EUROPEAN UROLOGY 63 (2013)

Transcription

1 EUROPEAN UROLOGY 63 (2013) available at journal homepage: Review Sexual Medicine Comparative Effectiveness and Safety of Oral Phosphodiesterase Type 5 Inhibitors for Erectile Dysfunction: A Systematic Review and Network Meta-analysis JinQiu Yuan a, RenJie Zhang a, ZuYao Yang a,b, Jack Lee c, YaLi Liu d, JinHui Tian d, XiWen Qin a, ZhengJia Ren a, Hong Ding e, Qing Chen e, Chen Mao a,b, *, JinLing Tang a,b, * a Division of Epidemiology, School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong; b Shenzhen Municipal Key Laboratory for Health Risk Analysis, Shenzhen Research Institute of The Chinese University of Hong Kong, Shenzhen, Guangdong Province, China; c Division of Biostatistics, School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong; d Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Lanzhou, China; e Department of Epidemiology, School of Public Health and Tropical Medicine, Southern Medical University, Guangzhou, China Article info Article history: Accepted January 14, 2013 Published online ahead of print on January 31, 2013 Keywords: Phosphodiesterase type 5 inhibitors Erectile dysfunction Systematic review Network meta-analysis Abstract Context: Phosphodiesterase type 5 inhibitors (PDE5-Is) are currently the first-line therapy for erectile dysfunction (ED), but available studies investigating the comparative effects of different PDE5-Is are limited. Objective: To compare the efficacy and safety of different classes of oral PDE5-Is for ED. Evidence acquisition: A systematic search was performed in PubMed, Cochrane Library, and Embase to identify randomized controlled trials that compared different PDE5-Is or PDE5-Is with a placebo for ED. The methodological quality of included studies was appraised with the Cochrane Collaboration bias appraisal tool, and the quality of evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation system. Evidence synthesis: A total of 118 trials ( individuals) were included. There was no major difference in the results between the traditional meta-analysis and the network meta-analysis. Network meta-analysis demonstrated that PDE5-Is were superior to placebo to improve erectile function. Compared with tadalafil (relative risk [RR]: 0.61; 95% confidence interval [CI], ) and vardenafil (RR: 0.63; 95% CI, ), avanafil was less effective on Global Assessment Questionnaire question 1. Tadalafil was more effective than vardenafil (mean difference [MD]: 1.49; 95% CI, ) and udenafil (MD: 1.84; 95% CI, 3.31 to 0.33) as measured by the erectile function domain of the International Index of Erectile Function. For all efficacy outcomes, the absolute effects and the rank tests indicated that tadalafil and vardenafil were the most effective agents. After adjusting for dosage, the conclusion remained the same. Safety analysis showed there was no major difference among different agents. Conclusions: In recommended doses, oral PDE5-Is are more effective than placebo for ED, and tadalafil seems to be the most effective agent, followed by vardenafil. PDE5-Is are generally safe and well tolerated, and there is no major difference on the safety profile. # 2013 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding authors. Division of Epidemiology, School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong. Tel / ; Fax: / addresses: maochen@cuhk.edu.hk (C. Mao), jltang@cuhk.edu.hk (J.L. Tang) /$ see back matter # 2013 European Association of Urology. Published by Elsevier B.V. All rights reserved.

2 EUROPEAN UROLOGY 63 (2013) Introduction Erectile dysfunction (ED), defined as the persistent inability to achieve or maintain an erection sufficient for satisfactory sexual intercourse [1], is one of the most common sexual disorders among men. Past surveys indicate that nearly 50% of men reported some degree of ED [2 4], and about 65% were not satisfied with the hardness of their erection [4]. ED also causes a huge economic burden to society. According to the US National Health and Nutrition Examination Survey, the annual costs of ED treatment in the United States could reach $15 billion if all patients sought medical care [5]. Current therapies for ED include phosphodiesterase type 5 inhibitors (PDE5-Is), hormones, vacuum constriction devices, intraurethral suppositories, intracavernosal injections, and surgery [6,7]. Oral PDE5-Is including sildenafil, tadalafil, and vardenafil are currently the first-line therapy for ED [6,7]. Four PDE5-Is (sildenafil, vardenafil, tadalafil, and avanafil) are approved worldwide, and two agents (udenafil and mirodenafil) are approved only in Korea [8]. Lodenafil, a new PDE5-I, is still undergoing clinical trials. PDE5-Is block the PDE5 enzyme that degrades cyclic guanosine monophosphate and thus results in the relaxation of smooth muscle in the corpus cavernosum, and finally increased blood flow and erection [9 11]. A large number of studies were conducted after the introduction of PDE5-Is (sildenafil) in These studies demonstrated that oral PDE5-Is are highly effective and well tolerated for ED patients [9,12]. However, available studies investigating the comparative effects of different PDE5-Is are limited. Given the variety of PDE5-Is available for prescription to ED patients and the limited evidence regarding the comparative efficacy of different PDE5-Is, it is hard for physicians to prescribe the best medicine. Network meta-analysis, in the context of a systematic review, is a meta-analysis in which multiple treatments are compared using both direct comparisons of interventions within randomized controlled trials and indirect comparisons across trials based on a common comparator [13,14]. In this study, we carried out a systematic review and network meta-analysis to compare the efficacy and safety between different PDE5-Is for the treatment of ED in a broad spectrum of the population. 2. Evidence acquisition 2.1. Data sources and searches We carried out an electronic search of Cochrane Library (Issue 4, 2012), PubMed (1966 to April 2012), and Embase (1984 to April 2012). The search strategy consisted of three parts (strategies for PDE5-Is, ED, and a specific filter for clinical trials) using the following keywords in combination with both Medical Subject Headings terms and text words: phosphodiesterase inhibitor, tadalafil, sildenafil, vardenafil, lodenafil, mirodenafil, udenafil, erectile dysfunction, impotence, and randomized controlled trial. There was no limitation on publication status or language. We also searched the metaregister and World Health Organization International Clinical Trials Registry Platform for ongoing studies. Reference lists of the included studies were checked manually to identify further studies Studies selection We included randomized controlled trials that compared different oral PDE5-Is or oral PDE5-Is versus placebo for ED. The patients in this study were limited to the broadspectrum population diagnosed with ED. Studies that examined the use of oral PDE5-Is in special population groups (eg, men with diabetes mellitus or hypertension) were excluded. The primary outcomes for this study were the Global Assessment Questionnaire question 1 (GAQ-1), and change from baseline to study end in the International Index of Erectile Function-Erectile Function domain score (IIEF-EF). The secondary outcomes included (1) change from baseline to study end in Sexual Encounter Profile question 2 (SEP-2), (2) change from baseline to study end in Sexual Encounter Profile question 3 (SEP-3), and (3) adverse events (AEs) that included the number of treatment-related adverse events, serious or severe adverse events, patients who experienced any adverse event (AE), and specific AEs. Trials were eligible if one of these outcome measures was reported. Study eligibility was independently determined by two authors. The authors evaluated the eligibility of remaining studies by examining the titles, abstracts, and full articles progressively. Discrepancies were resolved by discussion Data extraction and quality assessment Data were extracted independently by two authors using a standard form. Data extracted included study characteristics (eg, title, publication time, and patients number), patient characteristics (eg, age, height, weight, race, ED severity, and ED duration), intervention, control, method (eg, randomization, blinding, and loss to follow-up), and outcomes (eg, estimates, standard error, and p value). Discrepancies were resolved by discussion. The authors of original studies were consulted for missing information where necessary. Both methodological quality and quality of the evidence were assessed independently by two authors. The methodological quality of included studies was appraised with the Cochrane Collaboration bias appraisal tool [15]. The quality of evidence on GAQ-1, IIEF-EF, SEP-2, and SEP-3 was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system [16,17]. The quality of evidence was presented as follows: (1) high, indicating further research is very unlikely to change our confidence in the estimated effect; (2) moderate, indicating further research is likely to have an important impact on our confidence in the estimated effect and may change the estimate; (3) low, indicating further research is very likely to have an important impact on our confidence in the estimated effect and is likely to change the estimate; and (4) very low, indicating that we are very uncertain about the estimate [16,17].

3 904 EUROPEAN UROLOGY 63 (2013) Data synthesis and analysis The comparative effects were initially analyzed by the traditional pairwise meta-analysis method using Cochrane Collaboration review manager software. We applied a random-effects model that accounts for both within- and between-study variability to provide more conservative estimated effects. Heterogeneity among studies was assessed with the chi-square test and the I 2 index statistic. Low level of heterogeneity was defined as I 2 25%, accompanied by p > 0.10 for the chi-square test [15]. We then pooled the data of all PDE5-Is available with both fixed- and random-effects models within a Bayesian framework using WinBUGS [18]. Summary effect size was calculated as mean difference (MD), relative risk (RR), together with their 95% confidence intervals (CIs). We also calculated the absolute effects and the relative rank of different PDE5-Is to provide an overview of the efficacy and safety of all PDE5-Is. The Deviance Information Criterion was calculated to determine which model has better goodness of fit [18]. Apart from heterogeneity, the network meta-analysis also holds the assumption of similarity among trials and consistency of direct and indirect evidence [19 22]. To check the assumption of similarity, we used meta-regression analysis by adding other covariates (dosage, average age, weight, body mass index, ED duration, treatment duration, and baseline IIEF-EF score) to the network meta-analysis model [20,22]. Where direct and indirect estimates were combined, we applied an extension of the Bucher method to check the assumption of consistency [21]. We did sensitive analyses according to the following variables: dosage (including only studies within the therapeutic range) and quality of included studies (excluding studies of high risk of bias on any of the following items: randomization, allocation concealment, blinding, and incomplete outcome data). Publication bias was examined through visual inspection of funnel plot asymmetry [15]. 3. Evidence synthesis 3.1. Search results and study characteristics The literature search yielded 5938 citations, of which 5709 were excluded after review of titles and abstracts. The full texts of 229 remaining citations were screened, and finally 118 studies including patients were included (Fig. 1) (the references of the included studies are provided in Supplement 1). The included studies covered seven different PDE5-Is: sildenafil, tadalafil, vardenafil, udenafil, mirodenafil, avanafil, and lodenafil (Fig. 2). The dosages used in most included trials were within the recommended dose ranges. Table 1 The efficacy of different classes of phosphodiesterase type 5 inhibitors as measured by Global Assessment Questionnaire question 1 and the International Index of Erectile Dysfunction Erectile Dysfunction Interventions GAQ-1, risk ratio (95% CI) IIEF-EF, mean difference (95% CI) Network meta-analysis Direct comparison GRADE Network meta-analysis Direct comparison GRADE Placebo Sildenafil 3.20 ( ) 2.54 ( ) High 6.00 ( ) 6.03 ( ) High Tadalafil 3.31 ( ) 2.64 ( ) High 7.49 ( ) 8.07 ( ) High Vardenafil 3.22 ( ) 2.75 ( ) High 7.11 ( ) 7.05 ( ) High Udenafil 3.02 ( ) 2.82 ( ) High 5.66 ( ) 5.92 ( ) High Mirodenafil 3.05 ( ) 2.41 ( ) Moderate NA NA NA Avanafil 2.00 ( ) 1.91 ( ) Moderate NA NA NA Sildenafil Tadalafil 1.04 ( ) NA Moderate 1.49 ( ) 0.28 ( ) Low Vardenafil 1.01 ( ) 1.39 ( ) Low 1.11 ( ) 0.36 ( 1.84 to 2.56) Moderate Udenafil 0.95 ( ) NA Moderate 0.34 ( 1.74 to 1.13) Moderate Mirodenafil 0.96 ( ) NA Low NA NA NA Avanafil 0.63 ( ) NA Low NA NA NA Tadalafil Vardenafil 0.97 ( ) NA Moderate 0.38 ( 1.63 to 0.82) 0.16 ( 1.16 to 0.84) Moderate Udenafil 0.92 ( ) NA Moderate 1.84 ( 3.31 to 0.33) NA Moderate Mirodenafil 0.93 ( ) NA Low NA NA NA Avanafil 0.61 ( ) NA Low NA NA NA Vardenafil Udenafil 0.94 ( ) NA Moderate 1.45 ( 3.06 to 0.21) NA Moderate Mirodenafil 0.95 ( ) NA Low NA NA NA Avanafil 0.63 ( ) NA Low NA NA NA Udenafil Mirodenafil 1.02 ( ) NA Low NA NA NA Avanafil 0.67 ( ) NA Low NA NA NA Mirodenafil Avanafil 0.67 ( ) NA Low NA NA NA CI = confidence interval; GAQ-1 = Global Assessment Questionnaire question 1: While using the study medication, did you feel that your erections improved? ; GRADE = Grading of Recommendations, Assessment, Development, and Evaluation; IIEF-EF = International Index of Erectile Function-Erectile Function domain; NA = not applicable.

4 EUROPEAN UROLOGY 63 (2013) Table 2 The efficacy of different classes of phosphodiesterase type 5 inhibitors as measured by Sexual Encounter Profile question 2 and question 3 Interventions SEP-2, mean difference (95% CI) SEP-3, mean difference (95% CI) Network meta-analysis Direct comparison GRADE Network meta-analysis Direct comparison GRADE Placebo Sildenafil 8.73 ( ) 8.70 ( ) Moderate ( ) ( ) Moderate Tadalafil ( ) ( ) High ( ) ( ) High Vardenafil ( ) ( ) High ( ) ( ) High Udenafil ( ) ( ) High ( ) ( ) High Mirodenafil ( ) ( ) Moderate ( ) ( ) Moderate Sildenafil Tadalafil ( ) NA Low ( ) NA Low Vardenafil ( ) NA Low ( ) NA Low Udenafil ( ) NA Low ( 2.04 to 24.04) NA Low Mirodenafil 9.18 ( 4.04 to 22.23) NA Low ( ) NA Low Tadalafil Vardenafil 0.33 ( 5.85 to 5.16) NA Moderate 0.08 ( 6.27 to 6.78) NA Moderate Udenafil 1.52 ( 7.16 to 4.14) NA Moderate 7.88 ( to 0.13) NA Moderate Mirodenafil 9.83 ( to 1.90) NA Low 0.58 ( to 12.36) NA Low Vardenafil Udenafil 1.18 ( 7.09 to 4.76) NA Moderate 7.96 ( to 0.23) NA Moderate Mirodenafil 9.49 ( to 2.38) NA Low 0.66 ( to 12.67) NA Low Udenafil Mirodenafil 8.31 ( to 3.64) NA Low 7.30 ( 6.23 to 20.96) NA Low CI = confidence interval; GRADE = Grading of Recommendations, Assessment, Development, and Evaluation; NA = not applicable; SEP-2: Sexual Encounter Profile question 2: Were you able to insert your penis into your partner s vagina? ; SEP-3: Sexual Encounter Profile question 3: Did your erection last long enough for you to have successful completion of intercourse? The baseline characteristics of included studies are summarized in Supplement 1. The overall methodological quality was moderate. All included studies are randomized controlled trials, but most studies did not report the techniques for randomization and concealment. The quality of evidence varies in different outcomes and comparisons as measured by the GRADE system (Tables 1 and 2). The grade of quality was downgraded primarily due to indirectness and imprecision (see details in Supplement 2). The network meta-analysis results reported in the main text are based on random-effects models because they generally showed better goodness of fit and more conservative estimated effects compared with fixed-effects models. There was no major difference between the traditional meta-analysis results and the network metaanalysis results. The results of fixed-effects models are available in Supplements 3 and Results Global Assessment Questionnaire question 1 A total of 69 studies including six different classes of PDE5-Is (sildenafil, tadalafil, vardenafil, udenafil, mirodenafil, and avanafil) contributed to the analysis of GAQ-1 ( patients). Network meta-analysis indicated that all of these PDE5-Is were associated with significantly higher GAQ-1 positive responses than placebo. Compared with sildenafil (73% vs 46%; RR: 0.63; 95% CI, ), tadalafil (75% vs 46%; RR: 0.61; 95% CI, ), vardenafil (73% vs 46%; RR: 0.63; 95% CI, ), and udenafil (69% vs 46%; RR: 0.67; 95% CI: ), avanafil was associated with significantly lower GAQ-1 positive responses (Fig. 3). The absolute effects and rank test suggested that tadalafil was the most effective PDE5-I in terms of GAQ-1 among the PDE5-Is compared, followed by vardenafil (Table 3; Fig. 4) International Index of Erectile Function-Erectile Function A total of 39 studies including four different classes of PDE5-Is (sildenafil, tadalafil, vardenafil, and udenafil) contributed to the analysis of IIEF-EF ( patients). Network metaanalysis demonstrated that all PDE5-Is included were associated with a significant higher improvement in IIEF- EF than placebo. Tadalafil was more effective than sildenafil (9.21 vs 8.39; MD: 1.49; 95% CI, ) and udenafil (9.21 vs 7.53; MD: 1.84; 95% CI, 3.31 to 0.33) in improving IIEF-EF (Fig. 3). The absolute effects and rank test showed that tadalafil was the most effective PDE5-I in terms of IIEF-EF among the PDE5-Is compared, followed by vardenafil (Table 3; Fig. 4). Inadequate reporting prevented us from including mirodenafil, avanafil, and lodenafil in our network metaanalysis. However, results from original trials suggested that mirodenafil [23,24], avanafil [25,26], and lodenafil [27,28] were more effective than placebo as measured by IIEF-EF Sexual Encounter Profile question 2 A total of 17 studies including five different classes of PDE5- Is (sildenafil, tadalafil, vardenafil, udenafil, and mirodenafil) contributed to the analysis of SEP-2 (4252 patients). Network meta-analysis indicated that all PDE5-Is included in the model were associated with significant higher mean changes in SEP-2 than placebo. In addition, sildenafil was less effective than tadalafil (10.48 vs 29.70; MD: 19.0; 95%

5 906 [(Fig._1)TD$FIG] EUROPEAN UROLOGY 63 (2013) Fig. 1 Flowchart of study selection. ED = erectile dysfunction; PDE5 = phosphodiesterase type 5; RCT = randomized controlled trial. CI, ), vardenafil (10.48 vs 29.22; MD:18.67; 95% CI, ), and udenafil (10.48 vs 28.02; MD: 17.49; 95% CI, ) in improving SEP-2 (Fig. 3). The absolute effects and rank test suggested that tadalafil was the most effective PDE5-I in terms of SEP-2 among all PDE5-Is compared in the network, followed by vardenafil (Table 3; Fig. 4). Inadequate reporting prevented us from including avanafil and lodenafil in our network metaanalysis. Results from original trials suggested that avanafil [25,26] and lodenafil [27,28] were superior to placebo as measured by SEP Sexual Encounter Profile question 3 A total of 17 studies including five different classes of PDE5-Is (sildenafil, tadalafil, vardenafil, udenafil, and mirodenafil) and 4445 patients contributed to the analysis of SEP-3. Network meta-analysis showed that all PDE5-Is included in the model were associated with

6 [(Fig._2)TD$FIG] EUROPEAN UROLOGY 63 (2013) Fig. 2 Comparison network of included study. significant higher mean changes in SEP-3 than placebo. Sildenafil was less effective than tadalafil (29.10 vs 48.07; MD:18.92; 95% CI, ), vardenafil (29.10 vs 48.13; MD: 19.01; 95% CI, ), and mirodenafil (29.10 vs 47.48; MD: 18.34; 95% CI, ) in improving SEP-3 (Fig. 3). In addition, tadalafil significantly improved SEP-3 over udenafil (48.07 vs 40.23; MD: 7.88; 95% CI, to 0.13). The absolute effects and rank test suggested that tadalafil and vardenafil were the most effective PDE5-Is in terms of SEP-2 among all PDE5- Is analyzed (Table 3; Fig. 4). Inadequate reporting prevented us from including avanafil and lodenafil in our network meta-analysis. Results from original trials suggested that avanafil [25,26] and lodenafil [27,28] were superior to placebo on SEP Safety The number of studies included in the safety analysis on different outcomes range from 23 to 105. The adverse events caused by PDE5-Is were generally mild. Flushing, headache, and dyspepsia were the most common reported adverse events. The safety between different classes of PDE-5 inhibitors was similar, except tadalafil caused a higher incidence of myalgia than sildenafil (RR: 4.69; 95% CI, ). Figure 5 shows a summary of the safety of PDE5-Is for ED (only comparisons with a significant difference are shown). The detailed safety analysis results are available in Supplement Network assumptions, sensitivity analysis, and publication bias The test of heterogeneity in pairwise meta-analysis was generally moderate or small (Supplements 3 and 4). For similarity, meta-regression on GAQ-1 suggested that age (regression coefficient: 0.002; p = 0.55), weight (regression coefficient: 0.017; p = 0.95), body mass index (regression coefficient: 0.254; p = 0.97), ED duration (regression coefficient: 0.005; p = 0.28), treatment duration (regression coefficient: 0.254; p = 0.97), and baseline IIEF-EF score (regression coefficient: 0.017; p = 0.54) did not significantly contribute to the estimates. Meta-regression including dosage showed that dosage was statistically related to GAQ-1 (regression coefficient: ; p = 0.005). After adjusting for dosage, tadalafil and vardenafil are more likely to be statistically and significantly more effective than other PDE5-Is (see details in Supplement 3). The same change was found for other outcomes. Among all outcomes, there were 27 pairwise comparisons when direct and indirect data were combined together, of which two comparisons showed inconsistency by the extension of the Bucher test: sildenafil versus vardenafil on GAQ-1 ( p = ) and tadalafil versus sildenafil on flushing ( p = ). Sensitivity analysis of the drug dosage and the methodological quality of included studies did not show any major change on GAQ-1 and IIEF-EF (Supplement 3). Visual inspection suggested asymmetry in the sildenafil, vardenafil, and tadalafil funnel plots for GAQ-1, but funnel plots on other efficacy outcomes did not show any asymmetry (Supplement 3) Discussion There were four principal findings: (1) oral PDE5-Is are generally more effective in improving GAQ-1, IIEF, SEP-2, and SEP-3 compared with placebo; (2) in recommended doses, tadalafil seems to be the most effective ED treatment among all PDE5-Is analyzed, followed by vardenafil; (3) although oral PDE5-Is may cause more adverse events than placebo, they are generally mild and well tolerated; and (4) there is no significant difference on the safety profile of different classes of PDE5-Is. A large number of placebo-controlled trials have demonstrated the efficacy of oral PDE5-Is for patients diagnosed with ED. In this study we limited the patients to the broad-spectrum population, and the main findings favor PDE5-Is as well. Randomized controlled trials have also demonstrated the efficacy of oral PDE5-Is for the treatment of ED among patients with diabetes [29 32], hypertension [33,34], depression [35,36], and multiple sclerosis [37,38]. In2009,Tsertsvadzeetal.[12] reported a systematic review of 191 studies that also indicated that oral PDE5-Is improved erectile function and had similar safety profiles. Available comparative studies mainly investigated three traditional PDE5-Is: sildenafil, tadalafil, and vardenafil. Some randomized trials suggest that tadalafil is similar to sildenafil in improving IIEF-EF scores, and vardenafil is less effective than sildenafil and tadalafil [39 42]. There is evidence that patients prefer tadalafil to sildenafil and vardenafil for better efficacy [43 45]. Some studies also indicated that diet, especially fatty food intake, affected the pharmacokinetic profiles of sildenafil and vardenafil but not that of tadalafil [46]. Evidence from systematic reviews [12,47] suggests that oral PDE5-Is have similar efficacy profiles. However, these systematic reviews usually investigated only three PDE5-Is, and they may be affected by the

7 908 [(Fig._3)TD$FIG] EUROPEAN UROLOGY 63 (2013) Fig. 3 The comparative efficacy of phosphodiesterase type 5 inhibitors for erectile dysfunction (only comparisons with significance are shown). CI = confidence interval; GAQ-1 = Global Assessment Questionnaire question 1; IIFF-EF = International Index of Erectile Function-Erectile Function; MD = mean difference; RR = relative risk; SEP-2 = Sexual Encounter Profile question 2; SEP-3 = Sexual Encounter Profile question 3.

8 EUROPEAN UROLOGY 63 (2013) Table 3 The absolute treatment effects and the rank of efficacy of phosphodiesterase type 5 inhibitors for erectile dysfunction PDE5-I GAQ-1 IIEF-EF SEP 2 SEP 3 Absolute effect, mean (95% CI) Rank, mean, (95% CI) Absolute effect, mean (95% CI) Rank, mean (95% CI) Absolute effect, mean (95% CI) Rank, mean (95% CI) Absolute effect, mean (95% CI) Rank, mean, (95% CI) Sildenafil 0.73 ( ) 2.96 (1 5) 7.68 ( ) 3.30 (2 4) ( ) 4.92 (4 5) ( ) 4.94 (4 5) Tadalafil 0.75 ( ) 1.72 (1 4) 9.21 ( ) 1.15 (1 2) ( ) 1.80 (1 3) ( ) 1.99 (1 3) Vardenafil 0.73 ( ) 2.67 (1 5) 8.39 ( ) 2.15 (1 4) ( ) 1.95 (1 4) ( ) 1.98 (1 3) Udenafil 0.69 ( ) 4.12 (1 5) 7.53 ( ) 3.40 (2 4) ( ) 2.44 (1 4) ( ) 3.85 (3 5) Mirodenafil 0.70 ( ) 3.59 (1 6) NA NA ( ) 3.89 (1 5) ( ) 2.24 (1 4) Avanafil 0.46 ( ) 5.94 (5 6) NA NA NA NA NA NA Placebo 0.24 ( ) 6.99 (7 7) 1.64 ( ) 5 (5 5) 1.86 ( ) 5.99 (6 6) ( ) 6 (6 6) CI = confidence interval; GAQ-1 = Global Assessment Questionnaire question 1: While using the study medication, did you feel that your erections improved? ; IIEF-EF = International Index of Erectile Function-Erectile Function domain; NA = not applicable; PDE5-I = phosphodiesterase type 5 inhibitor; SEP- 2 = Sexual Encounter Profile question 2: Were you able to insert your penis into your partner s vagina? ; SEP-3 = Sexual Encounter Profile question 3: Did your erection last long enough for you to have successful completion of intercourse? relatively small patient numbers involved in the data analysis. So far there is hardly any large-scale study that compares the cost effectiveness of different PDE5-Is [48]. Available evidence suggests the costs of PDE5-Is are generally comparative [49]. Drug dosage is a potential determinant of treatment effect and may affect the conclusions of the comparisons between different drugs. We have observed that without considering dosage, drugs that are clinically recommended at a smaller dosage (eg, tadalafil and vardenafil) are always more effective than those with a larger dosage (eg, udenafil, mirodenafil, and avanafil). For example, tadalafil and vardenafil showed relatively better efficacy than others in all efficacy-related outcomes. This finding suggests that the former drugs are biologically more potent than the latter ones. Based on this, it can be extrapolated that drugs that are clinically recommended at a smaller dosage would become even more effective than those with a larger dosage if the dosages are adjusted the same. This was indeed confirmed in our meta-regression analysis including dosage as a covariate. After adjusting for dosage, the positive GAQ-1 rate for the PDE5-Is recommended at a smaller dosage was increased and those used at a larger dosage were decreased. As a result, the conclusions about the relative effectiveness between the drugs remain the same: tadalafil and [(Fig._4)TD$FIG] Fig. 4 The rank of the efficacy of phosphodiesterase type 5 inhibitors for erectile dysfunction. GAQ-1 = Global Assessment Questionnaire question 1; IIFF-EF = International Index of Erectile Function-Erectile Function; SEP-2 = Sexual Encounter Profile question 2; SEP-3 = Sexual Encounter Profile question 3. vardenafil are always the most effective agents whether the dosage is considered or not. If a higher dosage is used for drugs that are clinically used at the smaller dose, it could be more effective. However, dosages higher than the clinically recommended ones may not be suitable clinically because they may also increase the possibility of adverse effects [50 52]. In addition, in trials about the same drug, the dosages used were mostly within the clinically recommended range. Only four trials used a dosage beyond the clinically recommended range, and the result remained the same in the sensitivity analyses by excluding these trials. The strengths of this study are as follows. First, this study is the most comprehensive comparative study of oral PDE5- Is for ED to date. An exhaustive and contemporaneous search strategy was undertaken to ensure all eligible clinical trials were included, resulting in 118 studies including individuals. In addition, we used the relatively new approach of network meta-analysis to synthesize data. This method provided indirect effect estimates where direct comparisons were not available. Lastly, we evaluated the quality of the evidence with GRADE to provide a more comprehensive evidence-based review for practitioners. Limitations of this network meta-analysis primarily arise from the quality of the original trials reviewed. Poor reporting of these original studies is another weakness. Data like standard error were not provided in many original papers, so we were unable to include these studies in our network meta-analysis. Attempts to obtain missing data from original study authors were largely unsuccessful. Moreover, the test for consistency identified significant inconsistencies on GAQ-1 (sildenafil vs vardenafil) and flushing (tadalafil vs sildenafil) between direct and indirect comparisons. This may be due to the drug dosage. The trial that directly compared sildenafil with vardenafil in the GAQ-1 network meta-analysis model used relatively highdose sildenafil (100 mg); hence the results favored sildenafil. Lastly, visual inspection shows asymmetry in the sildenafil, vardenafil, and tadalafil funnel plots for GAQ-1, suggesting there may be a risk of publication bias. However, the asymmetry was eliminated after we replaced the relative estimates measurement risk ratio by risk difference or odds ratio. Funnel plots on other outcomes also showed no asymmetry, so the risk of publication bias was low.

9 910 [(Fig._5)TD$FIG] EUROPEAN UROLOGY 63 (2013) Fig. 5 The safety of different phosphodiesterase type 5 inhibitors for erectile dysfunction (only comparisons with significance are shown). RR = risk ratio; AE = adverse event; CI = confidence interval; SAE = severe adverse event; TRAE = treatment-related adverse event. 4. Conclusions This study has important implication for clinical practice and ED research. For practitioners, the findings indicate that, in recommended dosage, oral PDE5-Is are more effective than placebo for ED. Tadalafil is likely to be the most effective PDE5-I for ED, followed by vardenafil. PDE5-Is are generally safe and well tolerated, and there is no major difference among them. Because tadalafil also has many other advantages, such as patient preference, it may be considered the first choice for ED patients. For the three traditional PDE5-Is, sildenafil, vardenafil, and tadalafil, their efficacy has been confirmed, and future studies are unlikely to change the estimates. However, research on other PDE5-Is and comparison among different PDE5-Is are still required due to the limitations on the quality and quantity of the current available evidence. Author contributions: Chen Mao and JinLing Tang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Tang, Mao. Acquisition of data: Yuan, Qin, Ren, Zhang, Ding, Chen. Analysis and interpretation of data: Yuan, Jack, Mao. Drafting of the manuscript: Yuan, Zhang, Yang. Critical revision of the manuscript for important intellectual content: Tang, Mao.

10 EUROPEAN UROLOGY 63 (2013) Statistical analysis: Yuan, Lee, Mao. Obtaining funding: Tang, Mao. Administrative, technical, or material support: Tang, Mao. Supervision: Tang, Mao. Other (specify): Liu, Tian (quality assessment). Financial disclosures: Chen Mao and JinLing Tang certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: The study was funded by the Health and Medical Research Fund from the Food and Health Bureau of Hong Kong. The Food and Health Bureau of Hong Kong was not involved in the design or conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript. Acknowledgment statement: The authors thank Dr. Erin Armstrong (School of Public Health and Primary Care, The Chinese University of Hong Kong) for her help in editing the manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at j.eururo References [1] NIH Consensus conference. Optimal calcium intake. NIH Consensus Development Panel on Optimal Calcium Intake. JAMA 1994;272: [2] Bener A, Al-Hamaq AO, Kamran S, Al-Ansari A. Prevalence of erectile dysfunction in male stroke patients, and associated co-morbidities and risk factors. Int Urol Nephrol 2008;40: [3] Feldman HA, Goldstein I, Hatzichristou DG, Krane RJ, McKinlay JB. Impotence and its medical and psychosocial correlates: results of the Massachusetts Male Aging Study. J Urol 1994;151: [4] Mulhall J, King R, Glina S, Hvidsten K. Importance of and satisfaction with sex among men and women worldwide: results of the global better sex survey. J Sex Med 2008;5: [5] Wessells H, Joyce GF, Wise M, Wilt TJ. Erectile dysfunction. J Urol 2007;177: [6] Montague D, Jarow J, Broderick G, Dmochowski R, Heaton J, Lue T. Erectile Dysfunction Guideline Update Panel. Chapter 1: The management of erectile dysfunction: an AUA update. J Urol 2005;174: [7] Wespes E, Amar E, Hatzichristou D, et al. EAU guidelines on erectile dysfunction: an update. Eur Urol 2006;49: [8] Bell AS, Palmer MJ. Novel phosphodiesterase type 5 modulators: a patent survey ( ). Expert Opin Ther Pat 2011;21: [9] Aversa A, Bruzziches R, Pili M, Spera G. Phosphodiesterase 5 inhibitors in the treatment of erectile dysfunction. Curr Pharm Des 2006;12: [10] Ravipati G, McClung JA, Aronow WS, Peterson SJ, Frishman WH. Type 5 phosphodiesterase inhibitors in the treatment of erectile dysfunction and cardiovascular disease. Cardiol Rev 2007;15: [11] Blount MA, Beasley A, Zoraghi R, et al. Binding of tritiated sildenafil, tadalafil, or vardenafil to the phosphodiesterase-5 catalytic site displays potency, specificity, heterogeneity, and cgmp stimulation. Mol Pharmacol 2004;66: [12] Tsertsvadze A, Fink HA, Yazdi F, et al. Oral phosphodiesterase-5 inhibitors and hormonal treatments for erectile dysfunction: a systematic review and meta-analysis. Ann Intern Med 2009;151: [13] Caldwell DM, Ades AE, Higgins JP. Simultaneous comparison of multiple treatments: combining direct and indirect evidence. BMJ 2005;331: [14] Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 2004;23: [15] Higgins J, Green S. Cochrane handbook for systematic reviews of interventions, v Updated March Cochrane Collaboration Web site. [16] Guyatt GH, Oxman AD, Kunz R, Vist GE, Falck-Ytter Y, Schunemann HJ. What is quality of evidence and why is it important to clinicians? BMJ 2008;336: [17] Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336: [18] Dias S, Welton NJ, Sutton AJ, Ades AE. NICE DSU technical support document 2: a generalised linear modelling framework for pairwise and network meta-analysis of randomized controlled trials. National Institute for Health and Clinical Excellence Web site. Updated April [19] Caldwell DM, Welton NJ, Ades AE. Mixed treatment comparison analysis provides internally coherent treatment effect estimates based on overviews of reviews and can reveal inconsistency. J Clin Epidemiol 2010;63: [20] Dias S, Sutton AJ, Welton NJ, Ades AE. NICE DSU technical support document 3: heterogeneity: subgroups, meta-regression, bias and bias-adjustment. National Institute for Health and Clinical Excellence Web site. Updated April org.uk. [21] Dias S, Welton NJ, Sutton AJ, Caldwell DM, Lu G, Ades AE. NICE DSU Technical support document 4: inconsistency in networks of evidence based on randomised controlled trials. National Institute for Health and Clinical Excellence Web site. Updated April [22] Song F, Loke YK, Walsh T, Glenny AM, Eastwood AJ, Altman DG. Methodological problems in the use of indirect comparisons for evaluating healthcare interventions: survey of published systematic reviews. BMJ 2009;338:b1147. [23] Paick JS, Ahn TY, Choi HK, et al. Efficacy and safety of mirodenafil, a new oral phosphodiesterase type 5 inhibitor, for treatment of erectile dysfunction. J Sex Med 2008;5: [24] Paick JS, Choi HK, Kim SC, et al. Efficacy and safety of oral SK3530 for the treatment of erectile dysfunction in Korean men: a multicenter, randomized, double-blind, placebo-controlled, fixed dose, parallel group clinical trial. Asian J Androl 2008;10: [25] Zhao C, Kim SW, Yang DY, et al. Efficacy and safety of avanafil for treating erectile dysfunction: results of a multicentre, randomized, double-blind, placebo-controlled trial. BJU Int 2012;110: [26] Goldstein I, McCullough AR, Jones LA, et al. A randomized, doubleblind, placebo-controlled evaluation of the safety and efficacy of avanafil in subjects with erectile dysfunction. J Sex Med 2012;9: [27] Glina S, Toscano I, Gomatzky C, et al. Efficacy and tolerability of lodenafil carbonate for oral therapy in erectile dysfunction: a phase II clinical trial. J Sex Med 2009;6: [28] Glina S, Fonseca GN, Bertero EB, et al. Efficacy and tolerability of lodenafil carbonate for oral therapy of erectile dysfunction: a phase III clinical trial. J Sex Med 2010;7:

11 912 EUROPEAN UROLOGY 63 (2013) [29] Boulton AJ, Selam JL, Sweeney M, Ziegler D. Sildenafil citrate for the treatment of erectile dysfunction in men with type II diabetes mellitus. Diabetologia 2001;44: [30] Rendell MS, Rajfer J, Wicker PA, Smith MD. Sildenafil for treatment of erectile dysfunction in men with diabetes: a randomized controlled trial. Sildenafil Diabetes Study Group. JAMA 1999;281: [31] Safarinejad MR. Oral sildenafil in the treatment of erectile dysfunction in diabetic men: a randomized double-blind and placebocontrolled study. J Diabetes Complications 2004;18: [32] Stuckey BG, Jadzinsky MN, Murphy LJ, et al. Sildenafil citrate for treatment of erectile dysfunction in men with type 1 diabetes: results of a randomized controlled trial. Diabetes Care 2003;26: [33] Albuquerque DC, Miziara LJ, Saraiva JF, Rodrigues US, Ribeiro AB, Wajngarten M. Efficacy, safety and tolerability of sildenafil in Brazilian hypertensive patients on multiple antihypertensive drugs. Int Braz J Urol 2005;31: [34] van Ahlen H, Wahle K, Kupper W, Yassin A, Reblin T, Neureither M. Safety and efficacy of vardenafil, a selective phosphodiesterase 5 inhibitor, in patients with erectile dysfunction and arterial hypertension treated with multiple antihypertensives. J Sex Med 2005;2: [35] Fava M, Nurnberg HG, Seidman SN, et al. Efficacy and safety of sildenafil in men with serotonergic antidepressant-associated erectile dysfunction: results from a randomized, double-blind, placebocontrolled trial. J Clin Psychiatry 2006;67: [36] Tignol J, Furlan PM, Gomez-Beneyto M, et al. Efficacy of sildenafil citrate (Viagra) for the treatment of erectile dysfunction in men in remission from depression. Int Clin Psychopharmacol 2004;19: [37] Fowler CJ, Miller JR, Sharief MK, Hussain IF, Stecher VJ, Sweeney M. A double blind, randomised study of sildenafil citrate for erectile dysfunction in men with multiple sclerosis. J Neurol Neurosurg Psychiatry 2005;76: [38] Safarinejad MR. Evaluation of the safety and efficacy of sildenafil citrate for erectile dysfunction in men with multiple sclerosis: a double-blind, placebo controlled, randomized study. J Urol 2009;181: [39] Jannini EA, Isidori AM, Gravina GL, et al. The ENDOTRIAL study: a spontaneous, open-label, randomized, multicenter, crossover study on the efficacy of sildenafil, tadalafil, and vardenafil in the treatment of erectile dysfunction. J Sex Med 2009;6: [40] Conaglen HM, Conaglen JV. Investigating women s preference for sildenafil or tadalafil use by their partners with erectile dysfunction: the partners preference study. J Sex Med 2008;5: [41] Rubio-Aurioles E, Porst H, Kim ED, et al. A randomized open-label trial with a crossover comparison of sexual self-confidence and other treatment outcomes following tadalafil once a day vs. tadalafil or sildenafil on-demand in men with erectile dysfunction. J Sex Med 2012;9: [42] Eardley I, Mirone V, Montorsi F, et al. An open-label, multicentre, randomized, crossover study comparing sildenafil citrate and tadalafil for treating erectile dysfunction in men naive to phosphodiesterase 5 inhibitor therapy. BJU Int 2005;96: [43] Fusco F, Lembo A, Ludovico GM, et al. Tadalafil versus sildenafil citrate in the treatment of ED: Italian patients preferences and explanatory note. Urologia 2008;75: [44] Yoo C, Park J, Kim W, Hong B, Hong J, Ahn T. Comparison of the efficacy, safety and patient preference of the phosphodiesterase type 5 inhibitors for the patients with erectile dysfunction. Korean J Urol 2007;48: [45] Govier F, Potempa AJ, Kaufman J, Denne J, Kovalenko P, Ahuja S. A multicenter, randomized, double-blind, crossover study of patient preference for tadalafil 20 mg or sildenafil citrate 50 mg during initiation of treatment for erectile dysfunction. Clin Ther 2003;25: [46] Wright PJ. Comparison of phosphodiesterase type 5 (PDE5) inhibitors. Int J Clin Pract 2006;60: [47] Moore R, Derry S, McQuay H. Indirect comparison of interventions using published randomised trials: systematic review of PDE-5 inhibitors for erectile dysfunction. BMC Urol 2005;5:18. [48] Canadian Agency for Drugs and Technologies in Health. Phosphodiesterase 5 inhibitors: a review of the clinical effectiveness and cost-effectiveness. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH); [49] South Central Priorities Committee from Milton Keynes, Oxfordshire, Berkshire East, Berkshire West, Buckinghamshire (MOBBB). Treatments for erectile dysfunction. Updated Solutions for Public Health Web site. [50] Seftel A. High dose sildenafil citrate as a salvage therapy for severe erectile dysfunction. J Urol 2003;170:684. [51] Padma-Nathan H, McMurray JG, Pullman WE, et al. On-demand IC351 (Cialis) enhances erectile function in patients with erectile dysfunction. Int J Impotence Res 2001;13:2 9. [52] Hellstrom WJ, Gittelman M, Karlin G, et al. Sustained efficacy and tolerability of vardenafil, a highly potent selective phosphodiesterase type 5 inhibitor, in men with erectile dysfunction: results of a randomized, double-blind, 26-week placebo-controlled pivotal trial. Urology 2003;61:8 14.