Knowledge of surgical success and its predictors has a great importance for the decision-making process and adequate pre-operative counselling of patients. This is especially important when dealing with patients with previously failed treatment, as is frequently the case in men with urethral strictures. Guidelines concerning the treatment of strictures are developed only recently and are established on largely insufficient and inconclusive evidence. Urethroplasty seems to be an effective treatment for urethral strictures, however stricture recurrence is not exceptional even in experienced centres (1). The underlying causation of stricture recurrences is not well understood. The little research that has been conducted shows inconsistency in associated predicting factors. This presumably is caused by inhomogeneous study populations, small cohort sizes, short follow-up times and poorly defined criteria for stricture recurrence. These aspects degrade comparability of the individual studies and therefore impede the drawing of significant conclusions. 
The present study examines the long-term results of urethroplasty in a single centre cohort between 2011 and 2018. The objective is to identify the potential  factors that are correlated with the need for reintervention, thereby contributing to a better assessment of the likelihood of urethral stricture recurrence and improving pre-operative counselling.

Between 2011 and 2018, 250 urethroplasties were carried out in our centre. The operations were all performed in the same centre and by two surgeons. Individual patient files were checked manually for inclusion and exclusion criteria. Study inclusion criteria were defined as any male patient above the age of 18 years who underwent any type of urethroplasty as a result of urethral stricture disease. Patients undergoing urethroplasty for any other cause than urethral strictures (e.g. artificial urinary sphincter complication repair and patients with history of hypospadias) were excluded. Data were obtained from the electronic medical records. The collected data consisted of patient characteristics, stricture characteristics, patient’s medical history and surgical technique (table 1).  
Surgical techniques comprised predominantly of end to end anastomosis. Dependent on stricture length and location, dorsal and ventral techniques, both inlay and onlay, were used as well. After surgery, patients were hospitalised for four days considering the imposed bed rest. Approximately three weeks after surgery, the transurethral catheter was removed if voiding cystourethrogram showed no abnormalities (i.e. leakage, diverticulum). Patients were seen for evaluation and uroflowmetry, retrograde urethrogram and/or urethroscopy at three, six and twelve months after surgery. Thereafter patients were followed symptomatically. Surgical failure was defined as the need for any post-operative instrumentation due to stricture recurrence at any time during follow-up. These reinterventions formed the basis for the statistical analysis.
Statistical analysis was carried out using SAS statistical software version 9.4. Firth bias correction was used to overcome potential separation. The α value was set at 0.05 for all analyses, indicating a statistical significance for a two-sided p ≤0.05.

Of all 250 urethroplasties carried out between 2011 and 2018, 175 cases met the inclusion criteria. Of these 175 patients, 34 developed stricture recurrence, concluding an overall success rate of 80.3% in the total population. The median follow up of the population was 12 months (inter quartile range (IQR) 10 to 17 months). The median patient age was 43 (IQR 32 to 56). Urethral dilatation in outpatient clinic was the most applied reintervention, accounting for almost a quarter of all reinterventions. Repeated urethroplasty was needed in 13 patients, accounting for 38% of all reinterventions. 
In univariable logistic regression analysis, age at surgery (OR 1.00 [0.998-1.003]), combined surgical techniques (OR 9.72 [2.37-39.86]), multiple stricture locations (OR 11.41 [2.71-48.07]), iatrogenic strictures (OR 5.5 [1.57-19.22]), pelvic fracture- (OR 6.17 [1.87-20.30]) and straddle injury- induced strictures (OR 3.15 [1.05-9.44]) were identified as predictive factors for stricture recurrence. Multivariable analysis was performed using significant variables from univariable analysis as well as clinically important variables. Taking those variables into consideration, stricture aetiology proved to be most predictive of surgical failure (table 2). Specifically, iatrogenic strictures (OR 4.53 [1.08-19.07]), pelvic fractures (OR 8.65 [1.50-49.72]) and straddle injury induced strictures (OR 3.34 [1.02-10.96]) demonstrated to be significant factors for stricture recurrence. Age at surgery neared significance (p<0.07). The C-statistic for the multivariable prediction model was 0.7868. No multicollinearity was identified.

In the present study, the overall success rate  was  80.3%, at a median follow up of 12 months. This finding is somewhat lower compared to other literature (1). The main reason for the relatively wide range in success rates is caused by the widely varying definitions used to determine failure and success. Although the need for reintervention is most commonly used to define recurrence, extensiveness of the definition of ‘reintervention’ still remains. Since in the present study all forms of reintervention were used to define recurrence, this could explain our higher recurrence rate. Median time to stricture recurrence was 11 months. This knowledge is especially important in tailoring patient follow up.
The aim of this study was to identity factors associated with stricture recurrence. In the present series, iatrogenic strictures, pelvic fractures and straddle injuries appeared to be the most important predictive variables on multivariate analysis. Age at the time of surgery also suggested to be  noteworthy. In literature both similar and contradictive relations are described (2). Plausible explanations could be related to the preferences of the surgeon regarding surgical technique, discrepancies in stricture aetiology and studied populations. Further, lack of well-described definitions, inconsistent criteria of success and variable measurements, differences in duration of follow up and cohort size make it troublesome to make conclusive statements. Another important explanation for this inconsistency could be the absence of known potential confounding variables.
This study is not without limitations. A substantial drawback is the retrospective design. However, as a result of well determined and accurately followed standardized follow up protocols, we were able to prospectively maintain the database. Another limitation in our series is the length of follow up. This has potentially led to missing delayed recurrences. However, since our centre is a tertiary referral centre, it is reasonable to assume patients will return to our centre. The absence of patient reported outcome measures might also be seen as a constraint. We would argue that our criteria of success are fairly extensive and clinically important to both patient and practitioner. Nevertheless, no elements regarding quality of life are taken into account in our definition. Therefore it remains possible that a patient may not require dilatation postoperatively, but is unsatisfied due to complications or conversely.

In our centre, we found urethroplasty has a success rate of 80.3% at 12 months follow up in the overall study population. Patients with strictures caused by pelvic fractures, iatrogenic injury and straddle injury are found to have significantly higher risk on recurrences, needing reintervention. Additionally, higher age at the time of surgery may also be predictive of failure. Patients matching these characteristics should be counselled accordingly and perhaps followed more closely during follow up.

Meeks JJ, Erickson BA, Granieri MA, Gonzalez CM. Stricture recurrence after urethroplasty: a systematic review. J Urol. 2009; 182(4): p. 1266-1270.
Kinnaird AS, Levine MA, Ambati D, Zorn JD, Rourke K. Stricture length and etiology as preoperative independent predictors of recurrence after urethroplasty: A multivariate analysis of 604 urethroplasties. Can Urol Assoc J. 2014; 8: p. 296-300.