European Urology

1. Introduction

Benign prostatic hyperplasia (BPH) is a histologic diagnosis characterised by smooth muscle and epithelial cell proliferation in the prostate transition zone leading to nonmalignant prostate enlargement. It is widely recognised that BPH is not the exclusive cause of lower urinary tract symptoms (LUTS) [1] and [2]. However, clinical drug trials often enrol men based in part on a clinical diagnosis of non-neurogenic LUTS suggestive of BPH (LUTS/BPH).

The prevalence of bothersome LUTS/BPH increases with age, and epidemiologic and pathophysiologic links between LUTS/BPH and erectile dysfunction (ED) have been demonstrated [3] and [4]. Medical therapy for LUTS/BPH currently consists of α-blockers, 5α-reductase inhibitors, or combination therapy [1] and [2]. Although efficacious, these therapies have the potential for side-effects relating to sexual dysfunction [5]. Tadalafil is a phosphodiesterase type 5 (PDE5) inhibitor (PDE5-I) widely approved for the treatment of ED. Several placebo-controlled studies in men with LUTS/BPH have demonstrated improvements in International Prostate Symptom Scores (IPSS) with tadalafil [6], [7], [8], [9], and [10]. Tadalafil was recently approved in the United States for treatment of signs and symptoms of BPH (LUTS/BPH) and for the treatment of coexisting LUTS/BPH and ED. Although the mechanisms for improvements in LUTS with PDE5 inhibition have yet to be fully clarified, proposed contributors include inhibition of PDE5 isoenzymes present in the bladder, prostate, urethra, and supporting vasculature and consequent increases in intracellular nitric oxide–cyclic guanosine monophosphate concentration, relaxation of the smooth muscle cells in these structures, improved blood perfusion, and reduced afferent signalling from the urogenital tract [11], [12], [13], [14], and [15].

The primary objective of this study was to compare the effect of tadalafil 5 mg once daily with placebo on LUTS/BPH. Given that the α-blocker tamsulosin is often a first-line treatment for LUTS/BPH, tamsulosin was included as an active control, with a secondary objective of comparing tamsulosin 0.4 mg once daily with placebo. Although not designed for statistical testing of noninferiority or superiority between tadalafil and tamsulosin, the study was adequately powered for the comparison of each active treatment with placebo. This study provides for the first time data for both tadalafil and tamsulosin from a single large, randomised, placebo-controlled, international study.

2. Patients and methods

A double-blind, placebo- and active-controlled, parallel-design trial was conducted at 44 urology sites in Australia, Austria, Belgium, France, Germany, Greece, Italy, Mexico, The Netherlands, and Poland. Following screening (and a 4-wk wash-out for BPH, overactive bladder, or ED drugs, as needed), participants began a 4-wk single-blind placebo lead-in period, followed by randomisation (1:1:1 ratio) to once-daily tadalafil 5 mg, tamsulosin 0.4 mg, or placebo for 12 wk (Fig. 1). Dosing was to occur approximately 30 min after eating (per tamsulosin dosing recommendations [16]). The trial was performed in accordance with applicable laws and regulations, good clinical practices, and ethical principles as described in the Declaration of Helsinki. Institutional review boards for each site approved the trial. All men provided written informed consent before initiating any trial procedure. Enrolment began in December 2009, and the last subject completed the study in January 2011.

Fig. 1 Disposition of subjects. Subject Consolidated Standards of Reporting Trials (CONSORT) diagram.AE = adverse event.* One subject randomised to tamsulosin did not take at least one dose of study drug and was excluded from the efficacy analyses.

Eligible men were ≥45 yr of age who had had LUTS/BPH for >6 mo at screening and with IPSS ≥13 and maximum urinary flow rate (Q_max) ≥4 to ≤15 ml/s prior to the placebo lead-in period; subjects with improvements in IPSS or Q_max during placebo lead-in were not excluded. Compliance ≥70% during the placebo lead-in was required for randomisation. Men could not have used finasteride or dutasteride within 3 or 6 mo, respectively. Other inclusion and exclusion criteria were as described previously [9] in addition to tamsulosin-specific exclusions of men with planned cataract surgery; history of symptomatic orthostatic hypotension (especially upon initial α-blocker administration); or recurrent dizziness, vertigo, loss of consciousness, or syncope.

Efficacy measures included IPSS (primary measure) [17], BPH Impact Index (BII) [18], and International Index of Erectile Function–Erectile Function Domain (IIEF-EF) [19]. A week 1 IPSS (mIPSS) used questions beginning with “Since your last visit.” IPSS storage and voiding subscores, nocturia question (question 7), and IPSS Quality-of-Life (QoL) Index were also assessed. Measures were assessed at baseline and end point (12 wk or end of therapy) and also at screening, 1 and 4 wk for IPSS, and 4 wk for BII. The Patient and Clinician Global Impression of Improvement (PGI-I and CGI-I, respectively) instruments [20] and [21] and the subject-rated Treatment Satisfaction Scale–BPH (TSS-BPH) [22], evaluated from 0% (greater) to 100% (lower) satisfaction, were administered at end point.

Uroflowmetry was performed using standard calibrated devices at the screening, baseline, and end point visits. Valid Q_max measurements required prevoid total bladder volume (assessed by ultrasound) of ≥150 to ≤550 ml and voided volume (V_void) of ≥125 ml. Screening uroflowmetry was interpreted by investigators to assess eligibility; baseline and end point uroflowmetry were interpreted by a blinded central reader and used for analyses. Bladder capacity was calculated post hoc as the sum of V_void and postvoid residual (PVR) volume.

Safety was evaluated based on subject-reported adverse events (AE), PVR, clinical laboratory parameters (haematology, chemistry, and urinalysis), and vital signs. Treatment-emergent AEs (TEAE) were those first reported or worsening after randomisation.

Randomisation was stratified by baseline LUTS severity, geographic region, and patient query regarding prior ED diagnosis (yes or no). The primary efficacy objective was evaluating the change in total IPSS from baseline to end point for tadalafil versus placebo. Efficacy was analysed in all randomised subjects who started double-blind study drug. At least 151 subjects per treatment arm provided an estimated ≥80% power to detect a placebo-adjusted mean treatment difference in IPSS of 2.0 (assuming a standard deviation of 6 and a projected nonevaluable rate of 5%). The study was not powered to demonstrate noninferiority of tadalafil to tamsulosin or for direct comparisons between active treatment arms.

Continuous efficacy measures, uroflowmetry, and PVR were evaluated as change from baseline (randomisation) to week 12/last-observation-carried-forward (LOCF) end point. Analyses for 1 or 4 wk did not use LOCF imputation. Continuous efficacy measures were assessed using analysis of covariance (ANCOVA), with terms for treatment group, region, and baseline, and baseline-by-treatment interaction and treatment-by-region interaction (removed where p≥ 0.1).

A fixed-sequence testing procedure was implemented to control type I error in analyses of primary and key secondary outcomes for tadalafil using the following prespecified order: total IPSS at end point, total IPSS after 4 wk, BII at end point, mIPSS after 1 wk, and BII after 4 wk. Statistical significance was interpreted only if results of the preceding analysis were significant at the 0.05 level. Results were presented independent of the fixed sequence, as all end points tested under this procedure achieved statistical significance. All other efficacy analyses (including all tamsulosin results) were assessed at the 0.05 significance level without adjustment for multiplicity.

PGI-I and CGI-I treatment group differences were analysed using the Cochran-Mantel-Haenszel test adjusted for baseline LUTS severity. TSS-BPH treatment group differences were analysed using the van Elteren test for differences in medians stratified by region.

Treatment group differences for AEs were analysed using Fisher exact tests. Changes from baseline to end of therapy in Q_max, PVR, and clinical laboratory parameters were analysed using a ranked analysis of variance (ANOVA) with a term for treatment group. Treatment group differences for average urinary flow rate (Q_ave), V_void, and bladder capacity were performed as post hoc analyses.

3. Results

Of 652 subjects screened, 511 were randomised (safety population), and 510 started study drug (efficacy population); 88.8% of subjects completed the study (Fig. 1). Treatment groups were well balanced concerning demographic and clinical characteristics (Table 1). Mean age was 64 yr of age (10.2% ≥75 yr of age), and most patients were from Europe (71%). At randomisation, IPSS was ≥20 in 30% of subjects, and Q_max was <10 ml/s in 54%.

The change from baseline to week 12 (LOCF) relative to placebo in total IPSS was statistically significant for both tadalafil and tamsulosin (p = 0.001 and p = 0.023, respectively; Fig. 2A; Table 2). Least squares mean (LSmean) plus or minus standard error (SE) differences in IPSS versus placebo were significant for both tadalafil and tamsulosin at 1 wk (mIPSS: −1.5 ± 0.5; p = 0.003 and −1.5 ± 0.5; p = 0.005, respectively) and 4 wk (−2.2 ± 0.6; p < 0.001 and −2.3 ± 0.6; p < 0.001, respectively; Fig. 3A). Based on prespecified subgroup analysis, there was no significant treatment by previous α-blocker therapy interaction (p = 0.230) with respect to changes in total IPSS. Changes in IPSS subscores and nocturia are shown in Table 2.

Fig. 2 Differences from placebo in (A) total International Prostate Symptom Score and (B) Benign Prostatic Hyperplasia Impact Index. Numbers represent least squares mean treatment difference (95% confidence interval) versus placebo in the change from baseline to 12 wk (last-observation-carried-forward) and corresponding p values from analysis of covariance.IPSS = International Prostate Symptom Score; BII = Benign Prostatic Hyperplasia Impact Index; LS = least squares.

Fig. 3 Changes from baseline in (A) total International Prostate Symptom Score and (B) Benign Prostatic Hyperplasia Impact Index. Data represent the least squares mean change plus or minus standard error.LS = least squares; EP = end point; ANCOVA = analysis of covariance.* p < 0.05 versus placebo based on ANCOVA.

Differences from placebo in BII were statistically significant for both tadalafil and tamsulosin (p = 0.003 and p = 0.026, respectively; Fig. 2B). Differences from placebo in BII were also significant at 4 wk for both tadalafil and tamsulosin (LSmean ± SE: −0.8 ± 0.2; p < 0.001 and −0.9 ± 0.2; p < 0.001, respectively; Fig. 3B).

For the IPSS QoL Index, significant improvements compared with placebo at 12 wk were reported with tadalafil (p = 0.022) but not tamsulosin (p = 0.546; Table 3). The TSS-BPH overall satisfaction score at end point was significantly lower (indicating higher satisfaction) in the tadalafil group compared with placebo (p = 0.005), driven by greater satisfaction with efficacy (p = 0.003; data not shown). There was no significant difference between tamsulosin and placebo in TSS-BPH overall satisfaction (p = 0.457) or satisfaction with efficacy (p = 0.409; data not shown).

Between-treatment group differences compared to placebo in the distribution of subjects over the PGI-I and CGI-I response categories were significant for tadalafil (PGI-I: p = 0.001; CGI-I: p = 0.004) but not for tamsulosin (PGI-I: p = 0.114; CGI-I: p = 0.452; Table 3). In the tadalafil group, more subjects and their clinicians perceived improvements in LUTS at end point compared to the placebo group.

In comparison to placebo, LSmean ± SE change from baseline to end point in the IIEF-EF domain in men with ED who were also sexually active (approximately 60% of subjects) was significant with tadalafil (4.0 ± 1.0; p < 0.001), whereas the LSmean change with tamsulosin was not (−0.4 ± 1.0; p = 0.699).

Improvements in Q_max were significantly greater than placebo with tadalafil (p = 0.009) and with tamsulosin (p = 0.014; Table 4), as were increases in Q_ave (p = 0.002 and p = 0.023, respectively). Differences versus placebo were not statistically significant for V_void or bladder capacity.

There were no significant differences between the active treatment groups and placebo for any TEAE or in the incidence of serious AEs or discontinuations because of AEs (Table 5). The most common TEAEs with tadalafil were headache (n = 5) followed by nasopharyngitis (n = 5), back pain (n = 4), dizziness (n = 4), and dyspepsia (n = 4), while with tamsulosin the most common events were headache (n = 7) and dizziness (n = 6). One subject in the tamsulosin group reported two TEAEs related to ejaculatory dysfunction (retrograde ejaculation and semen volume decreased). There were no clinically significant changes in laboratory measurements or vital signs. For PVR, mean reductions from baseline compared with placebo were observed with both active treatments but were not statistically significant (Table 4).

4. Discussion

In this, the first international, placebo-controlled study evaluating tadalafil or tamsulosin (as an active control) for LUTS/BPH, tadalafil 5 mg once daily for 12 wk resulted in significant and clinically meaningful improvements in LUTS/BPH similar to tamsulosin 0.4 mg once daily. Based on analyses of secondary end points, this was also the first tadalafil study to show significant improvement in LUTS/BPH after 1 wk and a significant increase in Q_max at 12 wk. In addition, tadalafil but not tamsulosin significantly improved measures of LUTS/BPH QoL, global impressions of BPH symptom impact, and BPH treatment satisfaction at end point and improved erectile function in those men who also reported ED.

The magnitude of improvement in total IPSS at end point with tadalafil 5 mg in this study was consistent with several previous reports [7], [9], and [10], and the improvement was also comparable to that seen in previous studies with α-blockers [1] and with tamsulosin 0.4 mg here. Similar numerical reductions in total IPSS were also reported for tadalafil 5 mg or tamsulosin 0.2 mg once daily in a small pilot study of Korean men with LUTS/BPH [23]. Tadalafil and tamsulosin also similarly reduced the impact of symptoms on daily living (assessed by BII) at 4 and 12 wk.

In contrast with findings for total IPSS and BII, only tadalafil was significantly superior to placebo on secondary measures of satisfaction and improvement, including IPSS QoL, global impressions of improvement (CGI-I and PGI-I), and satisfaction with BPH treatment (TSS-BPH). Tadalafil also significantly improved ED compared with placebo in sexually active men with ED, while tamsulosin did not.

The finding of a significant improvement in Q_max with tadalafil in this study contrasts with the majority of previous studies on tadalafil, sildenafil, and vardenafil for LUTS/BPH, which typically identified only a numeric Q_max improvement (reviewed in Laydner et al. [24]). Although the statistically significant Q_max change in this study could be a random finding, a trend towards a dose-response relationship was reported in a large dose-ranging study with tadalafil [25], a small increase in Q_max that reached statistical significance was seen with tadalafil 2.5 mg (but not 5 mg) in another controlled study [10], and a significant increase in Q_max compared with placebo was observed with another once-daily PDE5-I previously under development [26]. Compared with previous tadalafil studies, the change in Q_max observed here was similar for placebo (1.1–1.2 ml/s previously) but greater for tadalafil 5 mg (1.6–1.7 ml/s previously) [6], [9], [10], and [25]. Although the current study population did not differ from previous studies for baseline demographics or clinical characteristics, a greater proportion of subjects were from Europe. In addition, baseline Q_max in the active treatment groups (tadalafil 9.9 ml/s and tamsulosin 9.4 ml/s) was lower than prior tadalafil studies (range for tadalafil 5 mg: 9.8–11.7 ml/s), which could allow more room for improvement and thus increase the probability of observing an improvement in Q_max. Several in vitro studies have reported smooth muscle relaxation in the human bladder neck and prostate when exposed to PDE5 inhibition [14] and [27], and the smooth muscle relaxant effect of tadalafil seems to be of similar magnitude as alfuzosin [28]. However, it remains unclear whether these in vitro findings translate into the observed small Q_max changes seen with tadalafil. Tadalafil at a higher 20-mg daily dose had no significant effect on the bladder outlet obstruction (BOO) index in a safety study of men with LUTS/BPH; however, these men were not required to have BOO at baseline [29]. In contrast to the small improvements in Q_max, which have not reached statistical significance in most other studies, studies of tadalafil for LUTS/BPH have consistently demonstrated significant and clinically meaningful IPSS improvement [7], [9], and [10]. This is aligned with recently updated BPH guidelines stating that there is a poor correlation between symptoms and Q_max[1]. There were no novel safety findings in this study, and the most common TEAEs were consistent with the known side-effect profiles of these agents.

A limitation of the present study is that it was not powered to assess noninferiority or superiority between tadalafil and tamsulosin. Nonetheless, it was fully powered to individually assess tadalafil or tamsulosin versus placebo. Although a minimum clinically relevant noninferiority margin for LUTS/BPH treatments has not been clearly established, by conservative estimates, the sample size required for a noninferiority trial would likely be 2- to 4-fold larger than the current study. Although this study was of standard duration for trials assessing LUTS/BPH efficacy, it did not address longer-term efficacy of tadalafil or effects on disease progression; such studies would be of interest in the future. A prior 1-yr open-label extension tadalafil study found tadalafil 5 mg once daily to be well tolerated with maintenance of efficacy [30]. Although the increase in Q_max observed with tamsulosin in this study was in the range of changes observed in prior tamsulosin studies [2], the increase observed with tadalafil was greater than in prior tadalafil studies. Caution should therefore be exercised in interpreting this finding.

5. Conclusions

Tadalafil 5 mg or tamsulosin 0.4 mg once daily resulted in significant and similar improvements versus placebo in LUTS/BPH symptoms as early as 1 wk and throughout the 12-wk treatment period. In addition, tadalafil and tamsulosin similarly improved Q_max through 12 wk. However, only tadalafil resulted in significant improvements in LUTS/BPH based on secondary measures of QoL and treatment satisfaction and in erectile function as measured by the IIEF-EF domain. AEs were few, and the AE profiles of tadalafil and tamsulosin were consistent with those previously reported for these drugs.

Author contributions: Matthias Oelke 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: Viktrup, Xu.

Acquisition of data: Oelke, Giuliano, Mirone.

Analysis and interpretation of data: Oelke, Giuliano, Mirone, Xu, Cox, Viktrup.

Drafting of the manuscript: Oelke, Cox.

Critical revision of the manuscript for important intellectual content: Oelke, Giuliano, Mirone, Xu, Cox, Viktrup.

Statistical analysis: Xu.

Obtaining funding: None.

Administrative, technical, or material support: None.

Supervision: Oelke.

Other (specify): None.

Financial disclosures: I 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: Matthias Oelke has received lecturer and/or consultant honoraria in the field of LUTS/BPH from Astellas, GlaxoSmithKline, Eli Lilly and Company, and Merckle-Recordati. Francois Giuliano is a consultant and lecturer for Eli Lilly and Company and a consultant and investigator for Bayer-Schering. Vincenzo Mirone is a consultant, investigator, and speaker for Eli Lilly and Company and Bayer Health-Care. Lei Xu, David Cox, and Lars Viktrup are employees and stockholders of Eli Lilly and Company.

Funding/Support and role of the sponsor: Eli Lilly and Company helped design and conduct and provided support for this study. Data collection and management and all statistical analyses were performed and retained by the sponsor. The corresponding author and coordinating investigator, Matthias Oelke, together with the Lilly study team and other co-authors, interpreted the data and participated in the preparation, review, and approval of the manuscript.

Acknowledgment statement: Thomas Melby (an employee of PharmaNet/i3, Indianapolis, IN, USA) assisted in the preparation of this manuscript. The authors would like to thank the trial participants and study investigators, without whom this work could not have been performed.