ABSTRACT: This post hoc analysis of data from a multicenter, randomized, double-blind study determined how many attempts were needed to record at least 1 successful penetration and maintenance of erection long enough for successful intercourse in a broad population of men with erectile dysfunction taking vardenafil at 5, 10, or 20 mg or placebo. The cumulative probability of achieving successful penetration and of maintaining an erection increased with the number of attempts for all 3 vardenafil groups. For the first attempt, the probability of achieving successful penetration was higher in all 3 vardenafil groups compared with placebo; 67% in the 5-mg vardenafil group, 77% in the 10-mg vardenafil group, and 74% in the 20-mg vardenafil group compared with 46% for placebo. By the third attempt, the probability of at least 1 success was 82% for 5, 88% for 10, and 85% for 20 mg vardenafil compared with 68% for placebo. The probability of maintaining an erection long enough to complete intercourse at the first attempt was 51% for 5, 69% for 10, and 61% for 20 mg vardenafil compared with 28% for the placebo group. By the third attempt, the probability of maintaining an erection was 66% for 5, 81% for 10, and 77% for 20 mg vardenafil in contrast to 53% for placebo. The results of this analysis indicate that patients without initial treatment success should continue treatment or increase the dose because the cumulative probability of success increases with additional attempts with vardenafil, with a plateau at about the fourth dose.