BJOG: An International Journal of Obstetrics & Gynaecology

Laparoscopic versus open Burch colposuspension: a randomised controlled trial

Authors


Dr M Carey, Suite D, Level 10, Frances Perry House, Cardigan Street, Carlton, Melbourne, Victoria 3053, Australia. Email mcarey@netlink.com.au

Abstract

Objective  To compare perioperative characteristics, short-term, and long-term outcomes for laparoscopic Burch colposuspension (LBC) and open Burch colposuspension (OBC) for the treatment of urinary stress incontinence.

Design  Randomised surgical trial with single blinding.

Setting  Three tertiary level teaching hospitals involving seven surgeons of varying skill levels.

Population  Two hundred women with urodynamic stress incontinence (USI).

Methods  The two groups were treated in identical fashion, except for the laparoscopic or open approach to surgery. Attempts were made to blind the subjects and the observers obtaining outcome data to treatment group. Analyses were adjusted for surgeon experience.

Main outcome measures  Absence of USI 6 months following surgery, postoperative pain, time spent in hospital, and time to return to activities of normal daily living.

Results  There were no significant differences in objective and subjective measures of cure and in patient satisfaction at 6 months, 24 months, or 3–5 years of follow up between laparoscopic and open colposuspension groups. Laparoscopic colposuspension took longer time to perform (87 versus 42 minutes, P < 0.0001) but was associated with less blood loss (P= 0.03), less pain (P= 0.02), and quicker return to normal activities (P= 0.01).

Conclusion  LBC has significant advantages over traditional OBC, without any apparent compromise in short-term and long-term outcomes.

Introduction

Surgery for stress incontinence is generally recommended when conservative treatments fail. The open Burch colposuspension (OBC) has often been referred to as the ‘gold standard’ surgical procedure for stress incontinence, with a cure rate of 82% at 5 years.1,2 The laparoscopic Burch colposuspension (LBC), first described in the early 1990s, has the proposed advantages of no laparotomy incision, less postoperative pain, shorter time in hospital, and a quicker return to normal activities.3–6

The aims of this study were to compare the perioperative variables, and the short-term and long-term outcomes of a randomised controlled trial comparing the LBC and OBC procedures.

Methods

Between January 1997 and December 1998, women with urodynamic stress incontinence (USI) were recruited from the Urogynaecology Departments at the Royal Women’s Hospital, Mercy Hospital for Women, and Monash Medical Centre, Melbourne, Australia. Ethics approval was obtained in each centre prior to commencing the study.

A standardised urogynaecological history was obtained and examination performed on each woman in the outpatient clinic at the initial presentation. Urodynamic studies, including assessment of urethral closure pressure, were performed on all women who were recruited into the study. Inclusion criteria included women with USI and failed conservative therapy. Exclusion criteria included: previous retropubic continence surgery; maximum urethral closure pressure of 20 cm H2O or less; medically unsuitable for laparoscopic or open surgery; and major degrees of coexisting pelvic organ prolapse, requiring surgery other than a simple rectocele repair. Coexisting idiopathic detrusor overactivity was not an exclusion criterion for entry into the study. Urinary urgency, urgency incontinence, and detrusor overactivity were assessed preoperatively and postoperatively.

Subjects were randomised the day before surgery into two groups, OBC and LBC. Randomisation lists were computer generated with a block size of six, stratified for each centre and for women undergoing concomitant rectocele repair for symptomatic rectocele. Lists were held by the nonsurgical investigator (A.C.). The surgeons and operating theatre staff were informed of the treatment group immediately prior to surgery.

The two senior surgeons (M.P.C., A.R.) had a combined experience of approximately 70 LBC procedures at the start of this study and usually performed or supervised the laparoscopic procedures. During the study, seven different surgeons performed the laparoscopic and open procedures. The surgeons aimed to perform the LBC and OBC procedures in an identical manner apart from the incisions. The OBC operation was performed as described by Tanagho.7 Number-0 braided polyester suture on a CT-2 needle (Ethibond; Ethicon, Somerville, NJ, USA) was used for the colposuspension. The sutures were tied with minimal tension resulting in a 2- to 4-cm suture bridge between the vagina and iliopectineal ligament (Figure 1). At least two sutures, and preferably three, were placed on each side. A third suture was often used if a cystocele was present. Cystoscopy, with a 70° cystoscope, was performed at the conclusion of surgery to exclude suture penetration and confirm ureteric patency. A urethral catheter was inserted at the end of surgery.

Figure 1.

LBC and OBC sutures were tied such that there was a 2- to 4-cm suture bridge between the vagina and iliopectineal ligament.

For the LBC procedure, a 10-mm operating laparoscope at the umbilicus and two lower, lateral abdominal wall ports of 5 mm were used. The 5-mm ports were sited using the landmarks of two fingerbreadths above and two fingerbreadths medial to the anterior superior iliac spine. The space of Retzius was accessed via a transperitoneal approach. The colposuspension then proceeded in a similar fashion to the OBC using identical suture type, number of sutures, placement of the sutures, and the leaving of a suture bridge.

At the completion of surgery, identical iodine-soaked dressings were applied to all subjects in an attempt to blind both patients and postoperative nursing staff to the procedure performed. Iodine was used to try to mask any bleeding through the dressings. The mode of surgery, open or laparoscopic, was not documented on the postoperative record.

The transurethral Foley catheter was removed at approximately 18 hours postoperatively and the subjects had postvoid residual urine volume checked on two occasions. If they were unable to void and/or had residuals of more than 150 ml, a transurethral catheter was re-inserted and they returned to the clinic in 1 week for a repeat trial of voiding. The total number of days of catheterisation was recorded for each subject.

For both treatment groups, a standardised general anaesthesia protocol and a standard postoperative pain relief protocol with intravenous patient-controlled analgesia and nonsteroidal anti-inflammatory suppositories were used. Postoperative pain was assessed with a visual analogue pain scale (VAS; 0–10) administered while the patient was at rest, at 4 hours postoperatively and twice daily until the morning of day 2 postoperatively.

Women were reviewed at 6 weeks, 6 months, 24 months, and again at 3–5 years postoperatively. The Short Form-36 (SF-36) was administered at both baseline and follow up.8 This compares eight scales that can be collapsed into two summary measures assessing physical and mental health, the Physical Component Summary (PCS) and Mental Component Summary (MCS), respectively. In addition, the SF-36 has a general health question (excellent, very good, good, fair, and poor). Only the summary scales and general health question are reported in this study. Summary scores are presented as T-scores with means of 50 and SD of 10 points.9 Lower scores indicate better general, physical, and mental health using the SF-36 survey. Women completed the Short Urinary Distress Inventory (SUDI) and Short Incontinence Impact Questionnaire (SIIQ) at baseline and follow up.10 These instruments assess symptom distress and life impact, respectively, of urinary incontinence.

At the 6-week postoperative review, data on resumption of normal activities were also collected via standard questioning. Patient satisfaction was assessed on a visual analogue scale (VAS; 0–100 where 100 represented being completely satisfied and 0 completely unsatisfied). A score of 80 or more was considered to be a successful outcome in terms of satisfaction with treatment.

At the 6-month review, urodynamics tests were repeated, and the Genito-Urinary Treatment Satisfaction Scale (GUTSS) was used to assess satisfaction with surgery.11,12 The GUTSS comprises ten items in two scales measuring care and outcome satisfaction. The scale range is 0–32, with higher scores indicating greater satisfaction.

We compared the LBC with the OBC in terms of long-term differences in subjective and objective urinary symptoms, patient satisfaction, operating time, perioperative complications, postoperative pain scores, duration of bladder catheterisation, and return to normal activities. The primary outcome was absence of USI at 6 months following surgery. A total of 182 women, 91 in each treatment group (‘rounded off’ to 200 subjects, with 100 in each treatment group) was required to detect at least a 20% higher in cure of USI at 6 months for the open group, with a significance level of 0.05 and power 0.80, given an expected cure of 80% for the open group.13

Any LBC converted to an OBC was included as LBC, for an intention-to-treat analysis. Treatment groups were compared using t test for independent means for continuous variables such as age, parity, weight, and body mass index. Chi-square tests (or Fisher’s exact tests where appropriate) were used to compare proportions of women in each group with particular characteristics such as urgency, urgency incontinence, detrusor overactivity, and intraoperative complications. Surgeon experience was included as a covariate for statistical comparisons of the surgical outcomes. Estimated blood loss was changed to a logarithmic scale because of outliers. Ordinal logistic regression with mode of operation and surgeon experience was used for postoperative pain. There were five women who had ‘not yet’ returned to ‘normal activity’ by the time of the postoperative review, and these data were assumed to be 42 days (6 weeks) for the purposes of analysing this variable.

For long-term outcomes, logistic regression was used including mode of surgery and surgeon experience in each model. Six-month incidence of stress incontinence or detrusor overactivity symptoms, and the presence of either at urodynamics tests were analysed using logistic regression. For comparison of preoperative and postoperative detrusor overactivity, we created a variable of the difference and analysed this via ordinal logistic regression. Twenty-four-month stress incontinence, urgency, and urgency incontinence symptoms were analysed using ordinal logistic regression. For the sensitivity analysis of stress incontinence symptoms at 24 months, we pooled ‘occasional’ and ‘frequent’ and added all missing values to this outcome and to the denominator.

For the incontinence self-reporting measures (the SUDI, SIIQ) and health status measure, missing scores were imputed using hot deck, where the deck was defined as the treatment group.14 For SF-36 items contributing to the PCS and MCS scales, horizontal mean imputation was used.15 Between groups comparisons at baseline and follow up were made using the independent t-teat test; for baseline analyses, the dependent t test was used, and for baseline/follow up by group analyses, analysis of covariance (ANCOVA) was used with the baseline scores entered as the covariate.

Results

Two hundred women were randomised, 104 into the OBC group and 96 into the LBC group. Participants’ progress through the study is detailed in Figure 2. There were no significant differences among preoperative characteristics in both treatment groups (Table 1). Only 3% of the study population were nulliparous. Hundred and twenty-four women (67%) complained of preoperative urgency incontinence and 20 (11%) had detrusor overactivity confirmed at urodynamic testing.

Figure 2.

Flowchart of subjects’ progress through the study for OBC and LBC groups.

Table 1.  Demographics and preoperative findings for OBC and LBC groups
VariableOBCLBCP value
  • *

    t test for independent means.

  • **

    Because of missing data values, the number of subjects included and (where relevant) the proportion of subjects affected for each variable in each treatment group is reported.

  • ***

    Urinary urgency and urgency incontinence symptoms were assessed with standard questionnaire and considered present if these symptoms occurred occasionally or frequently.

  • ****

    Yates-corrected chi-square test.

  • *****

    Detrusor overactivity at urodynamics was defined as a rise in vesical pressure greater than 15 cm H2O (in the absence of an increase in abdominal pressure) associated with urinary urgency and/or leakage.

Age (years)
Mean (SD)52.3 (10.6)51.0 (9.9)0.39*
No. of subjects analysed**10496
Parity
Mean (SD)2.6 (1.3)2.8 (1.3)0.50*
No. of subjects analysed**10094
Body mass index (kg/m2)
Mean (SD)28.0 (4.8)29.0 (5.7)0.24*
No. of subjects analysed**8076
Weight (kg)
Mean (SD)72.1 (12.6)73.3 (14.9)0.55*
No. of subjects analysed**9591
Urinary urgency symptoms (%)***71670.69****
Proportion**69/9762/92
Urinary urgency incontinence (%)***70640.50****
Proportion**67/9657/89
Detrusor overactivity (urodynamics) (%)*****1570.12****
Proportion**14/916/87

At 6 months postoperatively, there was no significant difference in the rates of stress incontinence, detrusor overactivity, or the presence of either at urodynamic assessment (Table 2). Combining the two treatment groups gave an overall cure rate for stress incontinence at 6-month urodynamics of 75% and demonstrated detrusor overactivity in 12%. Only 66% of the population did not demonstrate either stress incontinence or detrusor overactivity. Despite this, most women (88%) reported a satisfaction score of ≥80 for their surgical results.

Table 2.  Urodynamics and patient satisfaction outcomes at 6 months for OBC and LBC. P values adjusted for surgeon experience*
VariableOBCLBCP value***
  • *

    To try to control for surgical experience, whether the primary surgeon was M.P.C. or A.R. was recorded and included as a covariate in the analysis.

  • **Yates-corrected chi-square test.

  • ***

    Because of missing data values, the proportion of subjects affected and analysed for each variable in each treatment group is reported.

  • ****

    Assessed with visual analogue score (0 as ‘completely unsatisfied’ and 100 as ‘completely satisfied’). A score of ≥80 was considered as a successful outcome.

USI (%)22280.22
Proportion analysed***20/9223/83
Detrusor overactivity (%)13110.88
Proportion analysed***13/9810/87
USI and/or detrusor overactivity (%)32360.41
Proportion analysed***29/9230/83
Patient satisfaction****87900.52
Proportion analysed***71/8266/73

In terms of serious perioperative complications, there were no significant differences between the two treatment groups. There was one case of haemorrhage at OBC requiring a blood transfusion and one case of haemorrhage at LBC from laceration of the obturator vein. The lacerated obturator vein was managed laparoscopically and this woman did not require a blood transfusion. There were six cases of suture penetration into the bladder that were diagnosed at intraoperative cystoscopy, one at OBC and five at LBC (P= 0.11). For the five LBC cases, these sutures were removed and re-sited laparoscopically in three cases, and the LBC was converted to OBC to manage this complication in the remaining two cases.

Mean operating time was approximately twice as long for LBC compared with OBC, but surgeon’s estimates of blood loss and patient’s estimates of immediate postoperative pain at rest were significantly less after LBC (Table 3). LBC subjects reported return to normal activities, on average, 5 days earlier than OBC subjects (P= 0.01). There were no significant differences in duration of bladder catheterisation between the two treatment groups (P= 0.72).

Table 3.  Perioperative data for OBC and LBC groups. P values adjusted for surgeon experience*
VariableOBCLBCP value*
  • *

    To try to control for surgical experience, whether the primary surgeon was M.P.C. or A.R. was recorded and included as a covariate in the analysis.

  • **

    Time in minutes from initial skin incision until closure of skin incisions.

  • ***

    Because of missing data values, the number of subjects included for each variable in each treatment group is reported.

  • ****

    Primary surgeon’s subjective estimate of blood loss in millilitres.

  • *****

    Assessed with VAS (0 as ‘no pain’ and 10 as ‘severe pain’) by questioning the patient while at rest 4 hours postoperatively, and twice daily until the morning of day 2 postoperatively.

  • ******

    Number of days until return to normal activities as assessed by standard questionnaire at 6-week review.

Operating time**
Mean (SD)42 (25)87 (31)<0.0001
No. of subjects analysed***10394
Estimated blood loss****
Mean (SD)170 (199)126 (71)0.03
No. of subjects analysed***9691
Pain score*****—mean (No. of subjects)***
4 hours2.9 (93)2.1 (81)0.02
Day 1, 08.00 hours2.8 (94)2.4 (84)0.20
Day 1, 20.00 hours2.5 (92)2.3 (79)0.84
Day 2, 08.00 hours2.4 (94)2.1 (78)0.72
Days in hospital3.93.70.32
Return to normal activity******—mean (SD)24.6 (12.5)19.7 (12.6)0.01
No. of subjects analysed***9186

There was no association among preoperative and postoperative detrusor overactivity and mode of surgery. Twelve women had detrusor overactivity on urodynamics before but not after surgery, seven from the OBC group and five from the LBC group. Fourteen women developed detrusor overactivity on urodynamics after surgery, five from the OBC group and nine from the LBC groups (P= 0.67).

At baseline, the open surgery group, when compared with those in the laparoscopic group, reported significantly better general health on the SF-36 question (Table 4). They also obtained better mental health status SF-36 scores. Table 4 shows that the baseline differences in general health between the OBC and LBC groups were maintained at follow up. There was no significant baseline to follow-up improvement in general health although the P values for both groups were suggestive. There were no significant differences on the SF-36 PCS score by treatment group at either baseline or follow up. For physical health, there was a significant baseline to follow-up improvement for those in the laparoscopic group, whereas there was no significant improvement for those in the open surgery group, although the P value was suggestive. The baseline differences in mental health between the OBC and LBC groups were maintained at follow up. There was no significant baseline to follow-up improvement in mental health status for either group. Finally, there were no significant differences on any measure at follow up between groups after controlling for baseline scores.

Table 4.  Analysis of self-reported measures at baseline and 6-month follow up by treatment cohort
 nBaselineFollow upStatistics*,**
MeanSDMeanSD
  • For SUDI, SIIQ, and general health (SF-36), the lower the score, the greater the level of improvement.

  • *

    Paired t test for baseline/follow-up differences.

  • **

    ANCOVA for differences between treatment cohorts at follow up for baseline scores.

  • ***

    Independent t test for differences between treatment cohorts at baseline and at follow up.

General health
Laparoscopic932.761.062.601.02P= 0.09
Open992.401.012.221.06P= 0.06
Statistics***P= 0.02P= 0.01P= 0.19
SF-36 PCS
Laparoscopic89 (82)46.2910.9549.5410.41P= 0.01
Open90 (81)46.7010.8349.1411.21P= 0.06
Statistics***P= 0.80P= 0.82P= 0.80
SF-36 MCS
Laparoscopic89 (82)42.1912.5742.7013.03P= 0.87
Open90 (81)48.2110.7048.3110.01P= 0.48
Statistics***P < 0.01P < 0.01P= 0.24
SUDI
Laparoscopic9350.2220.7323.9217.90P < 0.01
Open9948.3021.2921.5616.92P < 0.01
Statistics***P= 0.53P= 0.35P= 0.34
SIIQ
Laparoscopic9349.9823.9031.4023.83P < 0.01
Open9949.8824.1426.8729.36P < 0.01
Statistics***P= 0.99P= 0.24P= 0.24

Table 4 also shows that there were no significant differences on the SUDI or SIIQ scales by treatment group at either baseline or follow up. However, significant improvements were noted between baseline and 6-month follow up by SUDI and SIIQ scores for both treatment groups.

It is possible that the change in general health and mental health status acted as confounders on the effect of treatment as assessed by SUDI and SIIQ scores through differential bias. To investigate this, baseline general health and SF-36 MCS scores were entered as covariates into an ANCOVA model testing for group effects on the SUDI and SIIQ, respectively. The findings showed that there were no significant confounding effects (data not shown), and thus, the treatment effect measured by the SUDI and SIIQ was independent of these factors.

Satisfaction with treatment outcomes from the GUTSS at 6-month follow up was high in both groups, with no difference between treatment groups (Table 5). If the raw data in the table are converted to percentage scores, there was 89% satisfaction with the treatment outcomes and 88% satisfaction with the care received. The overall satisfaction was 87%.

Table 5.  Satisfaction with treatment outcomes at 6-month follow up: GUTSS scores
 Type of interventionStatistics*
LaparoscopicOpen
MedianIQRMedianIQR
  • IQR, interquartile range.

  • *

    Mann–Whitney U test.

  • **

    Scale range: 0−18. The higher the score, the greater the level of satisfaction.

  • ***

    Scale range: 0−16. The higher the score, the greater the level of satisfaction.

  • ****

    Scale range: 0−34. The higher the score, the greater the level of satisfaction.

Satisfaction with outcome**16.04.016.06.0P= 0.47
Satisfaction with care***14.04.014.54.0P= 0.64
Overall GUTSS satisfaction****29.56.329.08.0P= 0.46

At 24 months after surgery, there were no significant differences between the two treatment groups with respect to reporting urinary stress incontinence, urgency, urgency incontinence, or satisfaction score of ≥80 (Table 6). Across both treatment groups, by 24 months after surgery, 66% of women reported no stress incontinence, 38% reported no urgency, 47% reported no urgency incontinence, and 64% reported a satisfaction score of ≥80.

Table 6.  Subjective symptoms at 24 months of follow up for OBC and LBC. P values adjusted for surgeon experience*
VariableOBCLBCP value*,**
  • *

    To try to control for surgical experience, whether the primary surgeon was M.P.C. or A.R. was recorded and included as a covariate in the analysis.

  • **

    Ordinal logistic regression was used for symptoms, including operation type and surgeon experience in the model. Yates-corrected chi-square test used for patient satisfaction, adjusting for surgeon experience.

  • ***

    Standard questionnaire used to assess whether the symptom was present, and if present, whether the symptom occurred occasionally or frequently.

  • ****

    Because of missing data values, the proportion of subjects affected and analysed for each variable in each treatment group is reported.

  • *****

    Assessed with visual analogue score (0 as ‘completely unsatisfied’ and 100 as ‘completely satisfied’). A score of ≥80 was considered as a successful outcome.

Stress incontinence (%)***, proportion****
Occasionally(20), 18/90(31), 24/770.38
Frequently(10), 9/90(6), 5/77
Urinary urgency (%)***, proportion****
Occasionally(51), 44/87(40), 30/750.40
Frequently(10), 9/87(23), 17/75
Urge incontinence (%)***, proportion****
Occasionally(39), 34/87(39), 29/740.21
Frequently(10), 9/87(18), 13/74
Patient satisfaction (% score ≥ 80)*****, proportion****(70), 56/80(58), 43/740.10

Although we were able to obtain follow-up information on approximately 80% of all subjects by 24 months, we performed a sensitivity analysis, assuming that all women who did not complete the 24-month follow up had either occasional or frequent stress incontinence. With these assumptions, cure rates decreased to 61% for OBC and 50% for LBC. There was no significant difference between the two treatment groups, even when adjusted for surgeon experience (P= 0.08).

The study population was contacted by telephone for further follow up at a mean of 3.7 years (range 3–5 years) after surgery. This follow up was undertaken at a single point in time. A total of 162 women were contacted, 88 of whom underwent OBC and 76 of whom underwent LBC. There were no significant differences between the two treatment groups at 3–5 years after surgery. The percentages of women experiencing stress incontinence, urinary urgency, and urgency incontinence symptoms were similar to the 24-month data, as was the percentage of women rating their satisfaction score of ≥80 from surgery.

Discussion

The Burch colposuspension appears to be an effective and durable anti-incontinent procedure.16 LBC has been performed for over a decade with a relatively small number of reported prospective randomised trials.6,17 Most of the trials comparing surgical outcomes have been complicated by different surgical techniques for the open and laparoscopic approaches, such as different number or type of sutures.6,17 We attempted to treat both groups in an identical fashion, with the only difference being the laparoscopic or low transverse incisions for surgical approach. Our data show no difference in short-term and long-term cure rates assessed by both symptoms, self-completed questionnaires and urodynamics.

The LBC has several advantages over OBC: among perioperative outcomes, the primary surgeon’s estimates of blood loss were significantly different for the two treatment groups. Although the data were skewed by one case of significant haemorrhage at OBC requiring a blood transfusion, removal of this case still resulted in a significant difference between the two treatment groups in favour of LBC.

There was a significant difference in pain scores at 4 hours, and this likely represents a true difference because at this point, subjects and staff were probably still blinded to surgical approach. There was a trend for postoperative pain to be less for the LBC group during day 1 and day 2 postoperatively. The subjects treated by LBC returned to normal activities approximately 5 days sooner than the OBC subjects.

LBC can be technically more challenging to perform than OBC. Although the most experienced laparoscopic surgeons (M.P.C. and A.R.) had greater experience with LBC than the other surgeons, all other surgeons who performed LBC and OBC had fairly advanced operative skills in general and became proficient at the LBC techniques very quickly. Adjusting the analyses for surgical experience did not result in any additional significant differences between LBC and OBC.

The role of LBC in the treatment of urinary stress incontinence has changed with the introduction of the tension-free vaginal tape (TVT) procedure. From our data, rates of cure for stress incontinence at 6-month postoperative urodynamics were 78% for OBC and 72% for LBC. These were comparable to rates of cure for stress incontinence at 6-month postoperative urodynamics for a multicentre randomised controlled trial comparing OBC (67%) and TVT (81%) in greater than 300 subjects.18 Presently, there are only a few, small randomised controlled trials comparing LBC and TVT, with relatively small numbers and short follow-up times.19–21 TVT is also a minimally invasive procedure that is relatively quick to perform, requiring little equipment, and having a shorter learning curve than LBC. As more evidence is accumulated about the long-term success rates of TVT, it may ultimately become the first-line choice for stress incontinence surgery.

Conclusion

The LBC has significant advantages over traditional OBC, without any apparent compromise in short-term and long-term outcomes.

Acknowledgement

This study was supported by a research grant from The Royal Women’s Hospital, Foundation, Melbourne, Australia.

Ancillary