Introduction

Urinary incontinence is a highly prevalent dysfunction in middle-aged and elderly women, figuring as a social stigma that has a negative impact on quality of life [1].

Many epidemiological studies have linked obesity to a higher risk of urinary incontinence, related to both stress and urgency [2–4]. Considering the increasing rates of obesity worldwide [5, 6], this specific sub-population has to be taken into account whenever considering indications, risks, and outcomes of incontinence surgical procedures.

After the first description of tension-free vaginal tape (TVT) by Ulmsten and Petros in 1995 [7], midurethral slings (MUS) quickly became the most common surgical procedure performed for stress urinary incontinence, because of the reports of high success rates and low morbidity [8, 9].

Despite being the gold standard for treating genuine stress incontinence, the MUS of choice in complex incontinence situations, such as obesity, is still debatable. So far, conflicting data have been published regarding the outcomes of these procedures in obese women. As a number of studies have reported no differences in the outcome of midurethral slings according to BMI, others have suggested cure rates of 10 to 30 % lower than those seen in non-obese patients [10–24]. Importantly, significant limitations and potential confounders are reported in most of these studies, including small sample sizes, short-term follow-up analysis, and variable follow-up periods, the use of different body mass index (BMI) cut-offs to determine obesity, or the use of nonvalidated survey tools for primary outcome measurement.

As the present literature may be of little quality regarding the impact of BMI on the success of midurethral slings, these procedures are still lacking sustained validation in the obese sub-population.

The aim of our study was to evaluate the effectiveness and safety of a transobturator midurethral sling in incontinent obese women over a 4-year period. We hypothesized that the outcomes of a transobturator MUS, such as TVT-O, do not differ according to BMI.

Materials and methods

We conducted a retrospective comparative study of all women who underwent incontinence surgery using the inside-out mid-urethral transobturator sling TVT-O (Gynecare TVT™ Obturator System, Ethicon), between January 2004 and June 2012 at the Urogynecology Unit of a tertiary, University Hospital. Data were collected from an ongoing database designed for clinical research.

Written informed consent, approved by the Ethics Review Board (ERB) of our Institution, was obtained from all women at the preoperative consultation, concerning all medical interventions and the potential use of clinical data in future research studies. As this was a retrospective data analysis, with no interference in any medical decision or intervention, no additional ERB approval was needed.

All women had a chief complaint of involuntary urinary loss related to stress. Cases of mixed urinary incontinence were included whenever stress was the major symptom. A positive cough test on pelvic examination was mandatory for inclusion.

Cases of occult urinary incontinence or previous failed MUS were excluded. We also excluded any patient with an active urinary or vaginal infection, medication or any major medical condition possibly related to voiding dysfunction, such as poorly controlled diabetes, multiple sclerosis or spinal cord injury.

Cases of concomitant pelvic organ prolapse (POP) on pelvic examination or concomitant pelvic floor reconstructive surgery for POP at the time of the indexed surgery were not excluded from the analysis.

According to their body mass index (BMI), women were stratified in two groups—obese (BMI ≥ 30 kg/m2) and non-obese (BMI < 30 kg/m2)—using the cutoff points recommended by the World Health Organization for the western population [25].

Preoperative evaluation included a detailed medical history with symptom severity assessment, a pelvic examination with a cough test, urinalysis, and a complete urodynamic study, involving filling and voiding cystometries, uroflowmetry, and urethral pressure profile.

The impact on quality of life was assessed pre- and postoperatively using the King’s Health Questionnaire (KHQ) validated for Portuguese [26], in which women are asked about urinary incontinence symptoms and quality of life related to urinary incontinence. With this instrument, a higher score indicates worse symptoms and poorer quality of life.

Postoperative follow-up was scheduled at 6, 12, 24, and 48 months. Symptom assessment, pelvic examination, and a cough test were performed at all visits. Only records with the preoperative and at least two postoperative visits were considered for the analysis.

Our primary outcome was the objective cure, defined as no urine leakage during the cough test performed with full bladder at each follow-up visit [27]. The presence of any leakage with the same maneuver was considered to indicate surgical failure.

Our secondary outcomes were perioperative complications and impact on quality of life. As perioperative complications we considered the presence of excessive bleeding (ΔHb > 2 g/dL), vaginal wall perforation, bladder perforation, urinary retention (postvoiding residue > 100 ml), de novo overactive bladder (OAB) and vaginal mucosal erosion. As for the impact on quality of life, we relied on the variations in KHQ scores between baseline (preoperative) and each follow-up visit, stratified into three classes: “cured” if the KHQ score was < 30; “improved” if the KHQ score  was ≥ 30, but with an improvement from baseline > 5 points; “not improved” if the KHQ score was ≥ 30, with an improvement from baseline ≤ 5 points or with any aggravation (KHQ score at follow-up above the baseline score).

Data were compared using the Chi-squared and Fisher’s exact test for categorical variables and Student’s t test or the Mann–Whitney test for continuous variables. GraphPad Prism 6 (GraphPad Software, San Diego, CA, USA) was used for the statistical purposes. A p value < 0.05 was considered statistically significant.

Results

During the study period, approximately 600 women underwent an MUS procedure at our institution. Our final study group included 281 patients, 122 in the obese group and 159 in the non-obese group. The number of patients considered for analysis at each study timepoint is presented in Fig. 1. Patients’ mean BMIs for the obese and non-obese groups were 33.2 ± 3.4 kg/m2 and 26.9 ± 2.2 kg/m2 respectively. Demographics and preoperative parameters of the groups were compared and no significant differences were found, except for BMI, as presented in Table 1.

Fig. 1
figure 1

Number of women considered for analysis at each study time point. All women who followed up provided a stress test on pelvic examination, but some did not provide a valid King’s Health Questionnaire (KHQ)

Full size image
Table 1 Baseline demographics and preoperative parameters
Full size table

For both groups the objective cure rate was higher than 90 % at each follow-up visit, with no significant differences (Table 2). At 48 months, our follow-up rate was 59 % (n = 96) and 60.4 % (n = 72) in the non-obese and obese groups respectively (p = 0.9).

Table 2 Objective cure rates according to the BMI group
Full size table

In terms of impact on quality of life, the improvement in the mean KHQ total scores after surgery was significant compared with the preoperative scores and there were no significant differences between groups (Fig. 2). Regarding the KHQ score variation by classes, we also found both groups to feel mainly “cured” or “improved” at all follow-up assessments, with no significant differences in the comparative analysis over the 4-year period (Table 3).

Fig. 2
figure 2

Mean KHQ total scores before and after surgery (at 6, 12, 24, and 48 months), according to BMI group. The improvement at each follow-up visit was significant compared with baseline (p trend <0.0001) and there were no significant differences between groups. Baseline: non-obese n = 159, obese n = 122; 6 months: non-obese n = 126, obese n = 93; 12 months: non-obese n = 114, obese n = 82; 24 months: non-obese n = 88, obese n = 64; 48 months: non-obese n = 40, obese n = 30

Full size image
Table 3 Variation in KHQ scores according to BMI group
Full size table

Concerning perioperative complications we registered in the non-obese group 1 case of excessive bleeding, 2 vaginal perforations, 1 case of de novo OAB, 2 vaginal mucosal erosions and no cases of urinary retention or bladder perforation. In the obese group there was 1 case of excessive bleeding, 3 vaginal perforations, 2 cases of de novo OAB, 1 case of urinary retention, 2 vaginal mucosal erosions, and no cases of bladder perforation. Comparing their cumulative rates, no significant differences were found in the incidence of complications among groups (non-obese 3.8 % vs obese 7.4 %, p = 0.19). We also registered a low re-operation rate, with 2.5 % of women (3 out of 122) in the obese group being subjected to a second related procedure (1 for mesh erosion, 1 for urinary retention, and 1 for surgical failure) and 1.3 % (2 out of 159) in the non-obese group (both for mesh erosion), p = 0.66.

Discussion

In this study the objective cure rate showed no significant difference between obese and non-obese women. Also, considering the impact of surgery on quality of life, the variation in KHQ scores showed no significant differences between groups. We registered a low complication rate, with no significant differences related to BMI, and our global erosion rate was set at 1.4 % (4 out of 281).

Although no differences were found between groups, our objective cure rates were higher than the subjective cures based on the KHQ score analysis. This discrepancy can be related to the influence of overactive bladder symptoms—present in approximately half of the women before surgery—on the KHQ scores, conditioning higher improvements in this analysis. Additionally, a limitation related to objective cure-established criteria could also be pointed out. As the cough stress test was systematically performed in a lithotomy position and in a clinical environment, it is expected that some women with a negative stress test may actually experience some degree of urine leakage in their daily life activities, with a potential impact on their KHQ scores. Nevertheless, it is interesting to notice that, considering both women who were “cured” and “improved” according to the KHQ class distribution, the rate of patient satisfaction comes closer to the objective cure, and this is true for both groups studied.

Obesity is a known independent risk factor for the development of stress urinary incontinence (SUI), although the exact underlying mechanism is not yet completely clear. Mommsen and Foldspang studied 2,589 women aged 30 to 59 years, and reported that, independently of other risk factors, BMI had a positive correlation with the incidence of urinary incontinence, most closely with stress urinary incontinence [3]. Obese patients experience higher levels of intra-abdominal pressure and damage to the pelvic floor functions may come from this chronic state [28].

Since the description of the TVT procedure by Ulmsten and Petros [7], minimally invasive MUS have been applied for the surgical treatment of SUI, with highly satisfactory results in index patients with genuine stress incontinence [8, 9, 29–31].

As slings also started to be used in obese women as a substitute for the previous invasive incontinence surgeries, concerns for lower cure rates and higher perioperative complications were raised, leading to several reports on the outcomes of these procedures in this at-risk population in the last 10 years [10–24]. Table 4 summarizes the features of the main studies previously published on this topic.

Table 4 Summarized features of previous studies evaluating the impact of obesity in the success of midurethral slings
Full size table

Liu et al. retrospectively compared TVT-O efficacy and safety in 29 normal, 58 overweight, and 32 obese women. At a median 24 months’ follow-up, they found no differences in objective and subjective cure rates or postoperative complications among the three groups [15]. By contrast, Haverkorn et al. compared 117 obese women (BMI > 30 kg/m2) with 161 normal controls (BMI < 30 kg/m2) at a minimum of 12 months after a MONARC procedure. Defining cure using both objective cough test and a validated questionnaire, they found a significantly lower cure rate in the obese population (81.2 % vs 91.9 %, p < 0.001); however, they also reported that obese women were able to achieve an improvement in quality of life scores similar to those of the normal weight population, suggesting that the cure rate difference might not be clinically significant [10].

More recently, Brennand et al. published the results from a secondary analysis of a randomized control trial aimed at evaluating the impact of BMI on objective and subjective cure rates 12 months after MUS surgery. With 61 women enrolled in the obese arm of the study and 121 in the non-obese one, they found lower objective and subjective cure rates in the obese group (67.8 % vs 85.6 %, p = 0.006 and 70.7 % vs 85.8 %, p = 0.016 respectively), although no differences between groups were found in the quality of life improvement, assessed by validated questionnaires [22].

Given the contradictory data published on the outcomes of MUS in obese women, it is not clear to date if BMI should be considered as an independent risk factor for surgical failure. Discrepancies in the study results may be explained by differences in the designs, the types of surgeries performed, the number of patients enrolled, the BMI cutoffs considered, or, more importantly, the established definitions of success and outcome measuring tools.

The merit of this study lies in the long-term follow-up analysis, enabling assessment of this procedure’s durability in obese women. Also, we highlight the use of large sample sizes and objective cure criteria, easily evaluated on pelvic examination. In terms of quality of life assessment, we consider that the established criteria helped to enrich the outcome analysis, as we compared not only mean KHQ total scores between groups, but also KHQ variations by pre-established improvement classes, enabling a broader awareness of the real impact of surgery in the daily life of these women.

Our patients were carefully selected, and despite the inclusion of cases of concomitant pelvic organ prolapse, all women had clinical symptoms of stress incontinence before surgery and urodynamic findings to confirm that.

Although this is not a prospective controlled trial, we point out that the variable in focus is also not liable to randomization, which limits the advantages of a truly prospective design. We did in fact rely on a prospective data collection for a clinical database, with standardized surgical protocols. In that way, we were able to ensure that all women had the same peri-operative and follow-up assessment, which in our opinion strengthens our outcome analysis, despite the retrospective design.

Our main limitation lies in the lost to follow-up rate seen at 48 months, with 41 % drop-out in the obese group and 39.6 % drop-out in the non-obese group. Although this may weaken the strength of our long-term findings, there were no significant differences in this rate between groups (p = 0.9).

Despite including women with mixed urinary incontinence and with concomitant POP repairs performed together with the index surgery, we found no differences in these distributions between groups, which in our opinion makes our samples reliable and at the same time more representative of the usual heterogeneous populations seen in clinical practice.

With this study we were able to demonstrate that in our clinical set, the outcomes of a transobturator sling in women with SUI can be similar, in spite of obesity. Both obese and non-obese women had sustained elevated cure rates at 4 years’ follow-up, achieving great improvements in quality of life after surgery, and with a low rate of complications. We conclude then that TVT-O seems to be a effective and safe procedure for stress incontinence in the long term, regardless of BMI.

As the published results still lack consistency, we consider it useful to clarify these findings in large prospective studies, particularly designed to address this matter.