Time to discontinuation of randomized treatment was defined as the number of days from randomization until the day the participant discontinued the randomized treatment. Discontinuation of randomized treatment was defined as switching between the 3 study treatments (tadalafil on demand, tadalafil once a day, or sildenafil citrate on demand) or discontinuing from all treatments. A change of dose within the same treatment was not considered switching of treatment. This outcome measure was estimated using the Kaplan-Meier product-limit method.

The analysis included all randomly assigned participants.

Self-reported EF score over past 4 weeks. Items 1-5 scores range from 0 (no sexual activity) to 5 (high EF). Item 15 score ranges from 1 (very low confidence to get/keep erection) to 5 (very high confidence). Total scores range from 1 to 30; lower scores denote greater erectile dysfunction severity. Least Squares Mean changes from baseline to endpoint for each visit from repeated measures analysis included terms for baseline score, treatment group, country, baseline*treatment (if p<0.10), visit, and visit*treatment. Correlation matrix for repeated observations assumed to be unstructured.

The analysis included all randomly assigned participants who had a baseline and post-baseline IIEF-EF observation.

Self-reported orgasmic function on the IIEF over past 4 weeks and consists of 2 questions (items 9 and 10). Each question is rated on a scale from 0 (no sexual stimulation) to 5 (almost always/always). Total scores range from 0 to 10; lower scores represent lower orgasmic function. Least Squares Mean changes from baseline to endpoint for each visit were from a repeated measures analysis and included terms for baseline score, treatment group, country, baseline*treatment (if p<0.10), visit, and visit*treatment. The correlation matrix for the repeated observations was assumed to be unstructured.

The analysis included all randomly assigned participants who had a baseline and post-baseline IIEF Orgasmic Function observation.

Self-reported sexual desire on IIEF over past 4 weeks; comprises 2 questions (items 11 and 12). Each question rated on a scale from 1 (almost never or low/no sexual desire) to 5 (almost always or very high sexual desire). Total scores range: 2 to 10; lower numerical scores denote lower sexual desire. Least Squares Mean changes from baseline to endpoint for each visit from repeated measures analysis and included terms for baseline score, treatment group, country, baseline*treatment (if p<0.10), visit, and visit*treatment. Correlation matrix for repeated observations assumed to be unstructured.

The analysis included all randomly assigned participants who had a baseline and post-baseline IIEF Sexual Desire observation.

Participant-assessed diary. Has 5 questions (Question[Q]1:erection achievement, Q2:successful penetration, Q3:successful intercourse, Q4:satisfied with erection, and Q5:satisfied with sexual experience) for each sexual encounter made over specified period of time. SEP Q1-Q5 scores determined as percentage of 'Yes' responses to each of 5 questions out of all sexual attempts recorded during the time period. Least Squares Mean changes from baseline to endpoint for each visit from a repeated measures analysis included terms for baseline score, treatment group, country, visit, and visit*treatment.

The analysis included all randomly assigned participants who had a baseline and post-baseline SEP observation.

The GAQ consists of 2 Yes/No/No Response (No Respo) questions. GAQ Question (Q)1: Has the treatment you have been taking during this study improved your erections? GAQ Q2: Has the treatment improved your ability to engage in sexual activity?

The analysis included all randomly assigned participants. The last available GAQ assessment for each participant was used in the analysis.

The number of times participants switched erectile dysfunction medication within the 3 treatments being studied (tadalafil on demand, tadalafil once a day, or sildenafil citrate on demand).

The analysis included all randomly assigned participants who switched erectile dysfunction medication and had a baseline and post-baseline observation.

Results are reported as the number of participants per sequence of study medications (tadalafil on demand, tadalafil once a day, or sildenafil citrate on demand) that were taken as a result of switching treatments. The Other Treatment Sequence reports the number of participants per sequence of study medications that were taken as a result of switching treatments more than once. The number of participants who did not switch is also reported.

All randomly assigned participants were included in the analysis.

The reported reasons for a decision to discontinue from initial randomized treatment prior to Week 24 are reported. Discontinuation of randomized treatment was defined as switching between the 3 study treatments (tadalafil on demand, tadalafil once a day, or sildenafil citrate on demand) or discontinuing from all treatments. A change of dose within the same treatment was not considered as switching of treatment.

The analysis included all randomly assigned participants.

The differences in time between the 8-week time point and the discontinuation of all study treatments (that is, discontinuation from the study and not switching to another treatment) are reported by the median (95% confidence interval). Duration was measured as the number of days from Week 8 to the date of the last dose of the study drug. This outcome measure was estimated using the Kaplan-Meier product-limit method.

The analysis included all randomly assigned participants who completed the 8-week randomized treatment.

Self-reported intercourse satisfaction score over past 4 weeks (1 intercourse attempt item, 2 intercourse satisfaction items). Each item range: 0 (no intercourse attempts/no satisfaction) to 5 (more attempts/high satisfaction). Total scores range: 0-15; lower scores=lower intercourse satisfaction. Least Squares Mean changes from baseline to endpoint for each visit from repeated measures analysis and included terms for baseline score, treatment group, country, baseline*treatment (if p<0.10), visit, and visit*treatment. Correlation matrix for repeated observations assumed to be unstructured.

The analysis included all randomly assigned participants who had a baseline and post-baseline IIEF Intercourse Satisfaction observation.

Self-reported overall satisfaction on the IIEF over past 4 weeks and consists of 2 questions (items 13 and 14), each rated on a scale from 1 (very dissatisfied) to 5 (very satisfied). Total scores range from 2 to 10; lower numerical scores represent lower overall satisfaction. Least Squares Mean changes from baseline to endpoint for each visit were from a repeated measures analysis and included terms for baseline score, treatment group, country, baseline*treatment (if p<0.10), visit, and visit*treatment. The correlation matrix for the repeated observations was assumed to be unstructured.

The analysis included all randomly assigned participants who had a baseline and post-baseline IIEF Overall Satisfaction observation.

The participant questionnaire consists of 11 questions. Each question is rated on a scale of 0 (extremely low treatment satisfaction) to 4 (extremely high treatment satisfaction). The EDITS summary score was obtained by adding each individual result for all questions, dividing by the number of questions answered (mean satisfaction score), then multiplying by 25, thus obtaining a score range from 0 (extremely low treatment satisfaction) to 100 (extremely high satisfaction). Least Squares Mean changes were adjusted for treatment group, country, visit, and visit*treatment.

The analysis included all randomly assigned participants who had a baseline and post-baseline EDITS observation.

The PAIRS is a 23-item scale that assesses broader psychological/interpersonal outcomes associated with erectile dysfunction and its treatment. Each question is rated on a Likert scale from 1 (strongly disagree) to 4 (strongly agree). The scale consists of 3 domains: Sexual Self-Confidence (items 1-6), Spontaneity domain (items 7-15), and Time Concerns (items 16-23). The average domain score for each domain was calculated by adding the nonmissing items for the respective domain, then dividing by the number of nonmissing items for the respective domain. Each average domain score ranged from 1 to 4. Higher scores represent the following: greater sexual self-confidence (better outcome); greater spontaneity (better outcome); higher time concerns (worse outcome). The Least Squares Mean changes were adjusted for treatment group, country, baseline IIEF-EF severity, baseline domain score, and baseline domain score*treatment (if p<0.10).

The analysis included all randomly assigned participants who had a baseline and post-baseline PAIRS observation.

SEAR assesses psychosocial outcomes in men with erectile dysfunction. Sexual Relationship domain consists of 8 items (items 1-8). Items 2-8 are rated on a scale of 1 (Never) to 5 (Always), whereas item 1 is reverse scored (1=Always and 5=Never). The domain score was computed by summing its respective items, then transforming it into a 0 (least favorable) to 100 (most favorable) scale. Transformed score = 100 x [(actual raw score - lowest possible raw score)/possible raw score range]. Least Squares Mean changes adjusted for baseline score, treatment group, country, visit, and visit*treatment.

The analysis included all randomly assigned participants who had a baseline and post-baseline SEAR Sexual Relationship observation.

SEAR assesses psychosocial outcomes in men with erectile dysfunction. Confidence domain measures improvement in confidence; 2 subscales (6 items: Self-Esteem [items 9-12]; Overall Relationship [items 13-14]). Each item range: 1 (Never) to 5 (Always); item 11 reverse scored. Domain score=sum of domain's respective items, then transformed into 0 (least favorable) to 100 (most favorable) scale. Transformed score=100x[(actual raw score-lowest possible raw score)/possible raw score range]. Least Squares Mean change adjusted for baseline score, treatment group, country, visit, and visit*treatment.

The analysis included all randomly assigned participants who had a baseline and post-baseline SEAR Confidence observation.

SEAR assesses psychosocial outcomes in men with erectile dysfunction. The Self-Esteem domain consists of 4 items (items 9-12), each rated on a scale of 1 (Never) to 5 (Always). Item 11 is reverse scored (1=Always and 5=Never). The domain score was computed by summing its respective items, then transforming it into a 0 (least favorable) to 100 (most favorable) scale. The transformed score = 100 x [(actual raw score - lowest possible raw score)/possible raw score range]. Least Squares Mean changes were adjusted for baseline score, treatment group, country, visit, and visit*treatment.

The analysis included all randomly assigned participants who had a baseline and post-baseline SEAR Self-Esteem observation.

SEAR assesses psychosocial outcomes in men with erectile dysfunction. The Overall Relationship domain consists of 2 items (items 13-14), each rated on a scale of 1 (Never) to 5 (Always). The domain score was computed by summing its respective items, then transforming it into a 0 (least favorable) to 100 (most favorable) scale. The transformed score = 100 x [(actual raw score - lowest possible raw score)/possible raw score range]. Least Squares Mean changes were adjusted for baseline score, treatment group, country, visit, and visit*treatment.

The analysis included all randomly assigned participants who had a baseline and post-baseline SEAR Overall Relationship observation.