Potentially relevant studies were identified by a computerized search of the Medline electronic database (PubMed, 1966–2008). Relevant text and keywords were: incontinence OR stress urinary incontinence OR female incontinence, AND randomized controlled trial OR controlled trial OR meta-analysis. The search was limited to the English-language literature. Several recently updated Cochrane reviews represented a starting point for the evidence-based evaluation of the treatment of female SUI. In their assessment, relevant studies were identified from a register of controlled trials gathered from Medline, CINAHL, the Cochrane Central Register of Controlled Trials, and hand-searching of journals and conference proceedings. Details regarding review methods, including identification of primary studies, quality assessments, and data extraction, are described in each Cochrane review. The findings in the Cochrane reviews were augmented by additional meta-analyses and randomized controlled trials (RCTs) located by a hand-search of the Medline database by the authors.

Despite RCTs and meta-analyses of RCTs representing the highest level of evidence, interpretation of these types of trials poses some challenges. First, lack of blinding procedures is common in the surgical literature and may not be practical in many trials. Second, randomization procedures may be inappropriate or simply not described. Third, outcome measures, such as nonvalidated questionnaires, may not be appropriate. Fourth, short-term follow-up periods may not be representative of durable, long-term cure rates. Fifth, significant heterogeneity may exist between trials, contributing to wide confidence intervals (CI) (95% CI). Additionally, dropout or failure-to-follow-up rates may be substantial in some trials. Furthermore, many trials are small and underpowered to detect a significant difference between two treatment arms.

Stress urinary incontinence literature presents additional unique challenges during interpretation. First, patient populations may be different between trials. Women may present with SUI only or with mixed (urge and stress) incontinence. Some women may have also undergone previous anti-incontinence surgery or have pelvic organ prolapse (POP) eligible for concomitant repair. Second, the type of SUI may be different, as well. Women may have the symptoms of SUI based on clinical evaluation alone or urodynamic stress incontinence. Additionally, some women may have “occult” SUI, which is diagnosed only after concomitant POP has been reduced. Third, extensive procedural variations may exist even for established surgeries, making a meta-analysis difficult. Fourth, the generalizability of the results of a particular operation may be brought into question, as the results of many trials represent the experience of a single surgeon or single institution. Additionally, the follow-up period may be variable. For the purpose of this chapter, follow-up periods are defined as short term (< 12 months), medium term (12–60 months), and long term (> 60 months).

Perhaps the most difficult issue in evaluating SUI literature is a consistent definition of success or cure. Measures may be subjectively based on a woman’s report or a validated questionnaire or objectively based on a voiding diary, pad tests, cough stress testing or urodynamics. At present, there is no consensus on which measures are vital to a trial’s validity, but it is widely accepted that both subjective and objective outcomes should be reported. In some trials, definitions of success may include improved patients as well, while in some studies, cure of SUI may be at the expense of worsened postoperative urge incontinence or an increased incidence of obstructive symptoms and urinary retention. Unfortunately, issues germane to outcomes of anti-incontinence procedures, such as a review of perioperative complications, will not be addressed in this chapter due to space constraints.

The evidence

Buttressing procedures such as an anterior vaginal repair (AR) involve the plication of the pubocervical fascia in the midline. A Cochrane review (updated 28 February 2007) summarized the outcomes of 10 trials encompassing 1012 women, of whom 385 underwent AR [5]. All but one of the trials excluded women who had previously undergone anti-incontinence surgery, and one trial comparing 16 women after AR versus pelvic floor muscle training did not provide enough data for statistical analysis.

In eight trials comparing AR to RBNS, 353 of 627 women were randomized to RBNS. For statistical analysis, women undergoing a Marshall–Marchetti–Krantz procedure (MMK) in one study were combined with the remainder of women who underwent a Burch. Subjective failure rates were lower for women undergoing RBNS at all follow-up periods: short term (10% vs 19%; relative risk (RR) 0.51, 95% CI 0.34–0.76), medium term (16% vs 36%; RR 0.43, 95% CI 0.32–0.57), and long term (28% vs 53%; RR 0.49, 95% CI 0.32–0.75). Likewise, objective cure rates were lower for women undergoing RBNS at all follow-up periods: short term (10% vs 25%; RR 0.41, 95% CI 0.27–0.63), medium term (19% vs 44%; RR 0.40, 95% CI 0.30–0.53), and long term (26% vs 54%; RR 0.48, 95% CI 0.31–0.73).

In six trials comparing AR with Burch in women with concomitant POP, women undergoing Burch had lower subjective failure rates than women undergoing AR, both in medium-term (RR 2.49, 95% CI 1.83–3.39) and long-term (RR 3.39, 95% CI 1.4–8.22) follow-up. More women who underwent AR (25 of 107, 23%) required repeat anti-incontinence surgery when compared with women undergoing Burch (4 of 164, 2%) (RR 8.87, 95% CI 3.28–23.94). In three trials, women undergoing AR (64 of 181, 35%) and TNS (50 of 156, 32%) had similar subjective incontinence rates after 1 year (RR for failure 1.16, 95% CI 0.86–1.56); however, these trials may have been underpowered to detect a statistical difference. There were no trials comparing AR with sham procedure, laparoscopic RBNS, PVS or MUS.

The evidence

In addition to the four principal TNS procedures (see Box 21.1), some surgeons have modified these operations by varying the site of initial approach (abdominal vs vaginal) and incorporating spacers or sheaths. A Cochrane database (updated 11 May 2008) included 10 trials encompassing 864 women, 375 of whom underwent one of six different TNS procedures [6]. In addition to the principal TNS, trials also included a Raz four-corner repair and a “modified” Raz procedure, where needle sutures were placed under direct vision [7]. A comparison of TNS with AR is included in Clinical Question 21.

There were no studies comparing TNS with a sham procedure or conservative intervention. In seven trials, TNS was compared with RBNS (five Burch, one obturator shelf repair, and one MMK) [6,8]. Of 570 women, 278 underwent RBNS. Fewer women did not achieve subjective cure after RBNS (10.8%) than after TNS (15.3%) in short-term follow-up (RR 0.66, 95% CI 0.42–1.03). A similar relationship was seen in medium-term (13.7% vs 22.7%; RR 0.56, 95% CI 0.39–0.81) and long-term (18.2% vs 56.7%; RR 0.32, 95% CI 0.15–0.71) follow-up. Likewise, fewer women did not achieve objective cure after RBNS than after TNS in short-term (8.7% vs 14.4%; RR 0.56, 95% CI 0.32–0.97), medium-term (12.9% vs 21%; RR 0.59, 95% CI 0.40–0.88), and long-term follow-up (18.2% vs 56.7%; RR 0.32, 95% CI 0.15–0.71) from a single trial. In three trials including only women who had not previously undergone an anti-incontinence surgery, fewer women failed after RBNS (23 of 190, 12%) than after TNS (57 of 218, 26%) in medium-term follow-up (RR 2.37, 95% CI 1.54–3.66). Four small trials that included some women who had previously undergone anti-incontinence surgery also revealed that more women failed after TNS (34 of 95, 36%) than after RBNS (24 of 107, 22%) (RR 1.61, 95% CI 1.04–2.49). A single comparison between TNS and PVS was underpowered to detect a statistical difference. One small trial compared a standard Raz suspension with a modified Raz procedure (“transvaginal Burch”) [7]. While outcomes were similar, the trial was underpowered to detect a statistical difference.

The evidence

Retropubic bladder neck suspensions have been evaluated extensively in a Cochrane database (updated 25 May 2005), encompassing 39 trials and 3301 women [8]. Dropout rates ranging from 1.3% to 28.4% have been reported in 19 of these trials. In two small trials, women undergoing RBNS had higher subjective (RR 0.24, 95% CI 0.08–0.71) and objective (RR 0.26, 95% CI 0.13–0.53) cure rates than women undergoing pelvic floor muscle training. A comparison of RBNS and AR is provided in Clinical Question 21.1, while a comparison of RBNS and TNS is in Clinical Question 21.2.