Abdominal sacrocolpopexy is an anatomically effective operation for vaginal vault prolapse,1–3 but its functional effects on the anterior compartment include a high risk of postoperative stress incontinence.4–7 There is little information available regarding effects of sacrocolpopexy on the posterior compartment, although that from retrospective series suggests a high rate of postoperative bowel dysfunction.8–10 Virtanen et al8 noted that 26% of 30 women developed constipation, 22% difficulty in defecation, and 26% pain and pressure during defecation after sacrocolpopexy. Pilsgaard and Mouritsen10 reviewed 33 sacrocolpopexies and found that ten women had difficulties evacuating the rectum postoperatively compared with six before the operation. Cundiff et al9 reported persistent constipation in three of 12 women who had sacral colpoperineopexies with mean follow-up of 11 weeks.
Sacrocolpopexy was designed to correct vaginal vault prolapse. It has been recommended that coincident cystoceles and rectoceles be repaired vaginally before abdominal sacrocolpopexy5,7,11; however, many cystoceles and rectoceles might be caused by traction or displacement1,12 and vaginal vault fixation might be sufficient for anatomic restoration. Transabdominal correction of rectoceles has been attempted by attaching a mesh to the entire posterior vaginal wall.9,13,14 Villet et al14 reported good anatomic results in 86% of 56 women with mean follow-up of 14 months.
The purpose of this study was to assess the effect of abdominal sacrocolpopexy with obliteration of the pouch of Douglas on the anatomy and function of the posterior compartment. The goal was to correct all pelvic organ support defects transabdominally.
Materials and Methods
We enrolled 33 consecutive women with symptomatic pelvic organ prolapse in whom abdominal sacrocolpopexies with pouch of Douglas obliterations were planned. Preoperative assessments and operations were done by BS. Interviews and investigations at follow-up were done by KB. The interview and all investigations were part of routine preoperative assessments. The sample was adequate to detect a clinically important reduction of postoperative rectoceles of 75% with a power of 80% (1 − α = .8) and β = .05. There was a routine postoperative follow-up at 12 months by interview and examination, and a further follow-up by interview, vaginal, and rectal examination, urodynamic studies, and pelvic floor fluoroscopy for the study.
A structured interview by a standardized questionnaire on bladder, bowel, and sexual function was given before and after surgery. Questions on bowel symptoms addressed defecation frequency, consistency of stool, subjective feeling of constipation, feeling of incomplete emptying, excessive straining at defecation, need for digitation to evacuate completely, incontinence of flatus, fecal urgency and incontinence, passive fecal incontinence (fecal soiling), and pain on defecation. Slow transit constipation was presumed if it was diagnosed previously or the following three criteria were met: defecation fewer than three times per week, frequent hard stool, and need to strain more than 25% of the time to initiate defecation. The term outlet constipation was used when the patient complained of incomplete bowel emptying and had to digitate or strain excessively to facilitate evacuation. Questions on sexual function included deep and superficial dyspareunia, change of sexual habits because of prolapse, and satisfaction with sexual intercourse. A visual analog scale evaluated patient satisfaction with outcome.
Pelvic organ prolapse was staged in the half-sitting lithotomy position using the Baden and Walker vaginal profile classification.15 The pelvic organ prolapse standardization of the International Continence Society was not available at the initiation of the study.16 Prolapse of the anterior vaginal wall, posterior vaginal wall, and cervix or vault was noted as 0 = no prolapse, 1 = prolapse halfway to hymen, 2 = prolapse progressing to hymen, 3 = prolapse halfway through hymen, and 4 = total vaginal prolapse. A rectovaginal examination determined size and location of a rectocele and enterocele (bowel palpable between vagina and rectum). Enteroceles bulging into the anal canal on digital examination and proctoscopy were classified as internal anterior rectal wall procidentia. If the enterocele protruded through the anal canal, an external anterior rectal wall prolapse was noted. If rectal wall procidentia was equivocal a colorectal consultation was arranged. Patients with circumferential prolapse of the rectum or intussusception were excluded. Outlet constipation was confirmed if the patient's sensation of incomplete emptying was reproduced during rectal examination.
Multichannel cystometrogram was carried out preoperatively in the sitting position with and without reduction of the prolapse. A swab on a long clamp was used to reduce the prolapse. To detect enteroceles, a pelvic floor fluoroscopy with opacification of the bladder, rectum, and vagina was performed at rest, during straining, and with pelvic floor contraction. Rectal evacuation was not included. Enteroceles were described if there was a gap seen between the opacified vagina and rectum, which enlarged during straining. Rectoceles were defined by rectovaginal examination and not by pelvic fluoroscopy.
For abdominal sacrocolpopexy, the abdominal cavity was entered through a midline or Pfannenstiel incision. The pararectal peritoneum was opened on the right and the right rectal pillar was divided, so that the rectosigmoid colon could be displaced to the left side (Figure 1) allowing a vaginal attachment almost horizontally in the exposed midline of the sacrum. The vagina was elevated using an obturator to facilitate dissection of the bladder and rectum off the vaginal wall. Sagittal sutures were placed but not yet tied to obliterate the pouch of Douglas (Halban culdoplasty). An expanded polytetrafluoroethylene mesh (Gore-Tex; WL Gore, Flagstaff, AZ) was sutured to the posterior vaginal wall and extended caudally as far as necessary to correct vaginal wall defects. The mesh was then sutured into the anterior longitudinal ligament at the level of S3 or S4 without tension. The opposite end of the mesh was attached to the anterior vaginal wall (Figure 2). If a concomitant hysterectomy was performed first, the vault was closed and the mesh attached to the vagina in the same fashion. The Halban sutures were tied, resulting in closure of the pouch of Douglas. Nonabsorbable material was used for fixation of the mesh and absorbable for obliteration of the pouch of Douglas. All women received a single shot of antibiotics intraoperatively and a small enema the night before surgery. Postoperatively, laxatives were given as needed.
A Burch colposuspension was performed if genuine stress incontinence was demonstrated during urodynamic studies. Paravaginal support defects were repaired as described by Richardson et al.17 Colposuspension and paravaginal defect repair were performed before sacrocolpopexy to avoid tension to the vaginal vault.
It was intended to correct posterior vaginal wall defects transabdominally by suspension of the vaginal vault to the sacrum and covering posterior vaginal wall defects with mesh. After abdominal sacrocolpopexy was completed, a digital vaginal and rectal examination was performed. If a persistent rectocele stage 2 or more was encountered, a colpoperineorrhaphy was performed assuming a defect in the endopelvic fascia rather than a traction rectocele. Colpoperineorrhaphy involved endopelvic fascia and perineal body plication. The anterior vaginal wall was well supported in all women after completed sacrocolpopexy and no anterior colporrhaphy was performed.
For preoperative and postoperative paired comparisons, Wilcoxon signed-ranks test was used for non-normally distributed variables, paired t test for normally distributed variables, and McNemar test for nominal variables. Cross tabulations and χ2 were used to assess associations between variables. The Pearson correlation coefficient was calculated to measure the relationship between the postoperative stage of a rectocele and length of follow-up. A linear stepwise regression model was developed to evaluate factors influencing the patient's satisfaction with outcome.
The mean age of the 33 women was 61 years (range 41–77 years). Median parity was three (range 0–7). There was one nulliparous woman who also had Marfan syndrome. Twenty-six women (79%) were post-menopausal and seven (21%) were perimenopausal. Eighteen (55%) were using hormone replacement therapy. Previous operations are listed in Table 1. Table 2 shows the operations performed at the same time as sacrocolpopexy. The mean follow-up time was 26 months (range 12–48 months). Thirty-one women were available for follow-up examination. One woman refused physical examination but was interviewed on the telephone. Another woman had advanced Alzheimer disease and could be neither interviewed nor examined. The results of the Baden and Walker prolapse classification are shown in Table 3. Preoperatively, 11 women had internal anterior rectal wall procidentia and two had an external anterior rectal wall prolapse.
All 33 women underwent abdominal sacrocolpopexy with obliteration of the pouch of Douglas and individually different extension of the posterior mesh. In 28 women the posterior vaginal wall was well supported at examination after completed sacrocolpopexy. A concomitant colpoperineorrhaphy was performed in the remaining five women in whom a rectocele persisted. The mean estimated blood loss was 358 mL (range 120–750 mL, median 350 mL). In one woman blood loss exceeded 500 mL. Intraoperatively, there was one intermittent atrium fibrillation. Postoperative complications included urinary tract infection in five patients, wound infection in one, wound hematoma in one, pulmonary embolism in one, and voiding difficulties in one. There was no infection or erosion of the Gore-Tex mesh noted.
Postoperatively, there was no recurrence of vaginal vault prolapse and no evidence of an enterocele or anterior rectal wall procidentia on rectovaginal examination (Table 3), proctoscopy, and pelvic floor fluoroscopy. The reduction of postoperative anterior, apical, posterior, and anterior rectal wall prolapse was significant (Wilcoxon test, P < .001). Rectoceles recurred in 16 of 28 women with preoperative rectoceles (57%). One rectocele occurred de novo. Four of the five women who underwent concomitant colpoperineorrhaphy were examined at follow-up, two had no rectocele, one had a grade 1 rectocele, and another woman had preoperative grade 3 rectocele reduced to grade 1. There was no significant correlation between the length of follow-up and the stage of the rectocele (P = .874). Postoperatively, only two women with a grade 3 rectocele were aware of prolapse. Previous and concomitant operations did not affect the postoperative recurrence of rectoceles or occurrence of outlet constipation (χ2 test; P > .05).
All women underwent pelvic floor fluoroscopy before surgery to detect enteroceles. Eight women declined the investigation postoperatively. Pelvic floor fluoroscopy did confirm the presence of an enterocele in 11 of 21 women in whom an enterocele was detected at preoperative vaginal examination (52%). In six of the 13 women with anterior rectal wall procidentia the enterocele indented the anterior rectal wall and bulged through the anal canal in two women. There was no circumferential intussusception seen.
Outlet constipation was present in 21 of 33 women (64%) preoperatively and in 12 of 32 (38%) postoperatively (McNemar test, P = .035). Outlet constipation associated with rectoceles and anterior rectal wall prolapse ceased postoperatively in 18. Nine of 32 interviewed women (28%) reported that outlet constipation symptoms had changed postoperatively. They described a normal urge to defecate, but they were unable to evacuate feces. They did not feel a lump in the vagina during attempted defecation. Instead, they felt the “obstruction” higher in the rectum. These symptoms could not be reproduced on rectovaginal examination. We referred to this condition as “high outlet constipation.” All women with postoperative outlet constipation had had preoperative defecation problems. Preoperative symptoms of outlet constipation were related to the size of rectocele (χ2 test, P = .002). This relation was not demonstrable postoperatively (χ2 test, P = .101). Eight women had complaints of both outlet and slow transit constipation preoperatively. In six of these women, bowel symptoms had resolved postoperatively (McNemar test, P = .031). In three women, presumed slow transit constipation was not combined with outlet constipation and no change in bowel symptoms was noted postoperatively.
Preoperatively, anterior rectal wall procidentia was accompanied by fecal urge incontinence in two and by incontinence of flatus in another two women. These symptoms ceased in two patients postoperatively. Two patients developed fecal urge incontinence for liquid stool and one patient reported new fecal soiling after sacrocolpopexy.
Nine of 23 sexually active women reported prolapse-related dyspareunia preoperatively (39%). They had reduced their frequency of intercourse because of embarrassment and pain caused by the prolapse. After sacrocolpopexy, dyspareunia ceased in all but one woman with a recurrent stage 3 rectocele (McNemar test, P = .07). Two women developed pain (presumably at the sacral mesh fixation point) that was present only during intercourse. They refused removal of the mesh and were advised to try different positions during intercourse. Colpoperineorrhaphy caused discomfort in one woman. One woman was widowed and stopped having intercourse.
Satisfaction with the outcome measured on the visual analog scale ranged from 3 to 10 with a mean of 7.2 and a median of 8. Postoperative symptoms of outlet constipation, urinary urgency, and urge incontinence were reflected in the linear regression model to explain 27% of the variation of the patient's satisfaction (R2 = .27, P = .01).
The Research Committee of the American Urogynecologic Society recently recognized the debilitating nature of defecatory disorders and assembled a multidisciplinary symposium to define knowledge deficits in pathophysiology, imaging, and evaluation and treatment.18 Our prospective study focused on anatomy and function of the posterior compartment after abdominal sacrocolpopexy, which has been largely ignored. Anorectal dysfunction is common in women with complex pelvic organ prolapse.19,20 We found 73% of our patients to have bowel dysfunction, predominantly outlet constipation. Outlet constipation was still present in 38% postoperatively. Virtanen et al8 noted that constipation had developed in 26% of patients and difficulty in defecation in 22% after abdominal sacrocolpopexy, although no definitions were given and it is unclear whether these numbers represent the same patients. In some women a rectocele might account for incomplete bowel emptying. Nine of our patients (28%) described that the stool seemed to stop higher in the rectum postoperatively. Splinting of the vagina or transanal or transperineal digitation did not facilitate evacuation any more and most of the women considered this a deterioration of bowel symptoms. This form of constipation is a known sequela in up to 47% of patients after different rectopexy procedures.21–23 It is thought to be related to denervation occurring during rectal mobilization and division of the lateral ligaments.22,24 Although division of the lateral ligaments provides optimal access to attach the mesh in an anatomically caudal position at the sacrum, based on our findings, we have abandoned this procedure.
In six patients with presumed slow transit and outlet constipation preoperatively, both conditions resolved postoperatively. This observation is consistent with the suggestion that anorectal outlet constipation may induce prolonged colonic transit.25,26 However, patients' symptoms do not seem to be specific enough to differentiate between slow colonic transit and outlet constipation27 and we did not perform colonic transit studies to substantiate our presumed diagnosis of slow transit constipation. Anorectal physiology tests or endoanal ultrasound were not performed preoperatively but these tests should be considered in symptomatic patients.
Abdominal sacrocolpopexy with obliteration of the pouch of Douglas resulted in excellent fixation of the vaginal vault comparable to other studies.1–3 Anterior rectal wall procidentia was also successfully treated. It is known that rectal prolapse might be caused by an enterocele bulging into the rectum.28,29 To treat both the vaginal and rectal prolapse sufficiently, a combination of vaginal vault fixation and pouch of Douglas obliteration appears to be necessary. Enteroceles may persist or develop behind the sacrocolpopexy mesh if the pouch of Douglas is not meticulously obliterated.1,3 Isolated obliteration of the pouch of Douglas for rectal prolapse resulted in high recurrence rates.30 In our small series no prolapse of the anterior rectal wall or enterocele recurred. The caudal sacral mesh fixation of the vaginal vault seems to contribute to the restoration of the normal vaginal axis and normal intraabdominal pressure distribution.
Intraoperative vaginal examination after completed abdominal sacrocolpopexy revealed good support of the posterior vaginal wall in most of the women. A preoperative traction rectocele was therefore assumed that appeared to have been corrected by suspension of the vaginal vault and posterior extension of the mesh. Thus, a concomitant colpoperineorrhaphy was performed in five women only. However, postoperative rectoceles were present in 55%. This finding suggests an additional defect in the endopelvic fascia rather than a traction rectocele related to the displacement of the vagina and rectum. In our series, abdominal sacrocolpopexy even with posterior extension of the mesh failed to correct rectoceles durably. This observation is in contrast to those of Villet et al,14 who described good anatomic results in 86%, although no precise definition of “good result” was given. We could report good results in 71% if grade 1 rectoceles were included. Most of the rectoceles that recurred remained asymptomatic. Because of the high recurrence rate of rectoceles, we stopped recruiting patients for transabdominal correction. We adjusted our operative management and started transvaginal defect-specific repair of rectoceles.
Most of the women included in this study had had previous incontinence or prolapse surgery and most of them underwent operations in addition to sacrocolpopexy, resulting in a heterogeneous group of patients. Pelvic organ prolapse was usually associated with pelvic floor dysfunction, which needed to be addressed surgically. Colposuspensions, which are known to increase the incidence of postoperative rectoceles and enteroceles,31 were performed in 36%. However, neither previous nor concomitant operations appeared to affect the anatomic and functional outcome of the posterior compartment.
Pelvic floor fluoroscopy showed important limitations, because the filled bladder and rectum compete to prolapse with the potential enterocele, particularly if the rectum is not evacuated during the examination. However, during vaginal examination either the anterior or posterior vaginal wall is prevented from protruding by a speculum while the other wall is explored. This fact may explain why vaginal examination and pelvic floor fluoroscopy do not correlate32 and why only half of the enteroceles detected by rectovaginal examination could be substantiated on fluoroscopy.
Successful restoration of prolapse symptoms including preservation of satisfactory sexual function was related mainly to the fixation of the vaginal vault and reconstruction of the normal vaginal axis. In this study cystoceles, but not rectoceles, were satisfactorily managed with abdominal procedures. Concomitant transvaginal rectocele repair should be performed when appropriate. Long-term results will be needed to evaluate whether abdominal repair of the posterior vaginal wall defect followed by extension of the mesh down to the perineal body in the rectovaginal space and obliteration of the pouch of Douglas9 is effective for rectoceles. Extended dissection and mobilization of the rectum should be avoided in order to minimize denervation and subsequent difficulties with defecation.
Patients may frequently be discontent with the surgical outcome because of functional problems, despite an excellent anatomic result. In our study, outlet constipation symptoms diminished the patients' satisfaction with the outcome. Although this study was limited in terms of number of patients, concomitant operations, and the lack of anorectal physiology tests, it emphasizes that abdominal repair of rectoceles is not easily achieved and may be confounded by persistent or even deteriorating bowel dysfunction.
1. Addison WA, Livengood CH, Sutton GP, Parker RT. Abdominal sacral colpopexy with Mersilene mesh in the retroperitoneal position in the management of posthysterectomy vaginal prolapse and enterocele. Am J Obstet Gynecol 1985;153:140–6.
2. Creighton SM, Stanton SL. The surgical management of vaginal vault prolapse. Br J Obstet Gynaecol 1991;98:1150–4.
3. Timmons MC, Addison WA, Addison SB, Cavenar M. Abdominal sacral colpopexy in 163 women with posthysterectomy vaginal vault prolapse and enterocele. J Reprod Med 1992;37:323–7.
4. Brubaker L. Sacrocolpopexy and the anterior compartment: Support and function. Am J Obstet Gynecol 1995;173:1690–6.
5. Snyder TE, Krantz KE. Abdominal-retroperitoneal sacral colpopexy for the correction of vaginal prolapse. Obstet Gynecol 1991;77:944–9.
6. Valaitis SR, Stanton SL. Sacrocolpopexy: A retrospective study of a clinician's experience. Br J Obstet Gynaecol 1994;101:518–22.
7. Varner RE, Plessala KJ, Richter H. Effects of sacrocolposuspension on the lower urinary tract. Am J Obstet Gynecol 1995;173:1684–9.
8. Virtanen H, Hirvonen T, Makinen J, Kiilholma P. Outcome of thirty patients who underwent repair of posthysterectomy prolapse of the vaginal vault with abdominal sacral colpopexy. J Am Coll Surg 1994;178:283–7.
9. Cundiff GW, Harris RL, Coates K, Low VHS, Bump RC, Addison WA. Abdominal sacral colpoperineopexy: A new approach for correction of posterior compartment defects and perineal descent associated with vaginal vault prolapse. Am J Obstet Gynecol 1997;177:1345–55.
10. Pilsgaard K, Mouritsen L. Follow-up after repair of vaginal vault prolapse with abdominal colposacropexy. Acta Obstet Gynecol Scand 1999;78:66–70.
11. Birnbaum SJ. Rational therapy for the prolapsed vagina. Am J Obstet Gynecol 1973;115:411–9.
12. Nichols DH. Surgery for pelvic floor disorders. Surg Clin North Am 1991;71:927–46.
13. Perez M, Lefranc JP, Blondon J. Rectocèles et trouble fonctionnels rectaux après traitement chirurgical d'un prolapsus génital. J Chir (Paris) 1991;128:465–9.
14. Villet R, Morice P, Bech A, Salet-Lizée D, Zafiropulo M. Approche abdominale des rectocèles et des élytrocèles. Ann Chir 1993;47:626–30.
15. Baden WF, Walker TA. Genesis of the vaginal profile: A correlated classification of vaginal relaxation. Clin Obstet Gynecol 1972;15:1048–54.
16. Bump RC, Mattiasson A, Bo K, Brubaker LP, DeLancey JOL, Klarskov O, et al. The standardization of terminology of female pelvic organ prolapse and pelvic floor dysfunction. Am J Obstet Gynecol 1996;175:10–7.
17. Richardson AC, Lyon JBL, Williams NL. A new look at pelvic relaxation. Am J Obstet Gynecol 1976;126:568–73.
18. Cundiff GW, Nygaard I, Bland DR, Versi E. Proceedings of the American Urogynecologic Society multidisciplinary symposium on defecatory disorders. Am J Obstet Gynecol 2000;182:S1–10.
19. Spence-Jones C, Kamm MA, Henry MM, Hudson CN. Bowel dysfunction: A pathogenic factor in uterovaginal prolapse and urinary stress incontinence. Br J Obstet Gynaecol 1994;101:147–52.
20. Weber AM, Walters MD, Ballard LA, Booher DL, Piedmonte MR. Posterior vaginal prolapse and bowel function. Am J Obstet Gynecol 1998;179:1446–50.
21. Mann CV, Hoffmann C. Complete rectal prolapse: The anatomical and functional results of treatment by an extended abdominal rectopexy. Br J Surg 1988;75:34–7.
22. Speakman CTM, Madden MV, Nicholls RJ, Kamm MA. Lateral ligament division during rectopexy causes constipation but prevents recurrence: Results of a prospective randomized study. Br J Surg 1991;78:1431–3.
23. Novell JR, Osborne MJ, Winslet MC, Lewis AAM. Prospective randomised trial of Ivalon sponge versus sutured rectopexy for full-thickness rectal prolapse. Br J Surg 1994;81:904–6.
24. Mollen RM, Kuijpers JH, van Hoek F. Effects of rectal mobilisation and lateral ligaments division on colonic and anorectal function. Dis Colon Rectum 2000;43:1283–7.
25. Karlbom U, Pahlman L, Nilsson S, Graf W. Relationships between defecographic findings, rectal emptying, and colonic transit time in constipated patients. Gut 1995;35:907–12.
26. Klauser AG, Voderholzer WA, Heinrich CA, Schindlbeck NE, Muller-Lissner SA. Behavioural modification of colonic function. Can constipation be learned? Dig Dis Sci 1990;35:1271–5.
27. Koch A, Voderholzer WA, Klauser AG, Muller-Lissner S. Symptoms in chronic constipation. Dis Colon Rectum 1007;40:902–6.
28. Moschcowitz A. The pathogenesis, anatomy, and cure of prolapse of the rectum. Surg Gynecol Obstet 1912;15:7–12.
29. Orgias R. Rectal prolapse and hernia of the pouch of Douglas in the female. Br J Surg 1965;52:889–92.
30. Palmer JA. Prolapse of the rectum: Treatment by the Moschcowitz–Graham operation. Can J Surg 1969;12:116–23.
31. Wiskind AK, Creighton SM, Stanton SL. The incidence of genital prolapse after the Burch colposuspension. Am J Obstet Gynecol 1992;167:399–404.
32. Kenton K, Shott S, Brubaker L. Vaginal topography does not correlate well with visceral position in women with pelvic organ prolapse. Int Urogynecol J Pelvic Floor Dysfunct 1997;8:336–9.