**3. Discussion**

This study reveals that urethral sphincter BoNT-A injection is an effective treatment option for refractory non-neurogenic functional voiding dysfunction in both genders. The general success (GRA - 2) rate after injection was 64%. Patients with a history of recurrent UTI and favorable baseline VE had better a subjective response after urethral sphincter BoNT-A injections. DU is a significant predictor for poor outcome. These findings sugges<sup>t</sup> that undertaking urodynamic assessment before the procedure is important for predicting treatment outcomes in patients considering urethral sphincter BoNT-A injection due to voiding dysfunction.

The concept of urethral sphincter BoNT-A injection in treating non-neurogenic voiding dysfunction originated from the positive experience of its usage in treating patients with detrusor sphincter dyssynergia [12]. It was assumed that the improvement of voiding dysfunction is related to lowering of the urethral resistance through chemical sphincterotomy which was induced by blocking the presynaptic release of acetylcholine in the neuromuscular junction of the urethral sphincter after injection [13]. Previous studies have reported about 60–70% overall response rate in non-neurogenic voiding function after the urethral sphincter BoNT-A injections [14,15]. With a greater sample size, our studies demonstrated a similarly successful result in such patients. Notably, high proportion of our patients had history of TUI-BN or TURP. In our practice, we performed TUI-BN in female patients who presented with insufficient bladder neck opening during voiding in the videourodyamic studies prior to urethral sphincteric BoNT-A injection. This treatment sequence could exclude the patients whose voiding dysfunction was attributed to anatomical or functional bladder neck dysfunction. A similar rationale was also applied to the male patients; TURP or TUI-BN were performed first if obstruction in the prostate urethra or bladder neck was suspected. In short, urethral sphincteric BoNT-A was injected in patients with refractory voiding dysfunction due to DU or urethral sphincter dysfunction. In the logistic regression analysis, history of TUI-BN or TURP did not pose significant adverse effects to the outcome of urethral sphincteric BoNT-A injection.

DU and urethral sphincter dysfunction are the two major etiologies of non-neurogenic voiding dysfunction. However, studies comparing the treatment efficacy between the two are lacking and the study subjects were often mixed with those with neurogenic voiding dysfunction [15,16]. DU was reported to be one of the causes of treatment failure after urethral sphincter BoNT-A injections [16]. In this study, both women and men had a significantly lower rate of treatment success in patients diagnosed with DU compared to those with urethral sphincter dysfunction. For DU patients, the major effect of urethral sphincter BoNT-A injection is to release BOO by lowering the urethral resistance while the impaired bladder contractility persisted despite the treatment. This could explain the inferior outcome in these patients. After adjusting for the possible confounding factors including gender difference and age, DU remains a predictive factor for poor treatment response in this study.

It is reasonable that the therapeutic efficacy of urethral BoNT-A might be affected by the severity of baseline pathophysiology of voiding dysfunction. Patients with history of urethral catheterization due to severe emptying failure in idiopathic or neurogenic etiology had been reported to respond poorly to the treatment compared to others [17]. As an index of bladder emptying ability, VE before the treatment might work as an outcome predictor as well. In fact, we found that the baseline VE was positively correlated with the successful outcome in both univariate and multivariate analyses in this study. The best cutoff value for baseline VE were 23% and 4 % for females and males respectively according to the Youden's index in the receiver operating characteristic (ROC) curve. Therefore, patients with DU and poor VE diagnosed in pre-operative urodynamic studies should be adequately informed of the risk of inferior treatment responses.

Aside from the therapeutic effect for voiding function, urethral sphincter BoNT-A injection might also be beneficial for recurrent UTI, a common bothersome nightmare resulting from incomplete urine emptying [18]. Urethral sphincter BoNT-A injection had

been reported to achieve a 50% reduction of UTI in spinal cord injury patients with detrusor sphincter dyssynergia [19]. Urethral sphincter BoNT-A injection also decreased UTI in neurologically normal patients with functional voiding dysfunction [20]. The benefit of UTI prevention might explain the finding of higher subjective response rates reported in our patients who had a history of recurrent UTI. As a result, urethral sphincter BoNT-A injection might be considered in those who suffered from refractory voiding dysfunction concomitant with recurrent UTI.

This study provides the treatment response rate of urethral sphincter BoNT-A injection and its predictive factors in patients suffering from functional, non-neurogenic voiding dysfunction with considerable subject numbers as well as complete and detailed urodynamic study before and after the treatment. However, there are still some limitations. First, since the majority of male voiding dysfunction is caused by anatomical obstruction, the number of men in our study is relatively small which makes it difficult to undertake further subgroup analysis. Second, the diagnoses of DU and urethral sphincter dysfunction were based on the image and pressure flow parameters from VUDS which might be somewhat subjective. Nevertheless, it is the most common way to differentiate the cause of voiding dysfunction in clinical practice. Third, the retrospective nature of this study made it difficult to avoid all possible biases during analysis despite our adjusting for the significant variables statistically. A prospective trial with specific inclusion criteria and pre-defined sub-group analysis is required to confirm the results of our study.
