This large registry study demonstrates good short-term continence outcomes for all evaluated sling types. However, the shorter slings Ajust™ and TVT-A was found to have higher failure rates 6–12 months after surgery than traditional slings. To our knowledge, only the shorter sling, TVT-Secure, has previously been shown to have inferior cure rates to traditional slings [24, 25]. TVT-Secure is no longer on the market because of this lack of effectivity. The shorter slings Ajust™ and TVT-A have on the contrary in several studies, such as Zullo et al. and Abdel-Fattah et al., shown equally good short-term cure rates compared with traditional slings [19, 26]. The difference between our findings and these randomized controlled trials (RCTs) may be related to sample size and study design. RCTs entail strict inclusion and exclusion criteria whereas a registry study by design implies more heterogeneity in patient characteristics, department sizes, and surgeons with different levels of training. Therefore, even though the RCT by Abdel-Fattah and co-workers was a multicenter study including 21 hospitals, a large retrospective study adds knowledge by its amount of data and patient diversity, which may arguably be closer to a real-life clinical situation [19].

In our study women operated on with shorter slings were significantly more often less than “very satisfied” with their treatment, answering “moderately satisfied” or worse (17.5–18.3% vs 14.3, p < 0.01). This is comparable with what has been shown in a large multicenter RCT by Rudnicki and co-workers, in which 72.7% Ajust™-operated women and 82.2% traditional sling-operated women answered that they were “very satisfied” with treatment [13]. However, our study showed significantly higher subjective satisfaction with the treatment than a similar study by Canel and co-workers (TVT-O 66% vs TVT-A 68%) [27]. This might be ascribed to differences in the validated questionnaires involved, as we used the validated questionnaire developed by the NFIR, whereas Canel et al. used the Patient Global Impression of Improvement (PGI-I) questionnaire [22, 27].

Much focus has been placed on the risk of prolonged pain after synthetic sling surgery [9], and prior studies have shown that shorter slings with less synthetic material may cause less postoperative pain [13, 27]. However, the present study demonstrates that even though the incidence of prolonged postoperative pain was low in all groups, this serious complication occurred just as often after shorter slings as after traditional slings. This implies that mechanisms other than the length of the sling are important for causing prolonged pain, such as proximity to nerves, mechanical rubbing against the periosteum and soft-tissue irritation. In our study, the highest incidence of prolonged postoperative pain was found among women after TVT-A (1.4%). This contrasts some other studies showing reduced postoperative pain with shorter slings [13, 27]. A possible explanation might be that because TVT-A is a modification of the TVT-O and both sling types are placed using a transobturator technique [14], the risk of prolonged pain is related to the obturator route and not the length of the tape. It is, however, surprising that the Ajust™ performed better than TVT-A concerning prolonged pain, as the Ajust™ is also placed in the obturator region and is anchored in the obturator membrane. One may suspect that pain is prevented because the Ajust™ sling is not passed further into the pelvis beyond the obturator membrane. The nerve proximity in the obturator foramen could explain why in several systematic reviews the TVT-O has been shown to cause more postoperative pain than retropubic slings [1, 8]. Even though we included TVT-O in the traditional sling group, this group was dominated in number by retropubic TVTs, which might have diluted the pain impact from obturator slings.

One might speculate that women with low BMI are more prone to prolonged pain after TVT-A because less fat protects the surrounding tissues and more synthetic sling material might be deposited lateral to the obturator membrane in much the same way as a traditional transobturator sling. In a publication by Cadish and co-workers in 2010 it was shown that high BMI lowered the risk of prolonged postoperative pain after transobturator sling application [28]. However, the same research group later found the opposite association when studying a different cohort of patients, in which women with high BMI more often had prolonged pain after transobturator slings [29]. From subgroup analyses of women with BMI < 25 kg/m2 and < 20 kg/m2 in our data, we were not able to reveal any increased incidence of prolonged pain among slim women for either sling type. Apart from prolonged postoperative pain, our study demonstrated fewer overall complications with shorter slings than with traditional slings, which is in line with other studies [13, 26]. More bladder perforations, and more urinary retention solved by sling pull-down and transection, were as expected in the traditional slings as a part of this group consists of TVT slings, and these are known complications of the retropubic route [1, 8]. We also found an increased rate of vaginal erosions after TVT-A application compared to TVT-O, and this is surprising as other study groups have not been able to demonstrate any difference in erosion rates comparing TVT-A with traditional slings  [26].

Our study showed higher levels of de novo urgency incontinence for all slings than several other studies [13, 18, 19]. As we wanted to show all potential impacts on bladder symptoms, we chose a very strict definition for de novo urgency in which any, not just bothersome, symptoms of urgency urinary incontinence were considered relevant. An urgency incontinence index of 1–2 according to the index generated by the NFIR questionnaire is rarely regarded as bothersome by the patient; thus, including them probably explains the high incidence of urgency. Other studies showing a lower incidence of de novo urgency incontinence have used other definitions of bothersome incontinence [13, 18, 19]. However, the choice of definition does not explain the increased incidence among Ajust™ compared with the other slings, and we have no plausible explanation for this.

A registry study does have inherent limitations, such as potentially incorrect entries and missing data for certain variables. However, the NFIR was shown to have high accuracy, completeness, and coverage in a previous study by Dyrkorn and co-workers [20]. No imputation statistics was performed for missing values. The missing values were similarly distributed among groups, and therefore the likelihood of missing values introducing selection bias was considered low. Another limitation is possible selection bias caused by the NFIR only releasing data on women who had consented to their data being used for scientific purposes. According to the NFIR administration, the overall rate of women not consenting is 15%. However, the rate is not evenly distributed among reporting hospitals, and therefore we cannot exclude this having an impact on the results. Furthermore, the registry has until recently only recorded re-operations at the same hospital department. This means that data on previous incontinence surgeries at other hospitals are not available. However, there are few women with previous incontinence surgeries and they would be evenly distributed among groups. For the above reason, it was not possible to compare the rate of new incontinence surgery as a measure of treatment failure.

The NFIR questionnaire only includes one question concerning pain, and because pain is a subjective perception, this may have limited the quality of data on prolonged postoperative pain. The registry only records prolonged postoperative pain as lasting more than 3 months and not whether this has been resolved later.

The most important strength of this study is the large number of women included, which gives power to the statistical analyses and reduces risks of selection bias. As our study is based on national registry data, the external validity is high because it contains data from multiple surgeons with different levels of surgical training and all types of patients. The use of a validated questionnaire further strengthens the validity of the findings.