Analysis of two sling procedures using polypropylene mesh for treatment of stress urinary incontinence

International Journal of Gynecology and Obstetrics 84 (2004) 133–141

Article

Analysis of two sling procedures using polypropylene mesh for treatment of stress urinary incontinence M.J. Hunga,*, F.S. Liua, P.S. Shenb, G.D. Chenc, L.Y. Linc, E.S.C. Hoa a

Department of Obstetrics and Gynecology, Taichung Veterans General Hospital Taichung, Taiwan b Department of Statistics, Tung Hai University Taichung, Taiwan c Department of Obstetrics and Gynecology, Chung Shan Medical University Hospital, Taichung, Taiwan Received 8 April 2003; received in revised form 21 July 2003; accepted 30 July 2003

Abstract Objectives: To evaluate and compare the surgical outcome between the innovative tension-free vaginal tape (TVT) and conventional pubovaginal sling (PVS) procedures using polypropylene mesh. Methods: Eighty consecutive women with urodynamic stress urinary incontinence (SUI), who chose to undergo either a TVT (ns23) or a PVS (ns57) procedure using polypropylene mesh based on financial consideration, were recruited for this study. The surgical results were analyzed and compared subjectively and objectively. Results: The mean follow-up interval was 23 months for the TVT and 20 months for the PVS procedure (Ps0.062). Postoperatively, SUI (91.3% vs. 93.0%), concomitant urge symptoms (85.0% vs. 85.3%) and the negative impact of incontinence and urogenital distress on patients’ quality of life (79.8% vs. 77.8%) (77.4% vs. 68.8%) had improved markedly. After a multivariable logistic regression analysis, the treatment outcome of SUI was found to be independent of the main effects of patient age, parity, concurrent gynecological surgeries, intrinsic sphincter deficiency, previous failed incontinence surgeries, and concomitant urge symptoms. However, it was significantly related to treatment procedures (TVT vs. PVS) and their interaction with patient body mass index (BMI). Based on the fitted logistic model, we see that TVT performs better than PVS when BMI is less than 27.27 kgym2, and the advantage of TVT decreases as BMI increases. Conclusion: Both TVT and PVS procedures using polypropylene mesh are effective treatment modalities for female SUI. However, TVT was not as effective in treating overweight or obese women as PVS. 䊚 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Stress urinary incontinence; Tension-free vaginal tape; Pubovaginal sling; Polypropylene mesh; Body mass index

1. Introduction The pubovaginal sling (PVS) procedure was *Corresponding author. 40705, No. 160, Section 3, Taichung Harbor Road, Taichung, Taiwan. Tel.: q886-4-2374-1258; fax: q886-4-2350-3021. E-mail address: [email protected] (M.J. Hung).

traditionally selected for treating type III stress urinary incontinence (SUI) or intrinsic sphincter deficiency (ISD) that commonly develops after pelvic irradiation therapy and anti-incontinence surgery w1x. As it has been suggested that all women with symptomatic SUI might have some degree of urethral sphincter weakness, PVS has

0020-7292/04/$30.00 䊚 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0020-7292(03)00345-X

134

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

been increasingly used as the primary procedure for all types of SUI w2,3x. PVS surgery is appealing because it is a durable, vaginal, anti-incontinence procedure that requires a relatively short operative time and hospitalization w4x. The advantage of PVS is more obvious when allograft cadaveric fascia and synthetic materials are used w5,6x. Recently, an innovative sling procedure, tensionfree vaginal tape (TVT), has been used increasingly for the treatment of female SUI with an excellent short-term and long-term success rate w7,8x. TVT is different from conventional PVS in that it is placed under the ‘mid-urethra’ with specifically designed instruments and friction is created once the tape traverses the pelvic tissue, which secures the tape in place. In contrast, the PVS is placed under the bladder neck and proximal urethra after extensive retropubic dissection and it has to be kept in place by suburethral and abdominal anchoring. The mechanism of TVT for the treatment of SUI is based on the Integral Theory (in terms of mid-urethra theory) w9x. Continence can be achieved by creating an artificial pubourethral ligament, by which the urethra is kinked and closed during stress w10x. In contrast, the PVS procedure aims to create a ‘hammock’ under the bladder neck and proximal urethra where the urethra is compressed during increasing intra-abdominal pressure w11x. In order to evaluate the effectiveness of TVT and PVS procedures using polypropylene mesh, and the possible differences between these two procedures, we conducted this study to investigate and compare the surgical results of patients who underwent either of these two procedures. 2. Materials and methods From March 1999 to January 2001, 80 women with urodynamic stress incontinence were enrolled consecutively for this study. After a full explanation, all patients signed an informed consent before surgery and chose either a TVT (ns23) or a PVS (ns57) procedure using polypropylene mesh based on financial consideration. The TVT procedure was performed on patients who requested it and who were willing to pay approximately 600 US dollars for the tape (Prolene, Ethicon䉸, Som-

erville, NJ, USA), which was not covered by the National Health Insurance in Taiwan. The PVS procedure, using a self-fashioned polypropylene mesh sling (Prolene, Ethicon䉸, Edinburgh, UK), was performed on the patients who did not want to pay the extra money. Prior to enrolling patients for this study, the surgical team successfully performed both procedures on at least 10 patients in each group. An approval for this clinical trial was obtained from the Ethics Committee at Taichung Veterans General Hospital. Specific preoperative assessment included a pelvic examination, a symptom and quality of life questionnaire containing the validated Incontinence Impact Questionnaire (IIQ-7) and Urogenital Distress Inventory (UDI-6) short form w12x, and a multi-channel urodynamic study. The IIQ-7 and UDI-6 short form were translated into Chinese and were readily understood and clear to our patients. A small-scale pilot study of 15 cases had been carried out before the study took place to ensure the validity of the questionnaire. A test and retest were conducted with a 2-week interval between tests. The correlation coefficient for the questionnaire was gs0.87 (P-0.05). The responses given in the IIQ-7 and UDI-6 short form were on a four-step ordered category scale from ‘not at all’ (0) to ‘greatly’ (3). The IIQ-7 scores were calculated by summation of the value from each of the four domains of physical activity, travel, social activity and emotional health and then multiplied in order to put scores on a scale of 0–100. A higher score indicated that incontinence had a greater impact on patients’ quality of life. A similar scoring system was used for the UDI-6 to evaluate its three categories of irritative, stress and obstructiveydiscomfort symptoms. The multi-channel urodynamic study was performed including a free uroflowmetry followed by residual urine measurement, a provocative, electronically subtracted water cystometry, a pressureyflow study with simultaneous surface electromyography, and a static and dynamic (cough profile) urethral pressure profilometry. The methods, definitions, and units conformed to the standards proposed by the International Continence Society w13x. Urodynamic stress incontinence was defined as the involuntary loss of urine during increased intra-

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

135

Table 1 Comparison of preoperative characteristics of patients who underwent tension-free vaginal tape (TVT) and pubovaginal sling (PVS) procedures using polypropylene mesh Patient characteristics

TVT (Ns23)

Mean age at operation (range) Median parity (range) Mean body mass index (range) % with failed incontinence surgeries % with concomitant urge symptoms % requiring additional gynecologic surgeries % with intrinsic sphincter deficiency

63.3 4 25.8 17.4

(46–85) (0–8) (18.8-32.0) (4y23)

55.0 4 26.1 7.0

(32–77) (0–9) (19.2–34.9) (4y57)

0.001(T)* 0.223(T) 0.678(T) 0.218(F)

91.3

(21y23)

71.9

(41y57)

0.114(Y)

30.4

(7y23)

77.2

(44y57)

0.0002(Y)*

13.0

(3y23)

3.5

(2y57)

0.069(F)

(T)

: T-test;

(F)

: Fisher’s exact test;

PVS (Ns57)

Statistical significance

(Y)

: Yates’ correction x2-test; *P-value-0.05.

abdominal pressure, without a change in the detrusor pressure. ISD was defined as a maximum urethral closure pressure at rest of less than 20 cm H2O. Under regional or general anesthesia, the TVT procedure was performed according to previously described methods by Ulmsten et al. w7x, with the exception of a Crede’s maneuver, with a full bladder after a cystoscopy, which was performed to express urine freely in order to verify that the tape was not obstructive w5x. The PVS procedure was performed according to the method described by Chaikin et al. w2x under regional or general anesthesia. Before the operation, a self-fashioned 1.5 cm=30 cm polypropylene mesh sling was prepared. After the placement of a Foley catheter, a middle longitudinal incision was made in the anterior vaginal wall 2 cm proximal and distal to the level of bladder neck. Two vaginal flaps were developed and the dissection continued laterally until the endopelvic fascia was perforated at its most lateral insertion into the ischiopubic ramus to join the retropubic and the vaginal spaces. Lysis of any adhesion within the retropubic space was done by finger dissection. A low transverse skin incision was made just above the upper edge of the pubic bone. A Sarot clamp was inserted from the abdominal to the vaginal incision under the guidance of the surgeon’s index finger (on the vaginal side). When the tip of the clamp was visible, one end of the long suture that

was attached to the prepared polypropylene mesh sling, was grasped and pulled into the abdominal wound. The procedure was repeated on the other side. The bladder was left full after a cystoscopy was performed to check that the bladder was undamaged. The free ends of the sling were pulled upward, one by one, while gentle downward pressure was applied on the scissors placed between the urethra and the tape until there was no slack. A Crede’s maneuver, to express urine freely, was performed in order to verify that the tape was not obstructive w5x. The sling was secured (underneath the bladder neck and proximal urethra) in place suburethrally and the vaginal incision was closed. The free ends of the sling were fashioned and sutured to the rectus fascia bilaterally with nonabsorbable sutures (2yO Ethibond, Somerville, NJ, USA), and the abdominal wound was closed. Postoperatively, all patients went through transurethral bladder drainage. A voiding trial began on postoperative day 1 to day 3 depending on individual recovery. The catheter was removed once the patient could void freely and the postvoid residual was less than 25% of the bladder volume on two occasions. Patients were followed up regularly at the urogynecology clinic by the principal author (MJH). For study purposes, a follow-up investigation was arranged for these patients around January 2002, where a pelvic examination, a symptom and quality of life questionnaire, a cough stress test with a comfortably

136

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

full bladder, and a chart review were conducted. Follow-up urodynamic study was not routinely performed but was done if indicated by urinary symptoms. The patients were considered to have ‘cured’ of SUI if they had a negative cough stress test result and there were no reports of urine leakage during stress. The patients who were considered to have ‘improved’ were those who did not leak on the cough stress test but may have had occasional urine leakage during stress. This occasional leakage did not, however, influence their daily activities. Patients who did not meet these criteria were considered to have ‘failed’. Univariate analysis was used to identify different variables between these two groups. A P-value0.05 was considered statistically significant. In addition, eight covariates: body mass index (BMI), treatment procedures (TVT vs. PVS), concurrent gynecological surgeries, ISD, age, parity, previous failed incontinence surgeries, and concomitant urge symptoms, were considered for a multivariable logistic regression analysis. The surgical response is categorized into two groups (cure vs. improvement or failure). Due to the sparseness of data in modeling, which may cause the non-existence of a maximum likelihood estimate, 19 (instead of 28) two-way interactions were included. The Goodness-of-Fit (P-values0.94) of the logistic model tested with the method described by Hosmer and Lemeshow w14x indicated that the model fits well. 3. Results Preoperative patient characteristics are shown in Table 1. There was no statistically significant difference between these two groups regarding the patients’ parity, BMI, previous failed incontinence surgeries, concomitant urge symptoms, and ISD. However, the TVT group was older and a lower percentage of them had conditions that required additional gynecological surgeries. Concurrent gynecological surgeries including total hysterectomy, McCall culdoplasty, sacrospinous vaginal vault fixation, sacrocolpopexy, anterior colporrhaphy, posterior colporrhaphy, and paravaginal repair of cystocele were performed in combination with 30.4% (7y23) of the TVT and 77.2% (44y57) of the PVS procedures, respectively (Ps0.0002).

The majority (94%) of these concurrent surgeries were transvaginal procedures except for four total abdominal hysterectomies and two sacrocolpopexies. Surgical results are summarized in Table 2. The mean follow-up interval was 23 months for the TVT procedure and 20 months for the PVS procedure (Ps0.062), respectively. There was no statistically significant difference with respect to the treatment outcome of SUI, concomitant urge symptoms, the negative impact of incontinence on patients’ quality of life, and the negative impact of urogenital distress on patients’ quality of life between these two groups by univariate analysis. Surgical failure was noted in two patients in the TVT group (case sequence no. 15 and 23) and in four patients in the PVS group (case sequence no. 24, 26, 54 and 55), which was not especially noted in the early phase of this clinical trial. The mean duration of postoperative bladder drainage was 3.3 (range 1–16) days for the TVT procedure and 4.3 (range 1–44) days for the PVS procedure (Ps 0.476). Normal voiding began within 1 week postoperatively in 91.3% (21y23) of the TVT patients and in 91.2% (52y57) of the PVS patients. After 2 weeks, 95.7% (22y23) of the TVT patients and 96.5% (55y57) of the PVS patients resumed normal voiding. A bilateral bladder perforation occurred during one of the TVT procedures. This complication was rectified by reinserting the TVT more laterally and by having the transurethral Foley drain for 5 days. Significant urinary retention, requiring sling revision, occurred in 8.7% (2y 23) of the TVT procedures and in 3.5% (2y57) of the PVS procedures, respectively (Ps0.574). Sling revision was conducted at a mean of 12 days (range of 7–17 days) from the initial surgery. The resumption of normal voiding, without the recurrence of urinary leakage, was noted in all of the patients soon after the revision and at the latest follow-up. Significant surgical complaints, which were defined as newly onset or as a worsening of symptoms, are shown in Table 3. Subjective voiding difficulty was the most common side effect of these two procedures (8.7% vs. 7.0%, Ps1.000). However, there was no objective evidence, such as significant amount of residual urine requiring

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

137

Table 2 Surgical results of patients who underwent tension-free vaginal tape (TVT) and pubovaginal sling (PVS) procedures using polypropylene mesh Surgical results

TVT (Ns23)

PVS (Ns57)

Statistical significance

Mean months follow-up (range) Mean days bladder drainage (range) Stress urinary incontinence % Cured % Improved % Persistent Concomitant urge symptoms % Curedyimproved % Persistent % De novo or worse Mean IIQ-7 scores Preoperative (range) Postoperative (range) % Improvement (SD) Mean UDI-6 scores Preoperative (range) Postoperative (range) % Improvement (SD) Immediate surgical complications % Bladder perforation % Urinary retention (sling revision)

23 3.3

(12–33) (1–16)

20 4.3

(12–33) (1–44)

65.2 26.1 8.7

(15y23) (6y23) (2y23)

71.9 21.1 7.0

(41y57) (12y57) (4y57)

81.0 19.0 00.0

(17y21) (4y21) (0y23)

85.4 14.6 3.5

(35y41) (6y41) (2y57)

39.7 6.7 79.8

(0–95.2) (0–38.1) (33.0)

33.1 5.7 77.8

(0–95.8) (0–37.5) (37.8)

0.266(T) 0.686(T) 0.830(T)

49.3 11.6 77.4

(0–72.2) (0–44.4) (24.4)

38.3 9.3 68.8

(0–66.7) (0–38.9) (48.5)

0.020(T)* 0.415(T) 0.428(T)

(0y57) (2y57)

0.287(F) 0.574(F)

0.062(T) 0.476(T) 0.839(P)

0.722(F)

4.3 8.7

(1y23) (2y23)

0.0 3.5

(T) : T-test, (P): Pearson’s x2 -test, (F) : Fisher’s exact test, *P-value-0.05. IIQ, Incontinence Impact Questionnaire; UDI, Urogenital Distress Inventory.

frequent catheterization or causing recurrent urinary tract infections in these patients. Other surgical complaints were rare. Table 4 shows the results of the multivariable logistic regression analysis of treatment outcome of SUI in this study. Most of the parameters are not significantly different from zero (Wald test) except the parameters b2, b12 and b15, which have

P-values of less than 0.05. The estimated parameters b2s130.7, b12sy4.79 and b15sy0.13 denote the effect of treatment procedures, the interaction between BMI and the effect of treatment procedures, and the interaction between BMI and age on the treatment outcome of SUI, respectively. Hence, it can be concluded that the treatment outcome of SUI is independent of the main

Table 3 Comparison of surgical complaints at the latest follow-up after tension-free vaginal tape (TVT) and pubovaginal sling (PVS) procedures using polypropylene mesh Surgical complaints

TVT (Ns23)

Voiding difficulty Urge symptoms Dyspareuniaa Vaginalysuprapubic pain Defective healing

8.7% 0.0% 0.0% 0.0% 0.0%

(F)

(2y23) (0y23) (0y15) (0y23) (0y23)

PVS (Ns57) 7.0% 3.5% 2.5% 1.8% 0.0%

(4y57) (2y57) (1y40) (1y57) (0y57)

Statistical significance 1.000(F) 1.000(F) 1.000(F) 1.000(F) –

: Fisher’s exact test. The symptom of ‘dyspareunia’ was evaluated in 15 sexually active patients after TVT and in 40 sexually active patients after PVS procedures. a

138

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

Table 4 The results of the multivariable logistic regression analysis of treatment outcome of stress urinary incontinence (SUI) on eight clinical variables of patients who underwent tension-free vaginal tape (TVT) and pubovaginal sling (PVS) procedures using polypropylene mesh Parameter

d.f.

MLE

S.E.

Wald test

P-value

b0 (Intercept) b1 (Body mass index) b2 (Procedures: TVT vs. PVS) b3 (Concurrent gynecologic surgery) b4 (previous failed SUI surgery) b5 (Age) b6 (Parity) b7 (Concomitant urge symptoms) b8 (Intrinsic sphincter deficiency) b12 (Interaction of b1 and b2) b13 (Interaction of b1 and b3) b14 (Interaction of b1 and b4) b15 (Interaction of b1 and b5) b16 (Interaction of b1 and b6) b17 (Interaction of b1 and b7) b18 (Interaction of b1 and b8) b23 (Interaction of b2 and b3) b28 (Interaction of b2 and b8) b34 (Interaction of b3 and b4) b35 (Interaction of b3 and b5) b36 (Interaction of b3 and b6) b37 (Interaction of b3 and b7) b38 (Interaction of b3 and b8) b47 (Interaction of b4 and b7) b56 (Interaction of b5 and b6) b57 (Interaction of b5 and b7) b67 (Interaction of b6 and b7) b68 (Interaction of b6 and b8)

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

102.40 y4.51 130.70 32.25 y31.03 y1.25 y12.31 55.83 y16.16 y4.79 y0.91 0.66 y0.13 y0.36 y2.38 0.49 14.53 y2.75 9.38 y0.57 3.00 y6.90 2.27 24.26 y0.01 y1.65 17.87 0.10

138.30 4.74 49.77 28.16 114.60 1.90 15.35 104.90 10.24 1.78 0.67 4.09 0.06 0.21 3.72 0.35 14.26 3.51 17.60 0.35 1.90 8.02 3.96 23.65 0.05 1.40 13.97 0.92

0.55 0.91 6.89 1.31 0.07 0.43 0.64 0.28 2.49 7.28 1.82 0.03 4.89 2.96 0.41 1.98 1.04 0.61 0.28 2.64 2.49 0.74 0.33 1.05 0.03 1.38 1.64 1.19

0.46 0.34 0.01** 0.25 0.79 0.51 0.42 0.59 0.11 0.01** 0.18 0.87 0.03** 0.08* 0.52 0.16 0.31 0.43 0.59 0.10 0.11 0.39 0.57 0.31 0.87 0.24 0.20 0.28

d.f., degree of freedom; **P-value-0.05; *marginal statistical significance; MLE, maximum likelihood estimate; S.E., standard error.

effects of patient age, parity, concurrent gynecological surgeries, ISD, previous failed incontinence surgeries and concomitant urge symptoms. However, it is significantly related to treatment procedures and their interaction with BMI. Since the interaction between BMI and the effect of treatment procedures is significant, the odds ratio (ORBMI) on the effect of treatment procedures (TVT vs. PVS) varies with BMI. From some simple calculations on the fitted logistic model, we see that TVT performs better than PVS when BMI is smaller than 27.27 kgym2 (i.e. ORBMI)1, namely, 130.7–4.7922 BMI)0), and the advantage of TVT decreases as BMI increases. The treatment outcome of SUI, by treatment procedures (TVT vs. PVS) and the cut-off value of BMI

(27.27 kgym2) is shown in Table 5. In addition, note that the P-value (0.03) of b15 (y0.13) was significant and the P-value (0.08) of b16 (y0.36) was marginal. This implies that when both BMI and age are high, the probability of surgical failure is significantly increased and when both BMI and parity are high, the probability of surgical failure could be significantly increased. 4. Discussion A basic limitation of this study was the nonrandomized selection of the two treatment arms (TVT vs. PVS), which resulted in groups not being truly comparable at baseline. Therefore, we had to express the results of this observational

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

139

Table 5 Treatment outcome of stress urinary incontinence (SUI), by treatment procedures (TVT vs. PVS) and patient body mass index (BMI) (cutoff values27.27 kgym2) SUI outcome

TVT (Ns23)

PVS (Ns57) 2

BMIF27.27(kgym )

Cure Improvement Failure

2

BMI)27.27(kgym )

BMIF27.27 (kgym2)

BMI)27.27 (kgym2)

Ns16

(70%)

Ns7

(30%)

Ns37

(65%)

Ns20

(35%)

87.4% 6.3% 6.3%

(14y16) (1y16) (1y16)

14.3% 71.4% 14.3%

(1y7) (5y7) (1y7)

75.7% 16.2% 8.1%

(28y37) (6y37) (3y37)

65.0% 30.0% 5.0%

(13y20) (6y20) (1y20)

TVT, tension-free vaginal tape; PVS: pubovaginal sling.

study by showing the crude numbers and percentages, and further by fitting a multivariable logistical regression model with important clinical variables to analyze and compare the treatment outcomes between groups. Due to the moderate size of this series, these conclusions might be preliminary. Further studies with a larger series and a longer follow-up are needed to confirm these findings. In this study, the results for TVT and PVS procedures, using polypropylene mesh slings in the treatment of female SUI, are comparable with what has been previously reported in the literature w1–8x. Our results also uphold previous findings that most concomitant urge symptoms can be resolved or improved after a successful sling operation w1–8x. Morgan et al. w3x reported a 74% cure rate of preoperative urge incontinence using the PVS procedure. However, they did not find any preoperative variable that predicted the resolution of urge incontinence postoperatively except for concomitant anterior colporrhaphy, which correlated most closely with the resolution of urge incontinence. In our study, we performed additional gynecological procedures in combination with the TVT and PVS procedures for patients who had concurrent pelvic relaxation or benign gynecological tumors. Since these problems have been correlated closely with bladder instability, our high success rates for treating concomitant urge symptoms might be attributed to these combination surgeries. Objective measures such as a cough stress test and an urodynamic study are essential for the evaluation of urinary incontinence, but they provide little information regarding the negative

impact of incontinence on women’s lives. It is perhaps for this reason that a recent proposal by the International Continence Society for the standardization of outcome measures in clinical trials of continence care has strongly suggested that the quality of life data be included in evaluations w15x. By using a previously validated symptom and quality of life questionnaire containing the IIQ-7 and UDI-6 short form w12x, we measured the changes in the negative impact of incontinence and urogenital distress on patients’ quality of life after the TVT or PVS procedure. We found that, preoperatively, patients in the TVT group had significantly higher mean UDI-6 scores than those in the PVS group (49.3 vs. 38.3, Ps0.02). This difference might be attributed to the older age of patients in the TVT group. Postoperatively, low IIQ-7 and UDI-6 scores and significant improvements were noted after both the TVT and PVS procedures. Our findings are consistent with the report of Morgan et al. w3x that low UDI-6 scores were obtained after a PVS procedure at a mean follow-up interval of 51 months with an 88% continence rate, indicating that life quality directly correlates to surgical outcomes. The most common immediate surgical complication in our study was urinary retention, which required sling revision. The rates of retention (TVT 8.7% vs. PVS 3.5%, Ps0.574) were slightly higher than what has been reported in the literature w1–8,16x. Klutke et al. w16x reported a rate of 2.8% (17y600) of urethral obstruction that lasted more than 1 week from the date of the TVT procedure. All 17 patients subsequently underwent a transvaginal release by making a simple incision at a mean interval of 64 days (range 6–228 days).

140

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141

All 17 patients voided to completion within 24 h of the release and 16 patients remained dry. In contrast, our surgical technique of sling revision was modified from what has been described by Ghoniem et al. w17x in that the sling was incised in the middle and a polypropylene mesh patch, 1– 2 cm in width, was sutured to the stump of the sling arms for elongation purposes. The integrity of the sling is resumed using our method, which is different from that described by Klutke et al. w16x. We believe that the integrity of the sling is key to the prevention of the recurrence of urine leakage, especially when a revision is done at an early postoperative phase with minimal paraurethral fibrosis formation. In comparison to Jarvis’ meta-analysis report w18x, which states that de novo urge symptoms have been reported in a mean of 16.6% (range 4– 29%) of patients after a sling operation, we found only two patients (3.5%) who complained of de novo or worse urge symptoms after the PVS procedure. Newly onset or worse urge symptoms after sling operations are considered to be a manifestation of urethral obstruction. Chaikin et al. w2x suggest that a PVS operation converts an incompetent urethral closure mechanism into an obstructive system, which in the long term would theoretically create bladder instability. This concern is further supported in the findings of the study by Amundsen et al. w19x. They found newly onset or worse urge incontinence to be the primary complaint of 93.7% of women with urethral obstruction after various sling operations, and after urethrolysis, they found urge symptoms were resolved in 67% and were significantly improved in 27% of these cases. The low incidence of newly onset or the worsening of urge symptoms in our study might be characteristic of our surgical technique, which focuses on avoiding sling tension and the early sling revision procedure. There is clear epidemiological support for the role of obesity in urinary incontinence. Data from several studies indicate that urinary incontinence in women is associated with higher BMI and greater weight w20,21x. It is believed that the added weight of obesity may bear down on pelvic tissues causing chronic strain, stretching and weakening of the muscles, nerves, and other structures of the

pelvic floor w22x. However, corroborating results have yet to be reported which support obesity as adversely affecting the results of anti-incontinence surgery. In a five-year follow-up study of 187 surgically treated patients with SUI, no statistical difference in the recurrence rate was noted between normal weight and overweight patients w23x. Several references in the literature support the efficacy of TVT in obese women. A publication by Mukherjee and Constantine w24x, with data on 242 consecutive women who underwent TVT, showed that TVT is as effective in women with a BMI of -25 kgym2 and with a BMI of G30 kgym2. Unfortunately, the result was subjective and was evaluated at a short follow-up interval of 6 months postoperatively. Therefore, the long-term effect of increasing BMI on efficacy of TVT was not evaluated. Another study of BMI and TVT by Rafii et al. w25x showed no statistically significant difference of the efficacy of TVT in groups of women of different BMI ranges at a mean followup interval of 27 months. However, there was a trend noted on the decreasing efficacy of TVT with increasing patient BMI by objective cure rates (93% vs. 88.7% vs. 82%; P-values0.1). Contrary to the conclusion of these reports, our study showed that TVT was not as effective as PVS was in treating overweight (25–30 kgym2) or obese women ()30 kgym2). Meanwhile, old age and high parity reinforce the adverse effect of obesity on the efficacy of TVT. The reason for this phenomenon was not fully understood from the present study. One possible explanation for the results is that these two procedures are different in the continence mechanisms and surgical techniques. That is, the mid-urethral kinking effect of TVT takes precedence over the ‘hammock’ effect of PVS in treating normal weight, incontinent women. However, the advantage of TVT deteriorates in treating overweight or obese incontinent women, in whom the increased intra-abdominal pressure w26x and less effective neuromuscular function of the pelvic floor w22x becomes a hindrance to the TVT procedure. Further research is warranted in order to address this finding. References w1x McGuire EJ, Lytton B. Pubovaginal sling procedure for stress incontinence. J Urol 1978;119:82 –84.

M.J. Hung et al. / International Journal of Gynecology and Obstetrics 84 (2004) 133–141 w2x Chaikin DC, Rosenthal J, Blaivas JG. Pubovaginal fascial sling for all types of stress urinary incontinence: long-term analysis. J Urol 1998;160:1312 –1316. w3x Morgan TO Jr., Westney OL, McGuire EJ. Pubovaginal sling: 4-year outcome analysis and quality of life assessment. J Urol 2000;163:1845 –1848. w4x Leach GE, Dmocowski RR, Appell RA, Blaivas JG, Hadley HR, Luber KM. Female Stress Urinary Incontinence Clinical Guidelines Panel summary report on surgical management of female stress urinary incontinence. J Urol 1997;158:875 –880. w5x Wright EJ, Iselin CE, Carr LK, Webster GD. Pubovaginal sling using cadaveric allograft fascia for the treatment of intrinsic sphincter deficiency. J Urol 1998;160:759 –762. w6x Kuo HC. Comparison of video urodynamic results after the pubovaginal sling procedure using rectus fascia and polypropylene mesh for stress urinary incontinence. J Urol 2001;165:163 –168. w7x Ulmsten U, Johnson P, Rezapour M. A three-year followup of tension free vaginal tape for surgical treatment of female stress urinary incontinence. Br J Obstet Gynecol 1999;106:345 –350. w8x Nilsson CG, Kuuva N, Falconer C, Rezapour M, Ulmsten U. Long-term results of the tension-free vaginal tape (TVT) procedure for surgical treatment of female stress urinary incontinence. Int Urogynecol J 2001;2(Suppl):S5 –S8. w9x Petros P, Ulmsten U. An integral theory and its method for the diagnosis and management of female urinary incontinence. Scan J Urol Nephrol 1993;153:1 –93. w10x Lo TS, Wang A, Horng SG, Liang CC, Soong YK. Ultrasonograpic and urodynamic evaluation after tension free vaginal tape procedure (TVT). Acta Obstet Gynecol Scand 2001;80:65 –70. w11x DeLancey JOL. Structural support of the urethra as it relates to stress urinary incontinence: the hammock hypothesis. Am J Obstet Gynecol 1994;170:1713 –1723. w12x Uebersax JS, Wyman JF, Shumaker SA, McClish DK, Fantl JA. Short forms to assess life quality and symptom distress for urinary incontinence in women: the Incontinence Impact Questionnaire and the Urogenital Distress Inventory. Continence Program for women Research Group. Neurourol Urodyn 1995;14:131 –139.

141

w13x Abrams P, Blaivas JG, Stanton SL, Andersen JT. The standardization of terminology of lower urinary tract function. Scan J Urol Nephrol 1988;114(Suppl):5 –19. w14x Hosmer DW, Lemeshow SA. Goodness-of-fit test for multiple logistic regression model. Commun Stat Ser A 1980;9:1043 –1069. w15x Mattiasson A, Djurhuus JC, Fonda D, Lose G, Nordling J, Storer M. Standardisation of outcome studies in patients with lower urinary tract dysfunction: a report on general principles from the Standardisation Committee of the International Continence Society. Neurourol Urodyn 1998;17:249 –253. w16x Klutke C, Siegel S, Carlin B, Paszkiewicz E, Kirkemo A, Klutke J. Urinary retention after tension-free vaginal tape procedure: incidence and treatment. Urology 2001;58:697 –701. w17x Ghoniem GM, Elgamasy AN. Simplified surgical approach to bladder outlet obstruction following pubovaginal sling. J Urol 1995;154:181 –183. w18x Jarvis GJ. Surgery for genuine stress incontinence. Br J Obstet Gynecol 1994;101:371 –374. w19x Amundsen CL, Guralnick ML, Webster GD. Variations in strategy for the treatment of urethral obstruction after a pubovaginal sling procedure. J Urol 2000;164:434 – 437. w20x Dwyer PL, Lee ETC, Hay DM. Obesity and urinary incontinence in women. Br J Obstet Gynaecol 1988;95:91 –96. w21x Burgio KL, Mattews KA, Engle BT. Prevalence, incidence and correlates of urinary incontinence in healthy middle-aged women. J Urol 1991;146:1255 –1259. w22x Cummings JM, Rodning CB. Urinary stress incontinence among obese women: review of pathophysiology therapy. Int Urogynecol J 2000;11:41 –44. w23x Zivkovic F, Tamussino K, Peiber D. Body mass index and outcome of incontinence surgery. Obstet Gynecol 1999;93:753 –756. w24x Mukherjee K, Constantine G. Urinary stress incontinence in obese women: tension-free vaginal tape is the answer. Br J Urol Int 2001;88(9):881 –883. w25x Rafii A, Darai E, Haab F, Samain E, Levardon M, Deval B. Body mass index and outcome of tension-free vaginal tape. Eur Urol 2003;43(3):288 –292. w26x Noblett KL, Jensen JK, Ostergard DR. The relationship of body mass index to intraabdominal pressure as measured by multichannel cystometry. Int Urogynecol J 1997;8:323 –326.