Our study confirms that at short term, TOT is a safe procedure with very few per-operative and early post operative complications. However, during the long term follow-up, occurrence of de novo urge symptoms, de novo dyspareunia, perineal pain, and vaginal erosions significantly reduced the satisfaction of patients.
The only relevant per-operative complication, in our study, was haemorrhage of more than 200 ml (5.2%), but there were no haematomas, and no transfusion was required. Different series of TOT procedures report similar risks of bleeding rates varying between 0.83 and 5.4% [16, 17]. In a large study the incidence of intra operative bleeding of more than 200 ml was 1.9% for the TVT procedure .
The incidence of urinary retention is low in our study and is similar to the rates of 1.5% reported by other authors [7, 16]. In a recent review  which compared retropubic and transobturator tapes, voiding lower urinary tract symptoms were less common with the transobturator route.
De novo urge symptoms have a high impact on quality of life . We observed 6.2% of women with persistent de novo urgency at long term follow-up which is similar to other studies [17, 20]. In comparison, the risk with TVT is reported to be higher in the short term (33% for TVT vs 8% for TOT) , but after a longer follow up period, the risk is similar (6.3%) .
In our study, most women were sexually active (67%) and among them 9% reported de novo dyspareunia after the operation. This complication has a potential high impact on the quality of life, as the overall proportion of women satisfied with the procedure drops dramatically when it occurs (76.8% compared to 42.9%, P = 0.005). In a recent study , dyspareunia was more frequent after TOT (19.2% compared to 16.2% before), but the finding was not considered statistically significant. In another study, dyspareunia was reported in eight over 78 sexually active women (10.3%) . After TVT no significant difference in the incidence of dyspareunia was found post-operatively .
Perineal pain is reported to occur in 2.3% to 5% after transobturator surgery, to be transient, resolving within the first month [16, 24]. We report the same rates, but with persistent pain on long term follow up. The risk of groin pain is higher with TOT and TVT-O, compared to the TVT (OR 8.28, 95% CI 2.7-25.4) .
In our study 17 women had vaginal erosions. The proportion of women satisfied with the procedure was significantly reduced when erosion occurred. The mean time to erosions varied, which emphasizes the need to pay attention when symptoms like vaginal discharge, pain or dyspareunia occur even after a long period. Since some erosion occur without symptoms and can not be detected through telephone contact, this complication might be underestimated in our study. The majority of women who developed erosions required a reintervention. One woman presented with an abscess of the obturator fossa 38 months after surgery. In a recent review of suburethral sling procedures complications , the frequency of erosions after TOT varied between 1.8% and 20.0%. In a recent meta-analysis , erosions were more common after tape insertion by the transobturator route (TVT-O and TOT) compared to the retropubic route (OR 1.96; 95% CI 0.87-4.39). After a systematic search in the Manufacturer and User Facility Device Experience Database (MAUDE), Boyles et al. concluded that erosions constituted 60% of the complications associated with the TOT and are probably underreported . In our analysis the only significant risk factor for erosion was the type of sling, with a very high risk (17.8%) for the Obtape® sling compared to the two others. The mean follow-up was different between the three types of slings. However, the difference in the occurrence of erosion remained significant after adjustment for the duration in a multivariable model. There were more per operative complications with Aris® (12) than with Obtape® (5) which suggest that per operative complications were not a risk factor for the development of vaginal erosions.
We believe that vaginal erosion might be secondary to three potential factors: the sling material, surgical technique, or individual patient factors. Our data confirm that the tolerance of vaginal tissue depends on the type of sling used. The three tapes used in our study are polypropylene monofilaments. The TVT-O® and Aris® slings are type I meshes because they are macroporous (>75 μm). The Obtape® sling is more of a type II mesh since the pores are smaller (50 μm). The incidence of erosion with this sling was very high in our study which is similar to that of other studies if duration of follow up is taken into account [9, 10, 28]. As the Obtape® sling carries an important risk of vaginal erosion, we have abandoned its use. Type I mesh is considered to limit the risk of erosions [13, 29] because of a lower risk of infection, a lower inflammatory response and a better incorporation in the surrounding tissue. Aris® and TVT-O® are both type I meshes, but they have different mechanicals characteristics (different sizes of pores, different elasticity). No erosions were diagnosed after TVT-O®, but the number of women was small. Other authors also reported a low risk using this device (0.9%-1.8%) [10, 30].
Another difference between Aris® and TVT-O® is the surgical technique, TVT-O® being an in-out procedure and Aris® an out-in one. However, there was no statistically significant difference for the risk of erosion between the two techniques in our study and we are not able to determine if the surgical technique itself plays a role in the risk of erosion. We found only one study in which the authors concluded that erosion rate was associated with the surgical technique. They showed that plicaturing the pubocervical fascia between the sling and the vaginal mucosa could reduce the occurrence of erosion .
Individual patient characteristics (age, estrogen status, concomitant surgery, sexual activity) may play a role in the occurrence of vaginal erosion. Unfortunately, our study does not have the power to show differences in these potential risk factors. A study reported that diabetes mellitus was the only individual patient characteristic to be associated with a higher risk of erosions (RR 8.3, 95% CI 1.6-43.0) .
The strength of this study was the availability of a continuous cohort of women with a mean follow-up of more than two years and few patients lost for follow-up (6/233, 2.6%).
The limitations of our study included those typical of studies relying on information collected using medical files and during a telephone contact, with the risk of information bias. To avoid these biases, standardized questions were asked to the patients. Our evaluation of satisfaction was limited by the fact that we asked women their views by telephone without using a more detailed questionnaire. TOT was accompanied by another surgical procedure in almost half of the cases which could confound the analysis. However, this is not the case, because there was no difference in the occurrence of early or late complications between women with or without concomitant other surgical procedure.