Posterior repair and sexual function

Abstract

OBJECTIVE

The purpose of this study was to determine the effect of posterior repair (PR) on sexual function in patients who have undergone incontinence and/or pelvic reconstructive surgery.

STUDY DESIGN

A cohort study of women who underwent incontinence and/or prolapse surgery was performed. Participants completed the pelvic organ prolapse urinary incontinence sexual questionnaire (PISQ) before and after the operation. PISQ scores were compared between women who underwent PR and women who did not.

RESULTS

Of 73 study participants, 30 women underwent PR; 43 women did not (no PR). Although there was no difference in dyspareunia between groups pre-op, dyspareunia prevalence post-op was significantly lower in the no PR group. Preoperative PISQ scores were similar between groups. After the operation, both groups significantly improved their PISQ scores, without a difference between groups.

CONCLUSION

Although the incidence of dyspareunia differed between PR and no PR groups, overall improvement in sexual function was reflected in improved total PISQ scores that occurred irrespective of PR performance.

Posterior compartment defects arise from abnormalities in rectovaginal connective tissue and its attachments within the pelvis.¹ Repair of this defect includes traditional posterior colporrhaphy, which is performed by plicating the patient’s native rectovaginal connective tissue in the midline, or site specific posterior colporrhaphy, which is performed by plicating individual breaks in rectovaginal tissue.^1,2 Controversy exists whether posterior repair (PR) is associated with diminished sexual function or dyspareunia. Some authors have reported increased rates of dyspareunia after PR, whereas other authors have not.^3–7 Some authors have concluded that PR causes sexual dysfunction as a result of its association with dyspareunia.^3–5 Our objective was to evaluate prospectively pre- to postoperative changes in sexual function in women who had undergone PR compared with women who did not.

MATERIALS AND METHODS

This analysis is based on a previously reported population of women enrolled in a randomized controlled trial that evaluated antibiotic prophylaxis for urinary tract infections after urinary incontinence and/or pelvic organ prolapse surgery.⁸ Four sites in this multicenter trial also recruited women to participate in a sexual function study, which has also been described previously.⁹ These women form the present study population. We studied sexual function using a condition-specific, validated sexual function questionnaire, the Pelvic Organ Prolapse Urinary Incontinence Sexual Questionnaire (PISQ).¹⁰ This cohort study compared women who underwent PR with the group of women who did not undergo PR (no PR). Institutional review board approval was given, and written informed consent was obtained.

Sexually active patients who had undergone pelvic reconstructive and/or incontinence procedures agreed to answer the PISQ both before and 3 and 6 months after the operation. Each PISQ question is answered on a 5-point scale; higher PISQ scores indicate better sexual function. The PISQ has 3 domains: behavioral/emotive, physical, and partner-related, which have been described in detail previously.¹⁰

The PISQ 5-point Likert scale responses range between “never” and “always.” Responses were dichotomized so that 2 favorable responses for the desired outcome were compared with 3 responses that represented adverse outcomes. For example, responses of “always” or “usually” for the question, “Do you feel pain during intercourse?” were compared with responses of “sometimes,” “seldom,” or “never.” These cut-points dichotomized groups as equally as possible when compared with other possible cut points that considered answers to all 31 PISQ questions.

Demographic information, surgical procedures performed, and incidence of complications were recorded. All women underwent detailed physical examination with measurements of pelvic support according to the Pelvic Organ Prolapse Quantification System (POPQ).¹¹ Both site-specific and traditional posterior colporrhaphy were performed at the discretion of the attending physician and were not recorded explicitly. No patient underwent repair with graft material for the PR.

Statistics

Operative and demographic data and cumulative PISQ and PISQ domain scores at baseline and at 3- and 6-month follow-up evaluations were compared between the PR and no PR groups. There were no differences between 3- and 6-month responses (all P = not significant); therefore, all results are reported as 3- to 6-month follow-up evaluations. This decision to combine 3- and 6-month scores has been discussed previously.⁹ Continuous variables were compared with the use of the unpaired t-test to evaluate differences between groups and the paired t-test to evaluate within group differences. In addition to the use of the unpaired t-test to evaluate differences in PISQ scores between groups, postoperative PISQ and domain scores were adjusted for possible differences in baseline values with the use of analysis of covariance (ANCOVA). POP-Q findings were ranked, converted to numeric scores, and analyzed with the Mann-Whitney-Wilcoxon-U 2 sample test. Stepwise and all subsets regression analysis were used to verify whether variables that differed between the groups were covariates.

Between group differences for individual PISQ questions and surgical and demographic data where appropriate, were compared using Fisher’s exact test. McNemar’s test was used to evaluate changes within each group. Significance was set at P < 0.05. Data analysis was performed using SAS/STAT Software (Version 8. 1999–2001. SAS Institute, Inc., Cary, NC).

RESULTS

One hundred two sexually active women who underwent pelvic reconstructive and/or incontinence surgery agreed to answer the preoperative PISQ. Seventy-five women completed this same questionnaire at 3–6 months after the operation. Two women had incomplete data regarding performance of posterior colporrhaphy, which left 73 women with information that was available for this analysis. Of the 73 women who were the focus of our study, 30 women (41%) had PR performed, and 43 women (59%) did not.

The characteristics of the women who declined to participate in the sexual function study and women who did not complete the questionnaires post-op have been described previously.⁹ In summary, nonparticipants were older, more likely to take anticholinergics, more likely to be postmenopausal, to be white, and more likely to have had hysterectomy. Women who completed the study were more likely to have vaginal apical suspension performed. Additionally, we evaluated differences in baseline PISQ scores between women who completed the sexual function study and those who did not. The women who completed the study had higher baseline PISQ scores than those who did not (92 vs 87; P = .02), which was primarily due to higher behavioral/emotive domain scores (41 vs 37; P = .005).

Demographic and physical examination information for the no PR and PR groups is listed in Table 1. They differed in 5 baseline characteristics. The PR group was older and thinner than the no PR group and had more descent of point C, the vaginal apex or cervix, than did the no PR group (all P < .05). The PR group was also more likely to undergo concomitant anterior colporrhaphy and vaginal apical suspension than the no PR group (Table 2). None of the patients had levatorplasty performed. The PR group was similar to the no PR group with respect to other demographic variables and the operations that had been performed and the occurrence of complications.

TABLE 1.

Demographics

Demographic	No PR (n = 43)	PR (n = 30)	P value
Age (y)*	43.8 ± 11.0 (23–81)	51.3 ± 10.3 (32–71)	.005†
Body mass index (kg/m2)*	31.7 ± 6.2 (20.8–45.5)	27.3 ± 51.1 (19.9–37.7)	.002†
Postmenopausal (n)	13 (31%)	15 (50%)	NS‡
Parity (n)*	2.6 ± 1.3	2.7 ± 1.9	NS†
Ethnicity (n)
White	18 (44%)	16 (53%)	NS‡
Hispanic	13 (32%)	10 (33%)
Other	10 (25%)	4 (13%)

NS, not significant.

^*Data are given as mean ± SD (range).

^†Two sample t-test.

^‡Fisher’s exact test.

TABLE 2.

Surgical procedures that were performed and postoperative complication rates: PR (n = 30) and no PR (n = 43) groups

Procedure	No PR (n)	PR (n)	P value*
Anterior colporrhaphy	6 (14%)	16 (41%)	<.001
Vaginal apex repair†	23 (54%)	26 (87%)	.005
Vaginal hysterectomy	17 (58%)	18 (60%)	NS
Total abdominal hysterectomy	5 (12%)	0	NS
Abdominal repair of vaginal apex	2 (5%)	1 (3%)	NS
Abdominal paravaginal defect repair	15 (35%)	12 (40%)	NS
Vaginal paravaginal defect repair	1 (2%)	3 (10%)	NS
Incontinence procedure	43 (100%)	27 (90%)	NS
Retropubic urethropexy (Burch & MMK)	29 (67%)	16 (59%)
Fascial sling	10 (23%)	7 (26%)
Other	4 (9%)	4 (15%)
Postoperative complications	8 (18%)	6 (20%)	NS

NS, not significant.

^*Fisher’s exact test

^†Includes uterosacral, iliococcygeus, and sacrospinous ligament suspensions.

There were no differences in mean preoperative PISQ scores between the PR and no PR groups (88.5 ± 12.7 and 89.6 ± 12.2, respectively; P = .72). Similarly, there were no differences in mean postoperative PISQ scores that compared the PR and no PR groups (93.5 ± 11.9 and 95.8 ± 11.9, respectively; P = .41). This finding was verified by ANCOVA to control for baseline imbalance in individual scores that may have existed between the 2 groups. With this method, there were still no group differences between the PR and no PR postoperative PISQ scores (P = .46). Mean postoperative PISQ scores improved after surgery in both the PR group (93.47 from 88.5; P =.01) and the no PR group (95.8 from 89.6; P = .002).

Regression analysis of variables that differed between the PR and no PR groups (age, body mass index, point C on POP-Q examination, and performance of anterior colporrhaphy or vaginal apical suspension) showed that none of these variables affected the change in PISQ scores (all P ≥ .07). Thus, the variables that differed between the PR and no PR groups did not confound PISQ scores.

Changes in domain scores for the 2 groups were analyzed. Both the PR group and no PR group had improved scores in the physical domain, which measures physical effects of incontinence and prolapse on sexual function (mean change, 3.51 ± 5.25 points and 5.0 ± 6.45 points, respectively; both P = .001). Only the no PR group had improvement in the partner-related domain (mean change, 1.26 ± 3.22 points; P = .01). Neither group showed improvement in the behavioral domain. There were no differences between groups in postoperative domain scores (all P = not significant). Again, ANCOVA was used to control for possible differences in baseline scores and confirmed that there were no differences between groups regarding postsurgical domain scores (all P = not significant).

The 31 questions in the PISQ were evaluated individually for differences between the PR and no PR groups. All questions with significant within-group changes are listed in Table 3. Postoperative responses to 4 questions showed improvement in both the PR and no PR groups. Three questions improved only in the PR group; 2 questions improved only in the no PR group. Responses to the question “Is your vaginal opening so tight that sexual intercourse cannot occur?” worsened after the operation for both groups. Nineteen women noted new onset “vaginal tightness” after the operation: 10 of 43 women (23%) in the no PR group and 9 of 30 women (30%) in the PR group (P = .59). When these 19 women were compared with the 49 women who did not complain of vaginal tightness before or after the operation, there were no demographic or surgical differences or differences in intraoperative or postoperative complications (all P > .05).

TABLE 3.

Preoperative to postoperative changes in numbers of women with positive responses to individual PISQ questions

Question (answer)	Group*	Preoperative (n/N)	Postoperative (n/N)	P value†
Improved in both groups
Does fear of incontinence (either stool or urine) restrict your sexual activity? (always/usually/sometimes)	No PR	17/42 (40%)	4/42 (10%)	.003
	PR	8/29 (28%)	0/29	.005
Does fear of embarrassment because of incontinence restrict your sexual activity? (always/usually/sometimes)	No PR	18/42 (43%)	3/42 (7%)	<.001
	PR	7/29 (24%)	1/29 (3%)	.014
Do you avoid sexual intercourse because of bulging in the vagina (the bladder, rectum, or vagina falling out)? (always/usually/sometimes)	No PR	14 (33%)	5 (12%)	.029
	PR	15/29 (52%)	2/29 (7%)	<.001
Are you incontinent of urine with sexual activity? (always/usually/sometimes)	No PR	19/41 (46%)	4/41 (10%)	<.001
	PR	9/29 (31%)	2/29 (7%)	.020
Worsened in both groups
Is your vaginal opening so “tight that sexual intercourse cannot occur? (extremely tight/pretty tight/somewhat tight)	No PR	3 (7%)	11 (26%)	.029
	PR	2 (7%)	11 (37%)	.003
Improved in PR group only
How frequently do you feel desire? This feeling may include wanting to have sex, planning to have sex, feeling frustrated because of lack of sex, and so forth. (daily/weekly)	No PR	30 (70%)	31 (72%)	NS
	PR	15 (50%)	21 (70%)	.014
How satisfied are you with the variety of sexual activities in your current sex life? (always/usually)	No PR	30/42 (71%)	30/42 (71%)	NS
	PR	17/29 (59%)	23/29 (79%)	.034
Do you think that your partner avoids sexual intercourse with you because of your problems with incontinence or bulging (either the bladder, rectum, or vagina falling out)? (always/usually/sometimes)	No PR	12/42 (29%)	6/42 (14%)	NS
	PR	9 (30%)	2 (7%)	.008
Improved in no PR group only
Do you feel pain during sexual intercourse?‡ (always/usually/sometimes)	No PR	23 (53%)	12 (28%)	.005
	PR	13 (43%)	17 (57%)	NS
Do you engage in anal or oral sex because vaginal activity is uncomfortable for any reason? (always/usually/sometimes)	No PR	12 (28%)	3 (7%)	.007
	PR	5 (17%)	2 (7%)	NS

NS, not significant.

^*PR group = 43 women; no PR group = 30 women, unless otherwise indicated.

^†McNemar’s test.

^‡Single question with postoperative differences between groups.

The single question that differed after the operation between the no PR and PR groups was the question pertaining to dyspareunia. The question, “Do you feel pain during sexual intercourse?” evaluates dyspareunia. Answers of “always,” “usually,” or “sometimes” were considered positive for dyspareunia, and “rarely” or “never” were considered negative. Equal numbers of women in the no PR and PR groups reported dyspareunia before the operation (23 of 43 women [53%] and 13 of 30 women [43%], respectively; P = .48). However, there was a difference in postoperative reports of dyspareunia between groups (12 of 43 women [28%] in the no PR group and 17 of 30 women [57%] in the PR group; P = .02); the no PR group’s decreased dyspareunia (53%-28%) was statistically significant, whereas the PR group’s trend toward increased dyspareunia (43%–57%) did not reach statistical significance (Table 3). Because others have found Burch procedures to be associated with dyspareunia,⁶ logistic regression was performed to evaluate the effect of retropubic urethropexy on postoperative dyspareunia, and it found no effect (P = .26).

COMMENT

In this study, both the PR and no PR groups had significant postoperative improvement in PISQ scores. Because the 2 groups had similar improvement in postoperative PISQ scores and did not differ in preoperative scores, there was no postoperative difference observed between the total scores of the 2 groups.

An analysis of specific questions revealed that the only item with between-group differences was in the report of dyspareunia. Dyspareunia decreased after the operation in the no PR group and trended toward an increase in the PR group. Levatorplasty, an historic method of correction of rectocoele that has been abandoned largely because of reports of significant dyspareunia,^5–7 was not performed in any of our patients. Previous studies describe varying results regarding the association of dyspareunia and PR that was thought, in part, to be due to the performance of levatorplasty. Previously, some investigators have reported decreased prevalence of dyspareunia after site-specific PR,⁷ whereas other investigators have reported increased dyspareunia after PR.^3–6 Post-hoc analysis of our study sample determined that changes in the prevalence of dyspareunia would need to exceed 33% to reach statistical significance. A larger sample size is needed to detect smaller changes in dyspareunia after the operation.

Notably, both the no PR and PR groups reported increased vaginal “tightness” after the operation. Previous studies that have evaluated discomfort or “tightness” after prolapse surgery have attributed these symptoms to colporrhaphy, particularly PR.^12,13 However, in our study, vaginal tightness precluding sexual intercourse increased, regardless of whether the women underwent posterior colporrhaphy, and was not related statistically to any other single surgical procedure. The increase in vaginal tightness may have been due to synergism between multiple procedures,⁶ the performance of perineorrhaphy, or psychologic factors such as apprehension.

In the evaluation of questions with a Likert scale, the choice of a cut-point can affect outcomes. Weber⁶ chose to define the presence of dyspareunia with answers of “usually” “or always” on another questionnaire with a 5-point Likert scale. We chose to define presence of dyspareunia with answers of “always,” “usually,” or “sometimes,” and its absence with answers of “seldom” or “never.” This broader definition explains the higher prevalence of preoperative and postoperative dyspareunia in our population, compared with percentages reported by Weber. We believe our cut-point is appropriate for this study both statistically and clinically. We chose to group the intermediate answer on the Likert scale (“sometimes”) with the poorer sexual function or the adverse outcome group. The favorable and unfavorable outcome groups were divided more equally by this method when we considered all answers to the PISQ. We also chose an inclusive definition that was sensitive to the detection of negative sexual function outcomes.

Despite the postoperative difference in pain with intercourse that was found between these 2 groups, improvement in total PISQ scores was similar. Although authors have equated the presence of pain with sexual dysfunction, overall sexual function seems to be more complex than this. Sexual function, like many other quality of life issues, is multifactorial. Similar to other concepts that are measured by quality-of-life scores, sexual function may be better assessed by the evaluation of its many components rather than a single one, such as dyspareunia. Most patients in both the PR group (90%) and no PR group (100%) underwent urinary incontinence surgery and PR. After the operation, both groups significantly improved in questions that related to urinary incontinence, as illustrated by improvement in the physical domain of the PISQ. The negative effect urinary incontinence has on sexual function has been noted previously,^4,14 but reported effects of surgical repair of incontinence have varied.^4,15,16 After the operation, our patients reported improvement in sexual function regarding questions that were related to incontinence, with no differences in improvement reported between women with and without PR.

A limitation of our study was that the performance of PR was not standardized, nor was perineorrhaphy performance specifically recorded. Although all the institutions that were involved in this study reported that perineorrhaphy is performed commonly at their sites, their number was not documented explicitly. Another limitation is the number of subjects who were available for analysis and the potential effect of confounders. We performed a post-hoc power analysis and found that our sample size had 80% power to detect a 7.5 point (8%) difference in total PISQ scores between the PR and no PR groups, which is a difference that we think is clinically significant. We also controlled for confounders of the total PISQ scores using logistic regression. We were not able, however, to control for confounders in our analysis of individual questions. Thus, the interpretation of individual questions must be tempered by the unknown effects of possible confounding variables.

A further limitation was our follow-up of 3–6 months. Long-term follow-up evaluation of postoperative sexual function is needed. Finally, there is a potential for selection bias in our population, because this is a secondary analysis of sexual function in women who participated in a randomized controlled trial. Strengths of the study include its prospective, multicenter design and the use of a validated and condition-specific questionnaire with intermediate length of follow-up visits.

In summary, our work demonstrates an improvement in sexual function after prolapse and/or incontinence surgery that is not affected by the performance of PR. Dyspareunia improves after surgery in patients who do not receive PR, but this difference does not lead to a significant additional improvement in sexual function scores when compared with those patients who receiving PR. We hypothesize that the positive effect of surgery on incontinence and prolapse may overshadow the smaller, potentially detrimental affects of PR.