Incidence rates of the major long-term complications according to the type of prosthetic material used. Mersilene (Ethicon Inc, Somerville, NJ) was significantly different for fistula formation (P=.007) and for infection (P=.04). See footnote to Table 2 for names and locations of other manufacturers.
Leber GE, Garb JL, Alexander AI, Reed WP. Long-term Complications Associated With Prosthetic Repair of Incisional Hernias. Arch Surg. 1998;133(4):378-382. doi:10.1001/archsurg.133.4.378
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1998
To determine whether the type of prosthetic material and technique of placement influenced long-term complications after repair of incisional hernias.
Retrospective cohort analytic study.
Two hundred patients undergoing open repair of abdominal incisional hernias with prosthetic material between 1985 and 1994.
Four types of prosthetic material were used and placed either as an onlay, underlay, sandwich, or finger interdigitation technique. The materials were monofilamented polypropylene mesh (Marlex, Davol Inc, Cranston, RI), double-filamented mesh (Prolene, Ethicon Inc, Somerville, NJ), expanded polytetrafluroethylene patch (Gore-Tex, WL Gore & Associates, Phoenix, Ariz) or multifilamented polyester mesh (Mersilene, Ethicon Inc).
Main Outcome Measures
The incidence of recurrence and complications such as enterocutaneous fistula, bowel obstruction, and infection with each type of material and technique of repair were compared with univariate and multivariate analysis.
On univariate analysis, multifilamented polyester mesh had a significantly higher mean number of complications per patient (4.7 vs 1.4-2.3; P<.002), a higher incidence of fistula formation (16% vs 0%-2%; P<.001), a greater number of infections (16% vs 0%-6%; P<.05), and more recurrent hernias (34% vs 10%-14%; P<.05) than the other materials used. The additional mean length of stay to treat complications was also significantly longer (30 vs 3-7 days; P<.001) when polyester mesh was used. The deleterious effect of polyester mesh on long-term complications was confirmed on multiple logistic regression (P=.002). The technique of placement had no influence on outcome.
Polyester mesh should no longer be used for incisional hernia repair.
INCISIONAL herniation is a complication that all abdominal surgeons encounter. Of the more than 2 million abdominal procedures performed annually in the United States,1 roughly 2% to 11% of patients will develop an incisional hernia.2- 5 After primary repair, recurrent herniation is reported to occur in 30% to 50% of cases.6,7 Creating a tension-free repair with a prosthetic material has lowered this reported recurrence rate to between 0% and 10%.7,8
Prosthetic material was introduced with steel mesh in the 1940s. Usher9 was the first to use plastic prosthetics, in 1958. These materials showed distinct advantages over steel mesh in their ease of use, pliability, and lack of disintegration with age. Monofilamented polypropylene mesh (Marlex, Davol Inc, Cranston, RI), a refinement of plastic prosthetics, became available in 1962.10 Marlex mesh has since become the most widely used prosthetic material for repair of incisional hernias.2,11 Several other prosthetic materials have since been developed and used for incisional hernia repair. Some of these include multifilamented polyester mesh (Mersilene, Ethicon Inc, Somerville, NJ), double-filamented polypropylene mesh (Prolene, Ethicon Inc), and expanded polytetrafluroethylene patch (Gore-Tex, WL Gore & Associates, Phoenix, Ariz).
The early reports by Usher showed no harm resulting from placing the mesh in direct contact with omentum or bowel and claimed a mechanical advantage to placing mesh in the subfascial position.9 In 1981, Kaufman12 first reported enterocutaneous fistula formation as a late complication of intraperitoneal placement of mesh and advised against this technique.
We have experienced many complications with prosthetic incisional hernia repair (PIHR), such as wound infection, chronically draining sinuses, enterocutaneous fistula, small-bowel obstruction, malnutrition, and recurrent herniation. There are few series in the literature reporting long-term complications after PIHR.11- 14 The purpose of this study was to estimate the incidence of long-term complications and to evaluate the factors contributing to these complications. In particular, we were interested in the type of prosthetic material used, the anatomical placement of the material, and measures taken to protect the bowel.
We conducted a retrospective review of incisional hernias repaired with prosthetic material at Baystate Medical Center, Springfield, Mass, between October 1985 and April 1994. Operative logs were used to identify all patients who had undergone PIHR during this interval. Complications were identified by reviewing all of the medical records for subsequent admissions and procedures performed for each patient after the initial PIHR until May 1996. We identified 227 cases, of which 200 medical records contained complete information. Twenty-seven patients had incomplete medical records and were therefore excluded from the analysis. Length of follow-up ranged from 2.7 to 10.6 years, with a mean of 6.7 years.
The type of prosthetic material used, its anatomical placement, and measures used to protect the bowel were noted for all patients. The method of PIHR was classified as either onlay, underlay, sandwich, or finger interdigitation technique. Mesh was placed on top of the anterior rectus fascia in the onlay technique and underneath the posterior rectus fascia in the underlay technique. The sandwich technique incorporated both the onlay and underlay techniques, with mesh placed on top of the anterior rectus fascia and underneath the posterior rectus fascia. Finger interdigitation is a variation of the underlay technique in which the mesh is cut in a starburst pattern and finger projections are brought through both layers of fascia and then sutured to the anterior rectus fascia. Bowel coverage was classified according to the degree to which the hernia sac or omentum was used between the mesh and bowel, and by the completeness of fascial closure.
The data collected also included those factors that could influence the effectiveness of the repair. These included patient demographics, predisposing comorbid factors, location of the incision and diameter of the defect, preoperative diagnosis, previous attempts at primary repair, operative classification (clean vs contaminated), use of drains, prophylactic antibiotics, type of suture material used, and short-term complications.
The Pearson χ2 technique was used to compare different prosthetic materials and techniques of repair on the incidence of long-term complications. Adjusted standardized deviates were used to isolate sources of variation among groups.15 Factors associated with the occurrence of major complications or fistulae were tested using multiple logistic regression.16 A forward stepwise procedure was used. A maximum likelihood procedure was employed to calculate the regression coefficients, and significance of the individual factors in the regression model was determined by the likelihood ratio criterion.17
Differences in the mean number of long-term complications or in the total length of stay resulting from complications, according to type of mesh or operative technique, were tested using the analysis of variance with the Tukey multiple comparisons procedure for testing pairwise differences between groups.18
The population of 200 patients was predominantly female (61%), with a mean age of 59.9 years and mean weight of 86 kg at the time of their PIHR. Of the predisposing comorbid factors, smoking was present in 33% of patients and diabetes mellitus in 21%. Other factors included previous wound infections (13%), chronic lung disease (18%), and steroid use (2%).
Hernia characteristics are listed in Table 1. Mesh repair was most commonly used for incisional hernias developing after obstetric and gynecological procedures, representing 24% of the study group (specifically total abdominal hysterectomies and cesarean sections). Other common primary operations included lower gastrointestinal tract procedures, biliary procedures, and previous ventral hernia repair. The most common incision requiring mesh repair was in the upper midline of the abdomen. Paramedian incisions were also common causes of hernia formation. There were only a few transverse incisional hernias that required mesh repair. The mean fascial defect was 9.8 cm in the largest diameter.
The type of prosthetic material used, technique of repair, and attempt at bowel protection are listed in Table 2. Knitted monofilament polypropylene mesh (Marlex) was by far the most commonly used material. The onlay repair was the most frequently used technique. Bowel protection from the mesh with use of the imbricated hernia sac was infrequent and complete omental coverage of bowel was clearly accomplished in only 24% of patients.
Complications relating to the PIHR are listed in Table 3. Early complications occurred within 1 month of the PIHR and long-term complications occurred any time after the first month. Early complications occurred in 36 patients (18%). Postoperative ileus and cellulitis were the most common. The overall incidence of long-term complications was 27%. Most of these were related to a high recurrence rate of 16.8%. Each long-term complication resulted in at least 1 hospital admission with an average length of stay of 3 days (95% confidence interval [CI], 1.5-4.5 days). The median time to long-term complication was 0.5 years for infection, 1.5 years for small-bowel obstruction, 1.7 years for recurrence, and 3.3 years for enterocutaneous fistula.
Figure 1 shows the incidence rates of the major long-term complications according to the type of prosthetic material used. Mersilene was associated with the highest rates for all types of major complications. Mersilene mesh was used in 5 of the 7 hernias complicated by an enterocutaneous fistula, representing a 15.6% incidence of fistula, vs 1.7% for Marlex and 0% for Gore-Tex and Prolene. On univariate analysis, there was a significant difference among groups in the incidence of fistula (χ2=17.0, df=3; P=.007) and infection (χ2=8.3, df=3; P=.04). Analysis of standardized deviates revealed that Mersilene accounted for the significant differences in both fistulae (z=4.1) and infection (z=2.5).
The technique of repair was not significantly related to long-term complications. The underlay, finger, and sandwich techniques are all variations of subfascial placement of mesh. Although the incidence of fistulae was higher in the subfascial group (5.2% vs 2.6% for the onlay group), the power to detect a statistical significance between these groups was low (<20%). The theoretical mechanical advantage of subfascial mesh in reducing the recurrence rate was not demonstrated in our study. The recurrence rate with subfascial placement of mesh was actually higher (19.5%) than for the onlay technique (14.8%).
The incidence of enterocutaneous fistula related to attempted bowel protection is presented in Table 4. The excision of the hernia sac, lack of omental interposition, and the presence of a fascial gap all had a higher incidence of enterocutaneous fistula formation. These differences were not statistically significant. However, the power to exclude a significant difference between these groups was also less than 20%.
Results of the multiple logistic regression analysis for long-term complications are listed in Table 5. After adjusting for other significant factors, the risk of developing a long-term complication was almost 4 times greater in patients with a preoperative diagnosis of a recurrent hernia, 3 times as great in hernias repaired using Mersilene mesh, and over twice as likely with smokers and with hernias located in the upper abdomen. When looking at specific factors in the development of enterocutaneous fistula, the use of Mersilene mesh, a preoperative diagnosis of partial small-bowel obstruction or incarcerated hernia, upper abdominal location of the hernia, and previous wound infection were all found to be significant (P<.05). The most significant factor in the development of enterocutaneous fistulae was the use of Mersilene mesh (P=.002). Owing to the small number of enterocutaneous fistulae in our series, the estimates for the adjusted relative risks and regression coefficients for significant factors associated with this complication had wide 95% CIs, rendering them of little use.
Mersilene had a significantly higher number of complications and greater length of stay than any other prosthetic material used. In patients who developed complications, the mean number of complications was 1.4 for Gore-Tex, 1.8 for Prolene, 2.3 for Marlex, and 4.7 for Mersilene. The use of Mersilene resulted in significantly more complications than Marlex (P<.05) and Gore-Tex (P<.05). There was a significant difference among groups in the mean length of stay from complications (F=6.6, df=3.49; P<.001). Length of stay was 29.8 days for Mersilene vs 7.4 days for Marlex (P<.01), 2.8 days for Gore-Tex (P<.01), and 3.0 days for Prolene (P<0.05)
The use of prosthetic materials for incisional hernia repair has significantly lowered the reported recurrence rates. However, recurrence remains a problem and there are many potential long-term complications directly related to PIHR. The goal of a successful repair is to minimize the recurrence rate with the lowest possible incidence of complications. To accomplish this, the surgeon should make every effort to keep bowel out of contact with the mesh. Although not statistically significant in our study, excision of the hernia sac, lack of omental interposition, and the presence of a fascial gap all had a higher incidence of enterocutaneous fistula. Better documentation may prove these differences to be significant. Extrafascial techniques, such as described by Lewis,13 that use the imbricated hernia sac for bowel protection have been shown to be effective methods of repair and may help prevent the many long-term complications we are now seeing with PIHR.
Usher's claims of a mechanical advantage to subfascial mesh9,19 were not supported in our study. There was actually a higher incidence of recurrent herniation in this group. Additionally, subfascial mesh had a higher incidence of enterocutaneous fistula formation, as predicted by Kaufman.12 These techniques require considerably more dissection and risk intra-abdominal or delayed bowel injury. The true incidence of enterocutaneous fistula formation is likely underestimated, because there are few long-term studies and the average time for fistulae to develop was 3.3 years in our study.
Studies have praised Mersilene for its supple and elastic nature, its grainy texture for gripping tissue, and its rapid fibroblastic response for ensuring fixation to surrounding tissues.20 These attributes may actually contribute to the long-term complications we have seen with its use in PIHR. In addition, multifilamented braided mesh such as polyester mesh (Mersilene) may exclude macrophages but not bacteria, resulting in infection, foreign-body granuloma, and sinus tract formation.21 Our study clearly shows that the incidence of complications from Mersilene is markedly higher than for Marlex, Gore-Tex, and Prolene. There is no advantage to its use (as seen by its higher recurrence rate), and it has an unacceptably high incidence of infection, small-bowel obstruction, and enterocutaneous fistula formation. Because these complications can be devastating to the patient and lead to significant additional hospital days for their management, we recommend discontinuing the use of Mersilene mesh in PIHR.
James C. Hebert, MD, Burlington, Vt: Dr Leber and his associates have embarked on a difficult but important task. It is time that we again look at appropriate ways to prevent and treat incisional hernias. Dr Leber and associates have identified that the risk of significant long-term complications, including recurrence, was 27%, with a recurrence rate of almost 17%. The use of Mersilene mesh seems to represent a significant risk for developing fistulae compared with the other types of mesh, although the anatomical location of the patch does not seem to matter in this series.
The study would imply that transverse incisions are at a lower risk for herniation. Is there any information in your series regarding the relative rates of incisional type?
Was obesity examined as a risk factor? It seems to represent a significant risk factor at least in our hands and as reported by others.
Do the authors have any more detail regarding the operative technique to try to better understand why the complications occur? Particularly for recurrent hernias, was there evidence that the abdominal wall was thoroughly explored looking for fenestrations other than the original hernia sac?
Do the authors have any data regarding how many patients developed attenuation of the abdominal wall musculature and chronic pain following repair, and could this be related to the technique of the repair?
Finally, I would like to hear the authors' recommendations for preventing incisional hernias. What do they feel is the best technique for closing the abdomen? Do they recommend that we should use transverse incisions whenever possible? I believe that paying more attention to the way we close abdomens and trying to prevent hernia complications needs more attention and more study by all of us.
James E. Barone, MD, Stamford, Conn: I have 2 questions. Was there any relationship between the surgeon and the recurrence after incisional hernia repair? You didn't say anything about size vs recurrence. I would guess that people with bigger hernias would have more of a problem with recurrence. Would it be possible that Mersilene was just chosen more frequently in that group?
David W. Butsch, MD, Barre, Vt: I would like to ask the authors why they chose the type of prosthetic material that they did.
Dr Leber: Starting with Dr Hebert's questions, the location of the incision was categorized as upper or lower abdomen and either midline, paramedian, or transverse. I have reported the rates of complications for each of these incisions. Given that this study was retrospective, there was no control for what type of incisional hernias required mesh for repair. It just so happened that there were fewer transverse incisional hernias repaired with mesh in our study. Transverse incisions have been shown in other studies to have a lower incidence of herniation, and this may be the reason why fewer of them were repaired with mesh in our study.
The mean weight for patients in our study was 191 lbs (86 kg). Obesity has been shown to be a significant factor for the development of recurrent incisional herniation. However, in our study, obesity was not shown to be a significant independent risk factor in the overall occurrence of long-term complications or rate of enterocutaneous fistula formation.
Abdominal wall attenuation and chronic pain following repair were inconsistently reported; therefore, we did not include them in our analysis. The best method of prevention of incisional herniation may be to perform transverse incisions when possible. Transverse incisions run parallel to most of the muscle and aponeurotic fibers of the abdominal wall. Vertical incisions run perpendicular to these natural lines of tension and, as a result, wound edges are distracted by abdominal wall tension. This is why transverse incisions have the lowest rate of incisional herniation.
Bowel protection is essential in an attempt to avoid long-term complications after prosthetic repair of incisional hernias. Minimizing dissection and attempting to perform a fascial onlay when possible may help accomplish this. The hernia sac should be imbricated when feasible and used as a further barrier between bowel and mesh.
In response to Dr Barone's questions: There was no correlation between the different surgeons and long-term complication rate. Also, the diameter of the fascial defect was looked at closely in our analysis. The size of the hernia defect was similar in the onlay and subfascial methods of repair. We also found no statistical difference between the types of prosthetic material used and the size of the hernia defect.
Finally, with regard to Dr Butsch's question: Individual surgeon preference dictated the type of prosthetic material used for incisional hernia repair. Given the retrospective nature of this study, these choices were made long before the study was undertaken.
Presented at the 78th Annual Meeting of the New England Surgical Society, Bolton Landing, NY, September 19, 1997.
Reprints: William P. Reed, MD, Department of Surgery, Baystate Medical Center, Springfield, MA 01199.