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 Table 2. In total, 5985 patients were grouped randomly among the 18 articles, in which 2443 patients received silodosin 8 mg/day; 93 silodosin 4 mg/day; 1387, placebo; 672, tamsulosin 0.4 mg/day; 320, tamsulosin 0.2 mg/day; 201, naftopidil 75 mg/day; 79, naftopidil 50 mg/day; 117, alfuzosin 10 mg/day; 51, combination therapy. Various medical treatments were assessed, including 4 of the 28 RCTs compared silodosin with placebo, 10 RCTs with tamsulosin, 4 RCTs with naftopidil, 2 RCTs with alfuzosin, 2 RCTs with different dosage or administration of silodosin and 1 RCT with combination therapy (silodosin 8 mg/day and propiverine 20 mg/day). All studies applied silodosin 8 mg/day in males with LUTS/BPH. There are 7 trials in total had double blinded allocation for both outcome assessors and patients, 3 trials reported single blinded allocation for patients, 6 trials claimed open-label and 2 trials didn’t mentioned the blinding method. The study follow-up time points included were 2, 4, 8 and 12 weeks. On the whole, 7 studies were conducted in Japan, 6 studies in India, one in Korea, one in Taiwan and the remaining 3 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. Baseline prostate volumes, QoL, PVR and Qmax values ranged from 31.5 to 62.66 ml, 2.37 to 4.9 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. 20 RCTs reported the total IPSS improvement and four kinds of medications were comprised in our meta-analysis. In total, silodosin demonstrated greater efficacy in the total IPSS improvement than placebo (MD = -2.69, 95%CI: [-3.10, -2.28], P < 0.05) (Supplementary Fig. 3a) while no statistically significant difference was found when compared with tamsulosin (MD = 0.41, 95%CI: [-0.50, 1.33], P = 0.38) (Supplementary Fig. 3b), naftopidil (MD = -0.89, 95%CI: [-2.08, 1.76], P = 0.15) (Supplementary Fig. 3c) and alfuzosin groups (MD = -0.16, 95%CI: [-2.39, 2.07], P = 0.89) (Supplementary Fig. 3d) as a whole. However, statistically significant difference existed between naftopidil and silodosin groups at 8 weeks from the start of the treatment (P < 0.05).
The IPSS Voiding Subscore
9 RCTs reported the IPSS voiding improvement. In total three kinds of medications were incorporated. The mean change from baseline to study end in IPSS voiding subscore is demonstrated the forest plots in Fig. 1. According to the total analysis, all the RCTs reported the same result that silodosin groups were associated with greater improvement in the IPSS voiding symptoms than placebo, tamsulosin and naftopidil groups, with MD = -1.79, 95%CI: [-2.06, -1.52], P < 0.05; MD = -0.47, 95%CI: [-0.90, -0.05], P = 0.03; MD = -1.17, 95%CI: [-2.18, -0.16], P = 0.02; respectively.
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 Fig. 2. According to the total analysis, silodosin was obviously more effective than placebo (MD = -0.87, 95%CI: [-1.05, -0.69], P < 0.05) and naftopidil (MD = -0.93, 95%CI: [-1.60, -0.26], P < 0.05) in the IPSS storage improvement whereas silodosin showed a non-inferior trend to tamsulosin but no statistically significant difference were found.
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 (Supplementary Fig. 4a) (MD = 1.24, 95%CI: [0.81, 1.67], P < 0.001) in Qmax improvement while no statistically significant differences occured between silodosin, tamsulosin (MD = -0.27, 95%CI: [-0.98, 0.44], P = 0.46) (Supplementary Fig. 4b), naftopidil groups (MD = -0.30, 95%CI: [-1.64, 1.03], P = 0.66) (Supplementary Fig. 4c) and alfuzosin (MD = 0.77, 95%CI: [-0.47, 2.02], P = 0.22) (Supplementary Fig. 4d).
QoL Changes
13 RCTs reported the QoL changes. In total four kinds of medications were included in our meta. Totally the QoL decreased more significantly after treatment in silodosin than in placebo (MD = -0.44, 95%CI: [-0.59, -0.28], P < 0.05) (Supplementary Fig. 5a) and alfzosin groups (MD = -0.44, 95%CI: [-0.83, -0.05], P = 0.03) (Supplementary Fig. 5d). On the contrary, there was no statistical significance found between silodosin and other two medications (tamsulosin and naftopidil) in QoL changes at all time points (P > 0.05) (Supplementary Fig. 5b and Supplementary Fig. 5c).
Prostate Size And PVR Changes
3 RCTs reported the prostate size and PVR changes. There were respectively one and two medications included in our meta. Silodosin was founded significantly effective in decreasing the prostate size (MD = -1.51, 95%CI: [-2.27, -0.76], P < 0.001) (Supplementary Fig. 6) and PVR (MD = -6.78, 95%CI: [-11.51, -2.06], P = 0.005) (Supplementary Fig. 7a) when compared with tamsulosin. Moreover, silodosin demonstrated apparent superiority over naftopidil in decreasing PVR (MD =-5.88, 95%CI: [-10.87, -0.89], P = 0.02) (Supplementary Fig. 7b).
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 (RR = 1.35, 95%CI: [1.19, 1.52], P < 0.00001) and alfuzosin groups (RR = 3.78, 95%CI: [1.00, 9.92], P = 0.05) with significant differences as shown in Supplementary Fig. 8, so as the incidence of drug-related adverse events when compared with two included medications (placebo, RR = 2.21, 95%CI: [1.20, 4.07], P = 0.01; tamsulosin, RR = 1.32, 95%CI: [1.16, 1.50], P < 0.00001) as shown in Supplementary Fig. 9. The most prevalent drug-related adverse events observed were retrograde ejaculation. The pooled results shown in Supplementary Fig. 10 demonstrated that retrograde ejaculation had a significantly higher incidence in silodosin than in placebo (RR = 24.39, 95%CI: [13.95, 42.64], P < 0.00001), tamsulosin (RR = 7.35, 95%CI: [4.55, 11.89], P < 0.00001) and naftopidil groups (RR = 6.54, 95%CI: [2.36, 18.15], P = 0.0003).
As for cardiovascular adverse events like dizziness and vertigo (Supplementary Fig. 11), statistically significant differences only exited between silodosin and placebo groups (RR = 2.26, 95%CI: [1.21, 4.21], P = 0.009). Conversely, incidence of headache was similar in placebo and silodosin groups, so as that of thirst and diarrhea when comparing silodosin with placebo (all P values > 0.05) (Supplementary Figs. 12, 13 and 14). However, the incidence 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 apparently lower in silodosin than in tamsulosin groups with statistical significance, indicating the better cardiovascular safety of silodosin (Supplementary Fig. 15).
In respiratory symptoms, silodosin could decrease more upper respiratory tract infection than tamsulosin (RR = 0.69, 95%CI: [0.50, 0.96], P = 0.03) (Supplementary Fig. 16) while increase nasal congestion significantly when compared to placebo (RR = 7.76, 95%CI: [1.80, 33.41], P = 0.006) (Supplementary Fig. 17). No statistical significance was attained between silodosin and placebo in nasopharyngitis (RR = 1.16, 95%CI: [0.54, 2.47], P = 0.71) (Supplementary Fig. 18).
Efficacy And Safety Results Of Administration-different And Combination Therapies
Due to limited number of articles and diverse treatment methods, the meta-analysis failed to be conducted among administration-different and combination therapies. For administration-different therapies, Choo et al.[33] compared application of silodosin at 8 mg once a day (QD) and 4 mg twice a day (BID) in Korean population in 2014. The results suggested the improvement and frequencies of AEs were similar in QD and BID group, which indicating silodosin in BID administration may not be superior to QD in improvement of treatment outcomes at a total dosage of 8 mg. In 2015, Seki et al.[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 administration-different and combination therapies are needed to be conducted for meta-analysis furthermore.