The combination of prolapse surgery with an incontinence procedure can reduce the incidence of stress urinary incontinence (SUI) after surgery, but may increase adverse events. We compared the effectiveness and safety of prolapse surgery versus combined prolapse and incontinence surgery in women with pelvic organ prolapse.
Design and setting
Pubmed, EMBASE, DARE, the Cochrane Library and the register of Current Controlled Trials were searched for randomised trials (restricted to Burch colposuspension and midurethral sling as incontinence procedure) from 1995 to 2013 limited to the English literature.
Two reviewers selected eligible articles and extracted the data. Pooling for SUI was based on three patient groups: (1) women with coexisting SUI; (2) women asymptomatic for SUI; and (3) women with occult SUI. For adverse events, pooling was based on incontinence procedure.
Main outcome measures
The primary outcome was SUI. The secondary outcomes were treatment for SUI, bladder storage symptoms, obstructive voiding and adverse events.
Seven trials were included. Pooling for women with coexisting SUI was possible for objective SUI with two studies and showed no difference. Statistical (I2 = 95%) and clinical heterogeneity was, however, high. The largest study showed a lower incidence of persisting SUI (5% versus 23%) and treatment for this (0% versus 57%) in women who underwent prolapse repair with a midurethral sling. The second study did not find a difference in women undergoing a sacrocolpopexy with or without Burch colposuspension. In asymptomatic women, combination surgery resulted in a lower incidence of de novo subjective SUI (two studies; 24% versus 41%; relative risk [RR], 0.6; 95% confidence interval [CI], 0.3–0.9; I2 = 36%) and the need for subsequent anti-incontinence surgery (three studies; 2% versus 7%; RR, 0.4; 95% CI, 0.2–0.8; I2 = 13%). For the outcome objective SUI, pooling was possible for five studies, but statistical heterogeneity was high (I2 = 82%) and the difference was not statistically significant. In the subgroup of women with occult stress incontinence, we found a lower incidence of objective SUI after combination surgery (two studies; 22% versus 52%; RR, 0.4; 95% CI, 0.3–0.8; I2 = 32%). There were no differences in bladder storage symptoms, urgency incontinence or long-term obstructive voiding symptoms. Adverse events (two studies; 15% versus 10%; RR, 1.6; 95% CI, 1.0–2.5; I2 = 0%) and prolonged catheterisation (three studies; 6% versus 1%; RR, 4.5; 95% CI, 1.5–13.3; I2 = 0%) were more frequent after vaginal prolapse repair with a midurethral sling.
Combination surgery reduces the risk of postoperative stress incontinence, but short-term voiding difficulties and adverse events were more frequent after combination surgery with a midurethral sling.
Women frequently report symptoms of stress urinary incontinence (SUI) after correction of pelvic organ prolapse (POP). About 40–50% of the women with POP also report SUI before surgery, and these women are at highest risk to encounter SUI postoperatively.[1, 2] Continent women can also develop SUI after surgery.[3-5] The risk is assumed to be highest (up to 80%) in the 20–30% of continent women with occult SUI, and is believed to be associated with kinking or compression of the urethra by the prolapse.[3-7] The combination of prolapse surgery with an anti-incontinence procedure to treat or prevent SUI is frequently considered in three patient groups: (1) women with coexisting SUI; (2) women asymptomatic for SUI; and (3) women asymptomatic for SUI with occult SUI.
There are two main methods to repair POP: (1) vaginally, e.g. with fascial repair or MESH reinforcement; or (2) abdominally with sacrocolpopexy. In the treatment of SUI, there are two procedures considered to be gold standard because of similar cure rates: (1) vaginally with the introduction of a midurethral sling (MUS); or (2) abdominally with a Burch colposuspension. When combination surgery is performed vaginally, repair is often combined with a MUS, whereas sacrocolpopexy is more likely to be combined with a Burch colposuspension.
Recent studies have reported a lower incidence of postoperative SUI after combination surgery relative to prolapse surgery alone.[8, 10, 11] However, studies on possible adverse events of combination surgery, such as overactive bladder symptoms, obstructive voiding and complications, have shown conflicting results.[12-20] Therefore, the balance of risk and benefit of combination surgery is unclear.
The aim of this systematic review and meta-analysis is to provide an overview of the evidence comparing prolapse surgery only with combination surgery in women with POP with respect to postoperative urinary incontinence (UI), bladder storage symptoms, voiding dysfunction and complications.
Randomised controlled trials (RCTs) that compared prolapse surgery only with combination prolapse and incontinence surgery were eligible for inclusion. Prolapse surgery was defined as vaginal repair (conventional or with MESH) or sacrocolpopexy. Studies that only included obliterative prolapse procedures, such as Le Fort colpocleisis, were excluded, because of the selective use of these procedures in women of very great age or with severe co-morbidity. Combination surgery was defined as prolapse surgery with incontinence surgery. We included Burch colposuspension or any synthetic MUS. Other incontinence procedures (such as Kelly plication) were excluded, because of the varying effectiveness of these procedures and because MUS and Burch colposuspension are the most practised procedures and the gold standards for the treatment of SUI. At least one of the following outcome measures should have been described as a sign or symptom: UI (e.g. SUI or treatment for UI); bladder storage symptoms (e.g. frequency or nocturia); voiding symptoms (e.g. incomplete emptying or retention); or adverse events or complications (e.g. bladder injury or blood loss). Follow-up had to be at least 3 months.
Information sources and search strategy
We searched Pubmed, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), the Database of Abstracts of Reviews of Effectiveness (DARE) and the register of Current Controlled Trials. The search strategies consisted of MESH/Emtree terms and keywords related to POP and SUI (Appendix S1). Limits were used for language (English), humans and type of article (RCTs only). We also limited for publication date (1 January 1995 to 31 December 2012) because the MUS procedure was introduced around 1996. Reference lists of eligible primary studies, reviews and meta-analyses were checked for additional articles.
Two reviewers (AvdS and JMvdP) independently screened the titles and/or abstracts of all articles retrieved. Of the potentially relevant articles, the full text was assessed for eligibility. Any disagreement about inclusion or exclusion was resolved by a consensus meeting between the reviewers.
AvdS and JMvdP independently extracted data on patient and study characteristics, trial quality and outcomes using a data extraction sheet as advised by the Cochrane Consumer and Communication Review Group (Appendix S2). All authors of selected publications were asked by email and/or post whether any further outcome measures not reported were available, and the final data extraction sheet was presented to them as a final check.[10, 11, 23-27]
Data synthesis and pooling
We pooled the outcome measures related to UI based on preoperative incontinence status, because we assumed that this is the most important factor in developing postoperative UI. Based on incontinence status, we defined the following patient groups that closely relate to clinical practice: (1) women with symptomatic coexisting SUI; (2) women asymptomatic for SUI (with or without occult SUI); and (3) women asymptomatic for SUI with occult SUI. We pooled the outcomes for the different anti-incontinence procedures, because they are equally effective in the treatment of SUI.[9, 28, 29] We also pooled different methods of prolapse repair, because surgical techniques will differ much between studies and surgeons anyway. We assumed that adverse events and bladder emptying would be related more to surgical procedure than to preoperative incontinence status. We therefore stratified only for incontinence procedure (MUS or Burch).
As in most included RCTs, the primary outcome was SUI. The secondary outcomes were treatment for SUI, bladder storage symptoms, obstructive voiding and adverse events. Definition was dependent on the descriptions used in the studies. Prolonged catheterisation of the bladder after surgery was defined as at least 1 week. Where possible, we used terminology as proposed by the International Urogynecological Association (IUGA) and International Continence Society (ICS). When outcome data about more than one follow-up period were available, we used the longest follow-up period.
If at least two RCTs were available, data were pooled. For dichotomous outcomes, we reported relative risk (RR) with 95% confidence intervals (CIs). In the case of substantial heterogeneity, defined as I2 > 25%, a random effect model was applied. Review Manager 5.1 (The Cochrane Collaboration, Copenhagen, Denmark) was used for all analyses.
The risk of bias of the selected studies was judged on the basis of the ‘risk of bias’ table developed by the Cochrane Collaboration. It includes the domains sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting and other issues. Risk of bias is categorised as low, high or unclear risk of bias.
In Figure 1, the process of literature identification and study selection is summarised. Of 1522 abstracts, 23 articles were potentially relevant. We excluded eight studies because of study design, six after reading the full text[17, 19, 31-34] and two after communication with the author.[35, 36] The remaining 15 publications describing seven RCTs were included, reporting on 1115 women.[5, 10, 11, 23-27, 37-43]
Details of the studies are summarised in Table 1 and Appendix S3. Combination surgery was compared with prolapse surgery in two trials with women with symptomatic coexisting UI.[23, 25] One combined vaginal prolapse repair with MUS and one combined sacrocolpopexy with Burch colposuspension. In five trials, combination surgery was compared with prolapse surgery in continent women.[10, 11, 24, 26, 27] In one of these, only women with occult SUI were included and, in one, women without any form of objective or subjective UI. Two of these five studies combined sacrocolpopexy with Burch colposuspension and three studies combined vaginal prolapse repair with MUS. Although the definition differed widely, all studies reported on the prevalence of persistent or de novo (S)UI after surgery (Appendix S3).
The risk of bias of the trials is summarised in Appendix S3. Two studies were considered to be of high quality[10, 11] and one study was assessed to be of low quality. We had too little information about the study by Schierlitz et al. to adequately judge the quality of this study. Contact was achieved with all but one author. Four of the six authors confirmed the data presented in the extraction sheet.[10, 11, 23-25] From one study, no power analysis was available. The other studies were adequately powered.[10, 11, 23, 24, 26]
Five of the seven RCTs found significantly lower rates of postoperative (S)UI after combination surgery relative to prolapse surgery.[10, 11, 23, 26, 27] Below, we have divided the outcomes into three categories: (1) women with POP and coexisting SUI; (2) women with POP asymptomatic for SUI; and (3) women with POP asymptomatic for SUI with occult SUI.
Women with POP and coexisting SUI
Two studies compared combination surgery with prolapse surgery in women with coexisting SUI.[23, 25, 40] The trial by Borstad et al. compared combination surgery with a two-step approach, in which, 3 months after prolapse surgery, women with persisting SUI were offered a MUS. In the intention-to-treat analysis, combination surgery significantly reduced the risk of postoperative SUI at 1 year of follow-up (5% versus 23%). In the two-step arm, 57% (54/94) of women had persisting SUI for which a MUS was performed within one year, 15% (14/94) had SUI but declined further surgery and 28% (26/94) no longer had SUI. Costantini et al.[25, 40] compared Burch colposuspension with sacrocolpopexy versus sacrocolpopexy only in women with POP and symptomatic UI. The preoperative incontinence status was diverse (13 SUI, 30 mixed UI and four occult SUI). They observed no difference in postoperative SUI after a median follow-up of 82 months: 30% (7/23) versus 27% (6/22). Pooling the studies in Figure 2 showed no significant difference in objective SUI (10% [11/110] versus 63% [73/116] in two studies; RR, 0.3; 95% CI, 0.0–5.6; random effects model). However, statistical heterogeneity was high (I2 = 95%) and there were important differences between the studies: (1) women with objective and symptomatic SUI versus women with mixed UI; (2) protocol tension-free vaginal tape (TVT) at 3 months in the case of persisting SUI versus additional therapy conforming to daily practice; (3) vaginal prolapse surgery with MUS versus a sacrocolpopexy with Burch; and (4) a relatively large study versus a relatively small study.
Women with POP asymptomatic for SUI
As shown in Figures 2 and 3, pooled analysis of all studies in continent women showed that combination surgery reduced the risk of postoperative objective SUI (9% [32/373] versus 20% [81/399] in five studies; RR, 0.5; 95% CI, 0.2–1.4; random effects model), subjective SUI (24% [45/190] versus 41% [80/194] in two studies; RR, 0.6; 95% CI, 0.3–0.9; random effects model) and the need for subsequent anti-incontinence surgery (2% [9/389] versus 7% [27/405] in four studies; RR, 0.4; 95% CI, 0.2–0.8; fixed-effect model). For the outcome objective SUI, statistical heterogeneity was high (I2 = 82%) and the difference was not statistically significant because of a counter-intuitive result in the trials by Costantini et al. and Brubaker et al. The study by Costantini et al.[24, 39] was relatively small, had the longest follow-up period and was the only study in which all women were completely dry preoperatively (no symptom, no sign and no occult SUI). The outcome in the CARE trial is remarkable because all other outcome measures in this study concerning SUI were less frequent after combination surgery.
The number needed to treat (NNT) was six to prevent one woman developing de novo subjective SUI after prolapse repair, and 20 to prevent one woman undergoing an additional MUS.
Women with POP asymptomatic for SUI with occult SUI
Three studies reported on postoperative (S)UI in women with occult SUI.[11, 26, 37] The prevalence of occult SUI was 27% in the CARE trial and 33% in the OPUS trial.[11, 37] Both studies performed a subanalysis of women with occult SUI and showed lower rates of postoperative (S)UI after combination surgery (Table 2). The OPUS trial only reports on UI and not on SUI, and is excluded from the pooled analysis. Pooling showed in Figure 2 that combination surgery significantly reduces the risk of objective SUI compared with prolapse surgery (22% [18/81] versus 52% [41/79] in two studies; RR, 0.4; 95% CI, 0.3–0.8; random effects model). Statistical heterogeneity was (moderately) high (I2 = 32%); however, all RR favoured combination surgery.
Table 2. De novo (stress) urinary incontinence ([S]UI) in patients asymptomatic for SUI preoperatively: with occult SUI versus without occult SUI
Postoperative SUI in women WITH occult SUI, % (n/N)
Postoperative SUI in women WITHOUT occult SUI, % (n/N)
ARR, absolute risk reduction; NA, not available; NNT, number needed to treat.
The NNT to prevent one woman with occult SUI from developing de novo objective SUI after prolapse repair was three (Table 2).
Postoperative bladder storage
All trials, except the study by Borstad et al. reported on bladder storage symptoms, but the definitions differed considerably. None found a significant difference between combination surgery and prolapse surgery only. The OPUS and CARE trials both reported on bladder storage symptoms measured with a validated questionnaire, and both found no increase in symptoms after combination surgery.[10, 11] As a result of differences in the definitions used, pooling was only possible for urgency UI (UUI) in continent women, which showed no statistically significant difference (6% [14/224] versus 10% [22/226] in three studies; RR, 0.7; 95% CI, 0.4–1.2; fixed effects model) (Figure 4). Statistical heterogeneity was low (I2 = 0%).
Postoperative bladder emptying
All but one study reported on voiding capability after surgery, and all found lower rates of prolonged catheterisation after prolapse surgery only. Figure 4 shows the results of pooling studies in women undergoing vaginal prolapse repair with or without a MUS procedure. It shows an increased risk of prolonged catheterisation after combination surgery versus prolapse repair only (6% [17/275] versus 1% [4/290] in three studies; RR, 4.5; 95% CI, 1.5–13.3; fixed effects model). These results were consistently found in all three studies and statistical heterogeneity was low (I2 = 0%). Three studies reported on surgery for persisting obstructive voiding and, overall, about 2% (7/295, range 0–7%) underwent any form of surgery for persistent voiding dysfunction after vaginal prolapse repair with MUS.[11, 23, 26] The CARE and OPUS trials reported on obstructive voiding symptoms at 1 year, and found no differences between combination surgery and prolapse surgery.[10, 11]
Other adverse events
Two studies reported on postoperative complications in a structured manner and found no difference in the percentage of women having one or more serious adverse events (SAEs).[10, 11] However, in the OPUS trial, rates of adverse events plausibly related to the MUS procedure were higher in the sling group than in the sham group: bladder perforation, 7% versus 0%; urinary tract infection, 31% versus 18%; major bleeding complications, 3% versus 0%. Borstad et al. and Costantini et al.[24, 25] did not predefine serious complications, but provided detailed data on the reported adverse events. In the study by Borstad et al., the rate of complications was initially 18% after combination surgery and 5% after prolapse surgery. However, at 1 year, after 53 women in the prolapse group had received a MUS as a second step, this difference decreased to a non-significant difference of 18% versus 11%. As shown in Figure 4, pooling the data showed borderline significantly more SAEs after combination surgery with a MUS relative to prolapse surgery only (15% [39/252] versus 10% [26/266] in two studies; RR, 1.6; 95% CI, 1.0–2.5; fixed-effects model) and no increased risk in SAEs after combining sacrocolpopexy with Burch colposuspension (30% [63/211] versus 33% [70/213] in three studies; RR, 0.9; 95% CI, 0.7–1.2; fixed-effects model). Voiding difficulties were excluded. Statistical heterogeneity was low for both analyses (I2 = 0%).
Two studies compared combination surgery with prolapse surgery only in women with POP and coexisting SUI. The largest study showed clear evidence for a reduced risk of postoperative SUI after combination surgery relative to vaginal prolapse surgery only. The other smaller study found no difference between sacrocolpopexy with or without Burch. We found a reduced risk of postoperative SUI in women asymptomatic for SUI. The higher risk of de novo SUI in women with occult SUI suggests that there is a particular benefit of combination surgery in this group. We found no evidence for a higher prevalence of bladder storage symptoms or UUI after combination surgery relative to prolapse surgery. For vaginal prolapse repair, the risk of prolonged bladder catheterisation is higher after combination surgery with a MUS than after prolapse surgery only. The rate of SAEs related to the insertion of a MUS also appeared to be higher after combination surgery. For abdominal sacrocolpopexy, pooling was not possible for prolonged bladder catheterisation, but the included studies showed no differences. We found no difference in SAE after sacrocolpopexy with or without Burch.
Strengths and limitations
Our study has some weaknesses that must be acknowledged. Although we performed two independent searches with wide search terms, we might have failed to locate studies. However, we also checked the Cochrane review published in 2013 and found no additional articles.
The selected studies had some flaws. Details on study and patient characteristics were not reported equally well and only two of the RCTs were assessed to be of high quality.[5, 11] Furthermore, as a result of important differences in population size and characteristics, statistical heterogeneity was sometimes high. This was especially the case for the outcome objective SUI in asymptomatic women. Further exploration of heterogeneity revealed, however, that most studies showed very similar results. The possibility to pool data was limited because of important differences in preoperative incontinence status and surgical procedure, and a lack of uniform outcome measures. We have addressed these differences in the review where possible. To our knowledge, this is the most comprehensive meta-analysis on this topic, in which the categorisation of women facilitates clinical application of the results.
Interpretation in the light of other evidence
This meta-analysis confirms earlier studies concluding that the addition of an incontinence procedure to prolapse surgery results in lower rates of postoperative SUI, and the results are also in line with the newest Cochrane review published in 2013.[8, 12, 35, 44, 45] Some important differences between our review and the Cochrane review can be recognised: (1) we only included studies that used a MUS or Burch colposuspension as an anti-incontinence procedure because, today, these procedures are considered to be gold standards in the treatment of SUI; (2) we only included studies that compared combination surgery with prolapse surgery and not with prolapse surgery with classic anti-incontinence procedures such as a Kelly plication; and (3) we not only focused on the benefits, but also systematically on the possible negative effects of combination surgery, making it possible to balance the risks and benefits.
We found that the NNT to prevent one woman developing de novo objective SUI was higher in all continent women (9% versus 20%, NNT 9) than in continent women with occult SUI (22% versus 52%, NNT 3) or in women with coexisting SUI (10% versus 63%, NNT 2). This supports the idea that combination surgery might be especially beneficial in women with coexisting or occult SUI. Earlier reviews and studies suggested the same, but evidence for this is still limited.[2, 46-48] It remains a problem that there is no consensus in diagnosing occult SUI, and different methods to detect occult SUI were used in the studies included in our review.
The percentage of women undergoing additional anti-incontinence surgery after prolapse surgery differs widely: 28% in women with coexisting SUI, 4–16% in asymptomatic women without occult SUI and 1.5–24% in women with occult SUI.[6, 7, 31, 35, 49, 50] Based on the five studies including asymptomatic women (19/85) and those included in our meta-analysis (27/124), we estimated roughly that, overall, about 22% (46/209) of the women with de novo SUI would request a subsequent anti-incontinence procedure.[6, 7, 10, 11, 24, 27, 31, 49, 50]
This meta-analysis showed that the incidence of adverse events related to tape placement was higher in the group undergoing combination surgery relative to vaginal prolapse surgery only. There are two possible reasons for this increased risk: (1) the presence of POP is an isolated risk factor for the development of a complication; and/or (2) with the addition of a surgical procedure, the risks also increase. Although several large observational studies found no differences in complications occurring in MUS procedures, POP surgery and combination surgery, it seems difficult to understand that the addition of a procedure does not result in an increase in complications.[13, 15, 19, 20] Other studies have confirmed this and found more voiding dysfunction[12, 16, 18] and more bladder perforation in women undergoing combination surgery compared with the MUS or POP procedure only.
Women planned for correction of genital prolapse should be counselled about the possibility of combination surgery to treat or prevent SUI. They should know that there is strong evidence that postoperative SUI is less frequent after combining prolapse and anti-incontinence surgery relative to prolapse surgery only. However, the NNT to prevent one woman needing an additional anti-incontinence procedure for postoperative SUI is probably considerable. In asymptomatic women, only 7% needed subsequent surgery for de novo SUI. Moreover, the rate of adverse events is very likely to be higher in combination surgery. This was shown in vaginal prolapse surgery with MUS. With an increasing risk of postoperative SUI and the need for an additional anti-incontinence procedure, the benefits might outweigh the risks of combination surgery. Therefore, we would consider combination surgery in women with coexisting and occult SUI, but would not advise it in continent women without occult SUI. Consequently, preoperative screening for coexisting and occult SUI seems wise.
Especially in women with coexisting and occult SUI, further RCTs are needed to estimate the positive effect of combination surgery. Furthermore, large observational studies in clear predefined patient groups may help to verify further how many women experience bothersome SUI, require additional anti-incontinence surgery and experience a complication.
Disclosure of interests
None to declare.
Contribution to authorship
All were closely involved in the study design. JMvdP and AvdS reviewed the literature. JMvdP contacted all authors for further information. JMvdP and KOR performed the data analysis. CHvdV and JPR assisted in the literature review and analysis. JMvdP wrote the first draft, and all were closely involved in revising the article and consented with the final version.
Details of ethics approval
Because this study is a review it was exempt from ethics approval.
We thank the following authors for providing extra information and checking the data extracted by us from their study: Ellen Borstad, Linda Brubaker, Elisabetta Costantini and Max Lazzeri, and John Wei.