Subjects were recruited from adults who contacted the Cornell University Medical Center in response to a newspaper article on dysthymia. Treatment was free, but subjects paid for medication. Upon recruitment, written informed consent was obtained. Inclusion criteria were age greater than 17 years, a score of at least 12 on the Hamilton depression scale (7), and DSM-III-R diagnosis of chronic major depression, double depression, or dysthymia. Subjects who had undergone surgery, were taking other medications, or had a confounding medical illness, neurological illness, or history of alcohol or drug abuse were excluded. At baseline, trained raters administered the axis I portion of the Structured Clinical Interview for DSM-III-R (8) and the Hamilton depression scale. Subjects were randomly assigned to 6 weeks of treatment with sertraline or paroxetine and received standard clinical management at 2-week intervals. The Hamilton depression scale was readministered after the 6-week trial by the treating physician. Sexual function was measured by using an early version of the Arizona Sexual Experience Scale, a gender-specific, five-item, self-report measure (9). Subjects rate, on a 5-point Likert scale, their current level of sexual drive, psychological arousal, physiologic arousal (erection or vaginal lubrication), ease of orgasm, and orgasm satisfaction. Low scores depict sexual response to be "as strong, easy, or satisfying as ever," while high scores depict the absence of response, e.g., "no sex drive," "never aroused," or "never reach orgasm." Higher scores, therefore, indicate greater sexual dysfunction.
Pretreatment differences between men and women were assessed by using the Mann-Whitney test. Within-group changes in item and total scores on the Arizona Sexual Experience Scale from before to after treatment were analyzed by using the Wilcoxon test. The associations between changes in scores on the Arizona Sexual Experience Scale and the Hamilton depression scale were evaluated by using Spearman correlation coefficients. Internal consistency reliability of the Arizona Sexual Experience Scale was evaluated by using Cronbach's coefficient alpha.