This study compared cognitive-behavioral psychotherapy with leuprolide acetate (LA) to cognitive-behavioral psychotherapy with saline injections for 12 months. Five white male pedophiles (M age, 50 years; range, 36-58) volunteered for a two-year study. LA was administered by Depo injection (7 mg initially, then 22.5 mg every 3 mos) for 12 months, followed by saline placebo. Effects of LA on testosterone levels, sexual interest preference by visual reaction time (Abel Assessment), penile tumescence (Monarch PPG), as well strong urges and masturbatory frequency to children (polygraph), were measured every three months. Subjects were treated with weekly cognitive-behavioral psychotherapy. On LA, testosterone decreased to castrate levels. Because of the suppression of testosterone, physiologic arousal response as measured by penile plethysmography (penile tumescence) was significantly suppressed compared with baseline. However, sufficient response remained to detect pedophilic interest. This pedophilic interest was also detected by visual reaction times. All subjects self-reported a decrease in strong pedophilic urges and masturbation. When asked about having pedophilic urges and masturbating to thoughts of children, polygraph responses indicated subjects were not deceptive when they reported decreases. On placebo, testosterone and physiologic arousal eventually rose to baseline levels. At baseline and on placebo, subjects were consistently deceptive regarding increased pedophilic urges and masturbatory frequency as noted by polygraph. Interest preference, as measured by Abel Assessment and Monarch PPG, was generally unchanged throughout the study. Cognitive-behavioral psychotherapy augmented with LA significantly reduced pedophilic fantasies, urges, and masturbation, but did not change pedophilic interest during one year of therapy. Deceptive responses by polygraph suggested that self-report was unreliable. Follow-up utilizing objective measures is essential for monitoring efficacy of treatment in pedophilia. Our study supports the supposition that modification of pedophilic behavior is possible. LA may augment cognitive-behavioral psychotherapy and help break the sequence leading to a reoffense.

Hormone therapy is aimed at modification of the basic sex drive, whereas cognitive-behavioral psychotherapy targets cognitive processes that precede all events leading to pedophilic behavior. Obsession/compulsion and sex drive might affect the decision process. Because of a suggested relationship between obsessive compulsive disorder and paraphilias, SSRIs have been utilized. Most reports have been documented from individual case studies or uncontrolled, retrospective studies (Kafka, 2000).

Reports of the efficacy of drug therapy for suppression of sexual drive and modification of behavior continue to emerge. The anti-androgens, MPA and CPA, are the most widely used drugs. They provide variable testosterone suppression, dependent on dose. LHRH inhibitors and GNRH agonists also provide profound testosterone suppression. More recently, triptorelin, an agent similar to LA, (Rosler & Wilztum, 1998) and selective serotonin reuptake inhibitors (SSRIs) have also been investigated (Raymond, Grant, Kim,.& Colemen, 2002; Stein, Hollander, Anthony, Schneier, Fallon, Liebowitz, Klein, 1992).

Despite recent advances in the use of testosterone-suppressing drugs for the treatment of pedophilia, limited information is available in the form of well designed, controlled objective studies. The lack of standardized, objective outcome measures has hampered objective assessment of efficacy of interventions for pedophilia (Briken, Hill, & Berner, 2003; Hanson et al., 2002; Marshall & Barbaree, 1990). A multi-disciplinary team of experts is needed to assess concurrent, objective physiologic measures of response. Though some results of treatment are promising, a reliable, well-tolerated approach that can be objectively monitored for efficacy is lacking.

The specific aspects of behavior that are affected by therapy, as well as the efficacy, limitations, and side effects of treatment need to be determined (Briken, Hill, & Berner, 2003; Raymond, Robinson, Kraft, Rittberg, & Coleman, 2001). The present study assessed the impact of cognitive-behavioral psychotherapy alone versus cognitive-behavioral psychotherapy augmented with a potent testosterone-lowering drug, LA, for suppression of pedophilic urges and arousability. The psychotherapy program comprised cognitive and behavioral therapy. Objective assessments of sexual interest preference and physical arousal, as well as sexual urges and masturbation, were monitored before, during, and after therapy. The research objectives were to explore the relationship between testosterone level and sexual interest preferences, physiologic responses to sexual stimuli (auditory and visual), and self-reported sexual behaviors.

METHOD Subjects Five men (M age, 50 years; range, 36-58), with a diagnosis of pedophilia (American Psychiatric Association, 1994) and who were "out of denial" were recruited for study. Subjects were identified by a focus group of judges, attorneys, Probation and Parole officials, representatives from the District Attorney's office and psychologists who were invited to a seminar regarding the Lupron Project. After potential subjects were identified, the Principal Investigator screened case histories to determine eligibility for study inclusion (Table 1). These potential subjects were contacted by the Principal Investigator. They were interviewed, informed of the study, and invited to participate. Three additional subjects were screened, but excluded. These included one psychotic adult, one adolescent, and one elderly, nonerectile adult. The study was conducted during a two-year period in a medical research center in north western Pennsylvania. The study was reviewed and approved by the local Institutional Review Board. All subjects provided informed consent. Participation was voluntary and not a condition of probation or parole. The subjects were informed that they would receive either a drug or placebo without an indication of duration of dose, the sequence, nor if a sequence would occur. The subjects did not know one another. They lived in diverse geographical locations and their only contact was during group cognitive-behavioral psychotherapy.

IQ data were available for four of the five participants, all of whom had normal IQ's. The fifth participant, on whom IQ data were unavailable, was not considered to be intellectually impaired. Four of the five participants reported having been sexually abused during their childhood. Each subject was more than 16 years old and at least five years older than the victim when offending was initiated (Table II). At baseline, the self-reported numbers of victims for the five subjects totaled 329.

Four of five had served lengthy prison sentences (M, 7.8 years) for conviction of sexual crimes. All subjects had previous treatment in penitentiary or community-based sexual offender programs. Two had a history of major depression and two had a prior diagnosis of alcoholism. Two subjects reported bisexual pedophilic preference; two exclusively heterosexual preference; and one an exclusively homosexual preference. At baseline and throughout the study, all five subjects met the DSM-IV criteria for pedophilia, as well as antisocial personality disorder (American Psychiatric Association, 1994; Berger, Berner, Bolterauer, Gutierrez, & Berger, 1999).

Study Design Five male pedophiles entered a prospective, repeated measures, nonrandomized, masked study. All subjects and investigators, except the Principal Investigator (PI), Project Director, and Study Coordinator, were masked. The PI administered the drug or placebo; the PI and Project Director designed the study and were cognizant of the dosing schedule; and the Coordinator was responsible for supply of study drug. However, the PI, Project Director, and Coordinator were not involved in direct data collection or evaluation of the subjects. Each subject served as his own control. Subjects received weekly cognitive-behavioral psychotherapy for 24 months and LA depot injection, 7.5 mg at baseline and 22.5 mg at months 1, 4, 7, and 10. Flutamide (250 mg TID for 14 days) was only administered with the first injection.

Flutamide is an acetanilid, nonsteroidal, orally active anti-androgen that exerts its anti-androgenic action by inhibiting androgen uptake and/or by inhibiting nuclear binding of androgen in target tissues or both. Flutamide has been demonstrated to interfere with testosterone at the cellular level. This can complement medical castration achieved with LHRH agonists that suppress testicular androgen production by inhibiting luteinizing hormone secretion (Brogden, 1989).

After the initial injection, subjects were sequestered for one week for community protection. After one year of therapy, LA was discontinued and saline placebo substituted. At baseline, one month after initiation of LA therapy, and then at 3-month intervals, subjects underwent Abel Assessment, Monarch Penile Plethysmography (PPG), polygraph, physical examinations, laboratory studies, radiology studies, and psychological testing.

Abel Assessment (Abel Screening, Inc., Atlanta, Georgia) Abel Assessment, a combination of self-report and visual reaction time (VRT), was utilized according to manufacturer's instructions (Abel Screening, Inc., Atlanta, Georgia). Assessments were conducted by a trained clinical staff member (Abel, Huffman, Warberg, & Holland, 1998; Letourneau, 2002). A questionnaire was also administered. The combined questionnaire and VRT results included evaluation for 21 sexual interests, including paraphilias.

To assess VRT (the amount of time a subject viewed a picture), a slide projector was used to present still images of models onto a 12-inch computer screen. Models were clothed, generally in swimsuits, and none were in sexual positions. Twenty-one stimulus categories, representing 21 potential sexual interests with approximately seven slides in each category, were viewed by the subject. Slides depicted Caucasian and African American males and females comprising preschool and school-age children, adolescents, and adults.

An examiner met individually with each subject and instructed him on advancing and rating the slides. The subject viewed the slides in a dimly lit, quiet cubicle. After a practice session, the subject completed the VRT in private. After viewing each slide, participants rated each slide on a scale from 1 (extremely sexually repulsive) to 7 (extremely sexually interesting). A computer recorded the amount of time a subject viewed and rated each slide. Data from the second viewing time were utilized by Abel Screening, Inc. to compute the z-scores. These were provided to the investigator in graph form, with a score for self-reported interest. Per manufacturer's instructions, the Abel "rule of thirds" was utilized to define the predominant sexual preference. Because of the large number of sexual preference choices generated by the Abel method, a modification was adopted. For this study, predominant sexual preference was defined as a relative visual reaction time z-score  1 SD. The number of times each stimulus was  1 SD was recorded. If only one or none of the SDs were  1, the two highest scores were recorded. This modification resulted in an objective measure of only those selections that stimulated the longest VRT (Table III).

To illustrate the difference in the modified Abel and the Abel "Methods of Thirds", subject # 1 self- reported an interest in adolescent and adult males. His known victims were boys aged 12 to 17. Using the Abel "Method of Thirds" per manufacturer instructions, the interest preferences included young and adolescent females, and young, adolescent, and adult males. Using the modified Abel, the selection was truncated to adolescent and adult males, the exact interest preferences self-reported by the patient (Table III).

Z-scores for each stimulus set were reported at baseline, on LA, and off LA, respectively. Assessments were compared before, during, and after treatment to determine if VRT associated with pedophilic interest preferences changed with treatment.

Given that lowering testosterone to castrate levels would affect penile responsiveness and potentially affect the Monarch PPG monitoring, the Abel Assessment was incorporated into the study with the anticipation that pedophilic sexual preference could still be potentially detected by visual reaction time.

Monarch PPG (Behavioral Technology, Inc., Salt Lake City, Utah) The Monarch Adult Projective Audio Visual Set, Version 5 (Salt Lake City, Utah) was a commercially available set of standardized stimulus materials that was used to assess all subjects. The stimuli contained audio and visual depictions of males and females comprising preschool and grammar school children, adolescents, and adults. The audio scripts were based on a projective, rather than explicit, portrayal of sexual activity. Unlike most other PPG visual stimuli, the set contained clothed pictures of children (all child models were shown in bathing suits or underwear). No lewd or provocative poses were presented (Byrne, 2000).

During this procedure, the subject was seated in a reclining chair located in a private room equipped with an intercom. The audiovisual stimulus materials were presented via headphones on a 14-inch television monitor. Stimulus presentation and collection of penile, galvanic skin response (GSR), and respiration data were recorded and coordinated by a microprocessor-based Monarch Data Recording Device (DRD) (Behavioral Technology, Inc., Salt Lake City, Utah).

Penile circumference changes were measured using an indium gallium gauge supplied by Behavioral Technology, Inc. The signal from the indium gallium gauge was amplified by 1000; data retrieval was set at a rate of 10 samples per second. Each penile gauge was calibrated to a full range of 3.0 cm before each assessment to ensure that the correlation between circumference and deflection was linear. GSR electrodes and respiration belts were also used for deterrence and potential detection of faking or suppression attempts.

Data from the assessment were compiled and analyzed using Monarch Adult Male Software Version 3.22f. This software guided the collection, management, and initial analysis of the PPG data. The software analyzed the PPG data and provided maximum scores and area- under-the-curve scores for each of the 22 segments of the test. A maximum penile tumescence score was calculated for each subject for each category of stimuli. For the purposes of this study, maximum penile tumescence was defined as any response above 20 points (.60 cm), which is 20% of the calibration of the gauge. Murphy and Barbaree (1994) argued that a minimum level for significance was 20%. To meet this requirement, a response was considered to be significant if it was at least 20/100-scaled units from the beginning of the stimulus detumesced level. Kuban, Barbaree, and Blanchard (1999) also suggested that this criterion was conservative. Subject's PPG calibration and projected full-scale responding were set to 3.0 cm. Thus, the cutoff of 20 points for significant responding falls within the 20% minimum criterion for significant responding. The subject's PPG calibration and projected full-scale responding were set to 3.0 cm circumference. A repeated measures ANOVA was performed on the PPG data to assess significant changes in arousal level throughout the study ( = .05). To assess sexual preferences, raw data for each subject were transformed to z-scores. The z-scores for each stimulus were recorded at baseline, on LA, and off LA. All the responses  1 SD, or the two highest responses, were recorded. Finally, a subject's classification as pedophilic/nonpedophilic was determined by using a deviance differential (Harris, Rice, Quinsey, Chaplin, & Earls, 1992). This was calculated using z-score data and taking the highest response to adults (male and female segments) minus the highest response to infants and children (male or female preschool or grammar school age). The adolescent segment was excluded from analysis. For the purposes of data analysis, deviance differential scores greater than 0 were classified as nonpedophilic and scores less than 0 as pedophilic. A deviance differential of O would mean that the subject was just as aroused to children as to adults.

All subjects were assessed at baseline. One subject #1 did not respond to a level of clinical significance (20 points). Further inspection revealed that this client did not obtain a significant response at any point during the two-year study. He was removed from the analysis of the PPG data (repeated measures ANOVA) because low level data generated from a circumferential PPG transducer were not considered reliable (Kuban et al., 1999; Barbaree & Mewhort, 1994).

Polygraph

An Axciton Computerized Polygraph System (Axciton Systems, Houston, Texas) and the Lafayette Instrument LX-4000 (Lafayette Instrument Co., Lafayette, Indiana) were utilized. The examiner was certified by the Colorado Sex Offender Management Board. For purposes of score reliability and validity confirmation, the results were both scored by hand and scored using the Johns Hopkins Applied Physics Laboratory Polyscore computer scoring algorithm (Heil, Ahlmeyer, Simons, & English, 1999). Hand scoring is recognized by the American Polygraph Association as the most accurate and reliable method of scoring. It was the only method used to evaluate the data. The algorithm was used as a quality control check of the hand scoring method. Subjects provided a separate informed consent before the first polygraph examination. The test was administered in a private office. The examination took an average of two hours. Subjects were tested every three months. Polygraph was standardized by an American Polygraph Association approved test procedure and was administered by a single experienced examiner, who was blinded to the dosing protocol. The polygraph examination process contained pretest and posttest components. The examiner worked closely with the psychotherapist in developing appropriate polygraph questions for confirmation of the disclosed number of victims and offenses, discovery of new offenses, and or compliance with study conditions. This study examined past victim and offense admissions data, as well as responses evaluating 30 paraphilias (Table IV). The pretest established which questions the examiner would ask based on the subject's self report of new admissions of past offending or the absence of accountability statements for known offending behaviors (identified in prison histories). The subject's cardiovascular, respiratory, and galvanic skin resistance were assessed in response to three specific key questions:

1. In the past three months, have you masturbated to sexual thoughts about anyone under the age of 18?
2. In the past three months, have you purposely withheld important sexual information from your therapist?
3. In the past three months, have you had strong urges to initiate sexual contact with anyone under 18 years old? No consequences were imposed for deceptive responses. The therapist and polygrapher worked closely and debriefed the subject after each assessment to identify areas for counseling, behaviors to be monitored, and issues detected on polygraph.

Cognitive-behavioral psychotherapy The psychotherapy program comprised cognitive and behavioral therapy. The weekly sessions were co-facilitated by two therapists. After establishing the rules for group therapy, psychotherapists guided participants through a basic group curriculum that included breaking down defense mechanisms, identifying the typology of sex offenders, and discussing the effects of sexual abuse on victims. The counseling component involved insight-oriented cognitive-behavioral psychotherapy. This provided a dynamic understanding of unconscious conflicts and interpersonal issues as they related to each participant's offense cycle. Relapse prevention plans were constructed with and for each participant. This self-controlled program was designed to teach offenders how to recognize and intervene when they were behaving in a manner that would lead to a re-offense. Relapse prevention enabled the individual to maintain the newly adopted behavior pattern, such as refraining from deviant thoughts and actions. The relapse prevention plans described red flags, offense cycles, interventions, and changes in thoughts, feelings, attitudes, and behaviors.

Psychopathology The Hare Psychopathy Checklist-Revised (PCL-R) was utilized to assess subject's psychopathology (Hare, 1998). The Minnesota Sex Offender Screening Tool-Revised (MnSOST-R) identified the most violent offenders and those offenders likely to re-offend (Epperson et al., 1999). The Static-99, another sexual offender risk assessment, measured long-term risk potential for child molesters (Hanson & Thornton, 1999). The Yale-Brown Obsessive Compulsive Scale (Y-BOCS) was utilized because of an inferred relationship between obsessive compulsive disorder and sexual paraphilias (Goodman et al., 1989).

Inclusion Criteria:pedohilia admission,healthy,male,adult,ability to have penile response on plethysmography -

Exclusion Criteria:Denial of pedophilia, seizure disorder treated with dilantin or barbiturates, alcoholism, IQ less than 70

-