Tolerability and efficacy of paroxetine and naltrexone for treatment of compulsive sexual behaviour disorder

Compulsive sexual behaviour disorder (CSBD) has recently been introduced in the ICD‐11. However, despite increasing research on its psychological and neural mechanisms, little is known about the efficacy of pharmacotherapy in people with this condition ¹ .

To date, only some case reports and one small (28 males) randomized controlled trial (RCT) have provided some evidence for the efficacy of selective serotonin reuptake inhibitors (SSRIs) in the reduction of sexual compulsivity ² . Several case studies and one small (20 males) open‐label study reported the clinical usefulness of the opioid antagonist naltrexone in CSBD ³ . Most studies were conducted before CSBD diagnostic guidelines were proposed in 2019.

We aimed to assess the safety and efficacy of an SSRI (paroxetine) and of naltrexone in male patients seeking treatment at an outpatient sexology clinic who met the ICD‐11 diagnostic guidelines for CSBD. For this purpose, we conducted a 20‐week double‐blind and placebo‐controlled RCT, approved by the local ethics review board in accordance with the Declaration of Helsinki.

Among the 73 recruited heterosexual cisgender men (mean age: 35.7±8.1 years), 24 were randomly assigned to paroxetine (20 mg/day), 24 to naltrexone (50 mg/day), and 25 to the placebo condition. No significant group differences were observed with respect to CSBD symptoms or demographic characteristics prior to treatment.

Results from the trial confirmed that paroxetine and naltrexone represent safe treatment options for CSBD. The total discontinuation rate was 15.1%, with the following causes for stopping medication: adverse effects (five patients, 6.8%: two with paroxetine, three with naltrexone); lack of improvement or worsening of CSBD symptoms (two patients, 2.7%, both with placebo); irregular medication intake (one patient, paroxetine group). Three patients (4.1%) discontinued/failed to show up at follow‐up (two in paroxetine and one in naltrexone group). No difference in treatment non‐adherence was noted between groups (F_2,57=0.25, p=0.78).

The most bothersome and persistent side effects included sedation (29.2% with paroxetine, 37.5% with naltrexone, and 0% with placebo), apathy (8.3%, 8.3% and 0%, respectively), orgasmic dysfunction (2.8%, 0% and 0%, respectively), erectile dysfunction (12.5%, 0% and 8%, respectively), and weight gain (16.7%, 4.2% and 12%, respectively). No medication‐related serious side effects occurred during the trial.

We observed a significant effect of time on severity of CSBD symptoms using self‐report questionnaires: Hypersexual Behavior Inventory (F_1,55=83.59, p<0.001, η ²=0.60), Brief Pornography Screen (F _1,47=34.66, p<0.001, η ²=0.42) and Sexual Addiction Screening Test (F_1,47=17.06, p<0.001, η ²=0.27). However, there was no difference between the conditions at any time point, nor an interaction of time and condition. Self‐reported frequency of pornography consumption (F_1,57=28.69, p<0.001, η ²=0.34) and duration of pornography consumption (F_1,52=7.863, p<0.01, η ²=0.13) decreased over the time of treatment across all conditions. No condition or interaction (time x condition) effects were noted.

On the other hand, clinical interviews revealed that patients treated with paroxetine or naltrexone, compared to placebo, were more likely to achieve at least 30 days of cessation of any compulsive sexual behaviour at treatment week 8 (X²=7.097, p=0.029, Cramer's V=0.34); to have a reduced frequency of sexual binges at week 20 (X²=6.935, p=0.031, Cramer's V=0.34); and to have a decrease in frequency of CSBD symptoms at both time points (week 8: X²=12.250, p=0.016, Cramer's V=0.31; week 20: x²=8.208, p=0.017, Cramer's V=0.37). They also reported higher satisfaction with treatment effects at both time points (week 8: X²=15.801, p=0.003, Cramer's V=0.35; week 20: X²=1.886, p=0.018, Cramer's V=0.31).

Using smartphone‐administered daily ecological momentary assessment (EMA), we observed a significant interaction (time x condition) effect in craving for sexual activity (F_6,1011.57=3.12, p=0.005). Patients receiving paroxetine reported significantly less craving for sexual encounters in the last week of treatment (estimated marginal means, EMMs=3.71, SE=0.55) compared to baseline (EMMs=4.88, SE=0.48) (c=1.17, lower control limit, LCL=0.07, upper control limit, UCL=2.27, p=0.03). A significant interaction (time x condition) effect was also found in craving for pornography viewing (F_6,1020.12=2.54, p=0.002). Craving for pornography in the 20th week of treatment with paroxetine (EMMs=2.69, SE=0.48) was significantly lower compared to baseline (EMMs=3.97, SE=0.39) (c=1.28, LCL=0.07, UCL=2.49, p=0.03).

To summarize, our double‐blind placebo‐controlled RCT demonstrated that paroxetine and naltrexone are safe and well‐tolerated by men with CSBD. Patients usually reported mild and transient side effects with either medication, and most complaints were similar to reports on safety and tolerability profiles of paroxetine and naltrexone in their registered indications, except for a high incidence of sedation reported by naltrexone users. A 6.8% discontinuation rate due to adverse effects is relatively low compared to other studies4, 5.

Based on clinical interviews, both medications were found to be more effective than placebo in reducing CSBD symptoms. Such a superiority of both active treatment arms over placebo was visible at the 20th week, but as early as the 8th week. EMA provided support for higher effectiveness in reducing craving for sexual encounters and pornography viewing in the paroxetine condition. However, based on data from self‐report questionnaires and self‐reported pornography consumption, the superiority of paroxetine and naltrexone over placebo did not reach statistical significance. Therefore, the clinical efficacy of these drugs in CSBD should be confirmed by further studies.

The high effectiveness of placebo in CSBD may be related to such factors as disclosing the problem, motivation for change, and initiation of therapy while receiving external support from the study team. Prior research ⁶ has also demonstrated high placebo response rates in gambling disorder treatment. Such results warrant further attention to non‐specific factors related to therapy as meaningful for clinical improvement in CSBD.