- Email: [email protected]
Abdominal Wall Reconstruction with Concomitant Ostomy-Associated Hernia Repair: Outcomes and Propensity Score Analysis
Accepted Manuscript Abdominal Wall Reconstruction with Concomitant Ostomy-Associated Hernia Repair: Outcomes and Propensity Score Analysis Alexander F. Mericli, MD, Patrick B. Garvey, MD, FACS, Salvatore Giordano, MD, PhD, Jun Liu, PhD, Donald P. Baumann, MD, FACS, Charles E. Butler, MD, FACS PII:
S1072-7515(16)31673-8
DOI:
10.1016/j.jamcollsurg.2016.11.013
Reference:
ACS 8549
To appear in:
Journal of the American College of Surgeons
Received Date: 24 August 2016 Revised Date:
21 November 2016
Accepted Date: 22 November 2016
Please cite this article as: Mericli AF, Garvey PB, Giordano S, Liu J, Baumann DP, Butler CE, Abdominal Wall Reconstruction with Concomitant Ostomy-Associated Hernia Repair: Outcomes and Propensity Score Analysis, Journal of the American College of Surgeons (2017), doi: 10.1016/ j.jamcollsurg.2016.11.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Abdominal Wall Reconstruction with Concomitant Ostomy-Associated Hernia Repair: Outcomes and Propensity Score Analysis
RI PT
Alexander F Mericli, MD, Patrick B Garvey, MD, FACS, Salvatore Giordano, MD, PhD, Jun Liu, PhD, Donald P Baumann, MD, FACS, Charles E Butler, MD, FACS
SC
Department of Plastic Surgery, The University of Texas MD Anderson Cancer Center, Houston,
M AN U
TX
Disclosure Information: Dr Mericli received support from Acelity for travel to the European Association of Plastic Surgeons’ annual meeting to present this paper. All other authors have nothing to disclose.
TE D
Support: This work was supported in part by the NIH/NCI under award number P30CA016672 and used the Clinical Trials Support Resource.
Presented at the European Association of Plastic Surgeons Annual Meeting, Brussels, Belgium,
EP
May 2016
AC C
Correspondence address: Charles E. Butler, MD Professor and Chair, Department of Plastic Surgery The University of Texas MD Anderson Cancer Center 1400 Pressler Street Unit 1488 Houston, TX 77030 713-794-1247 Brief Title: Combined Ventral and Parastomal Hernia Repair
ACCEPTED MANUSCRIPT 2
Abstract: Background: The optimal strategy for abdominal wall reconstruction in the presence of a stomal-site hernia is unclear. We hypothesized that the rate of ventral hernia recurrence in
RI PT
patients undergoing a combined ventral hernia repair and stomal-site herniorraphy would not differ clinically from the ventral hernia recurrence rate in patients undergoing an isolated ventral
compared to reinforced repairs regardless of stomal hernia.
SC
hernia repair. We also hypothesized that bridged ventral hernia repairs result in worse outcomes
Study Design: We retrospectively reviewed prospectively collected data from consecutive
M AN U
abdominal wall reconstructions performed with acellular dermal matrix (ADM) at a single center during 2000-2015. We compared patients who underwent a ventral hernia repair alone (AWR) and those who underwent both a ventral hernia repair and ostomy-associated herniorraphy (AWR+O). We conducted a propensity score matched analysis to compare the outcomes
TE D
between the two groups. Multivariable Cox proportional hazards and logistic regression models were used to study associations between potential predictive/protective reconstructive strategies and surgical outcomes.
EP
Results: We included 499 patients (median follow-up, 27.2 months, interquartile range (IQR): 12.4-46.6 months), 118 AWR+O and 381 AWR. After propensity score matching, 91 pairs were
AC C
obtained. Ventral hernia recurrence was not statistically associated with ostomy-associated herniorraphy (adjusted HR=0.7, 95%CI=0.3-1.5;p=0.34). However, the AWR+O group experienced a significantly higher percentage of SSO (34.1%) than the AWR group (18.7%; adjusted OR=2.3, 95%CI=1.4-3.7;p<.001). In the AWR group, there were significantly fewer ventral hernia recurrences when the repair was reinforced compared to bridged (5.3% vs. 38.5%;p<.001).
ACCEPTED MANUSCRIPT 3
Conclusions: There was no statistically significant difference in ventral hernia recurrence between the AWR and AWR+O groups. Bridging was associated with an increased rate of hernia
RI PT
recurrence and should be avoided if possible.
Key Words: Ventral hernia; parastomal hernia; stomal-site hernia; abdominal wall
Abbreviations
M AN U
SC
reconstruction; bioprosthetic matrix; acellular dermal matrix
ADM (acellular dermal matrix), AWR (abdominal wall reconstruction), AWR+O (abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia), SSI (surgical site
AC C
EP
TE D
infection), SSO (surgical site occurrence)
ACCEPTED MANUSCRIPT 4
Introduction The development of a ventral hernia after laparotomy is a relatively common complication, affecting 11-22% of all laparotomies and 350,000 operations annually.1,2
RI PT
Parastomal hernias occur even more often, affecting up to 48% of all ostomies.3 These two types of hernias often coexist: a recent study identified the risk for ventral hernia to be seven times greater in those with an existing parastomal hernia.4 The presence of a ventral hernia and
SC
concomitant ostomy-associated hernia is a challenging scenario for the reconstructive surgeon. These abdominal wall defects are among the most complicated to repair, necessitating thoughtful
M AN U
consideration of multiple variables: the presence of a contaminated field; a scarred, multiply reoperated abdominal wall; poor musculofascial quality; a large surface-area defect; and the frequent presence of medical comorbidities.4-8 The optimal reconstructive strategy for abdominal wall reconstruction in the presence of an ostomy is unclear. Aspects to consider include staging
TE D
the reconstruction versus repairing both hernias simultaneously, the role of mesh and the technique of mesh placement, and the use of component separation.4-9 The influence that an ostomy-associated hernia repair has on the success and longevity of a concomitant ventral
EP
herniorraphy is also unknown.
Despite the fact that ostomy-associated hernias are commonly associated with factors that
AC C
are predictive of poor wound healing, based on our clinical experience and prior to our review of the data, we hypothesized that the ventral hernia recurrence rate in patients undergoing a combined ventral hernia repair and stomal-site herniorraphy would not differ from the ventral hernia recurrence rate in patients undergoing an isolated ventral hernia repair. Additionally, we hypothesized that patients reconstructed with primary fascial coaptation and bioprosthetic matrix
ACCEPTED MANUSCRIPT 5
reinforcement experience lower hernia recurrence rates compared to those with bridged matrix repairs, regardless of the presence of an ostomy.
RI PT
Methods Patients and Outcomes Evaluated
We performed an institutional review board-approved retrospective cohort study,
SC
evaluating all consecutive patients who underwent midline repair of an abdominal wall hernia or oncologic defect, for which the fascia could or could not be primarily closed without undue
M AN U
tension, at The University of Texas MD Anderson Cancer Center between February 2000 and November 2015. The necessary informed consent was obtained for a retrospective study of this nature. All data were obtained from our Department’s prospectively-maintained abdominal wall reconstruction database as well as our institution’s electronic medical record.
TE D
For this analysis, inclusion criteria included an age ≥18 years and the use of bioprosthetic acellular dermal matrix mesh (ADM) in the reconstruction. Exclusion criteria included use of synthetic mesh, lateral abdominal wall defects (defects lateral to the semilunar line), and/or
EP
defects that could be closed primarily without tension (no ADM). Patients with synthetic mesh were excluded since there were so few (<2%) (Figure 1).
AC C
Patients were divided into two groups: those who underwent abdominal wall
reconstruction for ventral hernia alone (AWR group), and those who underwent AWR for ventral hernia with simultaneous ostomy-associated herniorraphy (AWR+O group). For the purpose of this study, we considered both parastomal hernias (hernias associated with a functioning stoma) and stoma-site hernias (hernias located at the site of a past stoma) ostomy-associated hernias. We analyzed patient, treatment, and defect characteristics and directly compared surgical outcomes
ACCEPTED MANUSCRIPT 6
between the AWR and AWR+O groups. Subgroup analysis compared outcomes for bridged versus matrix-reinforced repair between the AWR and AWR+O groups. Electronic medical records were reviewed for patient baseline characteristics and
RI PT
comorbidities. Prior abdominal wall radiation was defined as radiation directly to the abdominal wall or radiation to intra-abdominal organs. Defect characteristics assessed included the Ventral Hernia Working Group (VHWG) classification10, defect width, and defect surface area. Obesity
SC
was defined as a BMI greater than or equal to 30 kg/m2. Any patient who smoked tobacco within 1 month of surgery was considered an active smoker. The primary outcome was ventral hernia
M AN U
recurrence, defined as a contour abnormality associated with a fascial defect; a contour abnormality without a fascial defect was considered bulging. Hernia and bulge were considered mutually exclusive conditions and were diagnosed by physical examination and/or CT imaging. Secondary outcomes were ostomy hernia recurrence, surgical site occurrence (SSO; i.e. one or
TE D
more of the following: hematoma, seroma, wound skin dehiscence, fat necrosis), reoperation due to complication, surgical site infection (SSI) occurring within 30 days of the reconstruction (cellulitis requiring antibiotics, abscess, intra-abdominal sepsis, mesh/matrix infection),
EP
enterocutaneous fistula, mesh/matrix infection alone, and mesh/matrix explantation.
AC C
Surgical Technique
All repairs were performed through a midline laparotomy. Our technique for isolated
ventral hernia repair and placement of underlay ADM, as either a bridge or reinforcement, has been previously described.11 For all techniques, #1 polypropylene sutures are routinely used for both fascial reapproximation as well as for securing the ADM. For ostomy-associated hernias, our technique differed depending on whether it was a parastomal hernia or stoma-site hernia. For
ACCEPTED MANUSCRIPT 7
stoma-site hernia repair, we removed the hernia sac and performed complete adhesiolysis. If primary closure was possible, the fascial edges of the previous stoma site were freshened and reapproximated at both the posterior and anterior rectus sheaths, closing the defect transversely
RI PT
with interrupted #1 polypropylene sutures. The defect was then reinforced using a single sheet of ADM in an underlay, intraperitoneal position, reinforcing both the stoma-site and ventral midline hernias. If the stoma-site hernia could not be closed primarily without excessive tension, the
SC
ADM was placed in a bridged fashion, spanning the overlying fascial defect. Parastomal hernias were repaired either primarily, using a keyhole ADM design,12 or with the Sugarbaker
surgeon’s individual preference.
M AN U
technique.13 The decision to perform a keyhole or Sugarbaker repair was dependent on the
For the keyhole repairs, an area of ADM was excised in agreement with the diameter of the segment of bowel in the stoma. The stomal musculofascial defect was repaired by
TE D
approximating the abdominal wall in two layers using interrupted #1 polypropylene sutures in the posterior and anterior rectus sheaths, leaving only enough space for the bowel to pass through. The bowel serosa was carefully secured to the posterior rectus sheath with interrupted
EP
polyglactin sutures. The ADM was then placed in an underlay fashion, coning it circumferentially around the bowel. Care was taken to not inset the matrix so tightly as to
AC C
constrict the ostomy. The matrix was inset to the abdominal wall using interrupted #1 polypropylene sutures placed through the full thickness of the abdominal wall and secured to the bowel with interrupted polygalactin sutures placed through the serosa (Figure 2). The Sugarbaker repair was performed as previously described.13 ADM was then inset in
an underlay fashion, placing interrupted #1 polypropylene sutures through the full thickness of the abdominal wall, except for where the bowel passed between the matrix and the abdominal
ACCEPTED MANUSCRIPT 8
wall. The ADM was inset just firmly enough to not obstruct the bowel. Primary repair was used if the stomal fascial defect was 2 cm or less, extending from the serosa of the bowel conduit to the fascial border of the hernia defect. In this situation, the hernia
RI PT
sac was resected and the defect obliterated by approximating the abdominal wall in two layers using interrupted #1 polypropylene sutures in the posterior and anterior rectus sheaths.
SC
Statistical Analysis
Descriptive statistics [mean, standard deviation (SD), median, interquartile range (IQR)]
M AN U
were used as appropriate to summarize continuous variables. Frequencies and percentages described the categorical variables. Fisher’s exact test or the Chi-squared test was used to compare study subgroups for categorical variables. Logistic regression models were used to evaluate associations between patient and reconstructive characteristics and postoperative
TE D
complications. The time to hernia recurrence was defined as a time interval from the surgery date to the date of hernia recurrence or the last follow-up date when the patient did not experience a hernia recurrence. Patients who did not experience a hernia recurrence by the last follow-up date
EP
were censored in the data analysis. Kaplan-Meier curves were used to estimate the time to hernia recurrence, using the product limit method. Univariate and multivariable Cox proportional
AC C
hazard (Cox PH) regression models assessed the associations between patient and reconstructive factors and the time to hernia recurrence outcome. The backward model selection method was used to choose candidate models. Akaike information criteria (AIC) was used to determine the best fit model. In addition, to reduce the effect of treatment selection bias and potential confounding in this observational study, propensity score matching was used to re-assess the results.11 The propensity score (PS) was calculated as a predictive probability using a logistic
ACCEPTED MANUSCRIPT 9
regression model. The nearest neighbor algorithm was used to identify the matched pairs. The caliper was 0.2 of the standard deviation of the logit of PS. Additional information regarding the propensity score technique can be found in the supplemental data section. McNemar’s test,
RI PT
paired t-test, and stratified log-rank test were used to evaluate the outcomes for the paired data.14 All tests were two-sided. A p value of <0.05 was considered significant. All analyses were
performed using SAS 9.4 (SAS Institute, Inc) and R 3.3.0 (The R Foundation for Statistical
SC
Computing).
M AN U
Results
We included 499 consecutive patients who underwent AWR (381 AWR vs. 118 AWR+O) by 23 different surgeons with a median follow-up time of 27.2 months (IQR: 12.4-46.6). The median
TE D
follow-up times were similar between the AWR and AWR+O groups (29.5 vs. 23.6 months, respectively; p = 0.41). AWR+O patients had significantly longer hospital stays (7 days vs. 8 days; p =0.04). Eighty-eight percent of patients analyzed underwent postoperative CT imaging
EP
for oncologic surveillance or suspicion of hernia recurrence (85.3% AWR vs. 89.6% AWR+O).
AC C
Patient, Defect, and Technical Characteristics Patients in the AWR+O group were more likely to be older, male, and have >1 medical
comorbidity. AWR+O patients were also more likely to be categorized as ASA 4, to have had prior abdominal radiation therapy, and to have a VHWG class 3 or 4 hernia10 (Table 1). A bilateral component separation with underlay ADM was the most commonly performed technique for both groups (Table 2). Bridging ADM was required in 10.2% of the AWR patients
ACCEPTED MANUSCRIPT 10
and 9.3% of the AWR+O patients. The most common ADM was porcine-derived (Strattice, LifeCell Inc., Branchburg, NJ), followed by bovine (SurgiMend, Integra LifeSciences Corporation, Boston, MA), and then human cadaveric (AlloDerm, LifeCell Inc., Branchburg,
RI PT
NJ). The most common variant of ostomy-associated hernia was a parastomal hernia (91.5%). In the AWR+O group, the keyhole was the most common technique (74.6%); porcine-derived
SC
bioprosthetic mesh (52.5%) was the most common variety of ADM..
Propensity Score Matching Analysis
M AN U
Potential confounding factors, such as age, sex, BMI≥30 kg/m2, ASA score, >1 medical comorbidity, exposure to radiation, VHWG score 3 or 4, and ventral hernia width, were used to estimate the assignment probability to AWR+O vs. AWR in a logistic regression model. Table 1 shows baseline covariates, such as patient characteristics, in the matched study groups. After
TE D
propensity score matching, potential confounding factors such as the ASA status, VHWG 3 or 4 (i.e. contamination), and pre-operative radiation were balanced between the two groups. (Table 1) In the supplemental data section, we present the distribution of the propensity score before and
EP
after PS matching (eFigure 1) and calculated the standard difference to check balance between matched pairs (eTable 1). We also compared the patient characteristics between selected data and
AC C
discarded data by matching (eTable 2).
Factors Associated with Hernia Recurrence The percentage of patients with a hernia recurrence for the AWR and AWR+O groups
was not different (Before matching: 8.7% vs. 6.8%, respectively; p=0.60; After matching: 7.7% vs. 6.6%, respectively; p>0.99) (Table 3). With a multivariable Cox PH regression model, after
ACCEPTED MANUSCRIPT 11
adjusting for bridged mesh repair and smoking status, the presence of an ostomy-associated hernia was not significantly associated with ventral hernia recurrence (adjusted HR=0.7; 95% CI=0.31-1.5;p=0.34). A bridged mesh repair remained strongly associated with ventral hernia
RI PT
recurrence (adjusted HR=9.3; 95%CI=4.8-17.9; p<0.001). Kaplan-Meier curves demonstrated the ventral hernia recurrence rate of the AWR and AWR+O groups to be 11.1% and 10.3% by 3
SC
years and 17.7% and 10.3 % of 7 years after surgery (Figure 3).
Factors Associated with SSO and SSI
M AN U
Both before and after propensity matching, the AWR+O group was significantly more likely to develop a SSO, SSI, or require a reoperation (Table 3). We conducted univariate logistic regression analyses to identify factors associated with the development of an SSO or SSI (Table 4). Factors significantly associated with the development of SSO included the presence of an
TE D
ostomy-associated hernia (OR=2.5; 95% CI=1.6-3.9; p<0.001) and possessing >1 medical comorbidity (OR=2.3; 95% CI=1.2-4.3; p=0.014). After adjusting the potential confounding factors, such as BMI, defect width, and >1 comorbidity, the presence of an ostomy-associated
EP
hernia was still a significant risk factor for the development of SSO (adjusted OR =2.3; 95%
AC C
CI=1.4-3.7; p<0.001).
Bridged Versus Reinforced Matrix Given the strong association between ventral hernia recurrence and the presence of a
bridged repair in both groups, a subgroup analysis was conducted. The ventral hernia recurrence rate in patients who had a bridged repair was not significantly different between the AWR and AWR+O groups, respectively (38.5% vs. 9.1%; p>0.99) (Table 5). The ventral hernia recurrence
ACCEPTED MANUSCRIPT 12
rates were the lowest among those with a reinforced matrix in the AWR and AWR+O groups (5.3% vs. 6.5%, respectively; p>0.99).
RI PT
Discussion
This study supports our hypothesis that the rate of ventral hernia recurrence was not clinically and significantly different for patients who underwent an isolated ventral hernia repair
SC
compared to those who underwent a combined ventral hernia and ostomy-associated hernia
repair. This finding is not intuitive, given the degree of contamination present in any abdominal
M AN U
wall reconstruction in which a stoma is present.
We found that SSOs, infections, reoperations, and an increased length of stay were more common for the AWR+O patients. These findings agree with previously published literature from our group and others, demonstrating that patients who undergo abdominal laparotomies
18
TE D
classified as “clean” experience fewer reoperations and early, wound-related complications.12,15More specifically, the AWR+O group experienced more infectious events, including
subcutaneous abscess, deep infection, organ site infection, and >30 day SSI. However, we did
EP
not find that these infectious complications led to later ventral hernia recurrences. This finding is contradictory to previously published studies reporting SSI to be predictive of an initial
AC C
incisional hernia after laparotomy.19,20 However, in those studies, the authors defined SSI as either a deep space or organ-associated infection only, thus selecting for the most serious types of infections; in our study, SSI was defined more inclusively and included diagnoses such as cellulitis.
Armed with the results of this study, surgeons can now counsel patients with concurrent ventral and ostomy-associated hernias to expect a greater incidence of local wound-healing
ACCEPTED MANUSCRIPT 13
complications and even reoperation, but that the likelihood of developing a ventral hernia recurrence after their repair is no greater than that of a patient with only an isolated ventral hernia. The explanation for this finding may be related to the experience of our surgeons with
RI PT
complex abdominal wall reconstruction, our surgeons’ adherence to key surgical principles
(gentle handling of the soft tissue, minimal tension on the repair, minimal disruption to soft tissue vascularity), and the frequent use of ADM. ADM has been shown to maintain durability
SC
and functionality even in the repair of contaminated wounds.15-18 Furthermore, it is less strongly associated with adhesion formation or the development of enterocutaneous fistulae compared to
M AN U
synthetic mesh.20,21
Timmermans and colleagues4 found that the presence of a parastomal hernia is associated with a 7-fold greater risk of developing a ventral incisional hernia. The authors found that the location of the ventral hernia was most commonly parallel to the stoma, suggesting a mechanical
TE D
influence. The authors confirmed these findings in a second study, in which they used a virtual reality system to measure rectus muscle atrophy and midline shift on postoperative CT scans.5 Ipsilateral rectus muscle denervation, and increased, imbalanced force originating from the
EP
abdominal wall contralateral to the stoma are suggested as possible explanations for the development of the incisional hernia. If the development of these two anatomically distinct
AC C
hernias (i.e. stomal and ventral/incisional) is due to a disruption in the normal biomechanical forces on the abdominal wall, then simultaneous repair of both hernias may help to restore the physiologic tension of the abdominal wall, potentially resulting in a more durable repair, as found in our study.
We employed a similar technique for repairing the ventral incisional hernia in both groups: underlay ADM with or without an anterior component separation. We used either
ACCEPTED MANUSCRIPT 14
porcine or bovine-derived ADM with a similar frequency for both patient groups. Early in our series, before its inferiority in structural reconstruction was understood, we employed human cadaveric ADM in a small number of patients. We have since found human ADM to be
RI PT
associated with a significant risk of developing a ventral hernia recurrence.12,15,22 This is likely due to the greater amount of elastin in human dermis compared to xenograft dermis. Most
surgeons now avoid the use of human ADM in abdominal wall reconstruction.23 Considering this
SC
finding, excluding the 26 patients in whom human ADM was used would likely result in an even more favorable hernia recurrence rate in both the AWR and AWR+O groups.
M AN U
Despite the uniformity in ventral hernia repair, the technique of parastomal hernia repair was distributed between two different mesh orientations: keyhole repair and Sugarbaker repair.13,14 Neither technique was independently associated with an increased risk of ventral hernia recurrence. Hansson et al.13 performed a meta-analysis of various techniques for
TE D
parastomal hernia repair and did not identify any difference in hernia recurrence rates in those who underwent an open Sugarbaker versus keyhole repair. In their review, primary parastomal hernia repair without mesh reinforcement had the greatest recurrence rate, with a nine times
EP
higher chance of hernia recurrence compared to mesh-reinforced repairs. After controlling for associated factors such as defect width, defect surface area, and use
AC C
of human ADM, we found that bridged ADM was the only independent predictor of ventral hernia recurrence, with a nearly 9-fold increased risk of ventral hernia recurrence. This finding is in agreement with previously published literature showing bridged ventral hernia repair to be associated with a 56% rate of ventral hernia recurrence, compared to an 8% rate for mesh reinforced repair, at a mean follow-up of 31 months and a 100% hernia recurrence rate at 48 months (HR=9.5; p=<0.001).12 As proposed by other authors, we believe that a change in
ACCEPTED MANUSCRIPT 15
abdominal wall biomechanics is responsible for the elevated ventral hernia recurrence rate seen in bridged repairs. In a bridged repair, the forces generated by the bilateral oblique muscles are
rectus sheath fascia, as is the case in a reinforced repair.12,24
RI PT
borne by only the matrix as opposed to having the tension shared by both the matrix and the
The strengths of this study include a large cumulative clinical experience in abdominal wall reconstruction, both with and without ostomy-associated hernias; a large number of
SC
consecutive patients at a major U.S. cancer center; and data obtained from a prospectively maintained patient database. Our data analysis is further strengthened by utilizing both
M AN U
propensity-matched pairs as well as multivariable regression. Additional strengths of this study include a median follow-up of 27.2 months and the availability of routine CT surveillance results for tumor and hernia recurrence in over 88% of the patients.
Limitations of this study include its retrospective design and potential for selection bias.
TE D
We attempted to control for this bias by performing propensity score matching. However, the results after propensity score matching may still be biased by variables not taken into account in the analysis. Another limitation was the fact that all of the patients in the study were part of the
EP
oncologic population; therefore, the data may not be appropriately extrapolated to different
AC C
populations, such as those who have had damage-control abdominal surgery due to trauma or sepsis. Furthermore, our surgeons use ADM far more than synthetic mesh, which differs from the practice of many other centers, where synthetic mesh is primarily used due to lesser patient complexity and initial cost pressures. At our dedicated cancer center, every effort is made to avoid postoperative wound complications that might delay the delivery of adjuvant oncologic therapy and compromise cancer survival. Twenty-three different surgeons contributed patients to this study. This large group of surgeons likely introduces an element of inter-surgeon variability
ACCEPTED MANUSCRIPT 16
that may confound the results. An additional limitation was our limited statistical power to adequately analyze the effect of repair technique on stomal site hernia recurrence. Further studies
RI PT
on this topic are warranted. Conclusions
We found that patients who underwent abdominal wall reconstruction in the setting of an
SC
ostomy-associated hernia experienced more overall complications, infections, local woundhealing complications, reoperations, and longer hospital stays. However, despite the higher rate
M AN U
of early surgical-site complications, the risk of ventral hernia recurrence was no greater. The only factor found to be a significant, independent predictor of ventral hernia recurrence in both groups was the presence of a bridged mesh repair. Armed with this information, we recommend that surgeons not overestimate the impact of a concurrent ostomy-associated hernia on the outcomes of abdominal wall reconstruction. Surgeons should employ any and all means
AC C
EP
reconstruction.
TE D
available to achieve a reinforced, primary fascial closure and avoid a bridged fascial
ACCEPTED MANUSCRIPT 17
References 1.
Mudge M, Hughes LE. Incisional hernia: a 10-year prospective study of incidence and
attitudes. Br J Surg 1985;72:70-75. Poulose BK, Shelton J, Phillips S, et al. Epidemiology and cost of ventral hernia repair:
making the case for hernia research. Hernia 2012;16:179-183.
RI PT
2.
Carne PW, Robertson GM, Frizelle FA. Parastomal hernia. Br J Surg 2003;90:784-793.
4.
Timmermans L, Deerenberg EB, Lamme B, et al. Parastomal hernia is an independent
SC
3.
risk factor for incisional hernia in patients with end colostomy. Surgery 2014;155:178-183. Timmermans L, Deerenberg EB, van Dijk SM, et al. Abdominal rectus muscle atrophy
M AN U
5.
and midline shift after colostomy creation. Surgery 2014;155:696-701. 6.
Egun AI, Hill J, MacLennan I, et al. Preperitoneal approach to parastomal hernia with
coexistent large incisional hernia. Colorectal Dis 2002;4:132-134.
Zhu X, Tian W, Li J, et al. Repair of concomitant incisional and parastomal hernias using
TE D
7.
a hybrid technique: a series of 32 patients. Med Sci Monit 2015;17:2079-2083. 8.
Powell-Chandler A, Stephenson BM. Avoiding simultaneous incisional and parastomal
9.
EP
herniation. Colorectal Dis 2014;16:1020-1021.
Rosen MJ, Reynolds HL, Champagne B, et al. A novel approach for the simultaneous
AC C
repair of large midline incisional and parastomal hernias with biological mesh and retrorectus reconstruction. Am J Surg 2010;199:416-420. 10.
Ventral Hernia Working Group, Breuing K, Butler CE, Ferzoco S, et al. Incisional ventral
hernias: review of the literature and recommendations regarding the grading and technique of repair. Surgery 2010;148:544-558. 11.
Austin PC. An introduction to propensity score methods for reducing the effects of
ACCEPTED MANUSCRIPT 18
confounding in observational studies. Multivariate Behav Res 2011;46:399-424. 12.
Booth JH, Garvey PB, Baumann DP, et al. Primary fascial closure with mesh
reinforcement is superior to bridged mesh repair for abdominal wall reconstruction. J Am Coll
13.
RI PT
Surg 2013;217:999-1009.
Hansson BME, Slater NJ, van der Velden AS, et al. Surgical techniques for parastomal
hernia repair: a systematic review of the literature. Ann Surg 2012;255:685-695.
Sugarbaker PH. Prosthetic mesh repair of large hernias at the site of colonic stomas. Surg
SC
14.
Gynecol Obstet 1980;150:576-578.
Garvey PB, Martinez RA, Baumann DP, et al. Outcomes of abdominal wall
M AN U
15.
reconstruction with acellular dermal matrix are not affected by wound contamination. J Am Coll Surg 2014;219:853-864. 16.
Garvey PB, Bailey CM, Baumann DP, et al. Violation of the rectus complex is not a
2012;214:131-139. 17.
TE D
contraindication to component separation for abdominal wall reconstruction. J Am Coll Surg
Kanters AE, Krpata DM, Blatnik JA, et al. Modified hernia grading scale to stratify
18.
EP
surgical site occurrence after open ventral hernia repairs. J Am Coll Surg 2012;215:787-793. Rosen MJ, Krpata DM, Ermlich B, et al. A 5-year clinical experience with single-staged
AC C
repairs of infected and contaminated abdominal wall defects utilizing biologic mesh. Ann Surg 2013;257:991-996. 19.
Murray BW, Cipher DJ, Pham T, et al. The impact of surgical site infection on the
development of incisional hernia and small bowel obstruction in colorectal surgery. Am J Surg 2011;202:558-560. 20.
Veljkovic R1, Protic M, Gluhovic A, et al. Prospective clinical trial of factors predicting
ACCEPTED MANUSCRIPT 19
the early development of incisional hernia after midline laparotomy. J Am Coll Surg 2010;210:210-219. 21.
Karakousis CP, Volpe C, Tanski J, et al. Use of a mesh for musculoaponeurotic defects
RI PT
of the abdominal wall in cancer surgery and the risk of bowel fistulas. J Am Coll Surg 1995;181:11-16. 22.
Campbell KT, Burns NK, Rios CN, et al. Human versus non-cross-linked porcine
SC
acellular dermal matrix used for ventral hernia repair: comparison of in vivo fibrovascular remodeling and mechanical repair strength. Plast Reconstr Surg 2011;127:2321-2332.
Clemens MW, Selber JC, Liu J, et al. Bovine versus porcine acellular dermal matrix for
M AN U
23.
complex abdominal wall reconstruction. Plast Reconstr Surg 2013;131:71-79. 24.
Dumanian GA. Abdominal wall reconstruction. In: Thorne CH, Beasley RW, Aston SJ,
AC C
EP
TE D
et al, eds. Grabb and Smith’s Plastic Surgery. 6th ed. Boston: Lippincott; 2007:670-675.
ACCEPTED MANUSCRIPT 20
Table 1. Demographics and Baseline Patient Characteristics for the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair Of An Ostomy-Associated Hernia Groups
2
Mean BMI, kg/m (SD) 2
BMI ≥30 kg/m , n (%)
AWR; n=381
AWR+O; n=118
58.22 (12.29) 29.5 (12.247.0) 170 (44.6)
61.44 (12.26) 23.6 (14.0-42.0)
0.41
72 (61)
<0.01
30.9 (6.97)
31.5 (6.75)
0.23
199 (52.2)
72 (61)
0.09
ASA Status, n (%)
AWR+O; n=91
p † Value
58 (11.5)
60 (13.1)
0.26
29.8 (10.8-48.4)
24.7 (15.6-44.3)
0.26
44 (48.4)
49 (53.8)
0.39
32 (7.9)
30.7 (6.8)
0.27
47 (54)
51 (58.6)
0.65
M AN U
3
4 (1.0)
0.97
1 (1.1)
1 (1.1)
62 (16.3)
1 (0.8) 9 (7.6)
10 (11.0)
9 (9.9)
301 (79)
96 (81.4)
76 (83.5)
77 (84.6)
14 (3.7)
12 (10.2)
4 (4.4)
4 (4.4)
306 (80.3)
103 (87.3)
0.02
76 (83.5)
76 (83.5)
>0.99
31 (8.1)
10 (8.5)
0.91
9 (9.9)
9 (9.9)
>0.99
82 (21.5)
49 (41.5)
<0.01
34 (37.4)
32 (35.2)
0.51
11.2 (5.1)
12. (6.5)
0.51
11.2 (5.1)
12.0(6.5)
0.51
180 (84-300)
240 (90-377)
0.05
180 (90-300)
240 (80-375)
0.07
1 or 2
339 (89)
64 (54.2)
65 (71.4)
64 (70.3)
3 or 4
42 (11)
26 (28.6)
27 (29.7)
4 >1 medical comorbidity, n (%) Smoker, n (%)
54 (45.8)
EP
Radiation, n (%) Mean ventral hernia width, cm (SD) 2 Defect surface area, cm , median (IQR) VHWG Class, n (%)
TE D
1 2
0.013
AWR; n=91
SC
Characteristic Mean age, y (SD) Follow-up, mo, median (IQR) Male, n (%)
After propensity matching
p Value * 0.01
RI PT
Before propensity matching
<0.01
*p Values based on Chi-squared, Fisher's exact test, or Wilcoxon rank sum test. p Values based on McNemar's test, Bowker’s test, exact marginal homogeneity test for binary
AC C
†
outcomes, and Wilcoxon signed rank test for length of follow up. AWR, abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia; IQR, interquartile range; ASA, American Society of Anesthesiology; VHWG, ventral hernia working group.
0.70
ACCEPTED MANUSCRIPT 21
Table 2. Reconstruction Characteristics of the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair Of An Ostomy-Associated Hernia Groups AWR+O; n=118 n (%)
RI PT
AWR; n=381, n (%) n
%
n
%
None
128
33.6
38
32.2
Unilateral
57
15.0
28
23.7
Component separation
196
51.4
52
44.1
342
89.8
107
90.7
Mesh-bridged
39
10.2
11
9.3
M AN U
Type of mesh Porcine
189
49.6
62
52.5
Bovine
175
45.9
47
39.8
Human
17
4.5
9
7.6
108
91.5
10
8.5
88
74.6
24
2.03
6
5.1
Parastomal hernia Stomal site hernia Technique Mesh-Sugarbaker Primary repair
TE D
Mesh-key hole
0.079
SC
Bilateral Mesh-reinforced
p Value
0.77
0.27
AWR, abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with
AC C
EP
concomitant repair of an ostomy-associated hernia.
ACCEPTED MANUSCRIPT 22
Table 3. Outcomes of the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair Of An Ostomy-Associated Hernia Groups Before and After Propensity Matching After propensity matching AWR; AWR+O; p n=91 n=91 Value† 7 (7.7) 6 (6.6) >0.99
RI PT
Before propensity matching AWR; AWR+O; p n=381 n=118 Value* 33 (8.7) 8 (6.8) 0.6 0 (0)
2 (1.7)
>0.99
0 (0)
1 (1.1)
--
7 (5-10)
18 (7-13)
<0.001
7 (5-9)
8 (7-11)
0.04
SSO, n (%)
73 (19.2)
44 (37.3)
<0.001
17 (18.7)
31 (34.1)
0.04
0.36
2 (2.2)
1 (1.1)
>0.99
0.08
3 (3.3)
7 (7.7)
0.34
<0.001
9 (9.9)
22 (24.2)
0.006
4 (1)
3 (2.5)
Seroma
14 (3.7)
9 (7.6)
Wound skin dehiscence
46 (12.1)
32 (27.1)
Fat necrosis
25 (6.6)
10 (8.5)
0.54
6 (6.6)
7 (7.7)
0.76
30-d SSI, n (%)
20 (5.2)
11 (9.3)
0.12
3 (3.3)
10 (11)
0.05
Cellulitis
19 (5)
11 (9.3)
0.12
1 (1.1)
10 (11)
0.007
Abscess
19 (5)
13 (11)
0.03
5 (5.5)
11 (12.1)
0.13
Superficial infection
22 (5.8)
14 (11.9)
0.039
3 (3.3)
11 (12.1)
0.033
Deep infection
11 (2.9)
11 (9.3)
0.008
2 (2.2)
9 (9.9)
0.035
Organ site infection
7 (1.8)
5 (4.2)
0.17
3 (3.3)
5 (5.5)
0.48
6 (1.6)
6 (5.1)
0.039
3 (3.3)
4 (4.4)
0.71
3 (0.8)
2 (1.7)
0.34
1 (1.1)
2 (2.2)
>0.99
TE D
M AN U
Hematoma
SC
Ventral hernia recurrence, n (%) Ostomy-associated hernia recurrence, n (%) Length of hospitalization, d, median (IQR)
Other complications, n (%) >30 day SSI Mesh infection Mesh exposure
EP
Mesh removal Enterocutaneous fistula
8 (2.1)
3 (2.5)
0.73
3 (3.3)
3 (3.3)
>0.99
4 (1)
2 (1.7)
0.63
1 (1.1)
1 (1.1)
>0.99
5 (1.3)
2 (1.7)
0.67
0 (0)
1 (1.1)
>0.99
15 (16.5)
0.025
†
AC C
Reoperation 24 (6.3) 19 (16.1) 0.002 5 (5.5) *p Values based on Chi-squared, Fisher's exact test, or Wilcoxon rank sum test.
p Values based on McNemar's test, exact marginal homogeneity test for binary outcomes, Wilcoxon
signed rank test for length of hospitalization, and stratified log rank test for ventral hernia recurrence. SSI, surgical site infection; SSO, surgical site occurrence.
ACCEPTED MANUSCRIPT 23
Table 4. Univariate Logistic Regression Analysis of Factors Associated with the Development of Surgical Site Occurrence and Surgical Site Infection
Human Demographics and comorbidities
73 (19.2) 44 (37.3)
2.5 (1.6 - 3.9)
<0.001
108 10
20 (33.3)
1.8 (0.98 - 3.2)
0.06
2 (22.2)
0.9 (0.2-4.5)
24 88 6
7 (36.8)
1.1 (0.4 - 2.9)
32 (36.4) 2 (33.3)
449 50
103 (22.9)
251 222
59 (23.5)
26
14 (28)
Male
242
11 (9.3)
1.9 (0.9 - 4.0)
0.11
7 (11.7)
2.3 (0.9 - 5.5)
0.07
SC
p Value
0.91
0 (0)
--
--
0.93
2 (10.5)
1.1 (0.2 - 5.5)
0.91
1.1 (0.4 - 2.7)
0.87
7 (8)
0.5 (0.1 - 1.8)
0.28
0.9 (0.2 - 4.9)
0.87
2 (33.3)
5.5 (0.9 - 33.9)
0.07
27 (6)
Ref.
4 (8)
1.4 (0.5 - 4.0)
19 (7.6)
Ref. 0.6 (0.3-1.4) 0.48 (0.063.78)
Ref.
1.3 (0.7 - 2.5)
0.42
51 (23)
Ref. 0.97 (0.6-1.5)
0.89
11 (5)
7 (26.9)
1.20 (0.48-2.99)
0.69
1 (3.8)
--
1.0 (0.99 - 1.0)
0.73
--
49 (20.2)
0.7 (0.4-1.1)
0.1
12 (5)
EP
Age
SSO, n (%)
M AN U
AWR AWR+O Ostomy hernia subtypes Parastomal hernia Stomal site hernia Ostomy hernia repair technique Sugarbaker Keyhole Primary repair Mesh placement Mesh-reinforced Bridged Mesh type Porcine Bovine
Odds ratio (95% CI) Ref.
n 381 118
TE D
Groups
Surgical site infection 30-d SSI, n Odds ratio p (%) (95% CI) Value 20 (5.2) Ref.
RI PT
Surgical site occurrence
0.99 (0.96 1.02) 0.65 (0.311.37) 1.4 (0.7 - 2.9)
AC C
1.4 (0.7 - 2.9) 0.39 19 (7) BMI ≥30 kg/m2 271 83 (30.6) >1 medical 2.3 (1.2 - 4.3) 0.014 26 (6.4) 1.1 (0.4 - 3.0) 409 105 (25.7) comorbidity 9 (22) 0.9 (0.4 - 1.9) 0.81 2 (4.9) 0.7 (0.2 - 3.3) Smoker 41 1.1 (0.7 - 1.8) 0.7 7 (5.3) 0.8 (0.4 - 2.0) Radiation 131 32 (24.4) AWR, abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with
concomitant repair of an ostomy-associated hernia; SSI, surgical site infection; SSO, surgical site occurrence.
0.59
0.24 0.49
0.38
0.25 0.39 0.8 0.7 0.7
ACCEPTED MANUSCRIPT 24
Table 5. Outcomes for Mesh-Reinforced and Bridged Reconstructions in the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair
Mesh-reinforced AWR; n=342 AWR+O; n=107 (%) (%)
AWR; n=39 (%)
p Value*
15 (38.5)
1 (9.1)
>0.99
0
1 (9.1)
>0.99
7 (6.5)
>0.99
0 (0)
1 (1)
>0.99
7 (5-9)
8 (7-12)
<0.001
8 (6-15)
10 (8-14)
0.23
63 (18.4)
40 (37.4)
<0.001
10 (25.6)
4 (36.4)
0.48
4 (1.2)
Seroma
14 (4.1)
Wound skin dehiscence
36 (10.5)
Fat necrosis
22 (6.4) 17 (5)
Cellulitis
18 (5.3)
Abscess
17 (5)
2 (1.9)
SC
18 (5.3)
Hematoma
30-day SSI, n (%)
Mesh-bridged AWR+O; n=11 (%)
p Value*
0 (0)
1 (9.1)
0.22
6 (5.6)
0.59
0.63
0 (0)
3 (27.3)
0.008
30 (28)
<0.001
10 (25.6)
2 (18.2)
>0.99
10 (9.3)
0.29
3 (7.7)
0 (0)
>0.99
10 (9.3)
0.1
3 (7.7)
1 (9.1)
>0.99
9 (8.4)
0.25
1 (2.6)
2 (18.2)
0.12
11 (10.3)
0.06
2 (5.1)
2 (18.2)
0.21
12 (11.2)
0.08
2 (5.1)
2 (18.2)
0.21
11 (10.3)
0.002
2 (5.1)
0 (0)
>0.99
M AN U
Ventral hernia recurrence, n (%) Ostomy-associated hernia recurrence, n (%) Length of hospitalization, d, median (IQR) SSO, n (%)
RI PT
Of An Ostomy-Associated Hernia Groups
20 (5.8)
Deep infection
9 (2.6)
Organ site infection
5 (1.5)
4 (3.7)
0.23
2 (5.1)
1 (9.1)
0.53
>30-d SSI
5 (1.5)
5 (4.7)
0.06
1 (2.6)
1 (9.1)
0.39
Mesh infection
2 (0.6)
2 (1.9)
0.36
1 (2.6)
0 (0)
>0.99
Mesh exposure Mesh removal Enterocutaneous fistula
3 (2.8)
0.24
4 (10.3)
0 (0)
0.56
4 (1.2)
4 (1.2)
2 (1.9)
0.63
0
0
--
3 (0.9)
2 (1.9)
0.34
2 (5.1)
0 (0)
>0.99
1 (9.1)
>0.99
EP
Other complications, n (%)
TE D
Superficial infection
AC C
Reoperation 21 (6.1) 18 (16.8) 0.001 3 (7.7) *p Values based on Chi-squared, Fisher's exact test, or Wilcoxon rank sum test.
ACCEPTED MANUSCRIPT 25
Figure Legends
Figure 1. Flow diagram illustrating inclusion and exclusion criteria for study design. AWR,
RI PT
abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia.
SC
Figure 2. Bioprosthetic mesh design and inset technique for keyhole parastomal hernia repair.
M AN U
Figure 3. Kaplan-Meier curves showing ventral hernia-free probability estimates for up to 96 months after surgery. Note that the hernia recurrence rates are similar and not statistically significant. One patient who underwent a ventral hernia repair was excluded in analysis because of missing value of time to recurrence. N, number of patients in risk set at each time point; E,
TE D
number of patients who had ventral hernia recurrence by each time point; %, Kaplan-Meier
EP
estimate of the cumulative recurrence rate.
eFigure 1. Histogram of propensity scores (PS) of patients who underwent abdominal wall
AC C
reconstruction (AWR), and patients who underwent abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia (AWR+O), showing the distribution of the propensity scores before and after matching. Blue bar, the propensity score for AWR+O; red bar, the propensity score for AWR; purple bar, the overlap of the propensity scores. The figure shows that the balance of baseline covariates were much improved after propensity score matching. The balance of the baseline covariates in the matched pairs were also assessed through standardized
ACCEPTED MANUSCRIPT 26
AC C
EP
TE D
M AN U
SC
RI PT
differences. For more information see Austin.11
ACCEPTED MANUSCRIPT 27
Precis Ventral hernia repair with concomitant ostomy-associated herniorraphy is not associated with an increased hernia recurrence rate compared to ventral hernia repair alone. Bridging mesh is
AC C
EP
TE D
M AN U
SC
presence and should be avoided if possible.
RI PT
associated with an increased rate of ventral hernia recurrence regardless of stomal hernia
ACCEPTED MANUSCRIPT
Enrollment
Assessed for eligibility (n = 551) Excluded (n = 52)
SC
RI PT
Primary fascial closure without mesh (n=29) Synthetic mesh (n= 23)
M AN U
TE D
AWR (n = 381)
EP
AWR (n = 91)
AC C
Propensity Score
Group Allocation
AWR study patients (n = 499)
AWR+O (n = 118)
AWR+O (n = 91)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Figure 2
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT 1 [COPYEDITOR: Leave this 2-paragraph as an introduction to the tables.] Supplemental Analysis for Propensity Score Method
SC
RI PT
The original database had 499 consecutive patients who underwent a ventral hernia repair (AWR) (381 AWR vs 118 patients who underwent both a ventral hernia repair and ostomyassociated herniorraphy [AWR+O]). To control for the nonrandom assignment of patients, we employed a logistic regression model to predict the likelihood of AWR+O (the propensity score = PS). The potential confounding factors were used as the explanatory variables in the logistic regression model, including age, sex, BMI≥30 kg/m2, ASA score, >1 medical comorbidity, exposure to radiation, Ventral Hernia Working Group (VHWG) score 3 or 4, and ventral hernia width. We applied a nearest neighbor algorithm to identify the matched pairs. The caliper was 0.2 of the standard deviation of the logit of PS. After propensity score matching, we obtained 91 pairs.
AC C
EP
TE D
M AN U
The data are summarized in eTable 1. The standardized difference ranged 0 to 17% for the baseline covariates and most of the covariates had a standardized difference less than 10%, which demonstrates a good balance for the covariates in matched pairs. To describe the data more completely, we summarized the patient characteristics for matched and unmatched patients in eTable 2.
ACCEPTED MANUSCRIPT 2 eTable 1 Standardized Differences for Baseline Characteristics in Matched Data
After PS matching AWR+O
91
91
Mean age, y (SD)
58.0 (11.5)
60.1 (13.1)
Mean follow-up, mo (SD)
31.4 (22.0)
32 (22.4)
44 (48.4)
49 (53.8)
32.0 (7.9)
30.7 (6.8)
n
Male, n (%) Mean BMI, kg/m2 (SD) BMI≥30 kg/m2, n (%)
Standardized difference
0.17 --
0.11
SC
AWR
RI PT
Variables
0.16
52 (57.1)
0.02
1
1 (1.1)
1 (1.1)
0
2
10 (11)
9 (9.9)
0.04
3
76 (83.5)
77 (84.6)
0.03
4 (4.4)
4 (4.4)
0
76 (83.5)
76 (83.5)
0
9 (9.9)
9 (9.9)
0
34 (37.4)
32 (35.2)
0.05
11.3 (4.9)
12.1 (6.7)
0.13
Clean/clean contaminated
65 (71.4)
64 (70.3)
0.024
Contaminated/infected
26 (28.6)
27 (29.7)
0.024
M AN U
51 (56)
4
Smoker (%)
AC C
Radiation (%)
EP
≥1 medical comorbidity, n (%)
TE D
ASA Status, n (%)
Ventral hernia width, cm (SD) VHWG Class, n (%)
ACCEPTED MANUSCRIPT 3
Mean follow-up, mo (SD)
Matched data, n=182
58.9 (12.4) 28.1 (12.046.6)
59.1 (12.3) 25.5 (14.446.8) 93 (51.1)
31.4 (7.4) 103 (56.6) 19 (10.4) 153 (84.1) 8 (4.4)
86 (47.3)
AC C
EP
TE D
M AN U
Male, n (%) 149 (47) 2 Mean BMI, kg/m (SD) 31.0 (6.6) 2 BMI≥30 kg/m , n (%) 168 (53) ASA Status, n (%) 2 52 (16.4) 3 244 (77) 4 18 (5.7) >1 medical comorbidity, n (%) 142 (44.8) Smoker, n (%) 23 (7.3) Radiation, n (%) 65 (20.5) Median ventral hernia width, cm (IQR) 10 (7.0-15.0) Median defect surface area, 180 (84.0300.0) cm2 (IQR) VHWG Class, n (%) Clean/clean contaminated 274 (86.4) Contaminated/infected 43 (13.6) VHWG, Ventral Hernia Working Group
SC
Mean age, y (SD)
Unmatched data, n=317
RI PT
eTable 2. Patient Characteristics between the Matched and Unmatched Patients
18 (9.9)
66 (36.3)
11 (8.0-15.0) 207 (89.3350.0)
129 (70.9) 53 (29.1)
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
eFigure 1
AC C
EP
TE D
Reference:
S1072-7515(16)31673-8
DOI:
10.1016/j.jamcollsurg.2016.11.013
Reference:
ACS 8549
To appear in:
Journal of the American College of Surgeons
Received Date: 24 August 2016 Revised Date:
21 November 2016
Accepted Date: 22 November 2016
Please cite this article as: Mericli AF, Garvey PB, Giordano S, Liu J, Baumann DP, Butler CE, Abdominal Wall Reconstruction with Concomitant Ostomy-Associated Hernia Repair: Outcomes and Propensity Score Analysis, Journal of the American College of Surgeons (2017), doi: 10.1016/ j.jamcollsurg.2016.11.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Abdominal Wall Reconstruction with Concomitant Ostomy-Associated Hernia Repair: Outcomes and Propensity Score Analysis
RI PT
Alexander F Mericli, MD, Patrick B Garvey, MD, FACS, Salvatore Giordano, MD, PhD, Jun Liu, PhD, Donald P Baumann, MD, FACS, Charles E Butler, MD, FACS
SC
Department of Plastic Surgery, The University of Texas MD Anderson Cancer Center, Houston,
M AN U
TX
Disclosure Information: Dr Mericli received support from Acelity for travel to the European Association of Plastic Surgeons’ annual meeting to present this paper. All other authors have nothing to disclose.
TE D
Support: This work was supported in part by the NIH/NCI under award number P30CA016672 and used the Clinical Trials Support Resource.
Presented at the European Association of Plastic Surgeons Annual Meeting, Brussels, Belgium,
EP
May 2016
AC C
Correspondence address: Charles E. Butler, MD Professor and Chair, Department of Plastic Surgery The University of Texas MD Anderson Cancer Center 1400 Pressler Street Unit 1488 Houston, TX 77030 713-794-1247 Brief Title: Combined Ventral and Parastomal Hernia Repair
ACCEPTED MANUSCRIPT 2
Abstract: Background: The optimal strategy for abdominal wall reconstruction in the presence of a stomal-site hernia is unclear. We hypothesized that the rate of ventral hernia recurrence in
RI PT
patients undergoing a combined ventral hernia repair and stomal-site herniorraphy would not differ clinically from the ventral hernia recurrence rate in patients undergoing an isolated ventral
compared to reinforced repairs regardless of stomal hernia.
SC
hernia repair. We also hypothesized that bridged ventral hernia repairs result in worse outcomes
Study Design: We retrospectively reviewed prospectively collected data from consecutive
M AN U
abdominal wall reconstructions performed with acellular dermal matrix (ADM) at a single center during 2000-2015. We compared patients who underwent a ventral hernia repair alone (AWR) and those who underwent both a ventral hernia repair and ostomy-associated herniorraphy (AWR+O). We conducted a propensity score matched analysis to compare the outcomes
TE D
between the two groups. Multivariable Cox proportional hazards and logistic regression models were used to study associations between potential predictive/protective reconstructive strategies and surgical outcomes.
EP
Results: We included 499 patients (median follow-up, 27.2 months, interquartile range (IQR): 12.4-46.6 months), 118 AWR+O and 381 AWR. After propensity score matching, 91 pairs were
AC C
obtained. Ventral hernia recurrence was not statistically associated with ostomy-associated herniorraphy (adjusted HR=0.7, 95%CI=0.3-1.5;p=0.34). However, the AWR+O group experienced a significantly higher percentage of SSO (34.1%) than the AWR group (18.7%; adjusted OR=2.3, 95%CI=1.4-3.7;p<.001). In the AWR group, there were significantly fewer ventral hernia recurrences when the repair was reinforced compared to bridged (5.3% vs. 38.5%;p<.001).
ACCEPTED MANUSCRIPT 3
Conclusions: There was no statistically significant difference in ventral hernia recurrence between the AWR and AWR+O groups. Bridging was associated with an increased rate of hernia
RI PT
recurrence and should be avoided if possible.
Key Words: Ventral hernia; parastomal hernia; stomal-site hernia; abdominal wall
Abbreviations
M AN U
SC
reconstruction; bioprosthetic matrix; acellular dermal matrix
ADM (acellular dermal matrix), AWR (abdominal wall reconstruction), AWR+O (abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia), SSI (surgical site
AC C
EP
TE D
infection), SSO (surgical site occurrence)
ACCEPTED MANUSCRIPT 4
Introduction The development of a ventral hernia after laparotomy is a relatively common complication, affecting 11-22% of all laparotomies and 350,000 operations annually.1,2
RI PT
Parastomal hernias occur even more often, affecting up to 48% of all ostomies.3 These two types of hernias often coexist: a recent study identified the risk for ventral hernia to be seven times greater in those with an existing parastomal hernia.4 The presence of a ventral hernia and
SC
concomitant ostomy-associated hernia is a challenging scenario for the reconstructive surgeon. These abdominal wall defects are among the most complicated to repair, necessitating thoughtful
M AN U
consideration of multiple variables: the presence of a contaminated field; a scarred, multiply reoperated abdominal wall; poor musculofascial quality; a large surface-area defect; and the frequent presence of medical comorbidities.4-8 The optimal reconstructive strategy for abdominal wall reconstruction in the presence of an ostomy is unclear. Aspects to consider include staging
TE D
the reconstruction versus repairing both hernias simultaneously, the role of mesh and the technique of mesh placement, and the use of component separation.4-9 The influence that an ostomy-associated hernia repair has on the success and longevity of a concomitant ventral
EP
herniorraphy is also unknown.
Despite the fact that ostomy-associated hernias are commonly associated with factors that
AC C
are predictive of poor wound healing, based on our clinical experience and prior to our review of the data, we hypothesized that the ventral hernia recurrence rate in patients undergoing a combined ventral hernia repair and stomal-site herniorraphy would not differ from the ventral hernia recurrence rate in patients undergoing an isolated ventral hernia repair. Additionally, we hypothesized that patients reconstructed with primary fascial coaptation and bioprosthetic matrix
ACCEPTED MANUSCRIPT 5
reinforcement experience lower hernia recurrence rates compared to those with bridged matrix repairs, regardless of the presence of an ostomy.
RI PT
Methods Patients and Outcomes Evaluated
We performed an institutional review board-approved retrospective cohort study,
SC
evaluating all consecutive patients who underwent midline repair of an abdominal wall hernia or oncologic defect, for which the fascia could or could not be primarily closed without undue
M AN U
tension, at The University of Texas MD Anderson Cancer Center between February 2000 and November 2015. The necessary informed consent was obtained for a retrospective study of this nature. All data were obtained from our Department’s prospectively-maintained abdominal wall reconstruction database as well as our institution’s electronic medical record.
TE D
For this analysis, inclusion criteria included an age ≥18 years and the use of bioprosthetic acellular dermal matrix mesh (ADM) in the reconstruction. Exclusion criteria included use of synthetic mesh, lateral abdominal wall defects (defects lateral to the semilunar line), and/or
EP
defects that could be closed primarily without tension (no ADM). Patients with synthetic mesh were excluded since there were so few (<2%) (Figure 1).
AC C
Patients were divided into two groups: those who underwent abdominal wall
reconstruction for ventral hernia alone (AWR group), and those who underwent AWR for ventral hernia with simultaneous ostomy-associated herniorraphy (AWR+O group). For the purpose of this study, we considered both parastomal hernias (hernias associated with a functioning stoma) and stoma-site hernias (hernias located at the site of a past stoma) ostomy-associated hernias. We analyzed patient, treatment, and defect characteristics and directly compared surgical outcomes
ACCEPTED MANUSCRIPT 6
between the AWR and AWR+O groups. Subgroup analysis compared outcomes for bridged versus matrix-reinforced repair between the AWR and AWR+O groups. Electronic medical records were reviewed for patient baseline characteristics and
RI PT
comorbidities. Prior abdominal wall radiation was defined as radiation directly to the abdominal wall or radiation to intra-abdominal organs. Defect characteristics assessed included the Ventral Hernia Working Group (VHWG) classification10, defect width, and defect surface area. Obesity
SC
was defined as a BMI greater than or equal to 30 kg/m2. Any patient who smoked tobacco within 1 month of surgery was considered an active smoker. The primary outcome was ventral hernia
M AN U
recurrence, defined as a contour abnormality associated with a fascial defect; a contour abnormality without a fascial defect was considered bulging. Hernia and bulge were considered mutually exclusive conditions and were diagnosed by physical examination and/or CT imaging. Secondary outcomes were ostomy hernia recurrence, surgical site occurrence (SSO; i.e. one or
TE D
more of the following: hematoma, seroma, wound skin dehiscence, fat necrosis), reoperation due to complication, surgical site infection (SSI) occurring within 30 days of the reconstruction (cellulitis requiring antibiotics, abscess, intra-abdominal sepsis, mesh/matrix infection),
EP
enterocutaneous fistula, mesh/matrix infection alone, and mesh/matrix explantation.
AC C
Surgical Technique
All repairs were performed through a midline laparotomy. Our technique for isolated
ventral hernia repair and placement of underlay ADM, as either a bridge or reinforcement, has been previously described.11 For all techniques, #1 polypropylene sutures are routinely used for both fascial reapproximation as well as for securing the ADM. For ostomy-associated hernias, our technique differed depending on whether it was a parastomal hernia or stoma-site hernia. For
ACCEPTED MANUSCRIPT 7
stoma-site hernia repair, we removed the hernia sac and performed complete adhesiolysis. If primary closure was possible, the fascial edges of the previous stoma site were freshened and reapproximated at both the posterior and anterior rectus sheaths, closing the defect transversely
RI PT
with interrupted #1 polypropylene sutures. The defect was then reinforced using a single sheet of ADM in an underlay, intraperitoneal position, reinforcing both the stoma-site and ventral midline hernias. If the stoma-site hernia could not be closed primarily without excessive tension, the
SC
ADM was placed in a bridged fashion, spanning the overlying fascial defect. Parastomal hernias were repaired either primarily, using a keyhole ADM design,12 or with the Sugarbaker
surgeon’s individual preference.
M AN U
technique.13 The decision to perform a keyhole or Sugarbaker repair was dependent on the
For the keyhole repairs, an area of ADM was excised in agreement with the diameter of the segment of bowel in the stoma. The stomal musculofascial defect was repaired by
TE D
approximating the abdominal wall in two layers using interrupted #1 polypropylene sutures in the posterior and anterior rectus sheaths, leaving only enough space for the bowel to pass through. The bowel serosa was carefully secured to the posterior rectus sheath with interrupted
EP
polyglactin sutures. The ADM was then placed in an underlay fashion, coning it circumferentially around the bowel. Care was taken to not inset the matrix so tightly as to
AC C
constrict the ostomy. The matrix was inset to the abdominal wall using interrupted #1 polypropylene sutures placed through the full thickness of the abdominal wall and secured to the bowel with interrupted polygalactin sutures placed through the serosa (Figure 2). The Sugarbaker repair was performed as previously described.13 ADM was then inset in
an underlay fashion, placing interrupted #1 polypropylene sutures through the full thickness of the abdominal wall, except for where the bowel passed between the matrix and the abdominal
ACCEPTED MANUSCRIPT 8
wall. The ADM was inset just firmly enough to not obstruct the bowel. Primary repair was used if the stomal fascial defect was 2 cm or less, extending from the serosa of the bowel conduit to the fascial border of the hernia defect. In this situation, the hernia
RI PT
sac was resected and the defect obliterated by approximating the abdominal wall in two layers using interrupted #1 polypropylene sutures in the posterior and anterior rectus sheaths.
SC
Statistical Analysis
Descriptive statistics [mean, standard deviation (SD), median, interquartile range (IQR)]
M AN U
were used as appropriate to summarize continuous variables. Frequencies and percentages described the categorical variables. Fisher’s exact test or the Chi-squared test was used to compare study subgroups for categorical variables. Logistic regression models were used to evaluate associations between patient and reconstructive characteristics and postoperative
TE D
complications. The time to hernia recurrence was defined as a time interval from the surgery date to the date of hernia recurrence or the last follow-up date when the patient did not experience a hernia recurrence. Patients who did not experience a hernia recurrence by the last follow-up date
EP
were censored in the data analysis. Kaplan-Meier curves were used to estimate the time to hernia recurrence, using the product limit method. Univariate and multivariable Cox proportional
AC C
hazard (Cox PH) regression models assessed the associations between patient and reconstructive factors and the time to hernia recurrence outcome. The backward model selection method was used to choose candidate models. Akaike information criteria (AIC) was used to determine the best fit model. In addition, to reduce the effect of treatment selection bias and potential confounding in this observational study, propensity score matching was used to re-assess the results.11 The propensity score (PS) was calculated as a predictive probability using a logistic
ACCEPTED MANUSCRIPT 9
regression model. The nearest neighbor algorithm was used to identify the matched pairs. The caliper was 0.2 of the standard deviation of the logit of PS. Additional information regarding the propensity score technique can be found in the supplemental data section. McNemar’s test,
RI PT
paired t-test, and stratified log-rank test were used to evaluate the outcomes for the paired data.14 All tests were two-sided. A p value of <0.05 was considered significant. All analyses were
performed using SAS 9.4 (SAS Institute, Inc) and R 3.3.0 (The R Foundation for Statistical
SC
Computing).
M AN U
Results
We included 499 consecutive patients who underwent AWR (381 AWR vs. 118 AWR+O) by 23 different surgeons with a median follow-up time of 27.2 months (IQR: 12.4-46.6). The median
TE D
follow-up times were similar between the AWR and AWR+O groups (29.5 vs. 23.6 months, respectively; p = 0.41). AWR+O patients had significantly longer hospital stays (7 days vs. 8 days; p =0.04). Eighty-eight percent of patients analyzed underwent postoperative CT imaging
EP
for oncologic surveillance or suspicion of hernia recurrence (85.3% AWR vs. 89.6% AWR+O).
AC C
Patient, Defect, and Technical Characteristics Patients in the AWR+O group were more likely to be older, male, and have >1 medical
comorbidity. AWR+O patients were also more likely to be categorized as ASA 4, to have had prior abdominal radiation therapy, and to have a VHWG class 3 or 4 hernia10 (Table 1). A bilateral component separation with underlay ADM was the most commonly performed technique for both groups (Table 2). Bridging ADM was required in 10.2% of the AWR patients
ACCEPTED MANUSCRIPT 10
and 9.3% of the AWR+O patients. The most common ADM was porcine-derived (Strattice, LifeCell Inc., Branchburg, NJ), followed by bovine (SurgiMend, Integra LifeSciences Corporation, Boston, MA), and then human cadaveric (AlloDerm, LifeCell Inc., Branchburg,
RI PT
NJ). The most common variant of ostomy-associated hernia was a parastomal hernia (91.5%). In the AWR+O group, the keyhole was the most common technique (74.6%); porcine-derived
SC
bioprosthetic mesh (52.5%) was the most common variety of ADM..
Propensity Score Matching Analysis
M AN U
Potential confounding factors, such as age, sex, BMI≥30 kg/m2, ASA score, >1 medical comorbidity, exposure to radiation, VHWG score 3 or 4, and ventral hernia width, were used to estimate the assignment probability to AWR+O vs. AWR in a logistic regression model. Table 1 shows baseline covariates, such as patient characteristics, in the matched study groups. After
TE D
propensity score matching, potential confounding factors such as the ASA status, VHWG 3 or 4 (i.e. contamination), and pre-operative radiation were balanced between the two groups. (Table 1) In the supplemental data section, we present the distribution of the propensity score before and
EP
after PS matching (eFigure 1) and calculated the standard difference to check balance between matched pairs (eTable 1). We also compared the patient characteristics between selected data and
AC C
discarded data by matching (eTable 2).
Factors Associated with Hernia Recurrence The percentage of patients with a hernia recurrence for the AWR and AWR+O groups
was not different (Before matching: 8.7% vs. 6.8%, respectively; p=0.60; After matching: 7.7% vs. 6.6%, respectively; p>0.99) (Table 3). With a multivariable Cox PH regression model, after
ACCEPTED MANUSCRIPT 11
adjusting for bridged mesh repair and smoking status, the presence of an ostomy-associated hernia was not significantly associated with ventral hernia recurrence (adjusted HR=0.7; 95% CI=0.31-1.5;p=0.34). A bridged mesh repair remained strongly associated with ventral hernia
RI PT
recurrence (adjusted HR=9.3; 95%CI=4.8-17.9; p<0.001). Kaplan-Meier curves demonstrated the ventral hernia recurrence rate of the AWR and AWR+O groups to be 11.1% and 10.3% by 3
SC
years and 17.7% and 10.3 % of 7 years after surgery (Figure 3).
Factors Associated with SSO and SSI
M AN U
Both before and after propensity matching, the AWR+O group was significantly more likely to develop a SSO, SSI, or require a reoperation (Table 3). We conducted univariate logistic regression analyses to identify factors associated with the development of an SSO or SSI (Table 4). Factors significantly associated with the development of SSO included the presence of an
TE D
ostomy-associated hernia (OR=2.5; 95% CI=1.6-3.9; p<0.001) and possessing >1 medical comorbidity (OR=2.3; 95% CI=1.2-4.3; p=0.014). After adjusting the potential confounding factors, such as BMI, defect width, and >1 comorbidity, the presence of an ostomy-associated
EP
hernia was still a significant risk factor for the development of SSO (adjusted OR =2.3; 95%
AC C
CI=1.4-3.7; p<0.001).
Bridged Versus Reinforced Matrix Given the strong association between ventral hernia recurrence and the presence of a
bridged repair in both groups, a subgroup analysis was conducted. The ventral hernia recurrence rate in patients who had a bridged repair was not significantly different between the AWR and AWR+O groups, respectively (38.5% vs. 9.1%; p>0.99) (Table 5). The ventral hernia recurrence
ACCEPTED MANUSCRIPT 12
rates were the lowest among those with a reinforced matrix in the AWR and AWR+O groups (5.3% vs. 6.5%, respectively; p>0.99).
RI PT
Discussion
This study supports our hypothesis that the rate of ventral hernia recurrence was not clinically and significantly different for patients who underwent an isolated ventral hernia repair
SC
compared to those who underwent a combined ventral hernia and ostomy-associated hernia
repair. This finding is not intuitive, given the degree of contamination present in any abdominal
M AN U
wall reconstruction in which a stoma is present.
We found that SSOs, infections, reoperations, and an increased length of stay were more common for the AWR+O patients. These findings agree with previously published literature from our group and others, demonstrating that patients who undergo abdominal laparotomies
18
TE D
classified as “clean” experience fewer reoperations and early, wound-related complications.12,15More specifically, the AWR+O group experienced more infectious events, including
subcutaneous abscess, deep infection, organ site infection, and >30 day SSI. However, we did
EP
not find that these infectious complications led to later ventral hernia recurrences. This finding is contradictory to previously published studies reporting SSI to be predictive of an initial
AC C
incisional hernia after laparotomy.19,20 However, in those studies, the authors defined SSI as either a deep space or organ-associated infection only, thus selecting for the most serious types of infections; in our study, SSI was defined more inclusively and included diagnoses such as cellulitis.
Armed with the results of this study, surgeons can now counsel patients with concurrent ventral and ostomy-associated hernias to expect a greater incidence of local wound-healing
ACCEPTED MANUSCRIPT 13
complications and even reoperation, but that the likelihood of developing a ventral hernia recurrence after their repair is no greater than that of a patient with only an isolated ventral hernia. The explanation for this finding may be related to the experience of our surgeons with
RI PT
complex abdominal wall reconstruction, our surgeons’ adherence to key surgical principles
(gentle handling of the soft tissue, minimal tension on the repair, minimal disruption to soft tissue vascularity), and the frequent use of ADM. ADM has been shown to maintain durability
SC
and functionality even in the repair of contaminated wounds.15-18 Furthermore, it is less strongly associated with adhesion formation or the development of enterocutaneous fistulae compared to
M AN U
synthetic mesh.20,21
Timmermans and colleagues4 found that the presence of a parastomal hernia is associated with a 7-fold greater risk of developing a ventral incisional hernia. The authors found that the location of the ventral hernia was most commonly parallel to the stoma, suggesting a mechanical
TE D
influence. The authors confirmed these findings in a second study, in which they used a virtual reality system to measure rectus muscle atrophy and midline shift on postoperative CT scans.5 Ipsilateral rectus muscle denervation, and increased, imbalanced force originating from the
EP
abdominal wall contralateral to the stoma are suggested as possible explanations for the development of the incisional hernia. If the development of these two anatomically distinct
AC C
hernias (i.e. stomal and ventral/incisional) is due to a disruption in the normal biomechanical forces on the abdominal wall, then simultaneous repair of both hernias may help to restore the physiologic tension of the abdominal wall, potentially resulting in a more durable repair, as found in our study.
We employed a similar technique for repairing the ventral incisional hernia in both groups: underlay ADM with or without an anterior component separation. We used either
ACCEPTED MANUSCRIPT 14
porcine or bovine-derived ADM with a similar frequency for both patient groups. Early in our series, before its inferiority in structural reconstruction was understood, we employed human cadaveric ADM in a small number of patients. We have since found human ADM to be
RI PT
associated with a significant risk of developing a ventral hernia recurrence.12,15,22 This is likely due to the greater amount of elastin in human dermis compared to xenograft dermis. Most
surgeons now avoid the use of human ADM in abdominal wall reconstruction.23 Considering this
SC
finding, excluding the 26 patients in whom human ADM was used would likely result in an even more favorable hernia recurrence rate in both the AWR and AWR+O groups.
M AN U
Despite the uniformity in ventral hernia repair, the technique of parastomal hernia repair was distributed between two different mesh orientations: keyhole repair and Sugarbaker repair.13,14 Neither technique was independently associated with an increased risk of ventral hernia recurrence. Hansson et al.13 performed a meta-analysis of various techniques for
TE D
parastomal hernia repair and did not identify any difference in hernia recurrence rates in those who underwent an open Sugarbaker versus keyhole repair. In their review, primary parastomal hernia repair without mesh reinforcement had the greatest recurrence rate, with a nine times
EP
higher chance of hernia recurrence compared to mesh-reinforced repairs. After controlling for associated factors such as defect width, defect surface area, and use
AC C
of human ADM, we found that bridged ADM was the only independent predictor of ventral hernia recurrence, with a nearly 9-fold increased risk of ventral hernia recurrence. This finding is in agreement with previously published literature showing bridged ventral hernia repair to be associated with a 56% rate of ventral hernia recurrence, compared to an 8% rate for mesh reinforced repair, at a mean follow-up of 31 months and a 100% hernia recurrence rate at 48 months (HR=9.5; p=<0.001).12 As proposed by other authors, we believe that a change in
ACCEPTED MANUSCRIPT 15
abdominal wall biomechanics is responsible for the elevated ventral hernia recurrence rate seen in bridged repairs. In a bridged repair, the forces generated by the bilateral oblique muscles are
rectus sheath fascia, as is the case in a reinforced repair.12,24
RI PT
borne by only the matrix as opposed to having the tension shared by both the matrix and the
The strengths of this study include a large cumulative clinical experience in abdominal wall reconstruction, both with and without ostomy-associated hernias; a large number of
SC
consecutive patients at a major U.S. cancer center; and data obtained from a prospectively maintained patient database. Our data analysis is further strengthened by utilizing both
M AN U
propensity-matched pairs as well as multivariable regression. Additional strengths of this study include a median follow-up of 27.2 months and the availability of routine CT surveillance results for tumor and hernia recurrence in over 88% of the patients.
Limitations of this study include its retrospective design and potential for selection bias.
TE D
We attempted to control for this bias by performing propensity score matching. However, the results after propensity score matching may still be biased by variables not taken into account in the analysis. Another limitation was the fact that all of the patients in the study were part of the
EP
oncologic population; therefore, the data may not be appropriately extrapolated to different
AC C
populations, such as those who have had damage-control abdominal surgery due to trauma or sepsis. Furthermore, our surgeons use ADM far more than synthetic mesh, which differs from the practice of many other centers, where synthetic mesh is primarily used due to lesser patient complexity and initial cost pressures. At our dedicated cancer center, every effort is made to avoid postoperative wound complications that might delay the delivery of adjuvant oncologic therapy and compromise cancer survival. Twenty-three different surgeons contributed patients to this study. This large group of surgeons likely introduces an element of inter-surgeon variability
ACCEPTED MANUSCRIPT 16
that may confound the results. An additional limitation was our limited statistical power to adequately analyze the effect of repair technique on stomal site hernia recurrence. Further studies
RI PT
on this topic are warranted. Conclusions
We found that patients who underwent abdominal wall reconstruction in the setting of an
SC
ostomy-associated hernia experienced more overall complications, infections, local woundhealing complications, reoperations, and longer hospital stays. However, despite the higher rate
M AN U
of early surgical-site complications, the risk of ventral hernia recurrence was no greater. The only factor found to be a significant, independent predictor of ventral hernia recurrence in both groups was the presence of a bridged mesh repair. Armed with this information, we recommend that surgeons not overestimate the impact of a concurrent ostomy-associated hernia on the outcomes of abdominal wall reconstruction. Surgeons should employ any and all means
AC C
EP
reconstruction.
TE D
available to achieve a reinforced, primary fascial closure and avoid a bridged fascial
ACCEPTED MANUSCRIPT 17
References 1.
Mudge M, Hughes LE. Incisional hernia: a 10-year prospective study of incidence and
attitudes. Br J Surg 1985;72:70-75. Poulose BK, Shelton J, Phillips S, et al. Epidemiology and cost of ventral hernia repair:
making the case for hernia research. Hernia 2012;16:179-183.
RI PT
2.
Carne PW, Robertson GM, Frizelle FA. Parastomal hernia. Br J Surg 2003;90:784-793.
4.
Timmermans L, Deerenberg EB, Lamme B, et al. Parastomal hernia is an independent
SC
3.
risk factor for incisional hernia in patients with end colostomy. Surgery 2014;155:178-183. Timmermans L, Deerenberg EB, van Dijk SM, et al. Abdominal rectus muscle atrophy
M AN U
5.
and midline shift after colostomy creation. Surgery 2014;155:696-701. 6.
Egun AI, Hill J, MacLennan I, et al. Preperitoneal approach to parastomal hernia with
coexistent large incisional hernia. Colorectal Dis 2002;4:132-134.
Zhu X, Tian W, Li J, et al. Repair of concomitant incisional and parastomal hernias using
TE D
7.
a hybrid technique: a series of 32 patients. Med Sci Monit 2015;17:2079-2083. 8.
Powell-Chandler A, Stephenson BM. Avoiding simultaneous incisional and parastomal
9.
EP
herniation. Colorectal Dis 2014;16:1020-1021.
Rosen MJ, Reynolds HL, Champagne B, et al. A novel approach for the simultaneous
AC C
repair of large midline incisional and parastomal hernias with biological mesh and retrorectus reconstruction. Am J Surg 2010;199:416-420. 10.
Ventral Hernia Working Group, Breuing K, Butler CE, Ferzoco S, et al. Incisional ventral
hernias: review of the literature and recommendations regarding the grading and technique of repair. Surgery 2010;148:544-558. 11.
Austin PC. An introduction to propensity score methods for reducing the effects of
ACCEPTED MANUSCRIPT 18
confounding in observational studies. Multivariate Behav Res 2011;46:399-424. 12.
Booth JH, Garvey PB, Baumann DP, et al. Primary fascial closure with mesh
reinforcement is superior to bridged mesh repair for abdominal wall reconstruction. J Am Coll
13.
RI PT
Surg 2013;217:999-1009.
Hansson BME, Slater NJ, van der Velden AS, et al. Surgical techniques for parastomal
hernia repair: a systematic review of the literature. Ann Surg 2012;255:685-695.
Sugarbaker PH. Prosthetic mesh repair of large hernias at the site of colonic stomas. Surg
SC
14.
Gynecol Obstet 1980;150:576-578.
Garvey PB, Martinez RA, Baumann DP, et al. Outcomes of abdominal wall
M AN U
15.
reconstruction with acellular dermal matrix are not affected by wound contamination. J Am Coll Surg 2014;219:853-864. 16.
Garvey PB, Bailey CM, Baumann DP, et al. Violation of the rectus complex is not a
2012;214:131-139. 17.
TE D
contraindication to component separation for abdominal wall reconstruction. J Am Coll Surg
Kanters AE, Krpata DM, Blatnik JA, et al. Modified hernia grading scale to stratify
18.
EP
surgical site occurrence after open ventral hernia repairs. J Am Coll Surg 2012;215:787-793. Rosen MJ, Krpata DM, Ermlich B, et al. A 5-year clinical experience with single-staged
AC C
repairs of infected and contaminated abdominal wall defects utilizing biologic mesh. Ann Surg 2013;257:991-996. 19.
Murray BW, Cipher DJ, Pham T, et al. The impact of surgical site infection on the
development of incisional hernia and small bowel obstruction in colorectal surgery. Am J Surg 2011;202:558-560. 20.
Veljkovic R1, Protic M, Gluhovic A, et al. Prospective clinical trial of factors predicting
ACCEPTED MANUSCRIPT 19
the early development of incisional hernia after midline laparotomy. J Am Coll Surg 2010;210:210-219. 21.
Karakousis CP, Volpe C, Tanski J, et al. Use of a mesh for musculoaponeurotic defects
RI PT
of the abdominal wall in cancer surgery and the risk of bowel fistulas. J Am Coll Surg 1995;181:11-16. 22.
Campbell KT, Burns NK, Rios CN, et al. Human versus non-cross-linked porcine
SC
acellular dermal matrix used for ventral hernia repair: comparison of in vivo fibrovascular remodeling and mechanical repair strength. Plast Reconstr Surg 2011;127:2321-2332.
Clemens MW, Selber JC, Liu J, et al. Bovine versus porcine acellular dermal matrix for
M AN U
23.
complex abdominal wall reconstruction. Plast Reconstr Surg 2013;131:71-79. 24.
Dumanian GA. Abdominal wall reconstruction. In: Thorne CH, Beasley RW, Aston SJ,
AC C
EP
TE D
et al, eds. Grabb and Smith’s Plastic Surgery. 6th ed. Boston: Lippincott; 2007:670-675.
ACCEPTED MANUSCRIPT 20
Table 1. Demographics and Baseline Patient Characteristics for the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair Of An Ostomy-Associated Hernia Groups
2
Mean BMI, kg/m (SD) 2
BMI ≥30 kg/m , n (%)
AWR; n=381
AWR+O; n=118
58.22 (12.29) 29.5 (12.247.0) 170 (44.6)
61.44 (12.26) 23.6 (14.0-42.0)
0.41
72 (61)
<0.01
30.9 (6.97)
31.5 (6.75)
0.23
199 (52.2)
72 (61)
0.09
ASA Status, n (%)
AWR+O; n=91
p † Value
58 (11.5)
60 (13.1)
0.26
29.8 (10.8-48.4)
24.7 (15.6-44.3)
0.26
44 (48.4)
49 (53.8)
0.39
32 (7.9)
30.7 (6.8)
0.27
47 (54)
51 (58.6)
0.65
M AN U
3
4 (1.0)
0.97
1 (1.1)
1 (1.1)
62 (16.3)
1 (0.8) 9 (7.6)
10 (11.0)
9 (9.9)
301 (79)
96 (81.4)
76 (83.5)
77 (84.6)
14 (3.7)
12 (10.2)
4 (4.4)
4 (4.4)
306 (80.3)
103 (87.3)
0.02
76 (83.5)
76 (83.5)
>0.99
31 (8.1)
10 (8.5)
0.91
9 (9.9)
9 (9.9)
>0.99
82 (21.5)
49 (41.5)
<0.01
34 (37.4)
32 (35.2)
0.51
11.2 (5.1)
12. (6.5)
0.51
11.2 (5.1)
12.0(6.5)
0.51
180 (84-300)
240 (90-377)
0.05
180 (90-300)
240 (80-375)
0.07
1 or 2
339 (89)
64 (54.2)
65 (71.4)
64 (70.3)
3 or 4
42 (11)
26 (28.6)
27 (29.7)
4 >1 medical comorbidity, n (%) Smoker, n (%)
54 (45.8)
EP
Radiation, n (%) Mean ventral hernia width, cm (SD) 2 Defect surface area, cm , median (IQR) VHWG Class, n (%)
TE D
1 2
0.013
AWR; n=91
SC
Characteristic Mean age, y (SD) Follow-up, mo, median (IQR) Male, n (%)
After propensity matching
p Value * 0.01
RI PT
Before propensity matching
<0.01
*p Values based on Chi-squared, Fisher's exact test, or Wilcoxon rank sum test. p Values based on McNemar's test, Bowker’s test, exact marginal homogeneity test for binary
AC C
†
outcomes, and Wilcoxon signed rank test for length of follow up. AWR, abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia; IQR, interquartile range; ASA, American Society of Anesthesiology; VHWG, ventral hernia working group.
0.70
ACCEPTED MANUSCRIPT 21
Table 2. Reconstruction Characteristics of the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair Of An Ostomy-Associated Hernia Groups AWR+O; n=118 n (%)
RI PT
AWR; n=381, n (%) n
%
n
%
None
128
33.6
38
32.2
Unilateral
57
15.0
28
23.7
Component separation
196
51.4
52
44.1
342
89.8
107
90.7
Mesh-bridged
39
10.2
11
9.3
M AN U
Type of mesh Porcine
189
49.6
62
52.5
Bovine
175
45.9
47
39.8
Human
17
4.5
9
7.6
108
91.5
10
8.5
88
74.6
24
2.03
6
5.1
Parastomal hernia Stomal site hernia Technique Mesh-Sugarbaker Primary repair
TE D
Mesh-key hole
0.079
SC
Bilateral Mesh-reinforced
p Value
0.77
0.27
AWR, abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with
AC C
EP
concomitant repair of an ostomy-associated hernia.
ACCEPTED MANUSCRIPT 22
Table 3. Outcomes of the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair Of An Ostomy-Associated Hernia Groups Before and After Propensity Matching After propensity matching AWR; AWR+O; p n=91 n=91 Value† 7 (7.7) 6 (6.6) >0.99
RI PT
Before propensity matching AWR; AWR+O; p n=381 n=118 Value* 33 (8.7) 8 (6.8) 0.6 0 (0)
2 (1.7)
>0.99
0 (0)
1 (1.1)
--
7 (5-10)
18 (7-13)
<0.001
7 (5-9)
8 (7-11)
0.04
SSO, n (%)
73 (19.2)
44 (37.3)
<0.001
17 (18.7)
31 (34.1)
0.04
0.36
2 (2.2)
1 (1.1)
>0.99
0.08
3 (3.3)
7 (7.7)
0.34
<0.001
9 (9.9)
22 (24.2)
0.006
4 (1)
3 (2.5)
Seroma
14 (3.7)
9 (7.6)
Wound skin dehiscence
46 (12.1)
32 (27.1)
Fat necrosis
25 (6.6)
10 (8.5)
0.54
6 (6.6)
7 (7.7)
0.76
30-d SSI, n (%)
20 (5.2)
11 (9.3)
0.12
3 (3.3)
10 (11)
0.05
Cellulitis
19 (5)
11 (9.3)
0.12
1 (1.1)
10 (11)
0.007
Abscess
19 (5)
13 (11)
0.03
5 (5.5)
11 (12.1)
0.13
Superficial infection
22 (5.8)
14 (11.9)
0.039
3 (3.3)
11 (12.1)
0.033
Deep infection
11 (2.9)
11 (9.3)
0.008
2 (2.2)
9 (9.9)
0.035
Organ site infection
7 (1.8)
5 (4.2)
0.17
3 (3.3)
5 (5.5)
0.48
6 (1.6)
6 (5.1)
0.039
3 (3.3)
4 (4.4)
0.71
3 (0.8)
2 (1.7)
0.34
1 (1.1)
2 (2.2)
>0.99
TE D
M AN U
Hematoma
SC
Ventral hernia recurrence, n (%) Ostomy-associated hernia recurrence, n (%) Length of hospitalization, d, median (IQR)
Other complications, n (%) >30 day SSI Mesh infection Mesh exposure
EP
Mesh removal Enterocutaneous fistula
8 (2.1)
3 (2.5)
0.73
3 (3.3)
3 (3.3)
>0.99
4 (1)
2 (1.7)
0.63
1 (1.1)
1 (1.1)
>0.99
5 (1.3)
2 (1.7)
0.67
0 (0)
1 (1.1)
>0.99
15 (16.5)
0.025
†
AC C
Reoperation 24 (6.3) 19 (16.1) 0.002 5 (5.5) *p Values based on Chi-squared, Fisher's exact test, or Wilcoxon rank sum test.
p Values based on McNemar's test, exact marginal homogeneity test for binary outcomes, Wilcoxon
signed rank test for length of hospitalization, and stratified log rank test for ventral hernia recurrence. SSI, surgical site infection; SSO, surgical site occurrence.
ACCEPTED MANUSCRIPT 23
Table 4. Univariate Logistic Regression Analysis of Factors Associated with the Development of Surgical Site Occurrence and Surgical Site Infection
Human Demographics and comorbidities
73 (19.2) 44 (37.3)
2.5 (1.6 - 3.9)
<0.001
108 10
20 (33.3)
1.8 (0.98 - 3.2)
0.06
2 (22.2)
0.9 (0.2-4.5)
24 88 6
7 (36.8)
1.1 (0.4 - 2.9)
32 (36.4) 2 (33.3)
449 50
103 (22.9)
251 222
59 (23.5)
26
14 (28)
Male
242
11 (9.3)
1.9 (0.9 - 4.0)
0.11
7 (11.7)
2.3 (0.9 - 5.5)
0.07
SC
p Value
0.91
0 (0)
--
--
0.93
2 (10.5)
1.1 (0.2 - 5.5)
0.91
1.1 (0.4 - 2.7)
0.87
7 (8)
0.5 (0.1 - 1.8)
0.28
0.9 (0.2 - 4.9)
0.87
2 (33.3)
5.5 (0.9 - 33.9)
0.07
27 (6)
Ref.
4 (8)
1.4 (0.5 - 4.0)
19 (7.6)
Ref. 0.6 (0.3-1.4) 0.48 (0.063.78)
Ref.
1.3 (0.7 - 2.5)
0.42
51 (23)
Ref. 0.97 (0.6-1.5)
0.89
11 (5)
7 (26.9)
1.20 (0.48-2.99)
0.69
1 (3.8)
--
1.0 (0.99 - 1.0)
0.73
--
49 (20.2)
0.7 (0.4-1.1)
0.1
12 (5)
EP
Age
SSO, n (%)
M AN U
AWR AWR+O Ostomy hernia subtypes Parastomal hernia Stomal site hernia Ostomy hernia repair technique Sugarbaker Keyhole Primary repair Mesh placement Mesh-reinforced Bridged Mesh type Porcine Bovine
Odds ratio (95% CI) Ref.
n 381 118
TE D
Groups
Surgical site infection 30-d SSI, n Odds ratio p (%) (95% CI) Value 20 (5.2) Ref.
RI PT
Surgical site occurrence
0.99 (0.96 1.02) 0.65 (0.311.37) 1.4 (0.7 - 2.9)
AC C
1.4 (0.7 - 2.9) 0.39 19 (7) BMI ≥30 kg/m2 271 83 (30.6) >1 medical 2.3 (1.2 - 4.3) 0.014 26 (6.4) 1.1 (0.4 - 3.0) 409 105 (25.7) comorbidity 9 (22) 0.9 (0.4 - 1.9) 0.81 2 (4.9) 0.7 (0.2 - 3.3) Smoker 41 1.1 (0.7 - 1.8) 0.7 7 (5.3) 0.8 (0.4 - 2.0) Radiation 131 32 (24.4) AWR, abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with
concomitant repair of an ostomy-associated hernia; SSI, surgical site infection; SSO, surgical site occurrence.
0.59
0.24 0.49
0.38
0.25 0.39 0.8 0.7 0.7
ACCEPTED MANUSCRIPT 24
Table 5. Outcomes for Mesh-Reinforced and Bridged Reconstructions in the Abdominal Wall Reconstruction Only vs Abdominal Wall Reconstruction With Concomitant Repair
Mesh-reinforced AWR; n=342 AWR+O; n=107 (%) (%)
AWR; n=39 (%)
p Value*
15 (38.5)
1 (9.1)
>0.99
0
1 (9.1)
>0.99
7 (6.5)
>0.99
0 (0)
1 (1)
>0.99
7 (5-9)
8 (7-12)
<0.001
8 (6-15)
10 (8-14)
0.23
63 (18.4)
40 (37.4)
<0.001
10 (25.6)
4 (36.4)
0.48
4 (1.2)
Seroma
14 (4.1)
Wound skin dehiscence
36 (10.5)
Fat necrosis
22 (6.4) 17 (5)
Cellulitis
18 (5.3)
Abscess
17 (5)
2 (1.9)
SC
18 (5.3)
Hematoma
30-day SSI, n (%)
Mesh-bridged AWR+O; n=11 (%)
p Value*
0 (0)
1 (9.1)
0.22
6 (5.6)
0.59
0.63
0 (0)
3 (27.3)
0.008
30 (28)
<0.001
10 (25.6)
2 (18.2)
>0.99
10 (9.3)
0.29
3 (7.7)
0 (0)
>0.99
10 (9.3)
0.1
3 (7.7)
1 (9.1)
>0.99
9 (8.4)
0.25
1 (2.6)
2 (18.2)
0.12
11 (10.3)
0.06
2 (5.1)
2 (18.2)
0.21
12 (11.2)
0.08
2 (5.1)
2 (18.2)
0.21
11 (10.3)
0.002
2 (5.1)
0 (0)
>0.99
M AN U
Ventral hernia recurrence, n (%) Ostomy-associated hernia recurrence, n (%) Length of hospitalization, d, median (IQR) SSO, n (%)
RI PT
Of An Ostomy-Associated Hernia Groups
20 (5.8)
Deep infection
9 (2.6)
Organ site infection
5 (1.5)
4 (3.7)
0.23
2 (5.1)
1 (9.1)
0.53
>30-d SSI
5 (1.5)
5 (4.7)
0.06
1 (2.6)
1 (9.1)
0.39
Mesh infection
2 (0.6)
2 (1.9)
0.36
1 (2.6)
0 (0)
>0.99
Mesh exposure Mesh removal Enterocutaneous fistula
3 (2.8)
0.24
4 (10.3)
0 (0)
0.56
4 (1.2)
4 (1.2)
2 (1.9)
0.63
0
0
--
3 (0.9)
2 (1.9)
0.34
2 (5.1)
0 (0)
>0.99
1 (9.1)
>0.99
EP
Other complications, n (%)
TE D
Superficial infection
AC C
Reoperation 21 (6.1) 18 (16.8) 0.001 3 (7.7) *p Values based on Chi-squared, Fisher's exact test, or Wilcoxon rank sum test.
ACCEPTED MANUSCRIPT 25
Figure Legends
Figure 1. Flow diagram illustrating inclusion and exclusion criteria for study design. AWR,
RI PT
abdominal wall reconstruction; AWR+O, abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia.
SC
Figure 2. Bioprosthetic mesh design and inset technique for keyhole parastomal hernia repair.
M AN U
Figure 3. Kaplan-Meier curves showing ventral hernia-free probability estimates for up to 96 months after surgery. Note that the hernia recurrence rates are similar and not statistically significant. One patient who underwent a ventral hernia repair was excluded in analysis because of missing value of time to recurrence. N, number of patients in risk set at each time point; E,
TE D
number of patients who had ventral hernia recurrence by each time point; %, Kaplan-Meier
EP
estimate of the cumulative recurrence rate.
eFigure 1. Histogram of propensity scores (PS) of patients who underwent abdominal wall
AC C
reconstruction (AWR), and patients who underwent abdominal wall reconstruction with concomitant repair of an ostomy-associated hernia (AWR+O), showing the distribution of the propensity scores before and after matching. Blue bar, the propensity score for AWR+O; red bar, the propensity score for AWR; purple bar, the overlap of the propensity scores. The figure shows that the balance of baseline covariates were much improved after propensity score matching. The balance of the baseline covariates in the matched pairs were also assessed through standardized
ACCEPTED MANUSCRIPT 26
AC C
EP
TE D
M AN U
SC
RI PT
differences. For more information see Austin.11
ACCEPTED MANUSCRIPT 27
Precis Ventral hernia repair with concomitant ostomy-associated herniorraphy is not associated with an increased hernia recurrence rate compared to ventral hernia repair alone. Bridging mesh is
AC C
EP
TE D
M AN U
SC
presence and should be avoided if possible.
RI PT
associated with an increased rate of ventral hernia recurrence regardless of stomal hernia
ACCEPTED MANUSCRIPT
Enrollment
Assessed for eligibility (n = 551) Excluded (n = 52)
SC
RI PT
Primary fascial closure without mesh (n=29) Synthetic mesh (n= 23)
M AN U
TE D
AWR (n = 381)
EP
AWR (n = 91)
AC C
Propensity Score
Group Allocation
AWR study patients (n = 499)
AWR+O (n = 118)
AWR+O (n = 91)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Figure 2
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT 1 [COPYEDITOR: Leave this 2-paragraph as an introduction to the tables.] Supplemental Analysis for Propensity Score Method
SC
RI PT
The original database had 499 consecutive patients who underwent a ventral hernia repair (AWR) (381 AWR vs 118 patients who underwent both a ventral hernia repair and ostomyassociated herniorraphy [AWR+O]). To control for the nonrandom assignment of patients, we employed a logistic regression model to predict the likelihood of AWR+O (the propensity score = PS). The potential confounding factors were used as the explanatory variables in the logistic regression model, including age, sex, BMI≥30 kg/m2, ASA score, >1 medical comorbidity, exposure to radiation, Ventral Hernia Working Group (VHWG) score 3 or 4, and ventral hernia width. We applied a nearest neighbor algorithm to identify the matched pairs. The caliper was 0.2 of the standard deviation of the logit of PS. After propensity score matching, we obtained 91 pairs.
AC C
EP
TE D
M AN U
The data are summarized in eTable 1. The standardized difference ranged 0 to 17% for the baseline covariates and most of the covariates had a standardized difference less than 10%, which demonstrates a good balance for the covariates in matched pairs. To describe the data more completely, we summarized the patient characteristics for matched and unmatched patients in eTable 2.
ACCEPTED MANUSCRIPT 2 eTable 1 Standardized Differences for Baseline Characteristics in Matched Data
After PS matching AWR+O
91
91
Mean age, y (SD)
58.0 (11.5)
60.1 (13.1)
Mean follow-up, mo (SD)
31.4 (22.0)
32 (22.4)
44 (48.4)
49 (53.8)
32.0 (7.9)
30.7 (6.8)
n
Male, n (%) Mean BMI, kg/m2 (SD) BMI≥30 kg/m2, n (%)
Standardized difference
0.17 --
0.11
SC
AWR
RI PT
Variables
0.16
52 (57.1)
0.02
1
1 (1.1)
1 (1.1)
0
2
10 (11)
9 (9.9)
0.04
3
76 (83.5)
77 (84.6)
0.03
4 (4.4)
4 (4.4)
0
76 (83.5)
76 (83.5)
0
9 (9.9)
9 (9.9)
0
34 (37.4)
32 (35.2)
0.05
11.3 (4.9)
12.1 (6.7)
0.13
Clean/clean contaminated
65 (71.4)
64 (70.3)
0.024
Contaminated/infected
26 (28.6)
27 (29.7)
0.024
M AN U
51 (56)
4
Smoker (%)
AC C
Radiation (%)
EP
≥1 medical comorbidity, n (%)
TE D
ASA Status, n (%)
Ventral hernia width, cm (SD) VHWG Class, n (%)
ACCEPTED MANUSCRIPT 3
Mean follow-up, mo (SD)
Matched data, n=182
58.9 (12.4) 28.1 (12.046.6)
59.1 (12.3) 25.5 (14.446.8) 93 (51.1)
31.4 (7.4) 103 (56.6) 19 (10.4) 153 (84.1) 8 (4.4)
86 (47.3)
AC C
EP
TE D
M AN U
Male, n (%) 149 (47) 2 Mean BMI, kg/m (SD) 31.0 (6.6) 2 BMI≥30 kg/m , n (%) 168 (53) ASA Status, n (%) 2 52 (16.4) 3 244 (77) 4 18 (5.7) >1 medical comorbidity, n (%) 142 (44.8) Smoker, n (%) 23 (7.3) Radiation, n (%) 65 (20.5) Median ventral hernia width, cm (IQR) 10 (7.0-15.0) Median defect surface area, 180 (84.0300.0) cm2 (IQR) VHWG Class, n (%) Clean/clean contaminated 274 (86.4) Contaminated/infected 43 (13.6) VHWG, Ventral Hernia Working Group
SC
Mean age, y (SD)
Unmatched data, n=317
RI PT
eTable 2. Patient Characteristics between the Matched and Unmatched Patients
18 (9.9)
66 (36.3)
11 (8.0-15.0) 207 (89.3350.0)
129 (70.9) 53 (29.1)
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
eFigure 1
AC C
EP
TE D
Reference: