Systematic review and meta-analysis of efficacy and safety of silodosin in treatment of benign prostatic hyperplasia patients with lower urinary tract symptoms CURRENT

Background: Silodosin is a new high-selective α1A-adrenoceptors antagonist. A systematic review of literature and meta-analysis were performed to compare the clinical efficacy and safety outcomes of silodosin with placebo, tamsulosin, naftopidil and alfuzosin in treating benign prostatic hyperplasia (BPH) males with lower urinary tract symptoms (LUTS). Materials and Methods: We systematically searched literature among EMBASE, PubMed, Cochrane Library, ScienceDirect and Web of Science databases until April 2019. 18 related randomized controlled trials were included according to eligibility criteria. Random-effects model were applied for data analysis. Results: 5,985 patients were included in our study. Silodosin presented superiority to placebo in improving LUTS and better efficacy than tamsulosin and naftopidil in improving IPSS void subscore and post-void residual urine volume with statistically significance (all P values < 0.05). Greater QoL index improvement were found in silodosin than alfuzosin groups (MD = -0.44, 95%CI: [-0.83, -0.05], P = 0.03) while no differences in total IPSS score and Qmax changes between these two groups. Retrograde ejaculation was significantly frequent in silodosin than placebo, tamsulosin and naftopidil groups (all P values < 0.05). Besides, silodosin increased incidence of upper respiratory tract infection compared to tamsulosin groups (RR = 0.69, 95%CI: [0.50, 0.96], P = 0.03). A higher rate of nasal congestion (RR = 7.76, 95%CI: [1.80, 33.41], P = 0.006) were found in silodosin than placebo groups while no difference for nasopharyngitis ((RR = 1.16, 95%CI: [0.54, 2.47], P = 0.71). Prevalence of headache (RR = 0.54, 95%CI: [0.27, 1.06], P = 0.07) and postural hypotension (RR = 0.14, 95%CI: [0.03, 0.77], P = 0.02) were lower in silodosin than tamsulosin groups, although dizziness and vertigo was more frequent in silodosin than placebo (RR = 2.26, 95%CI: [1.21, 4.21], P = 0.009). Conclusions: Our study demonstrated silodosin’s possible superiority to placebo and naftopidil while noninferiority to tamsulosin and alfuzosin in LUTS improvement of BPH males. Better cardiovascular safety was in silodosin groups, although incidence of retrograde ejaculation and respiratory adverse events were higher.

affinity than prazosin in human prostate, 214 times higher than that in human aorta [11] . Additionally, apparently higher affinity for α1A-adrenoceptors was found in silodosin than α1B-adrenoceptors (593fold) and α1D-adrenoceptors (57-fold) [12] . An in vivo study in 2019 has demonstrated that α1A-to-α1B binding ratio of silodosin was 9.8 times higher than that of tamsulosin despite they shared a similar affinity for α1A-adrenoceptors, clearly proving silodosin's highly selectivity for α1A-adrenoceptors [13] .
These results illustrated that silodosin should allow the maximization of profitable effects on the symptoms related to LUTS/BPH to produce more favorable uroselectivity and minimal adverse effects on cardiovascular system. Several previous meta-analysis and sympathetic reviews have been conducted but the grouping method is not detailed due to a limited number of articles [14][15][16][17] . To date, more and more researches have evaluated the difference of efficacy and safety among silodosin and other medications whereas the results are controversial [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35] .Therefore, we decided to perform this meta-analysis through a more detailed systematic review of those accessible RCTs to distinguish silodosin's superiority and non-inferiority to placebo and other medications in males with LUTS/BPH, which can provided more dependable guidance for the clinical application of silodosin.

Methods And Materials Search strategy
A systematic search of the published literature was conducted among databases including EMBASE, PubMed, Cochrane Library, ScienceDirect, and Web of Science databases from their earliest dates until April 2019 to identify relevant researches based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [36] . We used and combined terms in our search strategy as follows: ('silodosin' OR 'KMD-3213') AND ('benign prostatic hypertrophy' OR 'benign prostatic hyperplasia' OR 'prostatic adenomas') AND 'lower urinary tract symptoms'. Specific generic and trade drug names like 'rapaflo' were included as well in our search. The references of the selected articles were also reviewed to avoid possible omission of potentially relevant publications.

Study Eligibility
The inclusion criteria were established before our search. Only RCTs in which patients with LUTS/BPH were grouped randomly to treat with silodosin or another mono-and combination medical therapy were included. We excluded in vitro studies, randomized crossover studies, surgery-controlled studies and observational studies without a control group. In cases that studies with analyses from the same data or duplicate study samples were classified, we preferentially evaluated the research with a larger sample size. All articles not related to the research topic, comments, editorials, conference proceedings, sympathetic review, meta-analysis articles, case reports or isolated abstracts are not included either. We didn't restrict the search strategy by publication year, sample-size or language.

Quality Assessment
The Cochrane System Evaluator Manual 5.1.0 were applied to assess the quality of included publications. The assessed aspects contain selection bias (random sequence generation and allocation concealment); performance bias (blinding of participants and personnel); detection bias (blinding of outcome assessment); attrition bias (incomplete outcome data); reporting bias (selective reporting) and other bias. We also used Begg's and Egger's test to estimate publication bias.

Endpoints Of Interest And Data Extraction
Two investigators extracted data independently from the qualified publications including basic information (authors, year of publication, geographical region), demographic information (e.g. patient age, prostate volume) and post-treatment endpoints. The primary endpoints of interest were the changes from the start of treatment for the total International Prostate Symptom Score (IPSS), the IPSS voiding subscore and the IPSS storage subscore, which all reflect changes of LUTS among males with BPH. Secondary endpoints of interest included the changes from the start of treatment for quality of life indexes (QoL), maximum urinary flow rate (Qmax), prostate size, post-void residual urine volume (PVR), adverse events (AEs) including retrograde ejaculation, postural hypotension, dizziness and vertigo, diarrhea, thirst, headache, nasopharyngitis and nasal congestion.

Statistical analysis
We collated the data extracted from included publications in Excel (Microsoft Corporation, Redmond, WA, USA) and then pooled for the summary estimates. Review Manager 5.3 (The Cochrane Collaboration) was used to perform meta-analysis. We applied random-effects model to calculate the risk ratio for dichotomous data, mean differences (MD) and standard deviations (SD) for continuous data, both with 95% confidence intervals (CI) [37] . The ᵪ 2 statistic was applied to evaluate the heterogeneity among trials and I 2 statistic for the degree of inconsistency. In cases that data were not described by mean ± SD, the missing standard deviations for change scores were estimated using standard deviations from before and after the intervention [37] .

Results
In total 1,456 articles have been identified during the primary search, among which 204 from PubMed, 398 from Web of Science, 407 from EMBASE, 354 from ScienceDirect and 93 from Cochrane Library. We excluded 1420 articles according to eligibility criteria. Among the remaining 36 potentially relevant publications, 18 were excluded for incomplete text and duplicate publications. In all, 18 articles with 28 RCTs were included for data analysis ( Supplementary Fig. 1). The quality assessment was shown in Supplementary Fig. 2. Most of the P value available accessed by Begg's and Egger's test were higher than 0.05, demonstrating the low risk of publication bias in our meta-analysis (Supplementary Table 1).

Study Characteristics
The basic characteristics were summarized in Supplementary  in Europe or/and USA. Mean patients age ranged from 50.0 to 71.5 years. Baseline total IPSS, void and storage scores ranged from 14.1 to 29.00, 7.47 to 13.0 and 6.3 to 8.9 points, respectively. points, 28 to 92 ml and 7.297 to 15.9 ml/s, respectively.

Meta-analysis Of Efficacy The total IPSS
The detailed outcomes of efficacy were shown in Table 1 The IPSS Storage Subscore 9 RCTs reported the IPSS storage improvement. In total three kinds of medications were comprised.
The mean change from baseline to study end in IPSS storage subscore is shown the forest plots in Qmax Changes 17 RCTs reported the Qmax changes. In total four kinds of medications were included in our meta.
According to the total analysis, silodosin was only apparently more effective than placebo

Meta-analysis Of Safety
The detailed data of safety were shown in Table 2. The parameter we used to assess the safety of silodosin therapy is AEs. Overall adverse event rate was more frequent in silodosin than in tamsulosin [34] evaluated the treatment outcomes between silodosin 4 mg BID and 4 mg QD in Japanese LUTS/BPH patients, in which statistically significant differences were not found in efficacy and safety endpoints despite ejaculation disorder was more frequent in former group. Results above indicated silodosin at 4 mg QD was not inferior to 4 mg BID. For combination therapies, Matsukawa et al. [35] assessed clinical efficacy and safety of monotherapy with silodosin only or combination therapy with silodosin and propiverine, an anticholinergic agent, after a long-term follow-up in Japanese males. The combination therapy group appeared statistically significant improvement in QoL index and PVR while similar improvement in changes of other efficacy endpoints after a one-year treatment, demonstrating better clinical performance of combination therapy with silodosin and propiverine for LUTS/BPH patients.
In conclusion, there was no statistical significance among different administration of silodosin according to included articles. In contrast, combination therapy with silodosin and propiverine showed superiority in efficacy of treatment. Of course, more highly qualified RCTs about administrationdifferent and combination therapies are needed to be conducted for meta-analysis furthermore.

Discussion
BPH is the most pivotal and typical cause of LUTS in aged men and medicine treatment for BPH requires great effectiveness, affordable safety and reduced frequency of associated side effects, for which silodosin could be the potential ideal pharmacological therapy in recent years. As far as we know, this is the most comprehensive meta-analysis emphasizing on the clinical efficacy and safety outcomes of silodosin in treating aged BPH males accompanied with LUTS. Our study demonstrated that silodosin was effective and tolerable in treating LUTS related to BPH because it can apparently improve the total IPSS score, Qmax, IPSS storage and void subscore with statistically significant differences when compared with placebo. Besides, silodosin and tamsulosin shared the similar efficacy, except silodosin was superior in decrease of prostate size and improvement in IPSS void subscore and PVR. The better voiding symptoms improvement of silodosin may be for its higher α1Aadrenoceptors selectivity than tamsulosin as observed in some studies [13,38] .
The α1D-adrenoceptors predominate in bladder detrusor muscle, parasympathetic pathways and motor neurons in sacral ventral so blockage of them should result in storage symptoms improvement [39] . However, in our meta-analysis, silodosin was more effective in improvement of IPSS storage subscore, so as the IPSS void subscore and PVR in comparison of naftopidil, an α1Dadrenoceptors antagonist. The previous study has demonstrated that silodosin prohibits detrusor overactivity by suppressing C-fiber afferent activity in cerebral infarction rat models or increasing blood flow in the bladder of spontaneously hypertensive rat models [40,41] . Therefore, blockade of α1A-adrenoceptors may improve the storage symptoms. Nevertheless, more studies should be conducted to compare the ability in storage symptoms improvement between α1A-and α1Dadrenoceptors antagonist. With regard to alfuzosin, a non-selective α1-adrenoceptors antagonist, these two drugs shared similar efficacy in improvement of total IPSS score and Qmax. However, silodosin was inferior in improving QoL index with a statistically significant difference which may be accounted for the clearly higher incidence of AEs in silodosin than alfuzosin (RR = 3.15, 95%CI: [1.00, 9.92], P = 0.05).
The incidence of drug-related adverse events was apparently higher in silodosin groups compared with tamsulosin and placebo groups. Among the AEs, it's remarkable that retrograde ejaculation had a significant higher incidence in silodosin groups when compared with placebo, tamsulosin, and naftopidil, which mainly led to discontinuation of therapy. Retrograde ejaculation could result from loosening of smooth muscle in the urethra, prostate, vas deferens and bladder neck along with reduced sperm movement through the posterior urethra [42] . The bladder neck, seminal vesicles and vas deferens of human have a high density of α1A-adrenoceptors, of which the highly selective blockade from silodosin can lead to more retrograde ejaculation than other therapies [7] . Additionally, Homma et al. [43] found ejaculation disorder may be relevant to better efficacy of silodosin in BPH, suggesting existence of ejaculation disorder can be a marker for classifying BPH patients with better improvement, which corresponds to the results of this meta-analysis.
Moreover, we have evaluated respiratory AEs in this meta-analysis. Obviously, silodosin led to more frequent nasal congestion but similar nasopharyngitis compared with placebo groups. Besides, upper respiratory tract infection was significantly less frequent in silodosin than in tamsulosin groups. The flu-like symptoms above have been reported in all sorts of α1-blockers, in which nasal congestion was a common one, for the blockade of α1A-adrenoceptors effecting the regulation of nasal mucosa blood flow rather than increasing the viral infection risk in upper respiratory tract [44] . However, the data of respiratory AEs were insufficient since only two articles were included in this part. More comparison in effects on respiratory tract among α1-blockers are required.
In our meta-analysis, the incidence of cardiovascular AEs like headache and postural hypotension was much lower in silodosin than in tamsulosin groups, although that of dizziness and vertigo was significantly higher than placebo, indicating the better cardiovascular safety of silodosin as supposed due to the higher selectivity of silodosin for α1A-adrenoceptors. Converse to the finding of favorable cardiovascular safety of the non-selective antagonist alfuzosin [45] , we only observed a higher trend in incidence of dizziness and vertigo in silodosin compared with alfuzosin groups without a statistically significant difference. However, only two studies comparing silodosin and alfuzosin were included in this article, more high-quality clinical researches are needed to draw a reliable conclusion.
The heterogeneity of some variables greater than 70% in our meta-analysis is worthy of discussion.
The reason for the high heterogeneity observed may be as follows. variation may lead to the heterogeneity of our meta-analysis, although tamsulosin 0.2 mg/day was a reasonable option for patients with low body mass index like Asians [46] while the recommended dosage of naftopidil was controversial [47,48] . Therefore, the overall effect of dosage on our outcome analysis remains immeasurable. Moreover, the administration of silodosin differed from 4 mg twice a day to 8 mg once a day. Choo et al. [33] found no apparent statistical difference in efficacy and safety between silodosin 8 mg/day and 4 mg twice a day in Korean population. Therefore, we considered the administration of silodosin in all studies included as 8 mg per day regardless of the frequency.
However, the performance bias may exist despite the total dosage seems to be equivalent. Second, although all the articles included patients with mild to moderate IPSS, the baseline scores were different. Patients with IPSS ≥ 7 or 8 in 12 trials [18, 22-25, 29-32, 34, 35] were included while IPSS ≥ 13 in 4 trials [19-21,  From this meta-analysis, we found that silodosin is an effective and safe medication for LUTS/BPH patients. Afterall, several limitations were existed in this meta-analysis. First, great heterogeneity still exists in some variables despite we applied a random effects model which may undervalue the statistical errors and defaults to the arithmetic mean in cases large heterogeneity exits [37] . Although we have divided groups according to the dosage of medicine, administration method and follow-up time, there were no more adequate data allowing to conduct a more detailed subgrouping to find the exact source of heterogeneity. Moreover, we were unable to take administration-different and combination therapies into meta-analysis due to insufficient number of articles. Second, the therapy period we evaluated was only short as 12 weeks. The comparison of long-term efficacy among α1blockers requires more trials. Third, some included studies are of moderate quality. Therefore, after taking the heterogeneity of variables into consideration, our meta-analysis is necessary for assessment in silodosin's short-term clinical efficacy and safety at 8 mg per day compared with quality, especially meticulous long-term RCTs to demonstrate and confirm in the future.

Conclusions
In the current meta-analysis, we suggested silodosin may be superior to placebo and naftopidil while not inferior to tamsulosin and alfuzosin in improvement of LUTS relevant to BPH, and better cardiovascular safety was found in silodosin groups. However, the incidence of retrograde ejaculation was higher than in the placebo, tamsulosin, naftopidil and groups, as well as that of nasal congestion when compared with placebo groups. More stringent and long-term RCTs are required to confirm these conclusions.    Figure 1 Results of the meta-analysis on the change from baseline for the IPSS voiding subscore at different time points (silodosin versus placebo, tamsulosin and naftopidil, respectively).
Statistically significant differences were found in silodosin compared with tamsulosin and naftopidil at 12 and 8 weeks from the start of the treatment respectively (P < 0.05).

Figure 2
Results of the meta-analysis on the change from baseline for the IPSS storage subscore at different time points (silodosin versus placebo, tamsulosin and naftopidil, respectively).
Statistically significant differences were found in silodosin compared with naftopidil at 8 weeks from the start of the treatment (P < 0.05).

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.