Vaginal mesh erosion after abdominal sacral colpopexy

Transcription

1 Vaginal mesh erosion after abdominal sacral colpopexy Anthony G. Visco, MD, Alison C. Weidner, MD, Matthew D. Barber, MD, Evan R. Myers, MD, MPH, Geoffrey W. Cundiff, MD, Richard C. Bump, MD, and W. Allen Addison, MD Durham, North Carolina OBJECTIVE: Our goal was to compare the prevalence of vaginal mesh erosion between abdominal sacral colpopexy and various sacral colpoperineopexy procedures. STUDY DESIGN: We undertook a retrospective analysis of all sacral colpopexies and colpoperineopexies performed between March 1, 1992, and February 28, The patients were divided into the following 4 groups: abdominal sacral colpopexy, abdominal sacral colpoperineopexy, and 2 combined vaginal and abdominal colpoperineopexy groups, one with vaginal suture passage and the other with vaginal mesh placement. Survival analysis and Cox proportional hazards models were developed to examine erosion rates and time to erosion between groups. RESULTS: A total of 273 abdominal sacral vault suspensions were performed with the use of permanent synthetic mesh. There were 155 abdominal sacral colpopexies and 88 abdominal sacral colpoperineopexies. Among the 30 combined abdominal-vaginal procedures, 25 had sutures attached to the perineal body and brought into the abdominal field and 5 had mesh placed vaginally and brought into the abdominal field. Overall, mesh erosion was observed in 5.5% (15/273). The prevalence of mesh erosion was 3.2% (5/155) in the abdominal sacral colpopexy group and 4.5% (5/88) in the abdominal sacral colpoperineopexy group (P not significant). The rates of erosion when sutures or mesh was placed vaginally were 16% (4/25) and 40% (2/5), respectively, and were significantly increased in comparison with the rates for abdominal sacral colpopexy (hazard ratio, 5.4; 95% confidence interval, ; P =.005; vs hazard ratio, 19.7; 95% confidence interval, ; P <.001). These variables retained their significance after we controlled for other independent variables, including age, concomitant hysterectomy, concomitant posterior repair, and estrogen status. The median time to mesh erosion was 15.6 months for abdominal sacral colpopexy, 12.4 months for abdominal sacral colpoperineopexy, 9.0 months in the suture-only group (P <.005), and 4.1 months in the vaginal mesh group (P <.0001). CONCLUSIONS: The rate of mesh erosion is higher and the time to mesh erosion is shorter with combined vaginal-abdominal sacral colpoperineopexy with vaginal suture and vaginal mesh placement in comparison with abdominal sacral colpopexy. ( 2001;184: ) Key words: Mesh erosion, colpopexy, colpoperineopexy, vaginal vault prolapse The performance of abdominal sacral colpopexy for the surgical treatment of posthysterectomy vaginal vault prolapse or complete uterine prolapse is well established. Suspension of the prolapsed vaginal vault to the anterior surface of the sacrum with graft material was first reported by Lane. 1 In 1985, Addison et al 2 reported the results of the first 56 abdominal sacral colpopexy procedures in the ongoing series of >500 abdominal vault suspensions performed at our institution. The goal of our abdominal vault suspension is to correct vaginal vault prolapse by re-establishing continuity of From the Department of Obstetrics and Gynecology, Duke University Medical Center. Received for publication January 21, 2000; revised March 23, 2000; accepted June 21, Reprint requests: Anthony G. Visco, MD, Campus Box 7570, Department of Obstetrics and Gynecology, University of North Carolina, Chapel Hill, NC anthony_visco@med.unc.edu. Copyright 2001 by Mosby, Inc /2001 $ /1/ doi: /mob the posterior vaginal endopelvic fascia from the perineal body to the vaginal apex with suspension to the anterior longitudinal ligament of the sacrum. If an isolated apical or superior defect in vault support exists, this correction can be accomplished with the abdominal sacral colpopexy as it was originally described, by placement of synthetic mesh from the apex of the vaginal vault to the sacrum. When there is significant perineal descent, which is usually concomitant with an identifiable posterior fascial defect that coexists with vault prolapse, elevation of the apex only may actually accentuate or worsen a posterior fascial defect. With these findings we attempt to avoid worsening the defect by extending the mesh all the way down the posterior vaginal wall to the perineal body or attaching the mesh to the portion of posterior vaginal fascia that is contiguous with the perineal body. Our group described this modification as the abdominal sacral colpoperineopexy. 3 In patients with distal detachment of the posterior fascia from the perineal body or in cases in which it appears 297

2 298 Visco et al February 2001 that the posterior vaginal fascia is severely attenuated or absent, we have resorted to a combined abdominal and vaginal colpoperineopexy. To accomplish this, a posterior vaginal incision is made and then the perineum is suspended in one of two ways. Either permanent sutures are attached directly to the perineal body and brought into the abdominal operative field through the cul-de-sac or the synthetic mesh is sutured directly to the perineal body and brought through the cul-de-sac. The remainder of the procedure is performed abdominally. Abdominal vault suspensions in which synthetic suspensory materials are used have reliably good cure rates. 4-7 However, the use of synthetic materials to suspend any part of the vagina to the sacrum has the potential complication of mesh erosion into the vagina. Such erosion has been managed successfully from a vaginal approach by removing the exposed mesh or sinus tract and undermining the edges of the vagina, which are then oversewn in multiple layers. 8 Laparotomy with complete removal of the mesh may be required in recalcitrant cases or if the mesh becomes infected. Such procedures have been associated with significant morbidity, including severe hemorrhage. 9 Timmons and Addison 8 reported a mesh erosion rate of 3.3% in their series of 375 colpopexy procedures performed over a 23-year period from 1972 to Snyder and Krantz 6 reported a similar erosion rate of 2.7% (4/147) in patients who had undergone a sacral colpopexy. Kohli et al 5 recently reported a mesh erosion rate of 8.8% after abdominal sacral colpopexy with the use of synthetic mesh. The time to mesh erosion has been quite variable in studies that reported this outcome measure. Most seem to occur in the first 2 years; however, some have been reported many years after mesh placement. Timmons and Addison 8 reported that the time to the first symptom of vaginal mesh erosion varied from 6 weeks to 6 years with a mean of 21 months. Kohli et al 5 reported a mean time to recognized mesh erosion of 14.0 ± 7.7 months (range, 4-24 months) with an average length of follow-up of 19.9 months (range, months). The rates of mesh erosion associated with modifications of techniques intended to address posterior vaginal fascial defects and perineal descent are unknown. The primary objective of this study was to compare the rate of vaginal mesh erosion and the time to erosion of the various sacral colpoperineopexies with the previously described techniques of abdominal sacral colpopexy in a large series with long-term follow-up. The secondary objective was to identify risk factors for mesh erosion. Material and methods A retrospective analysis of all abdominal sacral colpopexies, abdominal sacral colpoperineopexies, and combined abdominal-vaginal sacral colpoperineopexies performed at our institution between March 1, 1992, and February 28, 1999, was performed. This project was granted an exemption by the institutional review board at our institution. The study follow-up period ended on June 30, All the abdominal vault suspensions were performed at Duke University Medical Center by members of the Division of Gynecologic Specialties. The surgical cases were identified from our divisional database and confirmed with the operating room database. We defined abdominal sacral colpopexy as a vault suspension in which the synthetic mesh is attached to the upper third of the vagina and suspended to the sacrum. We defined abdominal sacral colpoperineopexy as a procedure that attached the mesh either directly to the perineal body or to the rectovaginal fascia contiguous with the perineal body, as has been previously described by our group 3 (Fig 1). The two combined abdominal-vaginal colpoperineopexy procedures included a vaginal dissection of the posterior vaginal wall. With one technique permanent sutures were attached to the perineal body and brought into the abdominal operative field through the posterior cul-desac with a long needle or ring carrier. The mesh was then attached to these sutures transabdominally without the mesh entering the vaginal operative field. With the other technique the mesh was placed and attached vaginally directly to the perineal body with permanent sutures and brought into the abdominal operative field for attachment to the sacrum. Before operation the patients gave a detailed history and underwent physical examination and urodynamic evaluation when indicated. All patients received preoperative intravenous antibiotics. Charts were reviewed for demographic data, date of last follow-up, and occurrence of any mesh erosion. Because all our mesh erosions are managed in the operating room, we queried the operating room database to confirm that we had identified all cases of vaginal mesh erosion. We examined the following independent variables: age, weight, parity, estrogen status, abdominal-only procedure, combined abdominalvaginal procedure, method of combined procedure (passage of sutures or mesh), concomitant hysterectomy, and concomitant posterior repair. We analyzed categoric data with the χ 2 or Fisher exact test where appropriate. Cox proportional hazards models were developed to examine the association of the various independent variables with the probability and time to erosion between groups. Censoring of patients was done at the time of their last physical examination if they were erosion free. An event was defined as an erosion diagnosed by pelvic examination by a member of our division. A Kaplan-Meier survival curve was generated to compare (log-rank) overall differences in the rate of erosion and the time to erosion among the various colpopexy groups (Fig 2). Statistical analysis was performed with SAS (SAS Institute, Cary, NC) statistical software, version 6.12.

3 Volume 184, Number 3 Visco et al 299 Fig 1. Abdominal sacral colpoperineopexy with mesh attached to posterior vaginal fascia contiguous with perineal body. Results A total of 276 abdominal sacral vault suspensions with permanent synthetic mesh were performed over the 7- year study period. Three patients were excluded because they received all of their postoperative care from the referring physician, leaving a study population of 273 patients. Of these, 155 underwent an abdominal sacral colpopexy, 88 underwent an abdominal-only sacral colpoperineopexy, and 30 underwent combined abdominal posterior vaginal procedures. In 25 of these, the procedures were performed by vaginal attachment of sutures to the perineal body; these sutures were then brought into the abdominal operative field through the cul-de-sac with a long needle or ring carrier. Attaching the mesh to these sutures prevented exposure of the mesh to the vaginal operative field. In the other 5 cases the mesh was attached vaginally and passed into the abdominal surgical field. Mersilene (Ethicon, Somerville, NJ) mesh was used in all except 4 patients, who received Gortex (W.L. Gore, Flagstaff, Ariz) mesh. We used various monofilament and braided permanent suture to secure the mesh to the vagina and sacrum. The mean age of the patients was 60.6 ± 10.6 years (mean, years), the mean weight was 71 ± 12.4 kg, and the median parity was 2 (range, 0-13). Of the total patients, 85% were either premenopausal or receiving estrogen replacement therapy, 45 (17%) underwent a concomitant hysterectomy, and 84 (31%) had a concomitant posterior repair. The estimated blood loss (mean ± SE) was 460 ± 390 ml. Overall, mesh erosion was observed in 15 of the 273 patients (5.5%). The rate of mesh erosion was 3.2% (5/155) in the traditional abdominal sacral colpopexy group. This group represented our reference group against which all other vault suspensions were compared. The rate of mesh erosion was not significantly higher in the abdominal-only sacral colpoperineopexy group, 4.5% (4/88), compared with the group undegoing traditional abdominal sacral colpopexy. In the combined groups the rate of mesh erosion was 16% (4/25) with suture passage and 40% (2/5) with vaginal mesh placement, both of which were significantly higher than the rate in the traditional abdominal sacral colpopexy group (P <.005 and P <.001, respectively). A difference in the time to erosion was also observed among the various vault suspensions. The overall mean follow-up time was 12.3 ± 16.7 months, and the median length of follow-up was 5.8 months (range, 1-87 months). The median length of follow-up was 6.5 months (range, 1-87 months) for abdominal sacral colpopexy, 4.9 months (range, 1-45 months) for abdominal sacral colpoperineopexy, 6.1 months (range, 1-28 months) for vaginally passed sutures, and 6.6 months (range, 2-11 months) for vaginally placed mesh. The median number of months to appearance of mesh erosion was 15.6 (range, 2-33 months) in the abdominal-only sacral colpopexy group, 12.4 months (range, 3-40 months) in the abdominal-only sacral colpoperineopexy group, 9.0 months (range, 5-15 months) in the suture-only group, and 4.1 months (range, 2-7 months) in the vaginal mesh group.

4 300 Visco et al February 2001 Fig 2. Kaplan-Meier survival curve showing time to erosion with various abdominal vault suspensions. There was no significant difference in time to erosion between the group undergoing abdominal sacral colpopexy (dashed line) (reference group) and the group undergoing abdominal sacral colpoperineopexy (solid line) but significantly shorter time to erosion with vaginal suture (thin line with hatch marks) and vaginal mesh (thick line with hatch marks). Abdominal-only sacral colpoperineopexy was not associated with a significantly increased risk of vaginal mesh erosion in comparison with traditional abdominal sacral colpopexy (hazards ratio, 1.2; 95% confidence interval, ; P =.75). Combination abdominal-vaginal sacral colpoperineopexy with passage of sutures was associated with a significantly increased risk of vaginal mesh erosion in comparison with traditional abdominal-only abdominal sacral colpopexy (hazards ratio, 5.4; 95% confidence interval, ; P =.005). Vaginal placement of mesh was also found to be significantly associated with vaginal mesh erosion (hazards ratio, 19.7; 95% confidence interval, ; P <.001). These variables retained their significance after we controlled for other independent variables, including age, weight, concomitant hysterectomy, concomitant posterior repair, and estrogen status (Fig 2). The variables age, weight, parity, estrogen status, concomitant hysterectomy, and posterior repair were examined; however, none was found to be independently associated with mesh erosion. Because various types of permanent suture material were used to secure the mesh to the vagina, we divided the colpopexies into 2 groups those that used monofilament suture and those that used braided suture. Mesh erosion was independent of type of suture used. The location of the mesh erosion varied with the type of vault suspension performed. All 5 mesh erosions in the traditional abdominal sacral colpopexy group occurred at the vaginal apex. Among the 4 erosions in the abdominal-only sacral colpoperineopexy group, 3 occurred at the apex and 1 was along the distal posterior vaginal wall. In the combined colpoperineopexy with vaginal sutures procedures, 2 erosions occurred at the apex and 2 occurred distally. Both of the mesh erosions in the colpoperineopexy procedure with vaginal mesh placement occurred distally. Of the 15 patients who experienced vaginal mesh erosion, 12 were treated successfully with a single transvaginal outpatient excision, 2 required 2 vaginal procedures, and 1, who has previously been described, 9 required 2 transvaginal excisions, a transabdominal resection with profuse hemorrhage, and another transvaginal excision. All distal mesh erosions were successfully treated with a single transvaginal outpatient excision. Comment Abdominal sacral colpopexy that uses synthetic mesh for the surgical treatment of uterine or vaginal vault prolapse is highly successful. However, mesh erosion into the vagina is a recognized complication. The abdominal sacral colpopexy procedure has evolved over the last 30 years during which surgical repairs have been progressively directed toward site-specific defects in pelvic organ support. Abdominal sacral colpoperineopexy described by Cundiff et al 3 was designed to address the posterior vaginal defects and perineal descent that often coexist with vault prolapse. We have recently begun using a combined abdominal and vaginal approach in patients with vaginal vault prolapse when continuity of the posterior endopelvic fascia and mesh could not be achieved abdominally. In this study abdominal sacral colpoperineopexy performed entirely through the abdomen was not associated with an increased rate of mesh erosion in comparison with abdominal sacral colpopexy. The prevalence of vaginal mesh erosion with the newer combined abdominalvaginal sacral colpoperineopexy with either suture pas-

5 Volume 184, Number 3 Visco et al 301 sage or placement of vaginal mesh was significantly greater than the baseline rate associated with the traditional abdominal sacral colpopexy or colpoperineopexy. This has caused us to discontinue the practice of attaching vaginal mesh directly to the perineal body. We have further limited our use of the combined vaginal and abdominal sacral colpoperineopexy with suture passage to patients in whom the distal posterior vaginal fascia is absent or so severely attenuated that any other method of engagement is believed to be suboptimal. We also now counsel our patients that a combined procedure may be associated with a higher rate of mesh erosion. Kohli et al 5 reported no statistically significant difference in mesh erosion in the 47 patients in whom Marlex (C.R. Bard, Cranston, RI) mesh (11%) was used compared with the 10 patients in whom Mersilene mesh (20%) was placed. Their limited sample size and power precluded any meaningful conclusions regarding the risks of mesh erosion attributable to any particular type of synthetic material. Synthetic mesh materials remain the most commonly used suspensory bridge between the vagina and sacrum because of their strength and durability. Several types of synthetic material have been used, including Mersilene, 2, 5 Marlex, 8, 10, 11 Prolene (Ethicon, Somerville, NJ), 12 and Gortex. 13 We use Mersilene mesh almost exclusively in our vault suspensions and cannot comment reliably on the erosion rates of these other materials. Autologous and heterologous biologic graft materials have been used for colpopexy in part to avoid the risks of synthetic mesh erosion. There have been anecdotal reports of colpopexy failure from graft autolysis with such procedures. FitzGerald et al 14 recently reported autolysis of cadaveric fascia lata allografts resulting in sling and colpopexy failure. The success rate and longevity of such suspensions are currently unknown, and only long-term follow-up studies will be able to adequately address this concern. We chose to perform our statistical analysis by using a Cox proportional hazards model to take into account the time to erosion and not merely compare the overall rate of erosion among the various procedures. In this study the hazard function is the probability that an individual person experiences a mesh erosion at time t, which is conditional on the patient having survived erosion free to that time. This form of analysis is useful when outcomes that occur temporally distant from a procedure or event are being compared and when the follow-up period is several months to years. 15 The Cox model provides a more sensitive assessment of mesh erosion because it assesses not only how many patients experienced erosion but how long after the surgery the erosion occurred. The time to vaginal mesh erosion seems to be quite variable with most erosions occurring within the first 2 years after surgery but some occurring many years later. Kohli et al 5 reported a mean time to erosion of 14.0 ± 7.7 months, (range, 4-24 months). This time to erosion is consistent with the time to erosion in our abdominal-only group. Not only were erosion rates higher in our combined vaginal-abdominal groups but the time to erosion was much shorter. As a tertiary referral center for pelvic organ prolapse, we could have underestimated the rate of vaginal mesh erosion if some patients presented with erosion and were treated at an outside institution. However, choosing to analyze our data by means of survival analysis and censoring patients at their last physical examination limited this concern. In this way we did not assume that patients were erosion free for any period of time after their last physical examination. Mesh erosions may be the only clinical manifestation of a bacterial contamination. If this is true, it supports our finding that the rate of mesh erosion was found to be higher in operations with vaginal mesh placed compared with those in which vaginal sutures were placed. This may be explained by a greater exposure of the mesh material to the vaginal flora in the vaginal mesh group. A higher level of bacterial contamination in the former group could explain this finding. By definition, the eroded mesh is directly exposed to the vagina; therefore it is difficult to know whether a positive culture of the mesh antedated the erosion or whether it was the result of posterosion exposure of the mesh to the vaginal flora. In conclusion, both abdominal sacral colpopexy and abdominal-only sacral colpoperineopexy appear to have a relatively low and comparable rate of vaginal mesh erosion. Vaginal placement of mesh results in an unacceptably high rate of mesh erosion and a shorter time to erosion than any other form of vault suspension in this study. Vaginal passage of suture was also associated with an increased rate of erosion in comparison with abdominal sacral colpopexy but the rate was less than that observed with passage of vaginally fixed mesh. Therefore the use of this modification should be reserved for those patients in whom the benefit achieved by vaginal suture placement is thought to exceed the associated risk of mesh erosion. On the basis of our experience we believe that re-establishing continuity from the perineal body to the vagina, along its entire length, and then to the sacrum is an important principle. In the future the development of other types of synthetic mesh, the use of biologic and potentially less reactive materials, or the combined use of these materials may provide reliable and strong support with fewer erosions. REFERENCES 1. Lane FE. Repair of posthysterectomy vaginal vault prolapse. Obstet Gynecol 1962;20: Addison WA, Livengood CH 3d, Sutton GP, Parker RT. Abdominal sacral colpopexy with Mersilene mesh in the retroperitoneal position in the management of posthysterectomy vaginal vault prolapse and enterocele. 1985;153: Cundiff GW, Harris RL, Coates K, Low VH, Bump RC, Addison WA. Abdominal sacral colpoperineopexy: a new approach for

6 302 Visco et al February 2001 correction of posterior compartment defects and perineal descent associated with vaginal vault prolapse. 1997;177: Timmons MC, Addison WA, Addison SB, Cavenar MG. Abdominal sacral colpopexy in 163 women with posthysterectomy vaginal vault prolapse and enterocele. Evolution of operative techniques. J Reprod Med 1992;37: Kohli N, Walsh PM, Roat TW, Karram MM. Mesh erosion after abdominal sacrocolpopexy. Obstet Gynecol 1998;92: Snyder TE, Krantz KE. Abdominal-retroperitoneal sacral colpopexy for the correction of vaginal prolapse. Obstet Gynecol 1991;77: Drutz HP, Cha LS. Massive genital and vaginal vault prolapse treated by abdominal-vaginal sacropexy with use of Marlex mesh: review of the literature. 1987;156: Timmons MC, Addison WA. Mesh erosion after abdominal sacral colpopexy. J Pelvic Surg 1997;1: Sutton GP, Addison WA, Livengood CH 3d, Hammond CB. Lifethreatening hemorrhage complicating sacral colpopexy. Am J Obstet Gynecol 1981;140: Keighley MR, Fielding JW, Alexander-Williams J. Results of Marlex mesh abdominal rectopexy for rectal prolapse in 100 consecutive patients. Br J Surg 1983;70: Grundsell H, Larsson G. Operative management of vaginal vault prolapse following hysterectomy. BJOG 1984;91: Baker KR, Beresford JM, Campbell C. Colposacropexy with Prolene mesh. Surg Gynecol Obstet 1990;171: Iosif CS. Abdominal sacral colpopexy with use of synthetic mesh. Acta Obstet Gynecol Scand 1993;72: FitzGerald MP, Mollenhauer J, Bitterman P, Brubaker L. Functional failure of fascia lata allografts. 1999; 181: Kleinbaum DG. Survival analysis: a self-learning text. Statistics in the health sciences. New York: Springer; O N THE MOVE? Send us your new address at least six weeks ahead Don t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here. OLD ADDRESS: Affix the address label from a recent issue of the journal here. NEW ADDRESS: Clearly print your new address here. Name Address City/State/ZIP COPY AND MAIL THIS FORM TO: OR FAX TO: OR PHONE: Mosby (800) Subscription Customer Service Outside the U.S., call 6277 Sea Harbor Dr (407) Orlando, FL 32887