Modified “components separation” technique using bilateral transverse subcostal incisions to access the external oblique muscle and fascia. A, Using a narrow Deaver retractor and a Bovie cautery with an extender, the external oblique muscle and fascia are divided superiorly (above the rib cage) and inferiorly. B, At the caudal aspect of the midline incision, the cut edge of the external oblique muscle and fascia is delivered using manual traction for complete release.
“Components separation” technique with midline approximation of the rectus abdominis muscles. A, No mesh. B, Acellular cadaveric dermis underlay. C, Soft polypropylene mesh underlay.
A 41-year-old man with a history of a perforated appendix treated through a midline incision who later developed an incisional hernia. A, Preoperative oblique view after a hernia repair with polypropylene mesh by another surgeon. B, Preoperative computed tomography scan demonstrating the small bowel herniating to the right of the polypropylene mesh, with wide displacement of the rectus abdominis muscles. C, Six-month postoperative oblique view demonstrates restoration of abdominal wall continuity. D, Postoperative anterior view demonstrates stable midline closure and bilateral transverse subcostal incision scars.
Predictors of hernia recurrence and major and minor complications using logistic regression controlling for mesh type and follow-up duration. Error bars represent 95% confidence intervals. BMI indicates body mass index.
Overall “components separation” procedures based on mesh type. Only the first surgery (N = 200) was included in this series. Subsequent surgeries were performed for hernia recurrences, which were successfully repaired unless otherwise specified.
“Components separation” procedures performed yearly between 1996 and 2007.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Ko JH, Wang EC, Salvay DM, Paul BC, Dumanian GA. Abdominal Wall Reconstruction: Lessons Learned From 200 “Components Separation” Procedures. Arch Surg. 2009;144(11):1047–1055. doi:10.1001/archsurg.2009.192
To determine the efficacy and describe the evolution of the “components separation” technique for abdominal wall repair in 200 consecutive patients.
Retrospective medical record review.
Northwestern Memorial Hospital, Chicago, Illinois.
Two hundred consecutive patients who underwent ventral hernia repair using the components separation technique.
Biological and permanent meshes were used in select patients to augment the repair of the midline fascial closure but were not used as “bridging” materials.
Main Outcome Measures
Hernia recurrence rates and major and minor complication rates for the overall series and for the different techniques.
Primary components separation (n = 158) yielded a 22.8% recurrence rate. Closure of the midline tissues with augmentation of the repair using an acellular cadaveric dermis underlay (n = 18) had a 33.3% recurrence rate requiring a second operation, whereas intra-abdominal soft polypropylene mesh (n = 18) had 0% recurrence (P = .04). Elevated body mass index was a significant risk factor predicting hernia recurrence (P = .003). Contamination (P = .04) and enterocutaneous fistula (P = .02) at the time of surgery were associated with increased major complications, whereas body mass index (P = .01) and diabetes mellitus (P = .04) were associated with increased minor complications.
Large complex hernias can be reliably repaired using the components separation technique despite the presence of open wounds, the need for bowel surgery, and numerous comorbidities. The long-term strength of the hernia repair is not augmented by acellular cadaveric dermis but seems to be improved with soft polypropylene mesh.
Despite the relatively high incidence of incisional hernias, no consensus has been reached on the best method of closure of the abdominal wall. For primary repair, recurrence rates range from 24% to 54%,1-4 with seemingly high recurrence rates after mesh (24%) and suture (43%) repairs.5,6 Although mesh repairs have led to improved recurrence rates overall, synthetic mesh repair is associated with various morbidities.7,8 For situations in which the abdominal wall defect is massive, the wound site is contaminated, bowel surgery is required, or the patient has undergone previous incisional hernia repairs, the autogenous “components separation” (“separation of parts”) procedure seems ideal.9-12 The procedure relies on bilateral release of the external oblique muscle and fascia, thereby allowing medial movement of the rectus muscles toward the midline to achieve an innervated midline closure.
One problem with outcome analyses after hernia repair is the lack of a common starting point for patients. In other words, midline hernias can be of various sizes, and patients differ in age, weight, tissue quality, wounds, and the need for concurrent bowel surgery. In addition, multicenter studies involving many patients introduce the added variables of the abilities and judgment of the surgeons. Because it is almost impossible to have the same starting point for all patients in a series, we grouped the 200 patients in this series together based on a common “middle” point. Specifically, each patient in this study underwent an approximation of the rectus abdominis muscles at the midline for hernia closure using the components separation technique. The senior author (G.A.D.) performed all the repairs in this series. In this retrospective study using a common middle point, we examined whether the midline closure technique or any other baseline patient characteristics affected long-term clinical outcomes, includinghernia recurrence and major and minor complication rates, after incisional hernia repair. To our knowledge, this represents the largest published series to date regarding the components separation technique.
A comprehensive retrospective medical record review was performed of all patients who underwent the components separation technique for midline abdominal wall defects by a single surgeon (G.A.D.) at Northwestern Memorial Hospital, Chicago, Illinois, between August 2, 1996, and July 2, 2007. For all of the patients, the medial aspects of the rectus abdominis muscles were approximated in the midline without any additional releases or fascial turnovers. Two hundred patients (115 men and 85 women) met the inclusion criteria, and 13 patients were excluded owing to mesh being used as a bridging material. Patient characteristics, including age, body mass index (BMI), medical comorbidities, hernia size, operative details, and postoperative results, were examined, and follow-up data were obtained from analysis of the surgeon's office records, the patient's hospital and outpatient electronic medical records, and postoperative abdominal computed tomography (CT) findings. This study was approved by the institutional review board of Northwestern University.
When reconstructing the abdominal wall, the surgeon must solve 2 independent but related problems: how best to repair the abdominal wall and how best to handle the skin. A key element is to perform the abdominal wall surgery while not devascularizing or injuring the skin. Beginning in 1999, several researchers13-16 described techniques to access the semilunar lines for release of the external oblique musculature while maintaining overlying skin blood flow. Our current technique is as follows17:
The abdomen is entered and the abdominal viscera are widely cleared from the undersurface of both the hernia sac and the abdominal wall. The anterior rectus fascia is located and cleared for 4 cm from its free medial edge. To perform the releases of the external oblique muscles, bilateral 6- to 8-cm transverse incisions are made just inferior to the lowest aspect of the rib cage, and the soft tissues over the semilunar line are elevated bluntly. The external oblique muscle is then divided just as it inserts into the anterior rectus fascia, from above the rib cage down to the iliac crest, under direct vision. To complete the release inferiorly to the level of the pubic symphysis, a 4-cm tunnel is created bluntly over the anterior rectus fascia in the suprapubic area from the midline incision and is joined with the soft-tissue dissection created via the lateral incision. The cut edge of the external oblique muscle is confirmed by means of bimanual palpation and is delivered into the midline using manual traction to achieve complete release of the tissues (Figure 1). Blunt dissection is performed to widely separate the external and internal oblique muscles, thereby allowing sliding of the tissue planes. In this series, no other releases were performed. The midline closure of the rectus muscles is performed using 0-polypropylene sutures placed in a figure-of-eight manner. When bringing the fascia together at the midline, there is often redundant fascia or skin that creates a “dog-ear” at the superior and inferior aspects of the closure, and these are addressed using direct excision or imbrication.
In the middle years of the study, human acellular cadaveric dermis (AlloDerm; LifeCell Co, Branchburg, New Jersey) was used on consecutive patients as an intra-abdominal reinforcement of the midline closure. Pieces of cadaveric dermis greater than 2 mm thick and 8 cm in transverse width were sutured together lengthwise to have sufficient material to span the entire midline closure. The 0-polypropylene sutures were placed 4 cm from the medial edge of the rectus abdominis muscles bilaterally so that the sheet of cadaveric dermis spanned the entire midline closure, and the biological mesh was stretched during placement as instructed by the manufacturer. The aforementioned sutures were placed 2 cm apart, with each bite encompassing 5 mm of abdominal wall tissue to avoid potentially compromising the blood supply to the medial rectus abdominis muscles. The medial aspects of the rectus muscles were then approximated in the midline over the cadaveric dermis underlay using figure-of-eight 0-polypropylene sutures so that the biological mesh was totally covered by the rectus muscles.
In the recent years of this study, soft polypropylene mesh (Prolene Soft Mesh and Proceed Surgical Mesh; Ethicon Inc, Somerville, New Jersey) was used as an intra-abdominal reinforcement of the midline closure in clean cases without bowel injury and when the omentum could be situated under much of the mesh. When used, the mesh was fashioned 8 cm in transverse width and was inset with bites taken 4 cm from the medial aspect of the rectus muscles. Analogous to the cadaveric dermis closures, the rectus muscles in these patients were also completely closed over the soft polypropylene mesh underlay (Figure 2).
A minimum of 3 closed-suction drains are inserted, 1 in each lateral tunnel and 1 or 2 in the midline over the closed abdominal wall fascia. In some patients with an elevated BMI, skin handling creates large suprapubic dead spaces. In these patients, and in patients with heavily contaminated wounds, a 3 × 3-cm wound is left open for immediate treatment, with a subatmospheric pressure dressing placed in the operating room. The subatmospheric pressure dressing is continued until hospital discharge, after which the wound is allowed to close by secondary intention. Although not specifically calculated, the mean surgery time for a case without concomitant bowel surgery was generally less than 2½ hours.
Standard CT of the abdomen and pelvis was performed at the discretion of the senior author (G.A.D.) to identify potential intra-abdominal findings, such as ascites and neoplasia. Using digital radiology system measurement tools, the widest separation of the medial aspect of the rectus muscles was recorded for each patient who underwent preoperative CT. In patients lacking preoperative CT, no estimation of hernia size was made because intraoperative assessments are difficult to standardize and allow significant surgeon bias. For each group analyzed, more than 50% of the patients had preoperative CT performed (Figure 3).
We compared the baseline patient characteristics between the overall series and between the various mesh groups by performing an F test for continuous variables and a Fisher exact test for categorical variables. To determine differences in follow-up duration, recurrence rates, time to recurrence, and rates of major and minor complications, we performed an F test for continuous variables and a Fisher exact test for categorical variables. P < .05 was considered statistically significant.
Patient characteristics are summarized in Table 1. Most patients were male (57.5%) and had a BMI greater than 25 (calculated as weight in kilograms divided by height in meters squared) (81%); the mean patient age was 54.2years. One hundred fifty-eight patients (79.0%) had no mesh used during primary components separation repair, and 42 (21.0%) had primary underlay mesh repairs as follows: 6 patients (3.0%) with polypropylene mesh (Prolene; US Surgical Corp, Norwalk, Connecticut), 18 (9.0%) with human acellular cadaveric dermis, and 18 (9.0%) with soft polypropylene mesh.
Most patients in this series underwent primary repair of their abdominal wall defects (n = 109, 54.5%), and 91 patients (45.5%) had recurrent incisional hernias; 58 patients (29.0%) had previously placed mesh that was removed at the time of surgery. Concurrent surgery occurred in 82 patients (41.0%), with various incidences of contamination (n = 73, 36.5%), formal bowel surgery (n = 53, 26.5%), and intraoperative enterotomy (n = 14, 7.0%). One hundred eight patients (54.0%) had preoperative CT demonstrating a hernia defect ranging from 5.4 to 22.4 cm (mean = 12.3 cm) in transverse dimension. Patients with soft polypropylene mesh exhibited significantly smaller hernia defects compared with the other groups (P = .02), and patients with no mesh underlay had the greatest likelihood of having a simultaneous intra-abdominal procedure (46.8%, P < .001), formal bowel surgery (29.7%, P = .02), and contamination (42.4%, P < .001). There were no significant differences in age, sex, BMI, tobacco use, corticosteroid use, and prevalence of diabetes mellitus in patients in the various groups (Table 1).
Recurrence rates and major and minor complication rates were calculated in the overall series of 200 patients and in the different repair subgroups (Table 2). Follow-up ranged from 3 days to 74 months (mean, 10 months). The overall hernia recurrence rate was 21.5% (n = 43). Soft polypropylene demonstrated a significant decrease in hernia recurrence compared with the other groups (P = .04). Although soft polypropylene is the “newest” of the meshes used, there were no significant differences in follow-up duration among the 4 repair groups (P = .20). The mean time to recurrence for the overall series was 14.8 months.
Major complications (n = 48, 24.0%) included hematoma, infection that required incision and drainage, repeated operation for any complication, and any other complications that may be deemed significant (ie, myocardial infarction, pulmonary embolus, and death). Minor complications (n = 38, 19.0%) included cellulitis, seroma that required aspiration, skin sloughing, and wound breakdown. In this study, mesh type had no statistically significant relationship with either major or minor complication rates.
A subset analysis of noncontaminated, or clean, cases was performed to help control for the finding that acellular cadaveric dermis was used more often in contaminated situations (Table 3). Soft polypropylene still had significantly decreased hernia recurrence rates compared with the other subgroups when contaminated patients were excluded (P = .02), and the differences between subgroups in terms of major and minor complications remained not statistically significant.
Logistic regression analysis was performed to predict the risk of hernia recurrence and complications for numerous preoperative risk factors: sex, age, BMI, previous hernia repair by another surgeon, diabetes mellitus, smoking, corticosteroid use, incision type, concurrent surgery, bowel surgery, contamination, presence of an enterocutaneous fistula, intraoperative enterotomy, and hernia width. For each regression model, we also controlled for mesh type and follow-up duration (Figure 4). Elevated BMI demonstrated a significant effect on hernia recurrence (odds ratio [OR] = 1.06, P = .003), with previous hernia repair by another surgeon demonstrating borderline significance (OR = 1.87, P = .08). Hernia width, diabetes mellitus, smoking, and contamination had no effect on hernia recurrence. Contamination (OR = 2.26, P = .04) and the presence of an enterocutaneous fistula (OR = 3.67, P = .02) had a significant effect on major complications, and patients with an increased BMI (OR = 1.06, P = .01) and diabetes mellitus (OR = 2.38, P = .04) demonstrated significantly increased incidences of minor complications.
The flowchart in Figure 5 summarizes the breakdown of hernia repairs for each surgery group, with some patients undergoing multiple hernia repairs for recurrences; however, only the primary components separation procedures are included in this study for data analysis. Most of the aforementioned hernia recurrences were successfully treated by direct reapproximation of the rectus abdominis muscles in the midline, reinforced by a nonbridging intra-abdominal soft polypropylene mesh.
The components separation technique may be an ideal hernia repair for large defects because it weakens or loosens the contracted sides of the abdominal wall to augment the midline repair.18,19 Increased lateral wall compliance may reverse the lateral abdominal wall disuse atrophy and fibrosis seen in animal incisional hernia models.20 A hernia recurrence occurs when the midline repair ruptures before the lateral abdominal wall stretches; therefore, an increase in lateral abdominal wall elasticity and compliance may be significantly protective.
During the 11 study years, 200 consecutive patients underwent a modified components separation procedure for midline abdominal wall reconstruction by the senior author (G.A.D.), representing the largest reported series of this type of repair by a single surgeon at a single institution. The evolution of the procedure has been along 2 “tracts”: how to improve the strength of the midline and how to reduce problems with handling of the skin. For the abdominal wall, despite the lateral releases, the midline repair fails after primary closure of the hernia in 22.5% of cases, independent of whether the repair is performed in a contaminated field. As demonstrated in Table 3, even clean cases, with contaminated patients excluded, had similar hernia recurrence rates compared with the analysis in Table 2. The midline movement of tissue with the components separation technique permits the excision of all scarred and inflamed tissue, and, for this reason, it is likely that hernia recurrences are related more to the chronic forces on the abdominal wall across time than to bacteria at the time of the surgery, possibly explaining the significant risk that elevated BMI poses for hernia recurrence.
This series demonstrates the senior author's (G.A.D.) attempts to augment the strength of the midline closure while outlining his evolution in technique across 11 years (Figure 6). Heavyweight polypropylene mesh was used in the early years of the study but was abandoned due to the stiff feel of the prosthetic material rather than to a specific postoperative finding. In the middle years of the study, literature21,22 had emerged showing that acellular cadaveric dermis did not incite intestinal adhesions and could, thus, be used as a fascial replacement in abdominal wall reconstruction. In 2004, consecutive patients had their midline repair augmented with cadaveric dermis with the idea that even if the dermis was not long-lasting, its presence could shield and protect the repair in the early stages of healing. In fact, just the opposite was found: the hernia recurrence rate with a cadaveric dermis underlay was even higher than that for primary closure. Postulated mechanisms for this included stress shielding of the repair and devascularization of the rectus muscles from the lateral stitches (not found at reexploration). At the time of repeated operation (not included in this study), the cadaveric dermis was often difficult to find, and, when present, large holes in the material itself were often noted. In the absence of improved results using cadaveric dermis as a reinforcement material, its use in abdominal wall repair is not supported. We agree with Lowe23 that cadaveric dermis alone does not provide long-lasting or durable results in abdominal wall reconstruction and should, therefore, be reserved for contaminated wounds, where a prosthetic mesh is best avoided.
Continued dissatisfaction with hernia recurrences after components separation, seen with primary repair and cadaveric dermis, led to the subsequent use of soft midweight polypropylene mesh for augmentation of the strength of the midline closure. At mean follow-up of 13.8 months, no patient who has undergone a components separation procedure using intra-abdominal soft polypropylene mesh in this series has required a revision. There remains a concern regarding the placement of mesh intraperitoneally and the risk of adhesive bowel disease or enterocutaneous fistulae; however, numerous studies24-26 have demonstrated safe placement of polypropylene mesh in direct contact with the bowel. In addition, underlay techniques have been demonstrated to be optimal for incisional hernia repair,27,28 and, in the present series, during implantation of soft polypropylene mesh, efforts were made to interpose omentum between the bowel and the mesh whenever possible. In our experience with mesh removals, most adhesions and mesh-related complications occur in locations where the mesh is folded or has wrinkles. Therefore, in cases where mesh was used, the mesh was placed flat, tight, and without any wrinkles. In cases where soft polypropylene mesh was used, the flexibility of the material also seemed to aid in the placement of a flat mesh. The overlying muscle prevented bowing of the prosthetic material, and the geometry of the closure, located just posterior to the flat rectus muscles, aided in avoiding wrinkles in the mesh. No patient in the soft polypropylene group in this study has developed a bowel obstruction due to adhesive disease, has developed fistulae, or has experienced mesh extrusion. Compared with polypropylene, cadaveric dermis is generally thought to be more “bowel friendly” and to generate fewer adhesions. Although this may be true, the only patients in this series who were reexplored for bowel obstructions were in the cadaveric dermis group, and, in these patients, the bowel was obstructed at sites of recurrent hernias.
In this study, the soft polypropylene mesh group exhibited a statistically significant improvement in hernia recurrence compared with the other groups. Although soft polypropylene has been used only since 2005, its follow-up duration did not statistically differ from that of the other 3 groups. Most of the hernia recurrences (64%) occurred in the first year after surgery. One limitation of this study is the relatively few patients with each of the various mesh types compared with the group with no mesh use. Additional follow-up data will continue to be obtained for the soft polypropylene group, but the 0% recurrence rate in the 18 patients with soft polypropylene mesh in this study is encouraging nonetheless. Although not analyzed in this study, hernia recurrences in all the groups after an initial components separation procedure have been successfully treated with reapproximation of the rectus muscles in the midline using a soft polypropylene mesh underlay.
Overall, the hernia recurrence rate of 21.5% is comparable with that of other published series using this technique.12,15,29-35 Twenty-five of the 91 patients (27.5%) who had a previous repair by another surgeon had hernia recurrences after components separation compared with documented recurrence rates as high as 54% for recalcitrant hernias.2,36-38 In addition, although hernia size is often considered a risk factor for hernia recurrence,38,39 we did not find this to be the case in this series. This finding is based on 108 patients with preoperative CT data. In addition, in each major subgroup, more than 50% of patients underwent CT. We did not find any significant differences in demographic and clinical characteristics between those who had hernia size data and those who did not. In this study, elevated BMI demonstrated a statistically significant relationship with hernia recurrence, whereas a previous hernia repair by another surgeon showed borderline significance. In the 200 patients in this series, mesh type had no statistically significant impact on either major or minor complication rates.
A major lesson learned over the years is that handling of the skin is important, especially in patients with an elevated BMI. Wide undermining of the skin to release the oblique musculature disrupts the perforator blood flow to the midline abdominal skin, thereby contributing to wound complications in these patients. Modifications proposed by Maas et al,13 and as seen in laparoscopic components separation techniques, aim to resolve this issue by better maintaining and maximizing blood flow to the midline.14,16,40,41 Building on the aforementioned modifications, in 2002, the senior author (G.A.D.) adapted his surgical technique to perform the external oblique releases through bilateral transverse subcostal incisions to avoid wide undermining, an evolution of the technique of “periumbilical perforator preservation.”15 Releases take only 15 to 20 minutes to perform and avoid the setup of endoscopic equipment.23 Another skin-handling technique is to perform a panniculectomy at the time of components separation for morbidly obese patients with infraumbilical hernias (repairs of hernias that extend above the umbilicus are generally performed using vertical midline incisions). A third improvement for skin handling is the use of short-term subatmospheric pressure dressings as immediate postoperative dressings in patients with an elevated BMI, gross contamination, and large suprapubic dead spaces. This “semi-closed” technique for skin management has led to decreased seroma formation and infections in addition to allowing access to the midline fascial closure in the immediate postoperative period.
The components separation technique is an effective treatment for massive midline hernia defects. Complication rates are manageable, despite the comorbidities seen in this patient population, and initial hernia size did not significantly affect long-term outcomes. We believe that attention to detail when handling the skin is an important lesson learned during the past decade, and we present a modified operative technique that aims to minimize soft-tissue complications. Finally, these results indicate that acellular cadaveric dermis may not be an effective material for long-term incisional hernia repair reinforcement. Soft midweight polypropylene mesh seems to provide superior long-term strength and durability at this stage and is our current material of choice for use as a reinforcement of the midline closure of the rectus abdominis muscles. Further evaluation of both materials with improved follow-up is necessary and ongoing.
Correspondence: Gregory A. Dumanian, MD, 675 N St Clair, Ste 19-250, Chicago, IL 60611 (email@example.com).
Accepted for Publication: October 22, 2008.
Author Contributions: Dr Dumanian had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ko, Paul, and Dumanian. Acquisition of data: Ko, Paul, and Dumanian. Analysis and interpretation of data: Ko, Wang, Salvay, Paul, and Dumanian. Drafting of the manuscript: Ko, Wang, and Dumanian. Critical revision of the manuscript for important intellectual content: Ko, Wang, Salvay, Paul, and Dumanian. Statistical analysis: Ko, Wang, Salvay, and Dumanian. Obtained funding: Dumanian. Administrative, technical, and material support: Ko, Salvay, Paul, and Dumanian. Study supervision: Ko and Dumanian.
Financial Disclosure: None reported.
Previous Presentation: This study was presented at the American College of Surgeons 94th Annual Clinical Congress; October 15, 2008; San Francisco, California.