Hypersexuality: fluoxetine

Drug action

Fluoxetine is a selective serotonin reuptake inhibitor (SSRI) antidepressant (summary of product characteristics [SPC] for example: Prozac).

Regulatory status

In adults, fluoxetine is licensed for the treatment of major depressive episodes and obsessive compulsive disorder. It is also licensed for use in the treatment of bulimia nervosa (as a complement to psychotherapy) for the reduction of binge‑eating and purging activity (SPC: Prozac).

Fluoxetine is not licensed for the treatment of hypersexuality; therefore use for this indication is off‑label.

In line with the guidance from the General Medical Council (GMC), it is the responsibility of the prescriber to determine the clinical need of the patient and the suitability of using fluoxetine outside its authorised indications.

Cost

According to the Drug Tariff (prices excluding VAT, June 2015):

Clinical effectiveness

Winder et al. 2014

This observational study included 64 men convicted of sexual offences (average age 43 years) from a UK prison and who had been referred for medication to reduce hypersexuality, sexual compulsion or sexual preoccupation between 2010 and 2012. Criteria for referral for medication included evidence of 1 or more of: hyper‑arousal, intrusive sexual fantasies or urges, subjective reports of experiencing urges that are difficult to control, sexual sadism or other dangerous paraphilia's such as necrophilia or highly repetitive paraphilic offending such as voyeurism or exhibitionism. Over 90% of the men had committed sexual offences against children with an average of 5 previous contact offences and 2 previous non‑contact offences per offender. Approximately half (46%) of the group had an intelligence quotient (IQ) less than 80 (which is below the usual average IQ test score, for example the Wechsler Abbreviated scale of intelligence which was the scale used in the study).

Of the 64 men referred for treatment, 36 received fluoxetine, 5 received anti‑androgen medication (cyproterone), 7 received a combination of fluoxetine and cyproterone, 1 received triptorelin and 15 did not receive any medication or were still under assessment for medication. It was reported that when fluoxetine did not appear to be working, cyproterone was prescribed. However it is unclear how 'did not appear to be working' was defined. Fluoxetine was usually started at a dose of 20 mg daily with an increase to 40−60 mg daily as necessary. Participation in the study was voluntary.

Participants in the study met with a psychiatrist on a regular basis to assess measures of sexual compulsivity, hypersexuality and sexual preoccupation. Data measures were reported at baseline, 1 month after participants started medication (except the sexual compulsivity score) and after 3 to 4 months' of treatment. Results were reported for the group treated with fluoxetine and the group treated with cyproterone. It was reported that participants taking cyproterone in addition to fluoxetine were included in the cyproterone group. Statistical analyses tested for associations between the outcome measures and time or medication group.

Sexual compulsivity was measured using a 10‑item scale (validated in homosexual men with HIV who reported engaging in 'risky' sexual practices). Participants rated the extent to which they agreed with each statement in the scale for example, 'my sexual thoughts and behaviours are causing problems in my life' from 1 (not like me at all) to 4 (very much like me). The mean sexual compulsivity score for each item was reduced from 2.68 at baseline to 1.20 in the fluoxetine group after 3 months' treatment. In the cyproterone group it was reduced from 3.08 at baseline to 1.15 after 3 months' treatment. For the treatment groups combined, there was a statistically significant improvement after 3 months' treatment for the mean sexual compulsivity score (p=0.003).

Hypersexuality was assessed by asking participants how many days in the last week they had masturbated to orgasm. After 3 months' treatment, there was a reduction from 3.67 days at baseline to 1.10 days in the fluoxetine group and from 3.97 days at baseline to 1.70 days in the cyproterone group. For the treatment groups combined, there was a statistically significant improvement after 3 months' treatment for the measurement of hypersexuality (p=0.001). There was no statistically significant effect of medication type (fluoxetine or cyproterone) on the hypersexuality outcome (no p value provided).

To assess sexual preoccupation participants were asked 3 questions and asked to rate each question on a 7‑point scale (it is unclear if the scale was validated). Questions 1 and 2 were rated from 1 (low) to 7 (high) and question 3 was rated from 1 (easy) to 7 (difficult):

Kafka and Prentky 1992

This 12‑week US observational study included 20 men (10 with paraphilic and 10 with non‑paraphilic sexual addictions; modal age 36 years) who were treated with fluoxetine. Participants were self‑referred by responding to a newspaper advertisement inviting people to participate in a study evaluating the treatment of sexual addictions and compulsions. Men reporting distress accompanied by social, psychological, or work impairment, unwanted medical risks, legal consequences or significant associated financial cost, were eligible for study inclusion. Nineteen of the men reported chronic depressive symptoms lasting more than 2 years.

After baseline assessment, fluoxetine was initiated at a dose of 20 mg daily with dose titration at 4‑week intervals to a maximum of 60 mg daily if clinically indicated. The mean dose of fluoxetine at week 12 was 39 mg. Ten of the men in the study also received psychological therapies. Sexual behaviour was assessed at 4‑week intervals using a self‑assessment tool to document the incidence of sexual fantasies, urges and activities during a designated week. The self‑assessment tool used in this study was previously unpublished and unvalidated. Total sexual outlet (defined as the total number of sexual activities in the measured week that either led to orgasm or would have done so had voluntary restraint not been utilised) was calculated using data obtained using the self‑assessment tool.

Sixteen men completed the study. A statistically significant reduction in mean total sexual outlet was seen from 10.6 at baseline to 3.6 at week 12 (p<0.001).

Greenberg et al. 1996

This 12‑week retrospective observational study, using data taken from medical charts and records, included 58 men (mean age 36 years) who were being treated for paraphilias (74% paedophilia) with 1 of 3 SSRIs (sertraline [n=25], fluoxetine [n=17] or fluvoxamine [n=16]) at a sexual behaviours clinic in Canada between 1991 and 1995. Seventy‑nine percent of the participants were also receiving some receiving some form of psychosocial intervention. It was reported that 31% of participants had a personality disorder, 28% had depression, 17% had alcohol dependence and 16% had adjustment disorders. Outcomes measured included severity of illness and global improvement (using a 7‑point Clinical Global Impression Scale from 1 'very much improved' to 7 'very much worse') and frequency and severity of sexual fantasies. Severity of sexual fantasies was rated on a 5‑point scale from 0 (no fantasies) to 4 (extreme, also constant, very disturbing, unable to stop).

Results were presented for the whole group together rather than for each individual SSRI. Between weeks 4 and 12, nearly 30% of the participants dropped out of the study and no further data was collected for them. Data until week 12 is available for 41 participants. For the global improvement scale there were statistically significant improvements from baseline to week 4 (p<0.001), from week 4 to week 8 (p<0.001) and from week 8 to week 12 (p<0.05). The global improvement scale score was approximately 4 (no change) at baseline, approximately 3 (minimally improved) at week 4 and approximately 2 (much improved) at week 12. There were statistically significant decreases in the fantasy severity score from baseline to week 4 (p<0.001), from week 4 to week 8 (p<0.05) but not from week 8 to week 12 (no p value reported). The severity of fantasy score for the group was approximately 2 (moderate, frequent, disturbing, less able to stop) at baseline, approximately 1 (mild, occasional, not too disturbing, easily able to stop) at week 4 and approximately 0.75 at week 8. It was reported that there were no significant differences between the 3 SSRI groups for fantasy severity or global improvement scale scores but no statistical analysis is provided.

Safety and tolerability

Neither Winder et al. 2014 nor Kafka and Prentky 1992 reported any safety data. In Greenberg et al. 1996 adverse effects were reported for all 3 SSRIs (fluoxetine, sertraline and fluvoxamine) as a group rather than individually, although it was reported that no differences were seen between them. Forty percent of participants reported adverse effects at week 4, 30% at week 8 and 23% at week 12. The most common adverse effects included insomnia (n=8), delayed ejaculation (n=8), headaches (n=7), drowsiness (n=6), reduced sex drive (n=4), diarrhoea (n=3), nausea (n=3) and blurred vision (n=2). Seventeen people withdrew from the study, 3 because of adverse effects.

The summary of product characteristics (SPC) for fluoxetine (for example: Prozac) reports the following very common (1 in 10 or more) adverse reactions: fatigue, nausea, diarrhoea, headache and insomnia. Common (between 1 in 10 and 1 in 100) adverse reactions are reported to include: decreased appetite and weight loss, anxiety, nervousness, restlessness, sleep disorders and abnormal dreams, disturbances in attention, dizziness, lethargy, tremor, blurred vision, palpitations, vomiting, dry mouth, dyspepsia, flushing, rash and urticaria, arthralgia, frequent urination and chills.

Fluoxetine is contra‑indicated in people taking monoamine oxidase inhibitors and has a number of special warnings and precautions for use, which are discussed in the SPC. See the summaries of product characteristics for licensed fluoxetine products for further information on contraindications, potential interactions and adverse effects of fluoxetine.

Evidence strengths and limitations

No RCTs which evaluate the use of fluoxetine in the treatment of hypersexuality were found during literature searches conducted for the production of this evidence summary.

This evidence summary is based on 3 small observational studies (2 open label and 1 retrospective; n=64, 58 and 20 respectively) which provide limited information on the use of fluoxetine for hypersexuality. None of the studies included a control group and none were randomised or blinded, and are therefore subject to bias and confounding. Observational studies are prone to confounding, which limits the conclusions that can be drawn. Unlike in the setting of a randomised controlled trial, in 'real life', treatment plans are chosen, changed, or actively not chosen in the light of individual peoples' risk factors, preferences and tolerability or responses to other drugs. Thus observed differences in outcomes may be due to differences among the participants, not only the different treatments.

All of the studies were short‑term (around 3 to 4 months) and, therefore, provide no evidence for using fluoxetine to treat hypersexuality in the longer term. All of the participants in the studies were male and had an average age of between 36 and 43 years. Because of this, they provide little information on the use of fluoxetine to treat hypersexuality in a younger or older population, and no information on use in women. The majority of men in the studies had committed a sexual offence or had a paraphilia (approximately 93% [132/142]). Only 10 participants in Kafka and Prentky 1992 had a non‑paraphilic sexual addiction (3 of whom dropped out of the study), providing little evidence for using fluoxetine to treat this group of people.

In Winder et al. 2014 the average IQ of the population was lower than the average IQ of the general population, which may have influenced the relevance of the self‑reported outcomes. The other 2 studies did not report IQ scores but participants in Greenberg et al. 1996 had a mean duration in education of 12.2 years and in Kafka and Prentky 1992 the modal participant had achieved a college or postgraduate degree. The majority of men in Kafka and Prentky 1992 reported concomitant chronic depressive symptoms lasting more than 2 years. In Greenberg et al. 1996, 31% of participants had a personality disorder, 28% had depression and 16% had adjustment disorders. Winder et al. 2014 did not report on concomitant psychiatric illness. Kafka and Prentky 1992 assessed the effect of treatment on concomitant depressive symptoms; however the other 2 studies did not assess the effect of treatment on any concomitant mental health disorder.

Winder et al. 2014 included 64 men who had committed a sexual offence and were housed in a prison. Over 90% of the men had committed sexual offences against children with an average of 5 previous contact offences and 2 previous non‑contact offences per offender. The results of the study may not be applicable to other populations with hypersexuality. Outcome measures were assessed by asking the participants questions relating to sexual compulsivity, hypersexuality and sexual preoccupation. The study authors highlight the use of self‑reported data as a limitation of the study and raise the question of its validity and reliability. Specialists who commentated on this evidence summary highlighted that more objective measures of hypersexuality and sexual preoccupation may have been more appropriate for this group of people. The reliability of using self‑reported data where there may be vested interests in reporting positives responses to treatment in terms of progressing with sentencing is unclear. In addition, the average IQ of the population was less than 80, which is lower than the usual average IQ score; this may have also had an impact on the validity of self‑reported data. The sexual compulsivity scale used in this study had been validated in homosexual men who were HIV positive who had reported engaging in risky sexual practices, it is unclear if the scale is valid to use for the population included in this study.

The study was conducted over 12 weeks while the participants were still in prison and it provides no information on sexual re‑offending rates such as re‑convictions, self‑reports or cautions. This study was a preliminary study and further data on correctional infractions is currently being collected. Participants in Winder et al. (2014) were treated with fluoxetine, cyproterone or a combination of the 2 treatments. Participants included in the cyproterone group for analysis had either been previously treated with fluoxetine or were currently taking fluoxetine in combination with cyproterone. The study comments that some participants stopped taking their medication but it provides no further information on this such as details on which medication they were taking or if outcomes from these participants were still included.

Kafka and Prentky 1992 included 20 men with a paraphilic or non‑paraphilic sexual addiction. However, four men dropped out of the study and 12‑week data is only provided for the 16 men who completed the study. Sexual behaviour was assessed using a self-assessment tool which documented the incidence of sexual fantasies, urges and activities during a designated week. This tool had not been previously validated and was unpublished.

Greenberg et al. 1996 was a retrospective observational study in 58 men with paraphilias treated with fluoxetine, sertraline or fluvoxamine. However, nearly 30% of the participants dropped out of the study and data until week 12 is only available for 41 participants. Outcomes were measured using a clinical global impression scale and a self‑assessment scale to rate sexual fantasies. As with the other studies the use of self‑reported data raises questions on its validity and reliability. In this study, the results are presented for the whole group together rather than for each individual SSRI and therefore it provides limited evidence on the use of fluoxetine. It was reported that there were no significant differences between the 3 SSRI groups for the outcomes but no statistical analysis is provided.

The clinical significance of the outcomes in all of these 3 studies is unclear. Assessment of outcomes used self‑reported data or un‑validated scales and no minimum clinically important differences are provided. Overall, these 3 studies provided very limited data on the use of fluoxetine to treat hypersexuality.

Cost effectiveness

Cyproterone acetate (Androcur) is licensed for the control of libido in severe hypersexuality or sexual deviation in adult men. Triptorelin (Salvacyl) is licensed for the reversible reduction of testosterone to castrate levels in order to decrease sexual drive in adult men with severe sexual deviations.

Current drug usage

No estimate of the current use of off‑label fluoxetine for the treatment of people who have sexually offended or are at risk of offending in UK clinical practice was identified.

Expert advisers

Dr Bradley Hillier, Locum Consultant Psychiatrist, South West London Mental Health Trust. Treasurer, International Association for Forensic Psychotherapy.

Dr Pamela Walters, Consultant in Forensic and addiction Psychiatry Shaftesbury clinic medium secure unit, South West London Mental Health Trust and HMP Wandsworth.

Declarations of interest

No relevant interests declared.

Copyright

© National Institute for Health and Care Excellence, 2015. All rights reserved. NICE copyright material can be downloaded for private research and study, and may be reproduced for educational and not‑for‑profit purposes. No reproduction by or for commercial organisations, or for commercial purposes, is allowed without the written permission of NICE.

ISBN: 978-1-4731-1299-5