Prospective multicentre randomised trial of tension-free vaginal tape and colposuspension as primary treatment for stress incontinence

Introduction

Urinary incontinence is reported by 14% of women, and urodynamic stress incontinence—the involuntary leakage of urine during increased abdominal pressure in the absence of a detrusor contraction—is diagnosed in over half of the women presenting to hospital with urinary incontinence.1^–3 Cure rates of around 50% are reported with physiotherapy,4 and surgery is recommended for those women who fail to respond. Systematic reviews have shown that colposuspension has the best surgical results when compared with other treatments for urodynamic stress incontinence, with cure rates of up to 90% in women who have had no previous surgery for incontinence, although there are only limited data from randomised trials on which to base clinical practice. 5 6 Previous studies have been criticised for insufficient numbers of participants and statistical power as well as lack of standardised criteria for entry to the study and outcome measures.6 Although colposuspension remains the most popular choice for the treatment of stress incontinence, some authors have reported less than half of patients remaining dry and free of complications long term.7^–9 Complications include haemorrhage, haematoma, bladder injury, and urinary tract infection. Up to 20% of women may develop de novo detrusor overactivity 5 9; voiding dysfunction has been reported in 3% to 32% of women, and surgery for vaginal prolapse may be required in 2.5% to 26.7% after the procedure.5

The tension-free vaginal tape procedure is a relatively recent treatment for stress incontinence.10 A polypropylene tape is inserted suburethrally under local anaesthesia with sedation. The procedure is thought to work by providing a pubourethral “neoligament.” Increased intra-abdominal pressure results in a kink at the point of fixation, which prevents urine flow. Data from three early case series suggest objective cure rates of 84-100%, with few complications.10^–12 We compared the tension-free vaginal tape procedure with colposuspension in a prospective randomised manner, on a multicentre basis, using defined valid outcomes, including objective and subjective measures.

Methods

Women with stress incontinence, who had completed their family, were invited to take part. Exclusion criteria were detrusor overactivity, vaginal prolapse requiring treatment, previous surgery for prolapse or incontinence, a major degree of voiding dysfunction (defined at cystometry as a voiding pressure >50 cm H₂0, maximum flow <15 ml/s, and residual urine volume >100 ml), neurological disease, and allergy to local anaesthetic.

The trial was conducted at gynaecology or urology departments in 14 centres in the United Kingdom and Eire, including university teaching hospitals and district general hospitals. Women were recruited from outpatient clinics once surgery had been selected for their stress incontinence.

The sample size calculation was performed assuming a 90% cure rate with colposuspension, and that a 10% difference in cure rate between procedures would be clinically important.5 To detect this level of difference with 80% power would require 197 patients in each arm of the trial; to detect the same difference with 90% power, 262 patients were required in each arm. Calculations based on anticipated throughput of the collaborating units gave a pragmatic recruitment target of 436 patients. Recruitment was limited to the period May 1998 to August 1999 owing to logistic and financial constraints.

Researchers randomised participants via a telephone system, which allocated trial identification number and treatment group. Randomisation was computer generated using blocks of four and six. Owing to differences between the procedures in incision, anaesthesia, and catheterisation, it was not possible to blind investigators or participants to the treatment allocation.

The tension-free vaginal tape procedure was performed as described by Ulmsten, under local anaesthesia and sedation.10 The bladder was not routinely catheterised after surgery. When bladder perforation or injury occurred an indwelling catheter was placed for 48 hours. Colposuspension was performed according to the standard technique used by the units. Before the start of the trial all investigators underwent training in the vaginal tape procedure in a recognised centre. The number of patients required for training varied between surgeons, who began recruiting when they were satisfied with their own technique.

At initial assessment, urodynamic evaluation was performed by medium fill dual channel subtracted cystometry with simultaneous pressure and flow voiding studies or videocystourethrography; in some centres urethral pressure profilometry at rest and at stress was also carried out.13 Patients were asked to complete a urinary diary for one week, and a one hour perineal pad test was performed according to the recommendations of the International Continence Society.3

Patients' perceptions of changes in their symptoms and treatment outcome were measured with the generic SF-36 and the disease specific Bristol female lower urinary tract symptoms questionnaire. 14 15 Six weeks after surgery a postal questionnaire was sent out comprising the SF-36 and questions about recovery. Six months after surgery reassessment was undertaken with symptom review, clinical examination, the one hour pad test, and urodynamic studies. The SF-36 and Bristol female lower urinary tract symptoms questionnaires were completed by the patient in the clinic, and they were given a urinary diary to complete over one week and return by post.

Results

Overall, 344 women were randomised over 15 months (figure). Table 1 shows the baseline characteristics of the two groups. Five patients who underwent surgery violated the protocol exclusion criteria; two in each arm had detrusor overactivity and one in the vaginal tape arm had voiding dysfunction. These patients are included in the analysis. The baseline characteristics of the patients who withdrew before surgery were similar to the participants, with the exception of a lower change in pad weight in the colposuspension group (median weight change for withdrawals 6 g (interquartile range 1-20) v 16 (6-38) for those undergoing surgery; P=0.0185, Wilcoxon rank sum test).

Flow of participants through trial

• Download figure
• Open in new tab
• Download powerpoint

Operative complications were more common after the vaginal tape procedure (table 2), largely injury to the bladder and vagina. Operation times, blood loss, analgesic requirements, postoperative complications, and catheterisation were greater in the colposuspension group than the vaginal tape group.

Discussion

In the short term, the tension-free vaginal tape procedure is as effective as colposuspension for the treatment of primary stress incontinence. Operative complications were more common after the vaginal tape procedure; colposuspension was associated with more postoperative complications and longer recovery.

Blinding may be possible in a trial of surgery for stress incontinence, as shown in a trial of laparoscopic versus open colposuspension.17 We believed that patients in our trial would be aware of their allocated treatment owing to differences between the procedures in incision, anaesthesia, and catheterisation. Staff undertaking postoperative assessments were not blinded to the procedure for logistical reasons, and will have been vulnerable to observer bias.

Surgery and anaesthesia for the vaginal tape procedure was standardised and was performed as described by Ulmsten.10 Although colposuspension was performed according to that used by the units it was standardised for numbers of suspensory sutures, suture material, and postoperative catheterisation. The investigators all had expertise in continence surgery, but were from a mixed background including gynaecologists, urogynaecologists, and urologists and represented both general hospitals and university teaching hospitals. All surgeons had undergone similar training in the vaginal tape procedure, although they had variable experience before recruitment. The group therefore reflects the current practice in the United Kingdom and Eire and as such increases the external validity of the study.

The number of patients recruited fell short of the target determined by a sample size calculation owing to limitations of resources and time. Although differences between the procedures for cure of stress incontinence were not shown, the numbers were not sufficient to achieve the power required to assume equivalence. Given that the objective cure for colposuspension was lower than expected from previous literature, the numbers recruited would give only 50% power to detect a 10% difference or 80% power to detect a 15% difference in cure rates.

No differences were shown between the vaginal tape procedure and colposuspension for objective cure of urodynamic stress incontinence, as measured by the pad test and urodynamic testing. Objective cure rates are lower than those previously reported. This may be due to the stricter definition of cure in this study. There has been little consistency in the objective outcome measures used in previous studies, and cure rates are often quoted for single tests. The cure rates for both procedures as measured by single urodynamic tests range from 74% to 84% and are comparable to those reported in previous series. Most reports of incontinence surgery give no information on non-attenders or the handling of missing data. They have by default assumed that non-attenders are equivalent to attenders. We have considered non-attenders or patients with missing data as treatment failures.