retrospective studies (7–40%). For example, Rosen et al. and Diaz-Siso et al. reported recurrence rates of 31.3% and 7.9%, respectively [16, 17]. The rate of recurrence increased over time, which may explain the high rate in our series due to the long follow-up period of 6 years. Giordano et al. reported recurrence rates of 9.2% at 1 year and 18.3% at 2 years, while Itani et al. reported rates of 15% and 22% in the corresponding years [18, 19].

Nevertheless, there are clearly large disparities in the recurrence rates among different series. These differences may be explained at least in part by the following factors.

First, many studies have investigated the behaviour of cross-linked biological implants compared to that of non-cross-linked mesh. Thus, heterogeneity in grades of hernia, study designs, and total numbers of patients receiving each graft make it difficult to perform clear comparisons among the reports.

Second, the patient selection criterion also varies among different studies, particularly with regard to the Altemeier classification. Indeed, in a review of the literature, Slater et al. reported that Altemeier grades III and IV appear to be risk factors for recurrence (OR = 1.9, 95% CI 1.24–2.91) [20]. However, many studies have not included many cases of grade III and IV hernias. For example, in the study by Diaz-Siso et al., 60% of cases were classified as Altemeier grade I, 25% as Altemeier grade II, and only 15% had either contaminated or dirty surgical sites, i.e., Altemeier grade III or IV [17]. The rate of Altemeier grades III–IV was 26.6% in the European multicentre retrospective study from Giordano et al. [18], compared to 51.5% in the present study. In a multicentre prospective study with a follow-up of 2 years, Itani et al. reported a recurrence rate of approximately 30%, with 49% of the cases classified as grade II, 49% as grade III, and 2% as grade IV [19]. Nevertheless, in a retrospective study with a follow-up period of more than 5 years, Abdelfatah et al. reported a recurrence rate of approximately 66%, but 18% of the cases were classified as Altemeier grade III–IV [15]. These observations suggest that the Altemeier classification is not the only explanatory element. Moreover, the rate of patients lost to follow-up in the last series reached 11% (6/56). The variation in the number of patients lost to follow-up among the studies represents one of the explanatory elements for the heterogeneity of the results.

Third, a factor capable of explaining the disparities in recurrence rates include the duration of postoperative follow-up, as suggested by Rosen et al., who reported survival rates without recurrence of approximately 92% at 1 year, 77% at 2 years, and 51% at 3 years [16]. Giordano et al. and Itani et al. both had follow-up periods of only 2 years, and Diaz-Siso et al. had only a 3-year follow-up [17,18,19]. The longest follow-up reported to date is the study from Abdelfatah et al. (5 years), which also reported the highest recurrence rate in the literature [15]. Thus, our study covered the longest follow-up period to date (up to 6.5 years).

Fourth, abdominal wall closure seems to partly explain the high rate of recurrence in the present study. Our series had a bridged mesh placement rate of 57.3%, which is associated with a high rate of recurrence. According to the location of the mesh, the recurrence rate was also significantly lower in the "fascia approximated" group than in the "fascia bridging" group, but there was no difference between the "inlay bridging" group and the "intraperitoneal bridging" group. In a recent multicentre, observational study of 109 patients who had undergone CAWR with Permacol™, the rate of recurrence was 5.4% with fascial closure versus 18.2% without fascial closure [18]. Jin et al. recorded a recurrence rate of 67% in a group of 37 patients with bridged mesh placements compared to 33% in cases of midline closure [21]. These recurrence rates were as high as 88.9% in the study from Patel et al. [22]. More recently, Madani et al. reported a high risk of recurrence in cases of bridged mesh placements (OR = 10.67, 95% CI 2.42–76.08, p < 0.01) [23].

Many techniques achieve midline closure and are used in mesh placements. These techniques may involve simple discharge incisions [24], traditional separation of the abdominal components [25, 26], and/or separation of the abdominal components laparoscopically. The multicentre COBRA study in 2017 confirmed the trend of an increased risk of recurrence for intraperitoneal prostheses (OR = 3.4, 95% CI 1.098–10.59, p = 0.0339) and recurrence rates of 40% and 13% in the intraperitoneal group and retromuscular group, respectively (p = 0.0451) [27]. In the present study, most of the mesh was placed intraperitoneally (64%), which could partially explain our high rate of recurrence. Nevertheless, a previous meta-analysis reported reductions in the risks for recurrence and infection with intraperitoneal mesh [28]. Postoperative infections may be risk factors of recurrence [7, 27]. In the present study, we established a strong link between pSSIs and recurrence. The high incidence of surgical site infections likely explains the high rates of recurrence in our series. The factors that may explain the increase in recurrence after infections include inflammatory reactions related to infectious episodes. Indeed, infections result in the secretion of collagenases and enzymatic reactions that cause mesh degradation.

To the best of our knowledge, this was the largest case series study of IH treated with Permacol™ with the longest follow-up period reported to date.

Limitations

This study had several limitations. First, our results are not based on randomized data, leading to potential selection bias. Second, given the significant heterogeneous patient characteristics and indications for biologic mesh as well as the variations in technique, our findings may not be applicable to all clinical situations. Although we believe our study has brought to light a number of important findings, its use as a guide for mesh selection in particular scenarios should be undertaken with caution. A prospective multicentre study is required to analyse these risk factors and provide precise insights regarding the use of these biological mesh.

Conclusions

Permacol™ porcine dermal biological mesh is associated with an increased risk for recurrence in patients with surgical site infections. Given the additional costs associated with the use of biological mesh, it seems necessary to avoid situations that present a high risk of recurrence by either increasing the number of midline closures after mesh placement or by diagnosing surgical site infections earlier and providing treatment that allows the rescue of these implants. Overall, this study challenges the purported advantage of biologics in treating incisional hernia repairs.