Int Urogynecol J (2015) 26:57–63 DOI 10.1007/s00192-014-2454-2
ORIGINAL ARTICLE
TVT versus laparoscopic mesh colposuspension: 5-year follow-up results of a randomized clinical trial A. Valpas & S. Ala-Nissilä & E. Tomas & C. G. Nilsson
Received: 2 April 2014 / Accepted: 14 June 2014 / Published online: 24 July 2014 # The International Urogynecological Association 2014
Abstract Introduction and hypothesis Before the introduction of the tension-free vaginal tape (TVT) procedure for the treatment of female stress urinary incontinence, the colposuspension operation was regarded as the “gold standard” procedure. The laparoscopic variant of the colposuspension was introduced as a less invasive operation. The aim of the present trial was to compare the new minimally invasive TVT procedure with laparoscopic mesh colposuspension (LCM). Methods A multicenter randomized clinical trial conducted in six public hospitals in Finland including primary cases of stress incontinence. Objective treatment success criteria were a negative stress test and no retreatment for stress incontinence. Patient satisfaction was assessed by Patients Global Impression of Improvement, a visual analog scale, and the Urinary Incontinence Severity Score. Results Of 128 randomized patients, 121 underwent the allocated operation. At the 5-year follow-up 77 % in the TVT group and 84 % in the LCM group could be assessed according to the protocol. The objective cure rate was significantly higher in the TVT group (94 %) than in the LCM group (78 %). Subjective treatment satisfaction (completely satisfied A. Valpas Department of Obstetrics and Gynecology, South-Karelian Central Hospital, Lappenranta, Finland S. Ala-Nissilä Department of Obstetrics and Gynecology, Turku University Hospital, Turku, Finland E. Tomas Department of Obstetrics and Gynecology, University of Tampere, Tampere, Finland C. G. Nilsson (*) Department of Obstetrics and Gynecology, Helsinki University Central Hospital, HUS, POB 140, 00029 Helsinki, Finland e-mail: [email protected]
with the procedure) was significantly higher in the TVT group (64 %) than in the LCM group (51 %). Conclusions By per protocol analysis both objective and subjective cure rates were significantly higher in the TVT group than in the LCM group. If cases that were lost to follow-up were regarded as failures, the intension-to-treat analysis found no difference between the groups. Keywords Laparoscopic colposuspension . Randomized trial . Tension-free vaginal tape
Introduction After the introduction of the tension-free vaginal tape (TVT) procedure in 1996 [1] surgical treatment of female stress urinary incontinence changed dramatically. Prior to the introduction of TVT the laparoscopic modification of the open form of colposuspension was regarded as a minimally invasive procedure for the treatment of stress urinary incontinence (SUI). Simultaneously, the open colposuspension was regarded as the gold standard procedure for the treatment of SUI. Long-term follow-up studies of open colposuspension report cure rates between 63 % and 81 % after 5 to 20 years of follow-up [2–6]. Few randomized trials compare TVT with traditional antiincontinence procedures during long-term follow-up. The 5years follow-up results of the largest trial, which compares TVT with open colposuspension, suggests that there might be no significant difference in cure rates between the two [7]. A Cochrane review comparing laparoscopic colposuspension with TVT found that there is a trend in favor of TVT in the short term and that long-term follow-up data are needed as the follow-up time of the studies included in the review was only 18 months [8]. The Cochrane review concludes that the longterm performance of laparoscopic colposuspension remains
58
Int Urogynecol J (2015) 26:57–63
uncertain. Later a small RCT from the USA, comparing TVT with laparoscopic colposuspension, with 36 patients in each arm, has been published. The follow-up time was 5.4 years and no significant differences in efficacy between TVT and laparoscopic colposuspension was found [9]. In the present trial TVT is compared with laparoscopic mesh colposuspension (LCM). We have previously published the immediate postoperative results and complications and the 12-month follow-up results [10, 11]. The 12-month results showed a higher objective and subjective cure rate for the TVT procedure than for the LCM, even though recruitment fell short of the goal set by the power calculation. The trial was planned to evaluate results at 3 and 5 years after surgery. Because of the shortcoming of recruitment and the number of subjects lost to follow-up we decided not to analyze and publish the 5-year results. When one of the authors participated in a European guidelines panel on incontinence in 2013, it became obvious that by that year still no long-term RCTs comparing TVT with laparoscopic colposuspension had been published. This fact was the reason for this late analysis and publication of the 5-years results.
Table 1 Inclusion and exclusion criteria
Materials and methods
5 % and a type II error of 20 % is accepted and if a smallest noticeable difference of 10 % is assumed. Analysis was carried out by using both the per protocol (PP) and the intention-to-treat (ITT) principle for the primary outcome. In the ITT analysis all the women who were lost to follow-up were regarded as failures. P values 32 Contraindication for general anesthesia Contraindication for laparoscopy Unstable bladder in filling cystometry Urethral closure pressure 100 ml
Urodynamically verified stress urinary incontinence Positive stress test (300 ml saline in bladder)
Int Urogynecol J (2015) 26:57–63
was allowed to escape. Cystoscopy with a 70° optic lens was performed intraoperatively. For the LCM an extraperitoneal approach was used [10]. A preperitoneal balloon dilator (PDB®, Preperitoneal Distension Balloon System; Origin Medsystems, Menlo Park, CA, USA) was placed under the anterior rectus fascia, advanced into the space of Retzius and filled; thus, space for the operation was created. Two to three 5-mm trocars were inserted and two strips of pre-cut polypropylene mesh 12 mm wide and 4 cm long were attached with tackers to the paraurethral tissue and Cooper’s ligament. A catheter was left in place and removed the next morning. The women attended the 12-, 36-, and 60-month follow-up visits at the clinics as originally planned. They were assessed according to the protocol by a research nurse and the doctor in charge at the clinic, who often was not the surgeon who performed the initial operation. A thorough gynecological examination was performed in order to detect adverse tissue reactions. Because of the very difference in operation techniques (LCM in general anesthesia and TVT in local anesthesia) no blinding was possible. Primary outcome was objective treatment success. Criteria for objective success were a negative stress test performed in the same manner as preoperatively and no re-operation because of persistent or recurrent stress incontinence. Secondary outcome was subjective treatment success. Criteria for subjective success were a significant decline in VAS and UISS scores compared with preoperative scores. Any additional surgery performed for any reason was registered during the follow-up period. The independent samples t test was used for continuous variables for base-line statistics. For categorical variables Fisher’s exact test/Chi-squared test was performed .To overcome the lost to follow-up problem the 5-year follow-up data were analyzed using generalizing estimating equations (GEE) modeling [13]. Statistical significance between groups was tested by the t test, the Chi-squared test, or Fisher’s exact test when appropriate. Repeated measures were analyzed using GEE models with the unstructured correlation structure. GEE were developed as an extension to the general linear model (e.g., OLS regression analysis) to analyze longitudinal and other correlated data. GEE models take into account the correlation between repeated measurements in the same subject; models do not require complete data and can be fitted even if individuals do not have observations at all time points [13]. Also, a permutation type ANCOVA, with the baseline as covariate, was used to calculate the differences between VAS and UISS as outcome measures after 5 years of follow-up. STATA 12.1, StataCorp LP (College Station, TX, USA) statistical package was used for the analyses. To confirm the results medians were also compared using the Mann–Whitney U test. The Statistical Package
59
for Social Sciences (version 21; SPSS, Chicago, IL, USA) was used for statistical analysis. Methods used, definitions, and units conform to the terminology of the International Urogynecological Association and the International Continence Society joint report on the terminology for female floor dysfunction [14]. The trial was initiated by the responsible researchers and funded solely by Finnish governmental research funding. The surgeries in all the participating public hospitals were performed according to the routines of the hospitals. The manufacturers of the material used within the trial had no part in providing the material, the planning of the trial, the analysis of the data or preparation of the manuscript
Results The operations were performed between November 1999 and August 2001. The last 5-year follow-up evaluation was carried out in January 2007. One-hundred and twenty-eight women were randomized and 121 women underwent the allocated operation. Four patients refused operation for unknown reasons, 2 patients did not accept the result of the randomization, and 1 patient was operated upon in a hospital not participating in the trial. Because recruitment fell short and the randomization design was in blocks of 40 for each hospital the final number of women operated upon with the TVT procedure was 70 and those with the LCM procedure 51. At the 5-year follow-up 18 patients were lost to follow-up, 11 in the TVT group and 7 in the LCM group. One patient in the TVT group had died for reasons not related to the surgery and 4 patients refused to attend the 5-year follow-up visit as did one patient in the LCM group. Data on 54 out of 70 (77.1 %) in the TVT group and 43 out of 51 (84.3 %) in the LCM group could be analyzed. Figure 1 shows the flow chart of the trial. No demographic differences between the initial groups were seen (Table 2). After the index operation 2 patients in the TVT group had had a surgical intervention because of procedure-related problems or persistent or recurrent incontinence: one tape revision because of tape protrusion and one bulking agent injection with Zuidex® because of persistent incontinence. In the LCM group 4 patients had had a surgical intervention: 3 TVT procedures and one transobturator sling procedure owing to persistent stress incontinence. Three vaginal procedures were performed during follow-up in the LCM group (one vaginal hysterectomy, another vaginal hysterectomy combined with a posterior colporrhaphy, and an anterior colporrhaphy) and one in the TVT group (a vaginal hysterectomy combined with an anterior colporrhaphy). The primary outcome results, the stress test, are given as PP and ITT analysis in Table 3 and the results of the stress test analyzed according to PP are given for 1, 3, and 5 years in
60
Int Urogynecol J (2015) 26:57–63
Fig. 1 Flow chart of the trial Randomised (n = 128)
Refused operation (n = 4) Result of randomisation not accepted (n = 2) Operated on in hospital not participating in the study (n = 1)
Allocated to TVT (n = 70) Received treatment as allocated (n = 70)
Allocated to LCM (n = 51) Received treatment as allocated (n = 51)
Followed up at 1 year (n = 66) Lost to follow up (n = 3) Withdrew (n = 1)
Followed up at 1 year (n = 49) Lost to follow up (n = 2)
Followed up at 3 years (n = 62) Lost to follow up (n = 4) Withdrew (n = 3) Died (n = 1)
Followed up at 3 years (n = 47) Lost to follow up (n = 2) Withdrew (n = 1) Data missing for unknown reason (n =1)
Followed up at 5 years (n = 54) Lost to follow up (n = 11) Withdrew (n = 4) Died (n = 1; same person as above)
Followed up at 5 years (n = 43) Lost to follow up (n = 7) Withdrew (n = 1)
Fig. 2. The objective PP cure rate (negative stress test) was 94 % in the TVT group and 78 % in the LCM group, the difference being statistically significant (P
ORIGINAL ARTICLE
TVT versus laparoscopic mesh colposuspension: 5-year follow-up results of a randomized clinical trial A. Valpas & S. Ala-Nissilä & E. Tomas & C. G. Nilsson
Received: 2 April 2014 / Accepted: 14 June 2014 / Published online: 24 July 2014 # The International Urogynecological Association 2014
Abstract Introduction and hypothesis Before the introduction of the tension-free vaginal tape (TVT) procedure for the treatment of female stress urinary incontinence, the colposuspension operation was regarded as the “gold standard” procedure. The laparoscopic variant of the colposuspension was introduced as a less invasive operation. The aim of the present trial was to compare the new minimally invasive TVT procedure with laparoscopic mesh colposuspension (LCM). Methods A multicenter randomized clinical trial conducted in six public hospitals in Finland including primary cases of stress incontinence. Objective treatment success criteria were a negative stress test and no retreatment for stress incontinence. Patient satisfaction was assessed by Patients Global Impression of Improvement, a visual analog scale, and the Urinary Incontinence Severity Score. Results Of 128 randomized patients, 121 underwent the allocated operation. At the 5-year follow-up 77 % in the TVT group and 84 % in the LCM group could be assessed according to the protocol. The objective cure rate was significantly higher in the TVT group (94 %) than in the LCM group (78 %). Subjective treatment satisfaction (completely satisfied A. Valpas Department of Obstetrics and Gynecology, South-Karelian Central Hospital, Lappenranta, Finland S. Ala-Nissilä Department of Obstetrics and Gynecology, Turku University Hospital, Turku, Finland E. Tomas Department of Obstetrics and Gynecology, University of Tampere, Tampere, Finland C. G. Nilsson (*) Department of Obstetrics and Gynecology, Helsinki University Central Hospital, HUS, POB 140, 00029 Helsinki, Finland e-mail: [email protected]
with the procedure) was significantly higher in the TVT group (64 %) than in the LCM group (51 %). Conclusions By per protocol analysis both objective and subjective cure rates were significantly higher in the TVT group than in the LCM group. If cases that were lost to follow-up were regarded as failures, the intension-to-treat analysis found no difference between the groups. Keywords Laparoscopic colposuspension . Randomized trial . Tension-free vaginal tape
Introduction After the introduction of the tension-free vaginal tape (TVT) procedure in 1996 [1] surgical treatment of female stress urinary incontinence changed dramatically. Prior to the introduction of TVT the laparoscopic modification of the open form of colposuspension was regarded as a minimally invasive procedure for the treatment of stress urinary incontinence (SUI). Simultaneously, the open colposuspension was regarded as the gold standard procedure for the treatment of SUI. Long-term follow-up studies of open colposuspension report cure rates between 63 % and 81 % after 5 to 20 years of follow-up [2–6]. Few randomized trials compare TVT with traditional antiincontinence procedures during long-term follow-up. The 5years follow-up results of the largest trial, which compares TVT with open colposuspension, suggests that there might be no significant difference in cure rates between the two [7]. A Cochrane review comparing laparoscopic colposuspension with TVT found that there is a trend in favor of TVT in the short term and that long-term follow-up data are needed as the follow-up time of the studies included in the review was only 18 months [8]. The Cochrane review concludes that the longterm performance of laparoscopic colposuspension remains
58
Int Urogynecol J (2015) 26:57–63
uncertain. Later a small RCT from the USA, comparing TVT with laparoscopic colposuspension, with 36 patients in each arm, has been published. The follow-up time was 5.4 years and no significant differences in efficacy between TVT and laparoscopic colposuspension was found [9]. In the present trial TVT is compared with laparoscopic mesh colposuspension (LCM). We have previously published the immediate postoperative results and complications and the 12-month follow-up results [10, 11]. The 12-month results showed a higher objective and subjective cure rate for the TVT procedure than for the LCM, even though recruitment fell short of the goal set by the power calculation. The trial was planned to evaluate results at 3 and 5 years after surgery. Because of the shortcoming of recruitment and the number of subjects lost to follow-up we decided not to analyze and publish the 5-year results. When one of the authors participated in a European guidelines panel on incontinence in 2013, it became obvious that by that year still no long-term RCTs comparing TVT with laparoscopic colposuspension had been published. This fact was the reason for this late analysis and publication of the 5-years results.
Table 1 Inclusion and exclusion criteria
Materials and methods
5 % and a type II error of 20 % is accepted and if a smallest noticeable difference of 10 % is assumed. Analysis was carried out by using both the per protocol (PP) and the intention-to-treat (ITT) principle for the primary outcome. In the ITT analysis all the women who were lost to follow-up were regarded as failures. P values 32 Contraindication for general anesthesia Contraindication for laparoscopy Unstable bladder in filling cystometry Urethral closure pressure 100 ml
Urodynamically verified stress urinary incontinence Positive stress test (300 ml saline in bladder)
Int Urogynecol J (2015) 26:57–63
was allowed to escape. Cystoscopy with a 70° optic lens was performed intraoperatively. For the LCM an extraperitoneal approach was used [10]. A preperitoneal balloon dilator (PDB®, Preperitoneal Distension Balloon System; Origin Medsystems, Menlo Park, CA, USA) was placed under the anterior rectus fascia, advanced into the space of Retzius and filled; thus, space for the operation was created. Two to three 5-mm trocars were inserted and two strips of pre-cut polypropylene mesh 12 mm wide and 4 cm long were attached with tackers to the paraurethral tissue and Cooper’s ligament. A catheter was left in place and removed the next morning. The women attended the 12-, 36-, and 60-month follow-up visits at the clinics as originally planned. They were assessed according to the protocol by a research nurse and the doctor in charge at the clinic, who often was not the surgeon who performed the initial operation. A thorough gynecological examination was performed in order to detect adverse tissue reactions. Because of the very difference in operation techniques (LCM in general anesthesia and TVT in local anesthesia) no blinding was possible. Primary outcome was objective treatment success. Criteria for objective success were a negative stress test performed in the same manner as preoperatively and no re-operation because of persistent or recurrent stress incontinence. Secondary outcome was subjective treatment success. Criteria for subjective success were a significant decline in VAS and UISS scores compared with preoperative scores. Any additional surgery performed for any reason was registered during the follow-up period. The independent samples t test was used for continuous variables for base-line statistics. For categorical variables Fisher’s exact test/Chi-squared test was performed .To overcome the lost to follow-up problem the 5-year follow-up data were analyzed using generalizing estimating equations (GEE) modeling [13]. Statistical significance between groups was tested by the t test, the Chi-squared test, or Fisher’s exact test when appropriate. Repeated measures were analyzed using GEE models with the unstructured correlation structure. GEE were developed as an extension to the general linear model (e.g., OLS regression analysis) to analyze longitudinal and other correlated data. GEE models take into account the correlation between repeated measurements in the same subject; models do not require complete data and can be fitted even if individuals do not have observations at all time points [13]. Also, a permutation type ANCOVA, with the baseline as covariate, was used to calculate the differences between VAS and UISS as outcome measures after 5 years of follow-up. STATA 12.1, StataCorp LP (College Station, TX, USA) statistical package was used for the analyses. To confirm the results medians were also compared using the Mann–Whitney U test. The Statistical Package
59
for Social Sciences (version 21; SPSS, Chicago, IL, USA) was used for statistical analysis. Methods used, definitions, and units conform to the terminology of the International Urogynecological Association and the International Continence Society joint report on the terminology for female floor dysfunction [14]. The trial was initiated by the responsible researchers and funded solely by Finnish governmental research funding. The surgeries in all the participating public hospitals were performed according to the routines of the hospitals. The manufacturers of the material used within the trial had no part in providing the material, the planning of the trial, the analysis of the data or preparation of the manuscript
Results The operations were performed between November 1999 and August 2001. The last 5-year follow-up evaluation was carried out in January 2007. One-hundred and twenty-eight women were randomized and 121 women underwent the allocated operation. Four patients refused operation for unknown reasons, 2 patients did not accept the result of the randomization, and 1 patient was operated upon in a hospital not participating in the trial. Because recruitment fell short and the randomization design was in blocks of 40 for each hospital the final number of women operated upon with the TVT procedure was 70 and those with the LCM procedure 51. At the 5-year follow-up 18 patients were lost to follow-up, 11 in the TVT group and 7 in the LCM group. One patient in the TVT group had died for reasons not related to the surgery and 4 patients refused to attend the 5-year follow-up visit as did one patient in the LCM group. Data on 54 out of 70 (77.1 %) in the TVT group and 43 out of 51 (84.3 %) in the LCM group could be analyzed. Figure 1 shows the flow chart of the trial. No demographic differences between the initial groups were seen (Table 2). After the index operation 2 patients in the TVT group had had a surgical intervention because of procedure-related problems or persistent or recurrent incontinence: one tape revision because of tape protrusion and one bulking agent injection with Zuidex® because of persistent incontinence. In the LCM group 4 patients had had a surgical intervention: 3 TVT procedures and one transobturator sling procedure owing to persistent stress incontinence. Three vaginal procedures were performed during follow-up in the LCM group (one vaginal hysterectomy, another vaginal hysterectomy combined with a posterior colporrhaphy, and an anterior colporrhaphy) and one in the TVT group (a vaginal hysterectomy combined with an anterior colporrhaphy). The primary outcome results, the stress test, are given as PP and ITT analysis in Table 3 and the results of the stress test analyzed according to PP are given for 1, 3, and 5 years in
60
Int Urogynecol J (2015) 26:57–63
Fig. 1 Flow chart of the trial Randomised (n = 128)
Refused operation (n = 4) Result of randomisation not accepted (n = 2) Operated on in hospital not participating in the study (n = 1)
Allocated to TVT (n = 70) Received treatment as allocated (n = 70)
Allocated to LCM (n = 51) Received treatment as allocated (n = 51)
Followed up at 1 year (n = 66) Lost to follow up (n = 3) Withdrew (n = 1)
Followed up at 1 year (n = 49) Lost to follow up (n = 2)
Followed up at 3 years (n = 62) Lost to follow up (n = 4) Withdrew (n = 3) Died (n = 1)
Followed up at 3 years (n = 47) Lost to follow up (n = 2) Withdrew (n = 1) Data missing for unknown reason (n =1)
Followed up at 5 years (n = 54) Lost to follow up (n = 11) Withdrew (n = 4) Died (n = 1; same person as above)
Followed up at 5 years (n = 43) Lost to follow up (n = 7) Withdrew (n = 1)
Fig. 2. The objective PP cure rate (negative stress test) was 94 % in the TVT group and 78 % in the LCM group, the difference being statistically significant (P
