Synthetic versus Biologic Mesh for Complex Open Ventral Hernia Repair: A Pilot Randomized Controlled Trial

Patients and Methods

After obtaining Iinstitutional Review Board (IRB) approval, we performed a prospective, single-center, blinded pilot RCT of patients undergoing complex open VHR. Consolidated Standards of Reporting Trials (CONSORT) guidelines were followed, and the trial was registered in clinicaltrials.gov (NCT03091790).

All eligible patients who were determined to be candidates for complex open VHR in hernia clinics between September 2017 and January 2019 were approached. Inclusion criteria included any adult patient (≥18 years old) with a complex primary or incisional ventral hernia deemed most suited for open repair given a large size defined as defects >10 cm wide, contamination, or extensive adhesions. Exclusion criteria included those patients with an acute infection, such as an infected mesh from previous hernia repair with untreated drainage or abscess; patients with severe comorbidities limiting survival beyond two years after the operation based upon surgeon judgment (e.g., metastatic cancer or complicated cirrhosis); patients in whom the surgeon would not normally place a prosthetic mesh (e.g., planned second surgery such as an ostomy takedown); and patients with high likelihood of being lost to follow-up (e.g., resides out of state or no access to telephone). Patients with chronic mesh infections that had been drained and treated with antibiotic agents but with no active/acute mesh infections were eligible for enrollment. A surgical informed consent and a separate research consent were signed by those patients who agreed to be enrolled in the study. No modifications were made to the research protocol after enrollment of patients began.

Baseline demographic information was obtained through a complete medical history and physical examination during pre-operative clinic visits. Patients completed the modified Activity Assessment Scale (mAAS), a validated hernia-specific survey that investigates patients' abdominal wall quality of life (AW-QOL) through 12 questions [14,15]. Responses are measured on a 10-point Likert scale and then the cumulative survey results are normalized to a scale of 1–100. On this scale, a score of one is poor AW-QOL, 100 is perfect, 80 or greater is normal, and the minimally clinically important difference (MCID) is seven points [16,17].

Patients were randomly assigned intra-operatively through computer-generated variable block randomization to either biologic mesh (porcine acellular dermal matrix) or synthetic mesh (medium-density, macroporous polypropylene). Porcine acellular dermal matrix was selected because it is one of the most commonly used biologic meshes in the market and is readily available in our institution, whereas polypropylene was selected because it is the gold standard synthetic mesh used for uncomplicated VHR. At the time of abdominal closure and mesh placement the operating surgeon called a research coordinator not involved in data collection who performed the allocation using sequentially numbered, sealed, and opaque envelopes kept in the research coordinator's office. A one-to-one allocation ratio was used to ensure an equal number of patients in each group and patients were considered randomized when the sealed envelope was opened.

Only experienced hernia surgeons performing more than 50 open VHR per year and with equipoise on using biologic or synthetic mesh for complex ventral hernias participated in the study. Pre-operative skin decontamination with chlorhexidine solution, prophylactic antibiotic therapy, multimodal pain control regimen, and discharge criteria were standardized. To control for differences in operative techniques between surgeons, hernia repairs were performed following recommendations of several surgical societies.

In general, the abdomen was entered via a midline laparotomy. The incision was extended cranially and caudally to provide adequate access to all the margins of the fascial defect. The peritoneum was entered, and the sac contents were reduced into the abdominal cavity. All abdominal wall adhesions were divided. All previously placed mesh, if present, was explanted. When the midline could not be closed without undue tension or when there were lateral hernias, posterior component separation was performed. Prior to closure, we determined if primary fascial closure was achievable. Among patients in whom the fascia (anterior or posterior sheath) could not be closed with suture alone, the defect was bridged with in situ hernia sac or polyglactin 910 mesh.

Posterior fascial closure was performed with running 0-polyglactin 910 suture. The mesh was placed in the retromuscular position and secured with 0-polydioxanone sutures. Jackson Pratt 10-French drains were selectively placed anterior to the mesh based on surgeon judgement. Anterior fascial closure was performed with running 0-polydioxanone suture. All excess skin and subcutaneous tissue were excised, Scarpas and Camper's fascia were approximated with 0-polyglactin 910 suture, skin was closed with staples, and negative pressure wound therapy devices were applied.

The operating surgeon and research coordinator who determined the randomization allocation were not blinded. However, the patient, the statisticians (O.A.O. and C.S.B.), and the remainder of the research team including surgeons assessing post-operative outcomes were all blinded to the allocation group.

All patients were scheduled to return to the clinic after their procedure for follow up at one-month and one-year post-operative. A physical examination with special focus on diagnosing complications such as SSI, seromas, hematomas, wound dehiscence, hernia recurrence, and re-operations was performed by surgeons blinded to the treatment allocation group. Additionally, patients completed the mAAS survey during each visit to determine changes in AW-QOL.

Primary outcome was defined as major complications at one-year post-operative, a composite of mesh infection, hernia recurrence, and re-operation. Hernia recurrence was determined by clinical assessment and on-demand computed tomography (CT) scan. Re-operation was defined as any procedure involving the fascia, mesh, or intra-peritoneal cavity. Secondary outcomes included SSI, surgical site occurrence (SSO), re-admissions, Clavien-Dindo complication grade, days in hospital at 90-days post-operative (including post-operative length of stay and re-admission length of stay), AW-QOL measured through the mAAS, and pain scores measured through the visual analog scale (VAS). Surgical site infection was defined, per the U.S. Centers for Disease Control and Prevention (CDC) guidelines [18]. Surgical site occurrence included seromas, hematomas, and wound dehiscence. Seromas were defined as localized fluid collection around the wound without signs of infection, hematomas were defined as an accumulation of clotted blood around the incision, and wound dehiscence was defined as an opening of the surgical incision of any size. Re-operation was defined as an unplanned return to the operating room for complications associated with the initial surgery. Surgical site infection and SSO were assessed at one-month post-operative clinic visit, whereas electronic medical records were queried to determine re-admissions, re-operations, and other complications up to 90-days after the initial operation.

Primary outcome was analyzed in an intention-to-treat model using both frequentist and Bayesian statistics. Frequentist analysis was performed using a Fisher exact test. Because there are no studies assessing differences in major complications between biologic and synthetic mesh, a neutral, non-informative prior (mean = 0, standard deviation = 1) was used for the Bayesian logistic regression analysis. Secondary outcomes were analyzed with frequentist statistical methods including Fisher exact tests for categorical data, two-tailed Student t-test for parametric continuous data, and Mann-Whitney U test for non-parametric continuous data. Analysis of covariance (ANCOVA) was performed to analyze AW-QOL and pain scores [19]. Subgroup analyses of the primary outcome, planned a priori, were performed for the following subgroup variables one at a time: hernia type, hernia area, and wound class.

Multivariable logistic regression analysis for the primary outcome using the same subgroup variables (hernia type, hernia area, and wound class) was performed to allow risk-adjusted interpretation of the results. All statistical analyses were performed using the computing environment R version 3.6.3 [20].

We aimed to perform the largest feasible pilot trial to obtain the least biased estimate of treatment effect. We initially planned to assess results on 50 patients at one-year post-operative. However, because of an increase in the volume of cases in our center, we obtained IRB approval to increase our sample size to the largest sample feasible, an estimated 50–100 patients.

Results

A total of 190 eligible patients were approached. Of these, 87 patients were randomly assigned: 44 to biologic mesh and 43 to synthetic mesh (Fig. 1). The majority of patients were Hispanic or African American females. Most patients were obese, had at least one comorbid condition, had an incisional hernia, and had a medium or large size hernia (>4 cm width). Patient baseline demographics and hernia characteristics were similar between the two groups (Table 1). Most cases were wound class 2–4, all cases were classified as Ventral Hernia Working Group (VHWG) class 2–4, primary fascial closure was achieved in the majority of patients, and all repairs were performed with retromuscular mesh placement (Table 2).

OPEN IN VIEWER

FIG. 1.

Consolidated Standards of Reporting Trials (CONSORT) flow diagram.

Patients were followed up for a median of 12.4 months (interquartile range = 11.4–12.8 months; biologic mesh = 12.4 months versus synthetic mesh = 12.5 months). At one-year post-operative, 70 (81.4%) patients were evaluated, 33 in the biologic mesh group and 37 in the synthetic mesh group. Seventeen patients were unable to be contacted by telephone or e-mail and were therefore excluded from our analysis. Baseline characteristics of patients lost to follow-up were similar compared with those included in the analyses except for the percentage of smokers (Supplementary Table S1). A total of three patients, two in the synthetic mesh group and one in the biologic mesh group, had a hernia so wide that primary fascial closure was not feasible. In these cases, either the posterior or anterior fascia, or both, were bridged with hernia sac. In all three cases, the mesh was still placed in the retromuscular position.

The percentage of patients in the biologic mesh group who had a major complication at one-year post-operative was nearly double that of patients in the synthetic mesh group, (42.4% vs. 21.6%; relative risk [RR] = 1.96 [95% confidence interval {CI} = 0.94–4.08]; number needed to harm = 4.8; p = 0.071) (Table 3). On Bayesian logistic regression analysis, there was a 95% posterior predicted probability of open VHR with biologic mesh resulting in more major complications compared with synthetic mesh. Details regarding patients who had major complications are shown in Supplementary Table S2. There were five patients in the biologic mesh group who developed mesh infections in the early post-operative period. Three of these required re-operation for abdominal wound washout and excisional debridement, and the other two were able to be managed non-operatively with wound care, percutaneous drainage of associated fluid collection, and/or antibiotic therapy. Two patients in the synthetic mesh group experienced mesh infections, one required re-operation and subsequent wound care, and one was amenable to be managed non-operatively with percutaneous drainage and antibiotic therapy.

Subgroup analysis by hernia type, hernia width, and wound class revealed an increase in the risk of major complications with the use of biologic mesh as opposed to synthetic mesh (Supplementary Table S3). Biologic mesh was associated with nearly double the percentage of hernia recurrence at one-year post-operative (30.3% vs. 13.5%), SSI (15.9% vs. 9.3%), wound dehiscence (25.0% vs. 14.0%), and re-admissions (22.7% vs. 9.3%). There was no clear evidence of differences in the percentages of seromas, hematomas, or other Clavien-Dindo complications. However, results of these subgroup analyses should be interpreted as hypothesis generating because of the lack of statistical power.

At one-year follow up, mean AW-QOL scores of patients who underwent complex open VHR with biologic mesh improved by 20.3 points, whereas on average those in the synthetic mesh group experienced an increase of 27.9 points (Table 4). Adjusting for baseline AW-QOL, the estimated mean difference in follow-up AW-QOL between groups was 7.5 points. A greater percentage of patients in the biologic arm experienced a worsening in AW-QOL (18.2% vs. 13.9%; RR = 1.36 [95% CI = 0.40–4.65]; p = 0.726) whereas a larger percentage of patients in the synthetic group had an improvement in AW-QOL (67% vs. 75%; RR = 0.89 [95% CI = 0.65–1.21]; p = 0.609) (Supplementary Table S4). There was no clear evidence of a difference seen in pain scores between groups at one-year post-operative.

Discussion

This is the first RCT comparing clinical and patient-centered outcomes of biologic mesh and synthetic mesh for complex open VHR. Contrary to the common belief of improved outcomes, biologic mesh is associated with more major complication rates (42.4% vs. 21.6%; RR = 1.96 [95% CI = 0.94–4.08]) at one-year post-operative. Moreover, with increased rates of SSI in the early post-operative period, patients undergoing repairs with biologic mesh are at increased risk of mesh infection, re-admission, and re-operation. Although most complications were minor (Dindo-Clavien 1 or 2), nearly half of patients in our study experienced a complication demonstrating that complex VHR is associated with substantial morbidity.

It is estimated that nearly half a million VHR are performed each year in the United States, accounting for over $3.2 billion in healthcare costs [21]. Increasing costs per hernia repair have been observed as biologic mesh popularity has increased. A previous national database study found dramatic differences in estimated mean operating room services and supply costs with biologic mesh being three times more expensive than synthetic mesh for uncomplicated clean VHR ($21,106 vs. $7,067; cost ratio = 3.0, mean difference = $14,172 [95% CI = $11,451-$16,893], p < 0.001) [22]. Most of this cost difference is because of the cost of the mesh: biologic mesh on average, costs substantially more than synthetic mesh. At our institution, a 30 × 30 cm piece of biologic mesh costs the hospital $23,147 as opposed to $170 for a piece of synthetic mesh of the same size. For this reason, biologic meshes are infrequently used in uncomplicated cases. Our study further builds on these findings suggesting that the role of biologic mesh even in complicated elective cases does not seem to be justified as cost dominance (increased complications and increased costs) is likely demonstrated [23].

Although the results of our study do not support the use of biologic mesh in open elective complex ventral hernia repair, several questions remain unanswered by our pilot RCT. Primary fascial closure and placement of mesh in optimal locations are believed to be important tenets of ventral hernia surgery [24,25]. For our study, we opted to perform a retromuscular repair because placement of mesh in this anatomic layer is associated with the lowest hernia recurrence and SSI rates [26,27]. This approach can be technically challenging, may not be the preferred approach for all surgeons, and may not be indicated in some clinical settings (damaged posterior sheath, emergency setting, inability to achieve primary fascial closure, i.e., bridged repair, etc.). In these settings, a surgeon may prefer to place mesh intraperitoneal. The results of this study cannot inform the use of synthetic or biologic mesh in these situations. Additionally, there is still ongoing debate on the ideal management strategy of ventral hernias among patients undergoing minimally invasive operations in the setting of contamination.

Many medical devices introduced to the market, including meshes for VHR, are cleared through the Food and Drug Administration (FDA) 510(k) process. As opposed to the drug industry in which RCTs are required to prove safety and efficacy prior to approval, the 510(k) process allows devices to be cleared with demonstration of substantial equivalence to already marketed products. Biologic meshes, including porcine acellular dermal matrix meshes that were utilized in our study, became commercially available by showing substantial equivalence to synthetic mesh [9]. Several of these FDA-cleared devices have not been evaluated rigorously. Surgeons, healthcare workers, and patients should be aware that “FDA clearance” does not ensure that “evidence-based” safety or efficacy has been demonstrated. We advocate that new devices, including surgical meshes, should undergo RCTs prior to wide adoption.

Our present study has considerable strengths. The main strength is a study design with low risk for bias (RCT). Randomization of patients just prior to mesh placement and blinding of patients and post-operative outcome assessors to the allocation group were performed. Wide inclusion criteria were established to increase the generalizability of our findings to a broad population. In addition, a high follow-up rate was achieved [28]. Although this is typical for reports unveiling short-term outcomes, we are currently conducting three-year and five-year follow-up to assess long-term outcomes such as mesh infection, recurrence, re-operations, AW-QOL, and pain scores, with plans to publish these results.

Some limitations of our study need to be acknowledged. As a small single-center pilot trial, our study may be underpowered. Assuming the treatment effect estimates of our study are precise, an appropriately powered (α of 0.05 and a β of 80%) RCT would require at least 156 patients. As more hernias are being repaired through minimally invasive techniques, an appropriately powered study on open ventral hernia repairs would likely require collaboration between multiple high-volume centers. However, Bayesian analysis estimated a high probability of biologic mesh having worse clinical outcomes when compared to synthetic mesh. We acknowledge the importance of high follow-up rates of RCTs. All patients (87 patients, 100%) were seen in clinic for their one-month post-operative follow-up appointment. However, despite aggressive measures to avoid dropouts, 17 patients (11 in the biologic mesh group and 6 in the synthetic mesh group) were unable to be reached at one-year post-operative and were therefore excluded from the analysis of the primary outcome. Although our one-year follow up of 81.4% could be higher, the ideal follow-up of RCTs has been conventionally described as equal to or greater than 80% of the original cohort of patients [28,29]. Generalizability of our results may be limited to other centers with different patient populations. This study was performed by experts in abdominal wall reconstruction of complex patients with high risk for complications, therefore, the generalizability of our findings to surgeons with less case-volume and case-complexity may be reduced. Yet, given the relation between outcomes and volume of cases, it is reasonable to imagine surgeons with higher volume might have better outcomes. We opted to investigate the most commonly utilized biologic mesh type as this would have the greatest applicability and impact a larger number of patients. Although these outcomes may not be generalizable to other biologic or biosynthetic meshes (e.g., human and bovine acellular dermal matrix, porcine collagen intestinal submucosa derivatives, bovine pericardium derivatives, and poly-4-hydroxybutyrate), other products should demonstrate superiority to mid-density polypropylene mesh through RCTs.

Conclusions

This is the first RCT comparing clinical and patient-centered outcomes of biologic versus synthetic meshes. Regardless of the mesh type utilized, complex open ventral hernia repair is associated with a high rate of complications. Although long-term follow-up needs to be completed, the use of biologic mesh resulted in more major complications compared to mid-density polypropylene synthetic mesh.