The role of psychological and interpersonal factors in the treatment of erectile dysfunction (ED) with sildenafil or other oral therapies has not been sufficiently investigated. We conducted a pilot study of psychosocial predictors of pharmacotherapy treatment outcome and satisfaction in men with ED and their partners. Sixty-nine men with mild to moderate ED and their partners were enrolled in a multicenter, open-label, treatment trial with sildenafil. Treatment measures included a battery of validated self-report measures and questionnaires. Subjects also were interviewed according to a semistructured interview protocol. Partner assessments included self-report measures of sexual function, mood, and relationship satisfaction. Results indicated that, prior to treatment, patients had erectile function scores in the range of mild to moderate ED, with relatively low levels of concomitant depression, anxiety, and psychological stress and high overall levels of relationship adjustment. Partner sexual function was in the normal range of total Brief Index of Sexual Functioning for Women (BISF-W; Taylor, Rosen, Leiblum, 1994) scores, although more than one third of female partners had specific sexual complaints or problems. Among couples who completed one or both follow-up visits (N = 34), sildenafil treatment resulted in significant improvements in all aspects of sexual function in men, including sexual desire, orgasmic function, erectile function and overall sexual satisfaction. Significant improvements also were noted in partners' ratings of sexual function in most domains, including arousal, pleasure, and orgasm. Higher baseline levels of sex-specific anxiety were negatively associated with improvement in erections following treatment. Relationship adjustment at baseline, contrary to expectations, did not predict erectile or sexual satisfaction following treatment in the men or their partners but was significantly correlated with changes in sexual desire. Baseline levels of depression, anxiety, and stress generally were unrelated to efficacy or treatment satisfaction. However, we observed a curvilinear relationship in the men between baseline levels of stress and treatment discontinuation (i.e., subjects with moderate levels of stress were less likely to discontinue treatment). Because of a high number of dropouts, results of this pilot study await confirmation in a larger and more adequately powered clinical trial.