Colposuspension after previous failed incontinence surgery: a prospective observational study


Correspondence: Professor L. Cardozo, 8 Devonshire Place, London W1N 1PB, UK.


Objective To evaluate the outcome of colposuspension for genuine stress incontinence in women who had previously undergone bladder neck surgery.

Design Prospective observational study.

Setting Tertiary referral urogynaecology unit.

Participants Fifty-two consecutive women with recurrent genuine stress incontinence operated on one surgeon.

Main outcome measures Subjective and objective cure of stress incontinence. Complications one surgeon.

Results The mean age of the women was 51 years (range 28–72) and weight 72.7 kg (range 53–112). Sixty-five continence procedures had been performed previously, with 13 women (25%) having had more than one operation. Nine months post-operatively the subjective cure rate was 80% and objective cure rate 78%. Intraoperative complications were few but included two bladder injuries and one rectus muscle tear which required repair. Seven women (13%) developed voiding difficulties which required clean intermittent self-catheterisation, but only one needed to continue this for six months. None of the women developed detrusor instability.

Conclusions In this setting colposuspension after previous bladder neck surgery offers a high rate of success. However, long term follow up is needed to see if this effect is maintained.


The number of treatments for genuine stress incontinence increases each year, but the Burch colposuspension1,2. remains the ‘gold standard’ to which all operations are compared. There are a number of mechanisms by which continence is thought to be achieved: bladder neck elevation improves abdominal pressure transmission to the proximal urethra and allows compression of the proximal urethra against the symphysis pubis during increases in intra-abdominal pressure, suburethral support is increased and a degree of bladder outflow obstruction is produced3,4. Success rates vary between studies, and are influenced by different methods of patient selection, operative technique and assessment of efficacy. Objective cure of incontinence may be achieved in approximately 90% of women at short term follow up when colposuspension is performed as a primary procedure5,6, although this falls to 69% over 10–12 years7. Failure may be due to a number of different factors, including poor pre-operative urethral function8,9 or tissue quality10, intraoperative technical difficulties secondary to the woman's weight or excessive haemorrhage7, or the development of detrusor instability post-operatively11–13. It is also widely accepted that previous bladder neck surgery makes colposuspension both technically more difficult and more likely to fail to achieve continence6,7.

Some women who have had a failed continence procedure are unwilling to undergo further surgery and may choose to try conservative treatments such as pelvic floor exercises, electrical stimulation or urethral devices. However, the best prospect of regaining continence is often with further surgery. As a tertiary referral urogynaecology unit, we recognised that treating women after an unsuccessful operative procedure represented an increasing part of our workload. We therefore studied the effectiveness of repeat colposuspension for recurrent genuine stress incontinence in order to offer our patients a realistic assessment of the benefits and risks.


A series of consecutive women referred with recurrent stress incontinence following previous bladder neck surgery were evaluated by means of a structured questionnaire, physical examination and urodynamic studies. Those who had undergone at least one previous surgical bladder neck procedure for incontinence and were shown to have genuine stress incontinence on videocystourethrography were included in the study. All definitions conform to the International Continence Society (ICS) standards14. Prolapse was defined as mild when it did not reach the introitus and marked when it reached the introitus or beyond on examination.

Selection of the women

On referral all the women were sent a urinary symptoms questionnaire and a frequency-volume chart to complete prior to their first appointment. The questionnaire was based on a binary scale and has been used sucessfully within our unit for a number of years. Physical examination included an assessment of vaginal capacity, mobility, scarring, pelvic organ prolapse and position of the bladder neck. Each woman was asked to attend for urodynamic studies with a comfortably full bladder. After initial uroflowmetry, videocystourethography was performed using a standardised protocol. A 12 F filling catheter and fluid filled bladder and rectal pressure catheters (Mediplus, High Wycombe, UK) were inserted and the post-micturition residual volume of urine recorded. Detrusor pressure was calculated by subtraction of the rectal pressure from the intravesical pressure. The bladder was filled at 100 mL/minute with Isopaque Cyst0 100 mg/mL (Nycomed, Birmingham, UK) which had been stored at room temperature. Filling was stopped and the catheter removed when the woman developed a strong desire to void or 500 mL of contrast had been infused into the bladder. The X-ray table was then rotated so that she stood in the vertical position. Incontinence of urine was detected in all cases by screening the bladder while the woman gave a series of one, three and then five coughs with maximal effort. At the end of the test she was asked to void to obtain a pressure-flow study.

If the woman was suspected of having underlying detrusor over-activity which was not demonstrated during videocysturethrography, further investigations in the form of ultrasound estimation of bladder wall thickness and ambulatory urodynamics were arranged. Women found to have detrusor instability with these tests were excluded from the study.

Only women with genuine stress incontinence on urodynamic investigation were included in the study. All had been offered or had undergone physiotherapy as their first treatment. Women with a short, scarred vagina with very limited mobility were considered unsuitable for repeat colposuspension and were therefore excluded from the study. If treated surgically, these women were given periurethral injections of a bulking agent. However, a formal comparison between the two methods of treatment was not made.

Operative procedure

Each woman was admitted the night before surgery. Prophyiactic heparin 5000 units by subcutaneous injection was given twice daily until she was fully mobile. Prophylactic co-amoxyclav 1.2 g by intravenous injection was given, one dose intra-operatively and two doses post-operatively. All the colposuspensions were performed by the same surgeon (L.D.C.) under general anaesthesia.

The woman was placed in the Lloyd-Davis position. Her abdomen and perineum and vagina were cleaned with 0.1% w/v chlorhexidene solution; sterile drapes were then applied around the surgical field. A low transverse suprapubic incision was made and the old scar excised if present. The rectus sheath was opened, then the cave of Retzius entered by a combination of blunt and sharp dissection. The sutures from previous continence surgery were removed, if present, and the bladder neck and paravaginal tissues mobilised as much as possible. Adhesions around the bladder neck were divided using diathermy scissors, with the aim of restoring normal anatomical planes and bladder neck mobility. Haemostasis was maintained with sutures, diathermy and stainless steel clips when necessary. The paravaginal tissues were sutured to the ipsilateral ileopectineal ligament on each side with four number one polydroxane sulphate sutures. The operator's fingers were in the vagina while the sutures were inserted to ensure accurate placement, while the assistant's fingers were in the vagina elevating the bladder neck while the sutures were tied. A suction drain was inserted into the cave of Retzius and a suprapubic urinary catheter inserted into the bladder at the end of the procedure.

Post-operative management

The drain was removed within 48 hours of surgery and the suprapubic catheter left on free drainage until the second post-operative day when clamping was commenced. When the residual urine volume was less than 100 mL on two consecutive occasions, and the woman was passing good volumes of urine, the suprapubic catheter was removed and she was allowed home. If spontaneous voiding was not achieved by the fifth postoperative day the woman was discharged home with the suprapubic catheter on free drainage, and re-admitted five to seven days later for reclamping. Women with a residual urine volume persistently above 100 mL were taught clean intermittent self-catheterisation. Each woman was reviewed initially at six weeks and then nine months post-operatively when the urinary symptoms questionnaire, pelvic examination, uroflowmetry and videocystourethrography were repeated.

Statistical analysis

The data were analysed using SPSS software (SPSS Inc., Chicago, Illinois, USA). Subjective cure was defined as absence of the symptom of stress incontinence. Objective cure was recorded when there was no evidence of involuntary leakage of urine on videocystourethrography. Binary data from the urinary symptom questionnaires and examination findings were analysed using the McNemar test. Non-parametric skewed variables were compared using the Wilcoxon matched-pairs signed-rank test. P < 0.05 was considered to be significant.


Fifty-two consecutive women complaining of recurrent genuine stress incontinence underwent repeat colposuspension. Their mean age was 51 years (range 28 to 72) and their mean weight 72.7 kg (range 53 kg–112). Thirty-four women (65%) were postmenopausal, twenty-five (74%) of whom were on hormone replacement therapy. Seven women were hypertensive, two were being treated for depression, one suffered from asthma, one from non-insulin-dependent diabetes mellitus and one from sarcoidosis. A total of 65 previous continence procedures had been performed, the details of which are shown in Table 1. Only four women (8%) had undergone a single anterior repair previously. A hysterectomy had been performed in 33 women (63%), either at the time of their previous continence procedure, or independently.

Table 1.  Continence procedures previously undergone by the women. Thirteen women (25%) had undergone more than one continence operation; 65 operations altogether. Values are given as n (%).
Operative procedureNumber of operations
Colposuspension34 (52)
Anterior repair13 (20)
Marshall-Marchetti-Krantz8 (12)
Stamey procedure5 (8)
Raz procedure2 (3)
Sling procedure2 (3)
Periurethral teflon injection1 (2)

Operative data

The mean anaesthetic time was 67 minutes (range 30–120). Four polydroxane sulphate sutures were inserted as planned in all but two women: one had three sutures inserted on one side and the other five. Intraoperative complications were few: two women had a tear in the bladder wall and one had a tear in the rectus muscle. Each was identified and repaired at the time of the operation. A total abdominal hysterectomy was performed in 17 of the 19 women (89%) with a uterus. No woman suffered significant haemorrhage (> 500 mL).

Twenty-nine women (56%) went home after having successful removal of their catheter, but 23 (44%) were discharged with their catheter and were re-admitted 10 days later for reclamping. Seven women (13%) were taught clean intermittent self-catheterisation post-operatively. Six women performed clean intermittent self-catheterisation for less than three months, but one needed to continue for six months.

Subjective data

Thirty-seven of the 46 women (80%) re-assessed nine months post-operatively had a subjective cure. There was a significant improvement in the symptoms of stress incontinence, urgency and urge incontinence (Table 2). Six women (12%) failed to attend: two sent back their questionnaires and were symptomatically cured but thought they lived too far away to attend in person; two had moved and were untraceable despite contacting their previous general practitioner; one was living abroad and the other failed to attend two appointments.

Table 2.  The change in urinary symptoms occurring in response to repeat colposuspension. Results do not include the women who returned their questionnaires but did not attend for objective assessment. Values are given as n (%).
SymptomPre-operativeNine monthsP
  1. Comparison with pre-operative symptoms by the McNemar test.

Stress incontinence46 (100)4 (9)<0.001
Urgency29 (63)18 (39)0.017
Urge incontinence21 (46)8 (17)0.003
Voiding difficulty6 (13)7 (15)1.00

Objective data

Thirty-six women (78%) had an objective cure. None of the women developed post-operative detrusor instability. There was a significant increase in maximum voiding pressure and decrease in flow rate, suggesting that the mechanism of action of repeat colposuspension is, at least in part, achieved by bladder outflow obstruction (Table 3).

Table 3.  Urodynamic variables and examination findings before and after repeat colposuspension. Values are given as median [interquartile range] for urodynamic variables and n (%) for examination findings.
VariablePre-operativeNine monthsP
  1. Analysis of urodynamic variables using the Wilcoxon matched pairs signed rank test and examination findings using the McNemar test.

Flow rate (mL/s)20 [12–30]15 [5–25]0.84
Residual volume (mL)0 [0–10]0 [0–18]0.56
Maximum voiding pressure (cmH20)20 [15–36]38 [26–45]0.006
Maximum capacity (mL)500 [437–508]479 [400–500]0.06
Pressure rise on filling (cmH20)7 [5–10]7 [4–12]0.697
Mild10 (22)7 (15)0.23
Marked12 (26)0 (0)< 0.001
Mild5 (11)8 (17)0.581
Marked3 (7)4 (9)1.00


Bladder neck surgery probably offers the best chance for a lasting cure for genuine stress incontinence, but even in the most impressive surgical series there are a small number of women in whom the operation fails. Women who have had unsuccessful treatment will almost certainly want to know why they are still incontinent, and this question should be carefully addressed by the surgeon. The woman's urinary symptoms may give some indication of the probable underlying pathology, but investigation is almost always necessary. Inadequate elevation of the bladder neck is associated with persistent urinary incontinence15,16, and while this may be detected on clinical examination, videocystourethgraphy gives more information. The development of detrusor instability following continence surgery6,12 is a common cause of recurrent urinary symptoms, and it is therefore essential that this is excluded on cystometry before further surgery is planned.

It is sometimes difficult to determine what represents a failure of the initial operative procedure and what constitutes a recurrence of incontinence, and indeed there may be some overlap between the two. Immediate failure in women who have had a recognised continence procedure, such as a colposuspension or sling operation, may be secondary to mis-diagnosis or poor surgical technique. Recurrence of urinary symptoms in the early post-operative period may be secondary to too early a return to physical activities, or straining because of constipation, and in the longer term, poor tissue collagen. In addition, some women may not initially have had the most effective surgical procedure, such as anterior colporrhaphy. The low pressure urethra and intrinsic urethral sphincter incompetence are thought to be an important aetiological factor in the failure of continence We did not always perform urethral pressure profilometry in this study and it is therefore unclear if any of our women had a low pressure urethra. However, even though we did not exclude women with this adverse factor, our success rate for repeat colposuspension approaches that which we have previously reported for primary colposuspension17. A criticism of some studies is that the assessment of outcome is based on the surgeon's opinion18. The women having repeat colposuspension in this study completed their symptom questionnaires independently and did not have urodynamic follow up performed by the primary surgeon. Although the videocystourethrography was undertaken by urogynaecology registrars who were aware of the woman's previous treatment, we do not feel that this is a major source of bias.

The treatment offered to women with recurrent urinary incontinence will depend on the examination findings, the results of investigations and the previous therapy given. In addition the health and expectations of the woman must always be taken into account. To undergo a repeat colposuspension she must have adequate vaginal capacity and mobility, with genuine stress incontinence proven on urodynamic investigation. This study has clearly demonstrated that, in this setting, repeat colposuspension is a worthwhile technique, giving subjective and objective improvement of urinary symptoms with few complications. An alternative surgical approach would be to perform a sling procedure, but the complication rate is often higher19. We reserve this operation for relatively young women with a narrow, fixed and scarred vagina where repeat colposuspension is not technically possible. Older, less active women with significant vaginal scarring may be treated with periurethral injections of a bulking agent20,21. The objective cure rates of approximately 60% in this situation are much lower than those achieved by repeat colposuspension and sling procedures. However, periurethral injection is far less invasive, may easily be repeated and, because it can be performed as a ‘day care’ procedure, is potentially more cost-effective, despite the current high price of materials. The most important factor in determining success is the choice of the correct procedure for the individual woman.

Repeat colposuspension is often difficult, with scarring and fibrosis in the cave of Retzius distorting anatomical planes and increasing the risk of lower urinary tract injury. Adequate dissection and mobilization of the bladder neck is essential for the technique to be successful. Our approach is for the most experienced urogynaecological surgeon always to perform this procedure, and we recommend that women having repeat continence surgery are referred to a specialist centre. There are few reports of continence surgery performed in a general setting, but the complication rate is likely to be higher and may approach 55%22. The use of four polydroxane sulphate sutures on each side of the bladder neck, like Burch's original description using three sutures bilaterally1, corrects both genuine stress incontinence and anterior vaginal wall prolapse. Reassuringly, few women in this study had significant posterior vaginal wall defects either pre- or post-operatively. Total abdominal hysterectomy was performed in all but two of the women with a uterus, thereby reducing the risk that the benefit of the repeat colposuspension would be compromised by future pelvic surgery.

There was an initial high rate of voiding difficulties while the women were in hospital, with 56% requiring a suprapubic catheter when they went home, but this immediate problem appears to settle, with only a few women requiring clean intermittent self-catheterisation. However, urodynamic investigation performed nine months post-operatively confirmed that an element of outflow obstruction persists, with a rise in the mean maximum voiding pressure and a fall in flow rate. This information is important when counselling women about the potential benefits and risks of continence surgery. Older women are particularly at risk of early voiding difficulty after colposuspension, regardless of whether a hysterectomy is performed at the same time23. Our policy is to teach women with evidence of pre-operative voiding difficulties clean intermittent self-catheterisation before they have a colposuspension, as this is technically easier to learn before the bladder neck is elevated. However, none of the women in this study was identified as being at risk of post-operative voiding difficulties before their operation.

Women having repeat continence surgery have been found by some24, but not others7, to be at greater risk of development of post-operative detrusor instability than those having primary surgery. We have shown a significant improvement in symptoms of urgency and urge incontinence post-operatively and no woman had detrusor over-activity before or after surgery. At the time of this study, ambulatory urodynamics and measurement of bladder wall thickness were being evaluated in our unit as screening tests for detrusor instability25,26. Some women with a stable bladder during videocystourethrography will have been suspected by these tests of having underlying detrusor overactivity which was not demonstrated on the cystometrogram, and were therefore excluded from the study. This method of selection of the women may also have improved the surgical outcome of the women in this series.

In conclusion, repeat colposuspension offers a worthwhile treatment for genuine stress incontinence following previous failed bladder neck surgery. Women with recurrent incontinence, who have been carefully assessed and are felt to be suitable for this procedure, may be advised that there is a high rate of success with few intraoperative and post-operative complications. However, long term follow up is essential to determine if this effect is maintained.