Abstract Background Data does not exist to suggest a mesh to hernia defect area ratio (M:HDR) in AWR. Furthermore, the data are very limited in this respect for biologic mesh. Our aim was to evaluate whether biologic M:HDR or defect size influences outcomes in AWR. Method A prospectively maintained hernia database at a tertiary hernia center was queried for all open ventral hernias repaired with a biologic mesh (2003–2023). Patients were stratified by terciles based on defect size, <168 cm2 (HD1), 168–299cm2 (HD2), and ≥300 cm2(HD3). Univariate analysis and multivariable regression were performed. Results Overall, 408 patients were evaluated. There were 142, 132, and 134 patients in the HD3,2,1 groups respectively. There were differences in race (Caucasian:86.6% vs. 81.8% vs. 76.1%, P = 0.047), sex (female:43.7% vs. 50.8% vs. 66.4%, P = 0.0005), age (57.9, 61.3, 56.3 years, P = 0.003), ASA score (ASA 3: 54.9% vs. 54.5% vs. 41.8%, P = 0.004) and recurrent hernias (66.2% vs. 59.1% vs. 51.5%, P = 0.045). The rate of diabetes, COPD, overall comorbidities, smoking, and BMI were similar. There was no difference in wound class. HD3 had more component separations (72.5% vs. 52.3% vs. 39.6%, P <0.0001). Fascial closure was achieved in all patients. HD3 had larger mesh (804.1 vs. 616.9 vs. 416.9 cm2, P <0.0001), longer operative time (246.2 vs. 204.6 vs. 198.3 minutes), and smaller M:HDR (2.0 vs. 2.7 vs. 5.4, P <0.0001). HD3 patients had longer length-of-stay (11.0 vs. 8.3 vs. 7.6 days, P = 0.0002) and ICU admissions (6.3% vs. 0% vs. 2,2%, P = 0.007). Wound complication and respiratory-failure rates were not different. Hernia recurrence was similar between groups (8.5.vs. 8.2 vs. 3.8, p = 0.227) with 24.5-months follow-up. On logistic regression, the wound complication rate was an independent predictor of recurrence (OR = 3.95,CI [1.82,8.6], P = 0.0005). Conclusion HD3 were more complex, had larger hernia defects and mesh size, but smaller M:HDR. Despite this, wound complications and recurrence rate were similar between groups. Fascial closure with biologic mesh reinforcement provides a durable repair across defect sizes despite varying mesh overlap.

Abstract Background The robustness of RCTs in AWR remains underexplored. The fragility index (FI) measures the minimum number of events whose outcome would have to change to reverse the significance of a result. Reverse fragility index (RFI) calculates the minimum number to turn a non-significant result to significant. This study aimed to evaluate the fragility of RCTs comparing biologic versus synthetic mesh. Method A comprehensive search of RCTs comparing biologic and synthetic mesh outcomes was conducted. Of 274 studies reviewed, 4 RCTs met the inclusion criteria. Primary outcomes were recurrence (n = 2) and major complications (n = 2). FI, (RFI), fragility quotient (FQ), and reverse FQ (RFQ) were calculated. Results Four RCTs, enrolling 87–253 patients, were published in journals with a median impact factor of 13.78. There was heterogeneity in hernia characteristics, mesh choice, mesh placement, and fascial closure among the studies. The follow-up time ranged from 12.4–36 months. For recurrence, the FI was 1 and 7; FQ was 0.008 and 0.028. For major complications, there was one positive study with an FI of 6 and FQ of 0.024. The one negative study was deemed significant using Bayesian statistics, but on review with the frequentist approach, the P-value = 0.071 and FI = 0, indicating non-significance. Therefore, the RFI was calculated as 1 and RFQ was 0.014. Conclusion The fragility analysis underscores the vulnerability of RCTs comparing biologic and synthetic mesh in AWR. Despite RCTs favoring synthetic mesh, the low fragility indices and marked differences in study details should caution translating these findings to clinical practice.

Abstract Background This is a 62-year-old male with bilateral inguinal hernias (IH), the left being multiply recurrent after an open, anterior tissue-based approach 30 years ago and recently a robotic Transabdominal Preperitoneal Repair. He recurred within a year and was referred to our hernia center. Methods The patient underwent preoperative counselling, and we particularly discussed the risk of converting to an open approach if we could not safely carry out the dissection robotically. Results Upon entry, a recurrent left IH and primary right IH were noted. We incised the peritoneum across the abdominal wall and dissected medially, down to the pubis. Although this was a reoperative case, we successfully accessed the preperitoneal plane. On the left, we encountered the prior mesh, which appeared rotated medially and densely adhered to the inferior epigastric vessels, so the decision was made to leave it in place. The lateral preperitoneal dissection was completed, and the cord structures were dissected off the hernia sac. We repaired the contralateral defect following the same steps. The critical view of the myopectineal orifice was obtained bilaterally, and two pieces of mesh were introduced. Our mesh spanned two centimeters below the pubic tubercle and sat flush without the edges rolling. The mesh was secured at the pubic tubercle, Cooper’s ligament, medially, and at the lateral edge of the preperitoneal flap, protecting the neurovascular structures in the area. The preperitoneal flap was closed with a running absorbable barbed suture. Conclusions The patient was discharged postoperatively and is doing well on follow-up.

Abstract Background Prescribed physical limitations following AWR primarily focus on weight lifting restrictions. The questions of how much and for how long are poorly understood. Neither a consensus, guidelines, nor definitive data exist to validate surgeon recommendations. We sought to assess the literature on postoperative instructions for patients after AWR. Methods A systematic review was performed of the Wake Forest Library and PubMed in January 2024 for literature on postoperative physical recommendations after AWR. Three separate searches concerning restrictions following hernia repair yielded 573 publications. Results Of the 573 publications reviewed, only eight articles were directly related to postoperative instructions after AWR. Four were surveys; one was an ongoing randomized controlled trial; one compared two postoperative lifting programs; two were systematic reviews. Three additional studies measured intraabdominal pressure and abdominal wall elasticity in healthy patients with intentions to apply this to AWR patients. There was variation in the duration of convalescence and weight limitation. For laparoscopic VHR, convalescence recommendations ranged from 1 day to 6 weeks. For laparotomy, convalescence ranged from 1–12 weeks postoperatively. Actions such as coughing, vomiting, and jumping caused higher intraabdominal pressure than lifting. Only one study used recurrence as the primary outcome after comparing postoperative activity regimens. Conclusion Despite hernia repair being one of the most commonly performed operations globally, surgeons lack validated and uniform postoperative instructions. There remains a need for higher quality of evidence concerning AWR recovery to optimize patient quality of life while balancing safe return to normal activities.

Abstract Background Reviewing abstracts for inclusion in scientific conferences involves substantial time and effort. Recent innovations in Large Language Model (LLM)s, such as OpenAI’s ChatGPT, have the potential to reduce the burden on reviewers. This study aimed to test that potential by programmatically grading abstracts using custom-trained LLMs. Methods Abstracts and reviewer grades of American Hernia Society (AHS – 2021) and American Plastic Surgery Association (ASPS 2021–2022) abstracts were obtained. Two new models of the state-of-the-art ChatGPT-3.5-turbo-1106 were fine-tuned on corpuses of these abstracts and grades. We trained for three epochs with a standard loss function and four inputs: a “system” directive, the grading rubric, the abstract, and the average reviewer grade on a five-point scale. The models were evaluated with the 2023 abstracts and grades from both societies. The trained models graded the abstracts and ranked them using three custom algorithms: QuickSort, which iteratively compared two abstracts at a time; Quartile, which ranked four; and Bridging, which combined batches of fifteen. Mean differences and Spearman correlation coefficients were calculated between actual and predicted grades/rankings. Results The trained models successfully followed the 2023 rubrics and produced in less than 20 minutes predicted grades centered near the actual: the mean difference in groups was 0.0719 ± 0.881 for AHS and 0.104 ± 0.944 for ASPS. However, variance was high and the predicted rankings had at best weak rank correlations with the actual rankings. Conclusion Custom-trained LLMs present a novel method for evaluating abstracts efficiently, but further research is needed to fine-tune the models for more precise results.

Abstract Background For three decades there has been significant debate over LIHR versus OIHR regarding recurrence, pain, and morbidity. This study assessed the robustness of evidence supporting either approach through a fragility analysis. The fragility index (FI) measures the minimum number of participants whose outcome would have to change to turn a significant result non-significant. The reverse fragility index (RFI) calculates the fewest number of events that would convert a non-significant result. Methods A systematic search of PubMed, Wake Forest Library, and Google Scholar was performed for randomized controlled trials (RCTs) comparing OIHR and LIHR with dichotomous outcomes from 1995–2023. FI and RFI were calculated for significant and non-significant outcomes, respectively. FI/RFI ≤ 3 were considered fragile. Results Review of 319 studies yielded 28 RCTs with a median of 141.5 (Q1, Q3: 103.0, 283.0) participants and 47 primary or secondary outcomes, including recurrence (n = 24), postoperative complications (n = 16), and chronic pain (n = 7). The median FI of all outcomes was 1.0 (1.0, 4.0), and RFI was 4.0 (3.0, 6.0), exhibiting substantial fragility in the significant results. Individually, recurrence, complications, and pain demonstrated median FIs of 3.0, 1.0, and 4.0, respectively. Sixteen (57.1%) RCTs had more participants lost to follow-up (LTFU) than the FI/RFI. Conclusions Of the inguinal hernia RCTs that demonstrate significant outcomes, 64.3% draw fragile conclusions. The overall FI of 1.0 indicates that a change in a single outcome could reverse the significance of the trial. Further, 57.1% of RCTs had more participants LTFU than FI/RFI, further weaking these studies’ conclusions. Inguinal hernia RCTs are remarkably fragile.

Abstract Background Parastomal hernia is a common complication that can impact quality of life or lead to intestinal obstruction. Repairs are complex, have high recurrences, and are without a clear best technique. Though repairs may be performed minimally invasively, a complex subset of patients require open parastomal hernia repair (OPSHR). OPSHR outcomes at a high-volume center are described. Methods A prospectively maintained database was queried for patients who underwent OPSHR at a tertiary hernia center. Operative details and outcomes are analyzed. Results Of 97 OPSHRs, mean age was 61.9 ± 12.6 years, 56.7% were female, 24.7% were diabetic, and average BMI was 31.3 ± 6.5 kg/m2. Defect size averaged 125.3 ± 130.0 cm2 and 41.2% were recurrent. Stomas were colostomies (56.7%), ileostomies (30.9%), and urostomies (12.4%). Patients underwent concurrent ventral hernia repair (56.7%), panniculectomy (22.7%), component separation (30.9%), and stoma resiting/reversal (40.2%). When stomas were not resited or reversed, techniques included suture repair (n = 7, 12.1%) or mesh reinforcement (n = 51, 87.9%; biologic – 54.9%, synthetic – 45.1%) with Sugarbaker (66.0%; preperitoneal – 63.6%, intraperitoneal – 33.3%, retrorectus – 3.0%), keyhole (28.0%), or onlay (6.0%) approaches. Wound complications occurred in 18.6% (infection – 10.3%, breakdown – 9.3%, hematoma – 3.1%, cellulitis – 2.1%, seroma requiring intervention – 2.1%). Rates differed by colostomy (10.9%), ileostomy (23.3%), and urostomy (41.7%) (P = 0.031). Mean follow-up was 31.6 ± 35.9 months with a 20.6% recurrence rate. On univariate analysis, patients with recurrence were younger (55.4 ± 12.6 vs. 63.6 ± 12.1 years; P = 0.008) with smaller mesh (470.4 ± 336.6 vs. 680.2 ± 388.1 cm2; P = 0.045). Recurrence did not differ by technique (onlay – 25.0%, primary repair – 21.4%, Sugarbaker – 18.2%, keyhole – 10.0%; P = 0.077). When stomas were relocated (26.8%), prophylactic mesh (84.6%) did not significantly impact recurrence (27.3% vs. 50.0%; P = 0.563). Conclusions This study represents the largest institutional series describing OPSHR outcomes. Larger meshes and older patients had fewer recurrences. Optimal OPSHR technique remains unclear, but these results may inform preoperative discussions and tailor surgical planning.

Abstract Background Concurrent panniculectomy (CP) with hernia repair allows patients to have removal of their redundant skin and excess overlying fat in addition to herniorrhaphy. While providing excellent cosmesis and patient satisfaction, concerns have historically included longer operative times, higher rates of wound complications and infection. This study analyzed the rates of postoperative complications in patients undergoing CP with abdominal wall reconstruction (AWR). Methods A prospectively maintained institutional database was queried for patients who underwent both hernia repair and CP. All patients had the same fellowship trained surgeon, between 2017 and 2023. Primary outcomes included wound complications and hernia recurrence rates. Results A total of 149 patients were identified; 75.8% were female, mean age was 57.3 ± 12.2 years, median BMI was 32.2 kg/m2 (IQR:28.4, 36.5), 30.9% had diabetes, 14.8% were immunosuppressed, and 5.4% were smoking. Prehabilitation resulted in a median preoperative weight loss of 11.7 lbs (IQR:21.7, 1.0). Median hernia size was 216.0 cm2 (IQR:150.0, 300.0), and 97.3% were ventral hernias. Median mesh size was 900.0 cm2 (IQR:471.2, 1050.0), 57.7% were biologic, and 98.7% of all repairs had preperitoneal mesh placement. Intraoperatively, 83.2% of cases were clean or clean-contaminated. Median operative time was 134.5 minutes (IQR:110.0, 176.0). Postoperative wound complications included seroma requiring intervention (11.4%), cellulitis (7.4%), wound infection (6.0%), intraabdominal abscess (6.0%), wound breakdown (3.4%), hematoma requiring intervention (3.4%), and mesh infection (1.3%). Hernia recurrence rate was 2.0% over a median follow-up time of 16.4 months (IQR:5.7, 41.1). Conclusion CP with AWR does not compromise wound morbidity or hernia recurrence. Despite significant patient comorbidities, rates of postoperative wound morbidity were low.

Abstract Background Preoperative relaxation of the abdominal wall with Botulinum Toxin A (BTA) injection can assist with fascial closure and decrease the need for component separation (CST). Its efficacy in subxiphoid (M1) hernias is unclear with high rates of CST reported despite BTA. For M1 hernias requiring CST, we assessed whether use of preoperative BTA offered any additional benefit toward fascial closure or recurrence rates. Methods A tertiary hernia center’s prospectively maintained database was reviewed for M1 hernias, as defined on CT imaging. Patients who underwent subsequent CST were grouped based on whether preoperative BTA injections were performed. Results Of 67 M1 hernia patients, 30 (44.8%) received BTA and 37 (55.2%) did not. BTA versus non-BTA groups had similar age (56.0 ± 14.1 vs. 61.5 ± 11.8 years; P = 0.087), ASA score, diabetes history, and defect size (499.2 ± 185.5 vs. 416.1 ± 238.6 cm2; P = 0.144). BTA group had lower BMI (28.9 ± 5.1 vs. 32.7 ± 7.2 kg/m2; P = 0.018), fewer current and former smokers (0% vs. 10.8%, 40.0% vs. 62.2%; P = 0.006), more contaminated and dirty cases (20.0% vs. 5.4%, 33.3% vs. 13.5%; P = 0.008), and more biologic mesh used (62.1% vs. 35.1%; P = 0.046). Neither panniculectomy rates nor CST techniques differed (external oblique release: 80.0% vs. 62.2%; P = 0.179, posterior rectus sheath release: 70.0% vs. 67.6%; P > 0.999, transversus abdominus release: 26.7% vs. 35.1%; P = 0.598). Rates of unilateral versus bilateral CST were not statistically different. There were no differences in fascial closure (90.0% vs. 94.6%; P = 0.650), recurrence (6.7% vs. 2.7%; P = 0.583), or in-total wound complication rates (33.3% vs. 43.2%; P = 0.458) between BTA and non-BTA groups with follow-up of 12.7 ± 18.8 versus 24.1 ± 28.2 months (P = 0.062). Conclusions M1 hernia repair often requires CST despite BTA injections. BTA as a preoperative adjunct in this population does not benefit fascial closure rates, risk of recurrence, or frequency of bilateral CST.

Abstract Background Artificial intelligence (AI) has permeated academia, especially with the inception of Chat Generative Pretrained Transformer (ChatGPT), a large language model chatbot. The goal of this study was to evaluate ChatGPT’s ability to write and grade surgical research abstracts. Methods ChatGPT models 3.5 and 4.0 were trained to write ten different abstracts by providing it with ten previous abstracts presented at national meetings as models, background literature, and data with statistics. ChatGPT3.5’s and ChatGPT4.0’s versions were compared with junior resident’s and the senior author’s edited, submitted version for each data set. The 40 abstract versions were judged by experienced, blinded surgeon-reviewers, who graded them on a 10- and 20-point scale and ranked them 1st–4th. ChatGPT3.5 and ChatGPT4.0 also graded/ranked the abstracts. Standard statistics were performed. Results are reported in the order of resident, senior author, ChatGPT3.5, and ChatGPT4.0, respectively. Results The surgeon-reviewers could not differentiate between abstract authors; every reviewer ranked an AI-generated version 1st at least once. There was no difference in the surgeon-reviewers’ average 10-point scores (6.8 ± 1.5 vs. 6.9 ± 1.5 vs. 6.6 ± 1.4 vs. 6.9 ± 1.3; P = 0.611), 20-point scores (14.3 ± 3.0 vs. 14.6 ± 2.8 vs. 13.8 ± 2.7 vs. 14.3 ± 2.5; P = 0.500), or rank (2.5 ± 1.2 vs. 2.4 ± 1.1 vs. 2.8 ± 1.1 vs. 2.3 ± 1.1; P= 0.138). ChatGPT3.5’s grades were similar to those of the reviewers, but ChatGPT4.0 graded all four abstract versions higher on the 20-point scale than both the surgeon-reviewers (P = 0.023, P = 0.032, P = 0.002, P = 0.037) and ChatGPT3.5 (P = 0.003, P = 0.004, P = 0.003, P = 0.012). Conclusions When appropriately trained with background knowledge, data, and statistics, ChatGPT can write convincing medical research abstracts that are undifferentiable from resident’s or senior author’s drafts. ChatGPT3.5 graded abstracts similar to the surgeon-reviewers, while ChatGPT4.0 was more generous.