Secondary Logo

Journal Logo

Original Study

Long-Term Outcomes in Complex Abdominal Wall Reconstruction Repaired With Absorbable Biologic Polymer Scaffold (Poly-4-Hydroxybutyrate)

Buell, Joseph F. MD, MBA*; Flaris, Alexandros N. MD, MSc; Raju, Sukreet MD; Hauch, Adam MD, MBA; Darden, Michael PhD§; Parker, Geoff G. PhD

Author Information
doi: 10.1097/AS9.0000000000000032
  • Open

The surgical repair of large and complex abdominal wall hernias remains a challenging and often controversial area in general surgery. The literature is filled with innumerable short- and intermediate-term studies evaluating the efficacy of techniques and mesh materials with 18- to 24-month follow-up data.1–6 Unfortunately, there are only a limited number of long-term studies with adequate follow-up that are not criticized for the lack of standardized definitions or patient risk stratification.7–9 In the last few years, several prospective studies and or meta-analyses have questioned the benefits of primary tissue closure and the use of composite tissue advancement flaps, indicating that repair without mesh leads to high recurrence rates.10–13 This leads to the remaining question of what is the most appropriate mesh material for the performance of complex abdominal wall reconstruction across a myriad of clinical conditions.

In the United States, over 90% of abdominal wall defects are currently repaired with the aid of surgical mesh. The majority of these materials are synthetic, comprised of either polypropylene or polyesters.7,14,15 Synthetic meshes are highly reliable but have been associated with numerous complications including mesh migration, local infections, and even the development of enterocutaneous fistulas.7,16,17 Biologic mesh and biologic scaffolds comprise the minority of the current market and are frequently used in the setting of high-risk or contaminated fields, infected mesh removal, or in the most complex abdominal wall defects requiring bridging. Biologic mesh is derived from a variety of tissue including human dermis, pericardium, and small intestine, as well as bovine and porcine dermis with over a decade of clinical application in abdominal wall reconstruction. Unfortunately, biologic mesh has also suffered from variable success rates, particularly in the setting of bridging. While the concomitant use of mesh with composite separation techniques have proved superior, the requisite tissue flaps have been associated with higher wound complication rates including seroma and potential biofilm formation.18,19 The greatest disadvantage of biologic mesh remains its exorbitant cost arising from a complex production process. Several major biomedical companies have furthered this process by coating cadaveric allografts with bactericidal agents or antimicrobials, thus increasing the processing time and expense. Despite the lack of clear data favoring the use of biologic mesh, surgeons still overwhelmingly favor their use in high-risk fields.

An alternative to costly cadaveric allografts is poly-4-hydroxybutyrate (P4HB), which is an absorbable polymer scaffold derived from transgenic Escherichia coli resulting in a material similar to synthetic polyester, but free of metal catalyst. P4HB is metabolized by the Krebs cycle and beta-oxidation over 365 to 545 days with nontoxic byproducts with a physiologic pH.14 P4HB significantly differs from at least 2 of the alternative polymer scaffolds because of the longer breakdown period and physiologic pH of the byproducts, contrasted with other mesh products that result in acidic breakdown products that may contribute or even promote local inflammation.

The current study is a continuation of our initial experience and further evaluates the long-term performance of P4HB in complex abdominal wall reconstruction. P4HB grafts were directly compared with porcine cadaveric mesh in the repair of large complex abdominal wall defects. The analysis concentrated on the longevity of a definitive repair including the complication rate, the incidence, and patterns of recurrence, as well as the long-term global expense to the patient while evaluating the specific benefits of an absorbable polymer scaffold.

METHODS

This is a longitudinal long-term follow-up of a previously reported retrospective cohort study using 2 sequential continuous cohorts of abdominal wall reconstructions by a single surgeon, performed between September 2010 and August 2015. The senior surgeon had 10 years of prior experience with complex hernia repair before these cases included in this series. Clinical follow-up was achieved in all patients in 2019. This study was approved by the Institutional Review Board. Surgical repair was achieved by (A) primary fascial closure when achievable with biologic mesh overlay or (B) primary closure achieved by component separation when necessitated followed by biologic mesh overlay. No patient requiring a bridged closure was included in the study. No synthetic mesh was used in the practice for ventral hernia repair. The first group used porcine cadaveric biologic mesh scaffold, Strattice (Lifecell, Bridgewater, NJ), whereas the second group underwent abdominal wall reconstruction using the absorbable polymer scaffold P4HB, Phasix (Bard, Warwick, RI). Four surgical drains were placed in all cases. Drains were only removed when output was less than 25 mL/d/drain. Patients were educated prior to discharge on how to measure and record daily drain out, they were also instructed to communicate with our primary clinic nurse when the 25 mL/d/drain criteria for drain removal was met. Patients were seen in clinic within a week of discharge and then weekly based on our clinic schedule or could be seen on an ad hoc basis for drain removal when the 25 mL/d/drain criteria were met. Postdischarge patients were seen once a week in the outpatient surgical clinic, where drains were removed. Clinical outcomes, complications, and cost data were compared across these 2 sequential groups.

Patient demographic data and risk factors that were collected include age, sex, ethnicity, obesity, diabetes, smoking history, drinking history, immunosuppressed state, and prior hernia operation. A surrogate for graft failure risk termed sum risk factor was calculated by the addition of relevant risk factors including obesity, diabetes, current smoker, current drinker, and immunosuppression. Immunosuppressed patients included solid organ transplant recipients. Patients were considered immunocompromised when they were maintained on calcineurin inhibitors and antimetabolites. No patient was operated on while receiving a target of rapamycin inhibitor. Technique data were collected including hernia repair method, underlay versus overlay, type of mesh used, component separation, and thoracic release. Clinical outcome data collected included infections, complications, reoperations, length of stay, time to drain removal, readmission rate, reherniation rate, and pattern of failure (either along the Chevron incision or the umbilicus). Recurrence was defined by the presence of a defect on physical exam. Any concern or symptoms communicated to the surgeon or our staff would result in a clinic visit or an outpatient computed tomography (CT) in conjunction with a physician visit and physical exam. All asymptomatic patients were followed-up on a yearly basis in the surgery clinic, where they were questioned about any new or recurrent symptoms, the presence of any new or recurrent hernia symptoms, as well as undergoing a physical exam by the senior author. If the patient expressed any symptoms or concerns for recurrent hernia, a CT scan was performed. Follow-up was discontinued after a patient had a recurrent hernia defined by physical exam or CT scan or died.

Univariate and a multistep multivariate analyses of variables that could influence the outcomes were performed. A financial evaluation was then performed to identify the cost-effectiveness of P4HB compared with porcine cadaveric mesh. The analysis was performed using 2019 surrogate cost data to eliminate the impact of inflation and evaluate the expenses incurred in both sequential groups. An initial inpatient admission cost was calculated that included operating room expenses and initial inpatient hospital stay. A second postoperative period cost was calculated and included readmission and all other treatment costs within the first 90 days postoperatively. A final financial comparison was then performed between the complete inpatient and outpatient treatment expense, which included the initial hospital stay and all 90-day postoperative costs. A z test was used to compare proportions. A t test was used to compare means for continuous variables. A generalized linear model was used for our multivariate analyses. The significance level was set at α = 0.05. All data were gathered with Excel 2013 (Microsoft, Albuquerque, NM) and analyses were performed using R v3.5.1 for Windows (R Project).

RESULTS

A total of 73 patients underwent abdominal wall reconstruction between June 2010 and June 2015. Forty-two patients underwent reconstruction with the porcine cadaveric mesh, Strattice (Lifecell) and 31 underwent repair with P4HB, Phasix (Bard). There were no significant differences in age, sex, ethnicity, cirrhosis, immunosuppressed state, obesity, diabetes prevalence, hypertension prevalence, smoking status, drinking status, or prior hernia operation (Table 1).

Table 1. - Patient Demographics and Preoperative Risk Factors Among the Porcine Cadaveric Group (Strattice) Biologic Mesh and the P4HB (Phasix) Biologic Resorbable Mesh
Porcine Cadaveric P4HB P
No. patients 42 31
Age (years) 54 ± 11 57 ± 13 0.318
Male gender (%) 64 74 0.368
African American ethnicity (%) 26 32 0.571
BMI (kg/m2) 31.7 ± 7.0 29.9 ± 6.6 0.264
Obesity (ie, BMI > 30 kg/m2, %) 52 39 0.247
Hypertension (%) 74 77 0.724
Diabetes (%) 40 26 0.192
Drinker (%) 29 37 0.468
Drinker at time of surgery (%) 10 13 0.648
Smoker (%) 50 55 0.683
Smoker at time of surgery (%) 7 3 0.457
Cirrhosis present (%) 7 6 0.908
Liver disease (%) 60 42 0.137
Immunocompromised (%) 64 48 0.174
Prior abdominal surgery (%) 95 97 0.744
BMI indicates body mass index.

Clinical Performance

As was reported in the first study, the P4HB group experienced a significant decrease in time to drain removal, postoperative complications, and reherniation (Table 2), although no significant differences were identified in hospital length of stay and reoperation or readmission rate. Three patients died during the course of this study, but all 3 had already experienced a recurrent hernia prior to their deaths. After patients with recurrent hernia and deaths were excluded, there were no other patients lost to follow-up. This resulted in a 100% follow-up for all remaining patients at 5 years post-repair visit.

Table 2. - Clinical Parameters Including Size and Location of Ventral Hernia Defects and Surgical Outcomes Including Hernia Recurrence and All Clavien-Dindo Classified Complications
Porcine Cadaveric P4HB P
No. patients 42 31
Total recurrence (%) 38.1 12.9 0.017
Postoperative infections (%) 31.0 12.9 0.071
Complications (%) 45.0 19.4 0.021
Redo hernia repair (%) 48.0 52.0 0.736
Chevron hernia (%) 59.5 51.6 0.501
Midline hernia (%) 31.0 38.7 0.637
Renal transplant hernia (%) 9.5 6.5 0.490
Transverse defect (cm) 6.55 ± 1.1 8.01 ± 1.0 <0.001
Midline defect (cm) 4.69 ± 0.9 5.29 ± 1.1 0.013
Component separation (%) 14.0 42.0 0.008
Thoracic release (%) 0.0 3.0 0.241
Length of hospital stay (days) 6.5 ± 5.5 6.7 ± 4.8 0.886
Time to drain removal (days) 14.3 ± 5.6 10.0 ± 5.1 0.001
Abdominal wound (%) 16.7 6.5 0.190
VAC required (%) 7.1 3.2 0.467
Skin graft required (%) 4.8 3.2 0.744
Reoperation rate (%) 14.0 10.0 0.554
Readmission (%) 29.0 13.0 0.110
Mean follow-up (months) 72.7 ± 7.1 98.3 ± 11.1 <0.001
Interquartile range (months) 34–88.7 53.1–67.9
VAC indicates vacuum-assisted closure.

Incidence and site recurrence after ventral hernia repair are presented in Table 3. Repair performance data are presented as a Kaplan-Meier plot for patients with ventral hernias free of recurrence (Fig. 1). Multivariate analysis of our long term clinical data identified African American race (odds ratio [OR] = 14.60; P = 0.004), smoking (OR = 10.07; P = 0.004), drinking (OR = 5.94; P = 0.043), and the use of porcine cadaveric mesh (OR = 1/0.081 = 12.35; P = 0.003) as being significantly correlated to the development of complications (Table 1). Regression analysis for infection identified African American race (OR = 7.91; P = 0.017) and the use of porcine cadaveric mesh (OR = 1/0.109 = 9.17; P = 0.015) (Table 5). Multivariate analysis for recurrent herniation identified smokers at the time of repair (OR = 4.313; P = 0.034) and the use of porcine cadaveric mesh (OR = 1/0.150 = 6.67; P = 0.014) as having the most correlation with recurrence (Table 6). As was seen in the early study, regression analysis did not identify the use of component separation as being significantly correlated with lower recurrent herniation.

Table 3. - Incidence, Location, and Interval to Any Ventral Hernia Recurrence Documented
Recurrence Site Porcine Cadaveric P4HB P
Total recurrence (%) 16 (38.1) 4 (12.9) 0.0168
Midline chevron (%) 4 (25.0) 1 (25.0) 0.9572
Midline umbilicus (%) 3 (18.8) 2 (50.0) 0.1967
Flank (%) 9 (56.3) 1 (25.0) 0.2635
Mean time to recurrence (months) 24.3 (6–42) 20.8 (9–42) 0.7003

Table 4. - Stepwise Multivariate Analysis for Hernia Recurrence Versus Patient Demographics, Comparing Demographics Across Mesh Type, Component Separation (Closure Type), and Demographics Combined With Mesh Type and Closure Type
Recurrence No Yes OR (Univariable) Demographics Demographics and Mesh Type Demographics and Closure Type Demographics, Mesh Type, and Closure Type
OR (Multivariable) OR (Multivariable) OR (Multivariable) OR (Multivariable)
P4HB, N (%) 27 (50.9) 4 (20.0) 0.241 (0.071–0.816, P = 0.022) 0.203 (0.051–0.814, P = 0.024) 0.150 (0.033–0.680, P = 0.014)
Component separation (%) 13 (24.5) 6 (30.0) 1.319 (0.421–4.135, P = 0.635) 1.44 (0.391–5.329, P = 0.582) 2.789 (0.597–13.026, P = 0.192)
Mean age, years (SD) 55.8 (12.9) 53.4 (9.4) 0.982 (0.940–1.027, P = 0.432) 0.974 (0.920–1.032, P = 0.375) 0.976 (0.918–1.038, P = 0.439) 0.972 (0.916–1.030, P = 0.338) 0.970 (0.912–1.033, P = 0.343)
Gender, N (%) 36 (67.9) 14 (70.0) 1.102 (0.361–3.366, P = 0.865) 0.553 (0.137–2.237, P = 0.406) 0.705 (0.157–3.164, P = 0.648) 0.533 (0.130–2.183, P = 0.382) 0.669 (0.912–1.033, P = 0.343)
African American, N (%) 16 (30.2) 5 (25.0) 0.771 (0.239–2.483, P = 0.663) 1.090 (0.261–4.545, P = 0.906) 1.117 (0.248–5.035, P = 0.886) 1.084 (0.258–4.561, P = 0.912) 1.145 (0.251–5.220, P = 0.861)
BMI (kg/m2), mean (SD) 31.2 (7.0) 30.3 (6.7) 0.981 (0.907–1.061, P = 0.625) 0.979 (0.889–1.079, P = 0.675) 0.963 (0.865–1.072, P = 0.494) 0.983 (0.892–1.084, P = 0.737) 0.971 (0.872–1.082, P = 0.600)
Hypertension (%) 40 (75.5) 15 (75.0) 0.975 (0.297–3.205, P = 0.967) 1.089 (0.253–4.692, P = 0.909) 1.080 (0.239–4.888, P = 0.920) 1.100 (0.251–4.810, P = 0.900) 1.156 (0.242–5.511, P = 0.856)
Diabetes (%) 18 (34.0) 7 (35.0) 1.047 (0.355–3.085, P = 0.934) 0.969 (0.272–3.455, P = 0.962) 0.779 (0.200–3.037, P = 0.719) 0.982 (0.275–3.514, P = 0.978) 0.766 (0.193–3.044, P = 0.705)
Alcohol use, N (%) 16 (30.8) 7 (35.0) 1.211 (0.407–3.608, P = 0.730) 1.204 (0.304–4.769, P = 0.791) 1.562 (0.361–6.759, P = 0.551) 1.212 (0.307–4.775, P = 784) 1.633 (0.372–7.169, P = 0.516)
Smoker, N (%) 24 (45.3) 14 (70.0) 2.819 (0.940–8.459, P = 0.064) 3.473 (0.984–12.265, P = 0.053) 4.094 (1.074–15.606, P = 0.039) 3.563 (1.005–12.631, P = 0.049) 4.313 (1.118–16.632, P = 0.034)
Liver disease, N (%) 26 (49.1) 12 (60.0) 1.558 (0.548–4.426, P = 0.405) 1.447 (0.385–5.441, P = 0.584) 1.025 (0.253–4.143, P = 0.973) 1.507 (0.394–5.768, P = 0.549) 1.068 (0.259–4.400, P = 0.927)
Immunocompromised, N (%) 28 (52.8) 14 (70.0) 2.083 (0.695–6.245, P = 0.190) 1.933 (0.537–6.964, P = 0.313) 1.963 (0.522–7.382, P = 0.318) 1.820 (0.489–6.772, P = 0.372) 1.743 (0.441–6.895, P = 0.428)
BMI indicates body mass index.

Table 5. - Stepwise Multivariate Analysis for Infection Versus Patient Demographics, Comparing Demographics Across Mesh Type, Component Separation (Closure Type), and Demographics Combined With Mesh Type and Closure Type
Infection No Yes OR (Univariable) Demographics Demographics and Mesh Type Demographics and Closure Type Demographics, Mesh Type, and Closure Type
OR (Multivariable) OR (Multivariable) OR (Multivariable) OR (Multivariable)
P4HB, N (%) 27 (48.2) 4 (23.5) 0.33 (0.959–1.139, P = 0.079) 0.160 (0.033–0.779, P = 0.023) 0.109 (0.018–0.655, P = 0.015)
Component separation (%) 14 (0.25) 5 (0.29) 1.25 (0.374–4.175, P = 0.717) 1.346 (0.320–5.658, P = 0.685) 3.052 (0.531–17.547, P = 0.211)
Mean age, years (SD) 55.7 (12.2) 53.3 (11.7) 0.983 (0.938–1.030, P = 0.465) 1.017 (0.957–1.081, P = 0.589) 1.026 (0.961–1.095, P = 0.446) 1.015 (0.954–1.080, P = 0.634) 1.021 (0.954–1.092, P = 0.550)
Gender, N (%) 38 (67.9) 12 (70.6) 1.137 (0.348–3.716, P = 0.832) 1.375 (0.283–6.676, P = 0.693) 1.992 (0.343–10.758, P = 0.457) 1.310 (0.267–6.426, P = 0.739) 1.740 (0.314–9.655, P = 0.526)
African American, N (%) 14 (25.0) 7 (41.2) 2.1 (0.672–6.564, P = 0.202) 5.774 (1.213–27.474, P = 0.028) 7.895 (1.477–42.193, P = 0.016) 5.718 (1.193–27.407, P = 0.029) 7.909 (1.453–43.051, P = 0.017)
BMI (kg/m2), mean (SD) 30.4 (6.1) 32.6 (9.0) 1.046 (0.969–1.129, P = 0.248) 1.103 (0.989–1.230, P = 0.079) 1.098 (0.975–1.235, P = 0.122) 1.106 (0.991–1.235, P = 0.072) 1.112 (0.986–1.255, P = 0.084)
Hypertension (%) 43 (76.8) 12 (70.6) 0.726 (0.216–2.442, P = 0.604) 0.456 (0.094–2.211, P = 0.330) 0.467 (0.095–2.290, P = 0.348) 0.450 (0.092–2.208, P = 0.325) 0.470 (0.092–2.387, P = 0.362)
Diabetes (%) 20 (35.7) 5 (29.4) 0.75 (0.231–2.435, P = 0.632) 0.585 (0.137–2.494, P = 0.469) 0.379 (0.076–1.893, P = 0.237) 0.592 (0.139–2.520, P = 0.478) 0.349 (0.068–1.796, P = 0.208)
Alcohol use, N (%) 16 (29.1) 7 (41.2) 1.706 (0.552–5.269, P = 0.353) 2.713 (0.589–12.491, P = 0.200) 4.056 (0.773–21.266, P = 0.098) 2.652 (0.570–12.347, P = 0.214) 3.862 (0.708–21.058, P = 0.118)
Smoker, N (%) 28 (50.0) 10 (58.8) 1.429 (0.476–4.286, P = 0.525) 2.622 (0.674–10.200, P = 0.164) 2.717 (0.659–11.201, P = 0.167) 2.727 (0.690–10.776, P = 0.153) 2.997 (0.704–12.771, P = 0.138)
Liver disease, N (%) 31 (55.4) 7 (41.2) 0.564 (0.188–1.696, P = 0.308) 0.350 (0.074–1.663, P = 0.187) 0.218 (0.040–1.187, P = 0.078) 0.364 (0.076–1.749, P = 0.207) 0.224 (0.040–1.236, P = 0.086)
Immunocompromised, N (%) 32 (57.1) 10 (58.8) 1.071 (0.356–3.223, P = 0.902) 2.368 (0.579–9.695, P = 0.230) 2.466 (0.559–10.883, P = 0.233) 2.304 (0.552–9.613, P = 0.252) 2.362 (0.500–11.154, P = 0.278)
BMI indicates body mass index.

Table 6. - Stepwise Multivariate Analysis for All Classes of Clavien-Dindo Complications Versus Patient Demographics, Comparing Demographics Across Mesh Type, Component Separation (Closure Type), and Demographics Combined With Mesh Type and Closure Type
Complication No Yes OR (Univariable) Demographics Demographics and Mesh Type Demographics and Closure Type Demographics, Mesh Type, and Closure Type
OR (Multivariable) OR (Multivariable) OR (Multivariable) OR (Multivariable)
P4HB, N (%) 25 (52.1) 6 (24.0) 0.291 (0.099–0.854, P = 0.025) 0.128 (0.028–0.582, P = 0.008) 0.081 (0.015–0.437, P = 0.003)
Component separation (%) 12 (0.25) 7 (0.28) 1.167 (0.392–3.471, P = 0.782) 1.655 (0.415–6.600, P = 0.475) 4.314 (0.762–24.410, P = 0.098)
Mean age, years (SD) 55.9 (12.7) 53.8 (10.8) 0.985 (0.945–1.026), P = 0.479) 1.013 (0.956–1.074, P = 0.659) 1.020 (0.957–1.087, P = 0.546) 1.011 (0.122–2.387, P = 0.720) 1.009 (0.944–1.078, P = 0.794)
Gender, n (%) 33 (68.8) 17 (68.0) 0.966 (0.342–2.729, P = 0.948) 0.583 (0.134–2.532, P = 0.471) 0.839 (0.166–4.250, P = 0.832) 0.541 (0.122–2.387, P = 0.417) 0.789 (0.154–4.036, P = 0.776)
African American, N (%) 11 (22.9) 10 (40.0) 2.242 (0.788–6.380, P = 0.130) 8.627 (1.860–40.024, P = 0.006) 13.157 (2.352–73.609, P = 0.003) 8.745 (1.844–41.478, P = 0.006) 14.595 (2.386–89.278, P = 0.004)
BMI (kg/m2), N (%) 30.3 (6.0) 32.2 (8.2) 1.04 (0.969–1.116, P = 0.275) 1.096 (0.992–1.211, P = 0.073) 1.094 (0.979–1.222, P = 0.111) 1.103 (0.996–1.221, P = 0.059) 1.118 (0.994–1.259, P = 0.064)
Hypertension (%) 37 (77.1) 18 (72.0) 0.764 (0.254–2.302, P = 0.633) 0.537 (0.123–2.344, P = 0.409) 0.527 (0.118–2.357, P = 0.402) 0.527 (0.118–2.359, P = 0.402) 0.501 (0.100–2.501, P = 0.400)
Diabetes (%) 17 (35.4) 8 (32.0) 0.858 (0.307–2.398, P = 0.770) 0.658 (0.177–2.441, P = 0.531) 0.400 (0.090–1.779, P = 0.229) 0.674 (0.181–2.505, P = 0.556) 0.386 (0.084–1.769, P = 0.220)
Alcohol use, N (%) 13 (27.1) 10 (40.0) 1.744 (0.626–4.855, P = 0.287) 3.285 (0.784–13.771, P = 0.104) 5.474 (1.072–27.952, P = 0.041) 3.277 (0.769–13.968, P = 0.109) 5.944 (1.059–33.352, P = 0.043)
Smoker, N (%) 21 (43.8) 17 (68.0) 2.732 (0.99–7.543, P = 0.052) 6.709 (1.727–26.063, P = 0.006) 8.279 (1.915–35.786, P = 0.005) 7.207 (1.796–28.917, P = 0.005) 10.073 (2.101–48.300, P = 0.004)
Liver disease, N (%) 26 (54.2) 12 (48.0) 0.781 (0.297–2.057, P = 0.617) 0.725 (0.177–2.978, P = 0.656) 0.400 (0.077–2.093, P = 0.278) 0.776 (0.188–3.200, P = 0.725) 0.409 (0.078–2.137, P = 0.290)
Immunocompromised, N (%) 28 (58.3) 14 (56.0) 0.909 (0.343–2.413, P = 0.848) 1.662 (0.466–5.930, P = 0.433) 1.816 (0.452–7.299, P = 0.400) 1.536 (0.420–5.621, P = 0.516) 1.621 (0.378–6.958, P = 0.516)
BMI indicates body mass index.

F1
FIGURE 1.:
Kaplan-Meier plot for recurrence-free ventral hernia repairs; patients with recurrence or death were excluded. The number of evaluable patients is presented at each marked interval on the bottom of the plot, and both groups are graphed with the 95% confidence intervals.

Wound Breakdown Performance

The porcine cadaveric mesh group experienced a higher incidence (16.7%) of wound complications than the P4HB (6.5%) group, but the difference was not statistically significant (P = 0.190). A higher but not statistically significant percentage of wounds required negative pressure therapy (7.1% vs 3.2%; P = 0.467) or skin grafts (4.8% vs 3.2%; P = 0.744) to manage wound breakdown in abdominal composite reconstructions performed with porcine cadaveric mesh compared with P4HB (Table 2).

Patterns of Recurrence

There were no significant trends identified in the location of hernia recurrence when the P4HB group was compared with the cadaveric porcine allografts (Table 3). The median time to recurrence was not different between the 2 groups. No higher incidence of recurrence was observed in the P4HB group after the 2-year involution period anticipated in the bioresorbable mesh (Fig. 1). No differential was identified when the location of recurrence was considered. This included the occurrence of central and lateral locations, as well as centrally located chevron, midline, and Mercedes incisions were considered.

Financial Analysis

A surrogate financial analysis was performed to eliminate the effect of inflation during a consecutive contiguous group utilizing 2019 United States regionalized average cost data. This analysis identified that the P4HB group incurred a significantly lower cost than the porcine cadaveric group. The initial hospitalization cost for the P4HB group was $3945 less than the porcine cadaveric mesh group. To further assess the cost reduction with the use of P4HB in the procedure, all readmissions ($4039 vs $1725), reoperations ($308 vs $197), outpatient visits ($1009 vs $730), and postoperative wound care ($470 vs $174), including skin graft procedures (147 vs $94) were included in a long-term follow-up calculation ($67,752 vs $53,157). The sum of these cost savings in the use of P4HB was $10,595 per case and was statistically significant when compared with the porcine cadaveric mesh group (P = 0.005).

DISCUSSION

Complex ventral hernia defects requiring abdominal wall reconstruction as a whole are often poorly defined in general and subsequently remain a significant challenge in general surgery, not only in the United States, but worldwide. After the introduction of biologic mesh, there was a surge in the use of multiple cadaveric allografts intended to mitigate postsurgical infections. This experience was met with mixed results and a lack of clear clinical superiority. In the last few years, there appears to be a shift back towards a more liberal use of synthetic materials even in contaminated fields.20,21 This clinical shift appears to have been driven by the variability in cadaveric biologic performance in combination with the significant and often prohibitive cost of these allografts in the absence of any clear prospective randomized data to support the routine use of biologic scaffold material in high-risk patients. While there has been significant enthusiasm for the use of synthetic mesh in clean and clean-contaminated defects, the clinical outcomes have been varied. The incidence of surgical site occurrences has been as high as 31%, while the surgical site infection reached 14.2% in contaminated wounds. Surgical site infections were not the only untoward events that occurred in that series that also reported a 12% reoperative rate and a 4% mesh explanation rate. The most concerning data evolving from the synthetic repair arena are the 7% recurrence rate after only a median follow-up period of 10 months. Conversely, several propensity-matched studies have observed that biologic mesh had equivalent or even higher odds for seroma formation, mesh explanation, readmission, and recurrence than synthetic mesh.22,23 Cadaveric biologic allografts have been frequently criticized for rarely promoting collagen and neovascularity ingrowth across large areas of exposed mesh, especially in the setting of bridging.24 Concurrently, cadaveric allografts continue to be susceptible to variation across production batches, potentially altering their performance or even stimulating a regional immunologic response.25 This variability in clinical performance has been attributed to the development of local inflammation resulting in seroma, nonhealing, and local infection of the allograft with potential development of biofilm and destruction of the allograft integrity.

One approach to address these concerns has been to bio protect or coat biologic allograft materials with antimicrobial or bacteriostatic molecules to inhibit bacterial ingrowth during the incorporation period or for at least the first week or 2 of implantation.26,27 This technology, although attractive, may not address late bacterial infiltration by fastidious organisms nor address an established biofilm. Complexity in the manufacturing of these products requires significant lead time in a multistep workflow to produce the allograft and then coat them with antimicrobials or antibiotics, resulting in long production lead-times and, ultimately, significantly higher allograft production costs. This makes the use of resorbable biologic polymer scaffolds even more attractive, particularly ones that minimize postoperative seroma and completely involute within 2 years. It is this same degradation process that has led several groups to question and maybe even criticize the clinical integrity of P4HB, particularly after the 2-year involution period.28,29 This current study now addresses this concern, confirming the long-term integrity and performance of P4HB as a bioabsorbable hernia scaffold with the longest degradation profile available in this new class of mesh. Data also suggests that through this integral polymer’s ability to break down into water and carbon dioxide with neutral pH byproducts reduces complications and infections, potentially through a reduction in tissue inflammation. The current reported clinical performance also supports the conjecture that a prolonged involution period leads to a slow and continuous transfer of weight-bearing tension from the graft to the native abdominal wall allowing for collagen ingrowth while ensuring elimination of the foreign body over time. This opinion is strengthened by our current data, which shows that P4HB has superior clinical outcomes over porcine cadaveric biologic mesh. The absence of metalloproteins and a nonacidic environment may also contribute to P4HBs ability to reduce inflammation while minimizing seroma formation and subsequently decreasing postoperative complications such as tissue or mesh infections decreasing recurrent hernia formation. With the current series, we have confirmed P4HBs lowers both short- and long-term hernia recurrence rates. The principle concern over inadequate tissue ingrowth and abdominal wall integrity following complete resorption of a bioresorbable mesh such as P4HB biologic scaffold is not warranted and should be discarded.

Multivariate analysis for the 5-year data identified that smoking and the use of porcine cadaveric mesh were significant contributors to recurrence. African American race, smoking, the use of alcohol, and porcine cadaveric mesh were significant contributors to the development of complications. African American race and the use of porcine cadaveric mesh as having a significant impact on infection. The effect of African American race has not been previously described. This may correspond to an immunologic response to foreign antigens provoking inflammation and preventing mesh integration. Despite significant literature to support the use of component separation, a multistep regression analysis did not identify their use as reducing recurrent herniation.10,11 Our principal approach to complex abdominal wall defects has been to achieve tension-free primary tissue closure. Over the evolution of our program, we have gained experience with numerous myocutaneous advancement approaches. The effect of component separation on primary closure is present in our analysis, but there is no clear effect identified until the addition of mesh selection. This may suggest that P4HB potentially has a salutary effect on wound healing, decreasing reherniation through the inhibition of seroma formation and infection. Conversely, absorbable biosynthetic scaffolds appear to provide an even greater advantage over porcine cadaveric mesh through their innate gradual breakdown through natural pathways minimizing local reaction and inflammation. This does not negate the belief that innumerable surgeons advocate the use of component separation in achieving a tension-free closure regardless of the mesh used.

The limitation of our current study is the lack of a prospective randomization trial. This would diminish or potentially eliminate many confounding effects such as clinical experience and learning on the study’s outcomes. Hernia recurrence was defined by the presence of clinically detected defects on physical exam, whereas other studies have advocated patient-centric surveys concurrently with diagnostic imaging including ultrasound and body CT. Additionally, the retrospective nature of the study did not provide an opportunity to perform patient satisfaction questionnaires or quality of life surveys. As in our initial report, concern about over the confounding effects of component separation has been raised. We shared these concerns, as our balance table suggests significantly different levels of component separation were utilized during the study period. As a result, we have explored sensitivity to controlling for component separation. For all outcomes, our main parameter of interest is very similar with and without controls for component separation, which suggests that, conditional on all other variables in our model, component separation does not significantly affect our outcomes.

However, the analysis of our long-term data on complex abdominal wall reconstruction performed with a resorbable biologic scaffold, P4HB confirms the resilience of these complex abdominal wall reconstructions and a value-based procedure, making it a superior graft for contaminated as well as tension bearing abdominal wall defects. The analysis for the use of P4HB for a bridging needs further analysis especially in the long-term as will the long-term performance of antimicrobial-coated cadaveric allografts. At this point, we can conclude that P4HB is a cost-effective scaffold that minimizes the downside of a costly synthetic mesh infections while demonstrating a low recurrence rate when compared with other large ventral hernia series.

REFERENCES

1. Winder JS, Majumder A, Fayezizadeh M, et al. Outcomes of utilizing absorbable mesh as an adjunct to posterior sheath closure during complex posterior component separation. Hernia. 2018; 22:303–309
2. Roth JS, Anthone GJ, Selzer DJ, et al. Prospective evaluation of poly-4-hydroxybutyrate mesh in CDC class I/high-risk ventral and incisional hernia repair: 18-month follow-up. Surg Endosc. 2018; 32:1929–1936
3. Anderson B, Hart AM, Maxwell D, et al. The biosynthetic option as an alternative in complex abdominal wall reconstruction. Ann Plast Surg. 2020; 85:158–162
4. Sasse KC, Lambin JH, Gevorkian J, et al. Long-term clinical, radiological, and histological follow-up after complex ventral incisional hernia repair using urinary bladder matrix graft reinforcement: a retrospective cohort study. Hernia. 2018; 22:899–907
5. Wegdam JA, Thoolen JMM, Nienhuijs SW, et al. Systematic review of transversus abdominis release in complex abdominal wall reconstruction. Hernia. 2019; 23:5–15
6. Plymale MA, Davenport DL, Dugan A, et al. Ventral hernia repair with poly-4-hydroxybutyrate mesh. Surg Endosc. 2018; 32:1689–1694
7. Kokotovic D, Bisgaard T, Helgstrand F. Long-term recurrence and complications associated with elective incisional hernia repair. JAMA. 2016; 316:1575–1582
8. de Vries FEE, Hodgkinson JD, Claessen JJM, et al. Long-term outcomes after contaminated complex abdominal wall reconstruction. Hernia. 2020; 24:459–468
9. Limura E, Giordano P. Biological implant for complex abdominal wall reconstruction: a single institution experience and review of literature. World J Surg. 2017; 41:2492–2501
10. Slater NJ, van Goor H, Bleichrodt RP. Large and complex ventral hernia repair using “components separation technique” without mesh results in a high recurrence rate. Am J Surg. 2015; 209:170–179
11. Garvey PB, Giordano SA, Baumann DP, et al. Long-term outcomes after abdominal wall reconstruction with acellular dermal matrix. J Am Coll Surg. 2017; 224:341–350
12. Holihan JL, Nguyen DH, Nguyen MT, et al. Mesh location in open ventral hernia repair: a systematic review and network meta-analysis. World J Surg. 2016; 40:89–99
13. Holihan JL, Askenasy EP, Greenberg JA, et al.; Ventral Hernia Outcome Collaboration Writing Group. Component separation vs. bridged repair for large ventral hernias: a multi-institutional risk-adjusted comparison, systematic review, and meta-analysis. Surg Infect (Larchmt). 2016; 17:17–26
14. Buell JF, Sigmon D, Ducoin C, et al. Initial experience with biologic polymer scaffold (poly-4-hydroxybuturate) in complex abdominal wall reconstruction. Ann Surg. 2017; 266:185–188
15. Majumder A, Winder JS, Wen Y, et al. Comparative analysis of biologic versus synthetic mesh outcomes in contaminated hernia repairs. Surgery. 2016; 160:828–838
16. López-Cano M, Martin-Dominguez LA, Pereira JA, et al. Balancing mesh-related complications and benefits in primary ventral and incisional hernia surgery. A meta-analysis and trial sequential analysis. PLoS One. 2018; 13:e0197813
17. Hodgkinson JD, Maeda Y, Leo CA, et al. Complex abdominal wall reconstruction in the setting of active infection and contamination: a systematic review of hernia and fistula recurrence rates. Colorectal Dis. 2017; 19:319–330
18. Maloney SR, Schlosser KA, Prasad T, et al. The impact of component separation technique versus no component separation technique on complications and quality of life in the repair of large ventral hernias. Surg Endosc. 2020; 34:981–987
19. Kathju S, Nistico L, Melton-Kreft R, et al. Direct demonstration of bacterial biofilms on prosthetic mesh after ventral herniorrhaphy. Surg Infect (Larchmt). 2015; 16:45–53
20. Warren J, Desai SS, Boswell ND, et al. Safety and efficacy of synthetic mesh for ventral hernia repair in a contaminated field. J Am Coll Surg. 2020; 230:405–413
21. Carbonell AM, Criss CN, Cobb WS, et al. Outcomes of synthetic mesh in contaminated ventral hernia repairs. J Am Coll Surg. 2013; 217:991–998
22. Totten CF, Davenport DL, Ward ND, et al. Cost of ventral hernia repair using biologic or synthetic mesh. J Surg Res. 2016; 203:459–465
23. Sandvall BK, Suver DW, Said HK, et al. Comparison of synthetic and biologic mesh in ventral hernia repair using components separation technique. Ann Plast Surg. 2016; 76:674–679
24. Gruber-Blum S, Brand J, Keibl C, et al. Abdominal wall reinforcement: biologic vs. degradable synthetic devices. Hernia. 2017; 21:305–315
25. Badylak SF. Decellularized allogeneic and xenogeneic tissue as a bioscaffold for regenerative medicine: factors that influence the host response. Ann Biomed Eng. 2014; 42:1517–1527
26. Guillaume O, Pérez-Tanoira R, Fortelny R, et al. Infections associated with mesh repairs of abdominal wall hernias: are antimicrobial biomaterials the longed-for solution? Biomaterials. 2018; 167:15–31
27. Majumder A, Scott JR, Novitsky YW. Evaluation of the antimicrobial efficacy of a novel rifampin/minocycline-coated, noncrosslinked porcine acellular dermal matrix compared with uncoated scaffolds for soft tissue repair. Surg Innov. 2016; 23:442–455
28. Miserez M, Jairam AP, Boersema GSA, et al. Resorbable synthetic meshes for abdominal wall defects in preclinical setting: a literature review. J Surg Res. 2019; 237:67–75
29. Petro CC, Rosen MJ. A current review of long-acting resorbable meshes in abdominal wall reconstruction. Plast Reconstr Surg. 2018; 1423 suppl84S–91S
Keywords:

abdominal wall reconstruction; biologic mesh; biosynthetic mesh; polymer mesh

Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc.