| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Department of Obstetrics and Gynecology, University Hospital of Lund, Lund, Sweden.
Address reprint requests to: Jan Persson, MD, PhD, Department of Obstetrics and Gynecology, University Hospital, S-22185, Lund, Sweden; E-mail: jan.persson{at}gyn.lu.se.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: Between 1991 and 1998, gynecologists at our department performed 428 operations for stress urinary incontinence by inserting expanded polytetrafluoroethylene or polyethylene suburethral slings. After suitable exclusions, 386 women followed for at least 24 months after surgery remained for analysis. Of the 386 women, 47 (12.2%) had graft rejection or symptoms associated with rejection within 24 months after surgery, which led to later removal. We compared cases and controls using logistic regression analyses with forward selection to identify independent risk factors and risk markers for rejection.
RESULTS: The rejection rate declined sharply during the years 1993–1994 after introduction of routines that included preoperative prophylactic antibiotics active against anaerobes and repeated preoperative treatment of the vagina with chlorhexidine acetate. After adjusting for potentially confounding variables, surgery after 1993, the gynecologist performing the procedure, and adequate antibiotic prophylaxis remained significantly associated with a lower rejection rate. There was no significant association between rejection and sling material, age at surgery, year of surgery (in one year steps), or concomitant prolapse surgery.
CONCLUSION: Rejection of suburethral slings might be associated with bacterial contamination of the graft, and the rate was decreased with prophylactic antibiotics a repeated and vaginal disinfection.
Graft rejection and/or erosion to adjacent tissue are important problems associated with synthetic slings for stress urinary incontinence. For expanded polytetrafluoroethylene slings, rejection rates of 6–32% have been reported.1–5 A similar rejection rate has been reported for polyethylene slings.6 Ulmsten et al reported a lower rejection rate for polypropylene, even when controlled studies have not been performed.7 Some investigators hypothesize that improvement in surgical technique and antiseptic procedures might prevent rejection.8 (Ferry P, Laplanche B, Salzard C, Anguill C, Olleac C, Quentin M. Tolerance of a Gore-Tex band in suburethral suspension using the Mouchel technique [Letter]. J Gynecol Obstet Biol Reprod Paris 1994;23:929.) In our department, different sling materials, antiseptic routines, and protocols for antibiotic prophylaxis have been used. However, there are no prospective, randomized studies available to identify associations, if any, between specific routines or sling materials and sling rejection. To generate hypotheses as to possible risk factors and risk markers for rejection to be focused on in future randomized studies, we performed this retrospective case-control study.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
For the analysis, we included women with a follow-up of at least 24 months after surgery (median 38, range 24–96 months) or minimum until rejection was diagnosed. We excluded five women from whom the sling had to be removed for other reasons than rejection, two women who refused follow-up, and one woman because pertinent data were lacking. Furthermore, we excluded 23 women who had been followed for less than 24 months (median 13, range 2–22). For nine women who had been reoperated after sling rejection, only the first procedure was considered. For two women who had the sling replaced because of failure, only the second procedure was considered for the analyses. There was no indication of sling rejection among excluded women.
Thus, data from 386 procedures performed on 386 women (273 with polytetrafluoroethylene and 113 with polyethylene) remained for analysis; 225 of them had been performed by surgeon 1, 101 of them by surgeon 2, and the remaining 60 by different other gynecologists (miscellaneous surgeons).
We defined cases as those women who experienced sling rejection, symptoms indicating ongoing rejection, or chronic wound infection associated with rejection within the first 24 months after surgery (n = 47). In general, rejection occurred or was diagnosed shortly after surgery (median 3 months). However, in six women, sling rejection was diagnosed between 6 and 60 months after surgery. Those women had had symptoms indicating rejection within 24 months after surgery. One woman had a chronic seroma, another had a chronic sinus formation in the pubic region, and a third (who had not shown up for the 2-month follow-up visit) reported several months of vaginal discharge. Persistent vaginal discharge was also reported by three women for whom rejection of the sling was not seen until the second follow-up visit, 19, 20, and 22 months after surgery, respectively. One woman was asymptomatic until she had a vaginal hysterectomy 40 months after the sling procedure. She rejected the sling 2 months later but was nevertheless regarded as a control.
The rejected slings were mainly removed by pulling them out per vaginam. In three women in whom the sling perforated through the abdominal scar, only the suprafascial part was removed. No specific complications were associated with rejection of slings or their removal.
Preoperative medical history included questions regarding hormonal status (ie, whether they were premenopausal, postmenopausal with, or postmenopausal without hormonal replacement therapy).
Antibiotic prophylaxis was given at the surgeon’s discretion to the first 29 consecutive patients. Between the end of 1992 and August 1998, the following protocol for antibiotic prophylaxis was used: metronidazole should be administered rectally, 1 g the night before surgery, 1 g times three on the day of surgery, and 1 g times two on the first postoperative day. After August 1998, the protocol was simplified and based on one per oral dose of 400 mg of metronidazole 2 hours before surgery. Either of these regimens, with few minor deviations, was administered in conjunction with 341 of the following 357 procedures. For the analysis of the effect of antibiotic prophylaxis, we dichotomized the material as follows. Antibiotic prophylaxis was adequate if women had received antibiotics active against anaerobic micro-organisms the day of surgery, before, or during the first 30 minutes of surgery (n = 366). Antibiotic prophylaxis was inadequate if women had not received antibiotics at all, had received antibiotics later than 30 minutes after start of surgery, or had received antibiotics with no or inadequate effect on anaerobic microorganisms (n = 20). The rationale behind this grouping was our belief that contamination of the sling with anaerobic vaginal flora might be of pathogenic importance in rejection. Moreover, after Classen et al9 had shown that antibiotic prophylaxis is most efficient if given just before or during surgery, we were convinced optimal prophylaxis included antibiotics being given before application of the sling (usually within 30 minutes after start of surgery).
Polytetrafluoroethylene slings were cut under sterile conditions in our department from 1-mm sheets. The slings were made 8–10 mm wide, but a few (three cases documented) of the first slings were made broader in the middle part. Each sling was separately packed and re-sterilized after cutting. Polyethylene slings were prefabricated 5-mm strips.
In the operating room, before draping, the lower part of the abdomen and the vulvovaginal area were disinfected using 2 mg/mL of chlorhexidine acetate. In January 1994, we introduced a new routine of an additional, similar disinfecting procedure of the vulvovaginal area to be performed by the surgeon immediately before start of the operation. We did not soak the sling or the operative site with any antibiotic irrigant.
The standard surgical procedure was performed in the following way with small variations. A short incision under the urethra and a 5-cm transverse suprapubic incision reaching the rectus fascia were made including necessary dissection of the paraurethral loose tissue. The tip of a forceps was introduced through the vaginal incision and pushed through the abdominal incision on one side of the urethra to grasp the sling and pull it down. Similarly, a suture was pulled down on the other side of the urethra and used to bring the attached "vaginal end" of the sling up to the abdominal incision. For 40 of the polyethylene slings a specially made perforation instrument was used to introduce the sling through the vaginal incision. To rule out perforation of the bladder, cystoscopy was performed. The ends of the polytetrafluoroethylene slings were fixed to the rectus fascia or the Scarpa’s fascia with resorbable sutures on each side, but the polyethylene slings were left unfixed. Finally, the vaginal mucosa was sutured (usually Vicryl 2-0, Ethicon GmbH, Hamburg, Germany). In 40 (35%) of the women with the polyethylene sling, the vaginal mucosa was duplicated over the sling before it was sutured. Any concomitant anterior colporrhaphy was performed either using a longer incision in the anterior vaginal wall or through a separate incision. As a rule, the sling procedure was completed before any concomitant posterior colporrhaphy.
The indwelling catheter was left open and removed as soon as residual urine was measured to less than 100–125 mL (median 2 days after surgery, range 0–79). At discharge, the women were instructed to avoid activities that might possibly interfere with the healing process for 5–6 weeks, including tub bath, use of tampons, sexual intercourse, or bicycling.
Regular follow-up examinations were planned 2 and 12 months after surgery. In addition, women were asked to call if they experienced adverse symptoms attributable to the procedure. During follow-up visits, specific attention was paid to signs of sling rejection or persistent wound infection. Women for whom the hospital charts contained no evidence of sling rejection and who had not been examined in our department 24 or more months after the surgery were contacted by phone and interviewed according to a standardized questionnaire. Among other things, they were asked if they had been operated on for sling rejection in another hospital. Those who reported current symptoms of persistent vaginal discharge, vaginal or lower abdominal wall discomfort, and/or irritation of the abdominal scar were called in for evaluation.
Data regarding patients operated with sling procedure were in part retrieved from charts, in part prospectively collected, and stored in a database (Stat-View, SAS Institute Inc., Cary, NC) for quality assurance purposes. The Regional Information Security Officials consented to the establishment of the database. Approval for the study was obtained from the local institutional review board. For univariate analysis, using the 0.05 significance level, logistic regression analyses were performed to select variables for further analysis. Multiple logistic regression models were used in which rejection or no rejection of the suburethral sling was the dependent variable, and variables significantly associated with rejection in univariate analyses were independent variables. Although not univariately associated with rejection, concomitant prolapse surgery was considered a potential confounder and was included in the regression analysis. For the logistic regression models, a forward selection procedure was performed. However, the size of the material allowed a maximum of four independent variables to be entered in the model simultaneously.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
shows the number of operations and sling rejections by year of operation. The sling rejection rate changed drastically between 1993 and 1994, simultaneously with the introduction of a second vaginal wash. A less dramatic, nevertheless obvious, decrease in rejection rate occurred between 1992 and 1993, coinciding with the introduction of a metronidazole prophylaxis protocol.
|
and 3 show variables of the women and the surgical procedure by sling rejection (yes or no) and time period of operation. It is evident that women operated in 1993 or earlier were younger than those operated after 1993. The former also more often had a polyethylene sling than a polytetrafluoroethylene sling inserted, and more often underwent concomitant prolapse surgery. During the first, but not the second, time period, surgeon 1 performed most of the operations. The surgical technique to duplicate the vaginal mucosa was mainly used during the first time period.
|
|
To identify risk factors or risk markers independently associated with sling rejection, a series of multiple logistic regression analyses was performed. We entered a maximum of four independent variables simultaneously. The initial analyses selected the three most important independent variables, all of them negatively associated with rejection. These variables were (in order of importance) operation after 1993, surgery performed by surgeon 2, and adequate antibiotic prophylaxis. The other possible risk factors were entered one at a time to a model consisting of these three independent variables (Table 4).
|
). Among women operated before 1994, four of 115 (3.5%) had an abdominal wound infection compared with 23 of 271 (8.5%) women operated after 1993. Eleven of the 27 women rejected the sling (41%). After adjusting for "operation before 1994," surgeon 2, and antibiotic prophylaxis in a separate regression analysis, the odds ratios for rejection among women with wound infection was 15.3 (95% confidence interval 5.1, 45.8, P < .001).
Women without rejection had a longer follow-up compared with women with rejection (median 35 months compared with 20 months). Nevertheless, symptoms of recurrent stress urinary incontinence occurred in 38 of 44 (86%) patients with a totally removed sling compared with 93 of 334 (28%) controls (odds ratio for incontinence 16.3, 95% confidence interval 7.98, 33.3, P < .001).
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two new routines were introduced that might explain the decline in rejection rates. First, a uniform antibiotic prophylaxis based on metronidazole, applied before and during surgery was introduced at the end of 1992. Slings introduced after that year were far less often rejected than before. More than 1 year later, another routine was introduced (ie, a second vaginal wash with chlorhexidine immediately before start of the operation). A further sharp decline in rejection rates occurred after that.
The hypothetic effects of the second new routine, that of a repeated vaginal wash, are in line with those of antibiotic prophylaxis (ie, to reduce the concentration of bacteria in the vagina). Since the sling procedure was introduced in 1991, nurses’ aides have washed the vagina with chlorhexidine before surgery. However, the sling might have been inserted more than a half hour after the vaginal wash, thus probably allowing some species to recover. According to Monif et al,10 vaginal wash with povidone-iodine reduces the concentration of bacteria by a factor of 106. The same authors showed that the flora recovered to near baseline concentration within 30–120 minutes after the wash. Thus, the second vaginal wash in this study probably again reduced the concentration of bacteria. Even when the procedure was not always recorded in the charts, we have reason to believe that it was performed on almost all women who were operated on after the routine was introduced.
Antibiotic prophylaxis, as defined in our study, was independently negatively associated with sling rejection, contradicting the report by Bent et al,3 who failed to find any protection of prophylaxis. However, in that study, the type and timing of the antibiotic prophylaxis used was not stated. In a separate regression analysis, we also found abdominal wound infection to be significantly associated with rejection of slings. This association further implies infection to be the main cause of sling rejection.
The second independent risk factor for rejection was "surgeon," suggesting surgical habits to influence the risk of sling rejection. We believe that a meticulous antiseptic surgical technique is essential for decreasing the rate of rejection (eg, the sling material should be kept within the wound and not in contact with vulvar skin or vaginal mucosa). Moreover, the sling should preferably be handled with instruments. Furthermore, the vaginal mucosal edges should not be compressed with forceps but handled with care to prevent necrosis after the wound has been closed. Hypothetically, individual improvement of the surgical habits could have contributed to a part of the decline in sling rejection rate over the years. However, year of surgery (analyzed in 1-year steps) was not found to be independently associated with rejection.
This study also shows that sling rejection is strongly related to recurrent stress urinary incontinence symptoms. The difference found may be even larger considering the longer follow-up for women without rejection. Thus, it is important to minimize the rejection rate.
A strength of this study is the large number of consecutive procedures performed in the same unit mainly by either of two surgeons. The study has some weaknesses. We had no protocol designed to prospectively identify risk factors for rejection. Accordingly, different changes in patient management, including antibiotic prophylaxis and switch from polyethylene to polytetrafluoroethylene slings, were gradually performed during the years without randomization. Moreover, the first vaginal wash performed by the nurses’ aides was not standardized in timing or in technique. The second vaginal wash was not always recorded in charts. The time between the two washing procedures was not standardized. Finally, surgical technique was not completely identical over the years and between surgeons.
In conclusion, this study suggests infection to be the cause of sling rejection. It provides some evidence that prophylactic antibiotics and vaginal wash each have contributed to the decline in rejection rates from 43% to around 5.5%. Prospective, randomized studies are needed to confirm these hypotheses.
|
Footnotes |
|---|
Received January 30, 2001. Received in revised form August 29, 2001. Accepted September 6, 2001.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Summitt RL, Bent AE, Ostergard DA, Harris TA. Stress incontinence and low urethral closure pressure. Correlation of preoperative urethral hypermobility with successful suburethral sling procedures. J Reprod Med 1990;35: 877–80.[Medline]
3. Bent AE, Ostergard DR, Zwick Zaffuto M. Tissue reaction to expanded polytetrafluoroethylene suburethral sling for urinary incontinence: Clinical and histologic study (see comments). Am J Obstet Gynecol 1993;169:1198–204.[Medline]
4. Debodinance P, De Bievre P, Parmentier D, Dubrulle R, Querleu D, Crepin G. The "hazards" of using a Gore-Tex sling in the treatment of stress urinary incontinence. J Gynecol Obstet Biol Reprod Paris 1994;23:665–70.[Medline]
5. Errando C, Batista JE, Arano P. Polytetrafluoroethylene sling for failure in female stress incontinence surgery. World J Urol 1996;14(Suppl. 1):48–50.
6. Ulmsten U, Johnson P, Petros P. Intravaginal slingplasty. Zentralbl Gynakol 1994;116:398–404.[Medline]
7. Ulmsten U, Johnson P, Rezapour M. A three-year follow-up of tension free vaginal tape for surgical treatment of female stress urinary incontinence. Br J Obstet Gynaecol 1999;106:345–50.[Medline]
8. Norris JP, Breslin DS, Staskin DR. Use of synthetic material in sling surgery: A minimally invasive approach. J Endourol 1996;10:227–30.[Medline]
9. Classen DC, Evans SR, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. N Engl J Med 1992;326:281–5.[Abstract]
10. Monif GRG, Thompson JL, Stephens HD, Baer H. Quantitative and qualitative effects of Povidone-iodine liquid and gel on the aerobic and anaerobic flora of the female genital tract. Am J Obstet Gynecol 1980;137:432–8.[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
