Treatment of Inappropriate Sexual Behavior in Persons With Dementia: A Systematic Review

ABSTRACT

Background

Inappropriate sexual behavior (ISB), such as sexual touching and sexual exposure, occurs in up to 25% of people with dementia. This systematic review examines the effectiveness of pharmacologic and non‐pharmacologic interventions to manage ISB in persons with dementia.

Methods

Systematic review without meta‐analysis (PROSPERO: CRD42023469625). We searched MEDLINE, APA PsycInfo, Embase, JBI EBP Database, CENTRAL, CDSR, and Ageline databases, with no limits on study date or language, from inception until September 8, 2023. All study designs were eligible for inclusion if they examined the effectiveness of any pharmacologic or non‐pharmacologic intervention in adults with dementia and ISB. Two reviewers independently completed all study screening, data abstraction, and risk of bias assessments. The JBI Critical Appraisal Checklist was used to determine the quality of case studies and case series, and the Cochrane RoB 2 was used to appraise the one randomized controlled trial. Findings were synthesized using vote counting based on the direction of effect.

Results

We included 74 studies, of which 60 were case studies, 13 were case series, and one was a randomized trial. Most studies (64%) reported exclusively pharmacologic interventions. Non‐pharmacologic interventions (e.g., distraction, environmental modification) were associated with improvement or resolution of ISB in 33 (72%) instances; however, only five (21%) cases improved or resolved without co‐prescribed pharmacotherapy. Among men, hormonal treatments, including progestins and anti‐androgens, led to a reduction in ISBs more frequently than antipsychotics, antidepressants, or anticonvulsants.

Conclusions

Nonpharmacologic interventions can be effective at reducing ISB, though pharmacologic interventions are also frequently needed. Randomized trials of nonpharmacologic and pharmacologic intervention effectiveness and safety are needed to guide practice.

Summary.

• Key points

  • ○
    Non‐pharmacologic interventions like distraction and caregiver education can be effective in managing inappropriate sexual behavior in people with dementia.
  • ○
    Except for one small randomized controlled trial (n = 15) of conjugated estrogens used in nursing home residents, the evidence for pharmacotherapy to treat inappropriate sexual behavior in people with dementia is limited to case reports or case series.
  • ○
    Some hormonal therapies may be effective for males in situations where inappropriate sexual behaviors are refractory to other interventions.
• Why does this paper matter?

  • ○
    Up to one in four people with dementia will display inappropriate sexual behavior, which can be distressing to patients and their caregivers.
  • ○
    This systematic review provides a comprehensive synthesis of the evidence informing clinicians' management strategies for inappropriate sexual behavior.
  • ○
    It also underscores the urgent need for more robust research in this field to better understand and manage these behaviors.

1. Introduction

Inappropriate sexual behavior (ISB) occurs in 1.8%–25% of people with dementia [1, 2, 3]. Definitions and descriptions of ISB (e.g., sexual disinhibition, hypersexual behavior) vary widely within the literature [4]. We operationalized ISB as “sexual expressions of potential risk” in a recent clinical practice guideline and defined it as “disruptive verbal or physical acts of an explicit or perceived sexual nature, which is either intrusive or engaged in without the consent of those around the person living with dementia” [5]. We used this description in the clinical practice guideline to focus recommendations on sexual behaviors that may pose a risk to either the person or those around them, therefore warranting treatment. We adopt the term “inappropriate sexual behavior” in this manuscript to convey the full range of sexual behaviors beyond those that may be of “potential risk.” However, the description of the behaviors being referred to remains unchanged.

ISB frequently occurs in people with vascular dementia [2], behavioral‐variant frontotemporal dementia [6] or individuals with Parkinson's disease dementia who are treated with dopaminergic agents [7], but it can also be associated with other etiologies of dementia, such as Alzheimer's disease [2, 8]. These behaviors have diverse etiologies: dementia‐related factors include disinhibition from dementias affecting the frontal lobe; delusions; sensory impairments and misidentifications or misinterpretation of caregivers and care activities; substances and medications such as alcohol and dopaminergic agents; and environmental factors such as lack of privacy in communal care settings and absence of usual sexual partners [9]. ISB is distressing to caregivers and is associated with an increased risk of institutionalization [10]. As the prevalence of dementia increases in aging global populations [11], the prevalence of ISB and related consequences will also increase, necessitating immediate attention and robust research.

Despite the need for evidence informing pharmacologic and nonpharmacologic management of ISB in persons with dementia, data are sparse [5]. Developing recommendations on the assessment and management of ISB was prioritized during the Canadian Coalition for Seniors' Mental Health Behavioral and Psychiatric Symptoms of Dementia (BPSD) assessment and management clinical practice guideline development process [5]. When we conducted an overview of systematic reviews published from 2017 onward on the assessment and management of BPSD, we could not identify a systematic review on managing ISB [12], therefore, we conducted this systematic review to inform guideline recommendations [5].

The aim of this systematic review is to provide a comprehensive overview of the effectiveness of pharmacologic and non‐pharmacologic interventions for ISB in people with dementia. By synthesizing the existing evidence, we aim to guide clinical practice and inform future research in this important area.

2. Methods

The protocol for this systematic review was published on Open Science Framework and registered with PROSPERO (CRD42023469625) and followed the Preferred Reporting Items for a Systematic Review and Meta‐analysis and the Synthesis Without Meta‐analysis guidelines (PRISMA) [13, 14].

2.1. Data Sources and Searches

We searched MEDLINE, EMBASE, the Cochrane Library (of systematic review and controlled trials), PsycINFO, JBI EBP Database, and Ageline for citations published in any language from inception until September 6, 2023 (see Supporting Information for search strategies). The search strategy was peer‐reviewed using the PRESS checklist [15]. Backward citation searching was applied to reference lists of 39 articles identified during the abstract and title screen as having completed a non‐systematic literature review [16].

2.2. Study Selection and Inclusion Criteria

Our population of interest included people identified as having dementia and ISB by study authors. Studies were included if they reported the effectiveness of any pharmacologic or non‐pharmacologic intervention to any other intervention or standard of care. Studies of adults with dementia of any etiology (e.g., Alzheimer's, vascular, frontotemporal) were included. All settings were included (community, outpatient, hospital, or institutional care). Given the paucity of high‐quality evidence, our initial review protocol included all study types, including case studies [12, 17]. A pilot exercise was done at each stage of the review process to achieve 80% agreement between reviewers [18]. MA and NL independently screened all titles, abstracts, and full‐text articles to establish eligibility for inclusion [13]. A third independent reviewer resolved discrepancies regarding study inclusion [13].

2.3. Data Abstraction and Quality Assessment

Reviewer pairs (two of M.A., N.E.L., or S.H. for each article) independently extracted data from full‐text articles and published abstracts. Studies were appraised for quality with tools appropriate for the given study design. The JBI Critical Appraisal Checklist was used to determine the quality of case studies and case series [19], and the Cochrane RoB 2 was used to appraise the one randomized controlled trial [20]. The following study and participant characteristics were extracted: authorship, year of publication, study design, country of publication, language of publication, participant location of dwelling (community, institutional care, hospital), and setting of intervention (outpatient clinic, inpatient hospital setting, institutional care). We also extracted information on participants' age, race, sex, past occupation, level of education, type of dementia, and cognitive test scores. Types of ISB (inability to inhibit, oversharing, inappropriate comments, inappropriate exposure, overly flirtatious, inappropriate sexual touching [Table S1]) were extracted, as well as the number of different types of behaviors, how behaviors were described in the manuscript, whether they were precipitated by drug initiation and if so, what drug class. Types of ISB were based on categories defined by Chapman et al. in their 2019 review [4]. “Inability to inhibit” captured behaviors reflecting the inability to refrain from sexual activity (e.g., compulsive masturbation causing self‐injury), preoccupation with sex and insatiable sexual desire. In cases where people's behaviors met the criteria for “inability to inhibit,” which resulted in “inappropriate exposure” (e.g., exposing private body parts publicly) or “inappropriate comments,” (e.g., sharing sexually explicit fantasies), they were coded as expressing both. We added a new category—inappropriate sexual touching—to Chapman et al.'s five to capture grabbing or touching others' body parts in a sexual way, forceful or violent sexual acts. We felt this was a relevant distinction from the broader “inability to inhibit” category with which it was initially grouped, given our clinical experience that forceful or unwanted touching is often perceived as posing greater risks and more consistently associated with the use of physical and chemical restraint or change in place of care for people experiencing dementia (Table S1).

We also extracted the following details of interventions: whether they were pharmacologic, deprescribing, non‐pharmacologic, or combination interventions, and whether multiple interventions were initiated sequentially or simultaneously. We also collected data on the type of non‐pharmacologic intervention, including removing precipitating factors, distraction, diversion, or otherwise engaging, fulfilling the need for intimacy/connection in other ways, rear‐opening clothes, sensory/environmental stimulation, and patient or caregiver education. The following details of pharmacologic interventions were also extracted: drug class and specific drug, and dose. Outcome information was extracted, including duration of follow‐up, quantification of change in ISB, whether ISB was worsened, not changed, partially improved, or completely resolved after the intervention, and whether biomarker correlates of behavior change (i.e., testosterone levels) were recorded. Whether studies reported tolerance or side effects to medications, as well as any specific reported side effects, was also extracted for each study. Data abstraction and quality assessment discrepancies were discussed to reach a consensus within reviewer pairs (N.E.L. and M.A., and S.H.) or by a third reviewer (N.E.L.).

2.4. Data Synthesis

We assessed heterogeneity between studies by examining combinations of interventions, follow‐up times, care settings, and outcomes (Table S2) [14]. A meta‐analysis was not conducted due to substantial heterogeneity across these characteristics in the included studies [14]. Descriptive statistics summarized the proportion of females, treatment settings, frequency of different dementia diagnoses, and most frequently reported ISB. We synthesized findings using vote counting based on the direction of effect [21], the proportion of reported cases of ISB in which partial or complete resolution of symptoms was described, the intervention types most commonly associated with outcomes, outcomes by drug class, and the number of interventions trialed before symptom improvement. For analyses of outcomes by drug class and non‐pharmacologic intervention category, we coded instances of interventions as each time an intervention was trialed, such that multiple interventions could be trialed per person per study. If studies described symptoms as “resolving completely,” “never recurring,” or similar terms, we coded this as “complete resolution.” If studies described any improvement in ISB with treatment, we coded this as “partial resolution.” Data collection forms extracted data and data used for analysis are available from the author by request.

3. Results

3.1. Characteristics of Included Studies

A total of 2103 article titles and abstracts and 161 full‐text articles were identified. After removing 672 duplicated studies, 66 full‐text articles and eight published abstracts were included (Figure 1). Individual study characteristics are summarized in Table S2.

FIGURE 1.

PRISMA diagram.

Of the included studies, 60 were case studies, 13 were case series, and one was a randomized controlled trial of 15 participants. The majority of studies (64%) reported pharmacologic interventions exclusively (Figure S1). Included case series and studies had a mean JBI quality score of 3.5 (SD: 1.95), where eight represents the highest quality case studies and 10 represents the highest quality case series (Table S2); case series and case studies both had poor reporting of cases baseline history and timeline (omitted in 72% of included studies), description of post‐intervention clinical status (omitted in 80% of included studies), and adverse/unanticipated outcomes of intervention (omitted in 65% of included studies). The one included RCT was found to be at moderate risk of bias per the Cochrane RoB 2.

3.2. Participant Characteristics

Included studies reported outcomes among 152 persons with dementia, of whom 21 (14%) were female, as well as 32 caregivers of persons with dementia. No studies made any distinction between biological sex and gender. Among the 52 (68%) studies that reported the location of dwellings of participants and ISB at the time of intervention, 41% of studies examined people who lived in residential care versus 26% in community dwellings (Figure S2). Of the 99 (65%) participants whose type of dementia was reported, Alzheimer's dementia was the most common (65% of reported dementia etiologies), followed by vascular and frontotemporal dementia (Figure S3).

Most study participants had more than one type of ISB reported, with the most common reported ISBs being inability to inhibit (32% of 693 documented behaviors), inappropriate touching (26%), and inappropriate comments (22%) (Figure S4). Of 152 individuals included in the analysis, 79 (52%) were treated with multiple interventions trialed in combination or sequence; 37.9% of these patients achieved complete resolution of their symptoms, versus only 21.9% of patients who received a single intervention (Figure S5).

3.3. Non‐Pharmacologic Interventions

Of the 24 studies reporting on interventions that included non‐pharmacological components, only 5 reported exclusively non‐pharmacologic interventions; the remaining 19 reported on a combination of behavioral and pharmacologic interventions initiated simultaneously or in sequence (Table S2).

The most frequently used non‐pharmacologic interventions were patient or caregiver education and distraction, and diversion or otherwise engaging patients' hands; each was described in 10 instances (Figure 2) and associated with partial or complete resolution of ISB in 14 of 20 cases. Descriptions of non‐pharmacologic interventions are detailed in Table S3. We coded instances of interventions as each time an intervention was trialed, such that multiple interventions could be trialed per person per study. Of 33 instances of nonpharmacologic interventions, 21 (64%) were associated with partial improvement in ISB, and three (9.1%) were associated with complete resolution of ISB, but nine (27.3%) resulted in no change (Figure 2). Non‐pharmacologic approaches incorporating distraction, diversion, and otherwise engaging hands were associated with the highest frequency and fraction of instances (5 of 10) of complete resolution of behavior; the other 50% (5 of 10) of instances were associated with partial improvement of behavior. A case series of 32 caregivers to patients with Parkinson's disease dementia who received an educational intervention reported an overall improvement in tolerability of ISB from the caregiver perspective following the intervention [22].

FIGURE 2.

Types of non‐pharmacologic therapies used 33 times in 24 studies and their associated change in ISB. Non‐pharmacologic therapies were associated with partial or complete resolution of ISB 24 of 33 times (72.7%) they were employed across 24 studies.

Five of 24 instances of non‐pharmacologic interventions resulted in complete or partial resolution of ISB without pharmacotherapy co‐intervention (Table S4). No studies reported delirium as a side effect of non‐pharmacologic interventions.

3.4. Pharmacologic Interventions

The only included randomized controlled trial (n = 15) examined conjugated equine estrogens for ISB in two females and 13 males in residential care; it was at moderate risk of bias per the Cochrane Risk of Bias assessment and did not find a significant improvement in ISB among eight people treated with hormonal therapy versus the seven in the placebo group [23]. Atypical antipsychotics, selective serotonin reuptake inhibitors, and anti‐androgen medications were the most commonly prescribed medications (50, 35, and 31 instances of use, respectively, Figure 3). Hormonal therapies (cyproterone acetate, medroxyprogesterone, anti‐androgens, and finasteride) were associated with complete or partial remission of ISB in 92%–100% of 74 instances of use, all of whom were males except for one individual. Atypical antipsychotics, selective serotonin reuptake inhibitors (SSRIs), and anticonvulsants were associated with complete or partial remission of ISB in 56%–61% of 109 instances of use (Figure 3).

FIGURE 3.

Types of non‐pharmacologic therapies used 277 times in 67 studies and their associated change in ISB. Atypical antipsychotics and SSRIs were the most prescribed pharmacotherapies for ISB but were associated with partial or complete resolution of ISB only 60% and 57% of the time. Anti‐androgen therapy was the third‐most prescribed pharmacotherapy for ISB and was associated with partial or complete resolution of ISB 93% of the time.

Atypical antipsychotics and SSRIs were used more frequently than medroxyprogesterone or antiandrogen therapies, but the former were only associated with complete or partial remission of ISB symptoms about 55% of the time, versus approximately 95% complete or partial remission of ISB symptoms associated with the latter described hormonal therapies (Figure 4).

FIGURE 4.

Proportion of instances of use that pharmacotherapeutic interventions for ISB led to partial or complete resolution of symptoms, with bubble size representing the number of times each pharmacotherapy was used. Pharmacotherapies represented by large bubbles in the upper left quadrant (such as medroxyprogesterone and anti‐androgens) are those used most frequently and associated with ISB improvement more frequently, versus small bubbles in the lower right quadrant (such as benzodiazepines and trazodone) that were used infrequently and associated with behavioral improvement less than 20% of the time.

In 43 of 69 (62%) instances in which typical and atypical antipsychotics were prescribed to manage ISB, they either had intolerable side effects or did not achieve partial or complete remission, resulting in their discontinuation or the prescription of additional medications (Figure 5). In only eight of 50 (16%) instances of atypical antipsychotic use were they the only or final pharmacologic intervention trialed. In contrast, medroxyprogesterone or antiandrogens were the only or final medication trialed in a series of interventions in 46 of 51 cases (92%).

FIGURE 5.

Frequency of pharmacotherapeutic interventions for ISB being used as only intervention, en route to another more effective intervention, as the final medication at the end of a trial of multiple interventions or only having a benefit in combination with another psychoactive medication.

Medication tolerance and side effects were only reported for 17 of the 24 included drug classes (Figure S6). Antiandrogens were the most likely to have side effects reported (adverse side effects in 21 of 31 instances of use), including nausea, arthralgias, fatigue, and headache. Atypical antipsychotics were the second most frequently noted drug class to have adverse side effects (in 15 of 50 instances of use), including extrapyramidal symptoms, excessive sedation, and falls (Figure S6).

4. Discussion

This systematic review found that evidence describing the effectiveness of pharmacologic and non‐pharmacologic interventions for managing ISB is almost exclusively based on case studies and case series. Non‐pharmacologic interventions were associated with improvement or resolution of ISB in 72% of instances they were used. However, only five of 24 (21%) of these occurred without co‐prescribed pharmacotherapy, limiting our ability to draw conclusions about their individual effect. Within pharmacotherapies, hormonal therapies such as medroxyprogesterone and antiandrogen therapy were often effective after typical or atypical antipsychotics and SSRIs had failed to sufficiently improve behavior. Our findings suggest that while both hormonal therapies and antipsychotic medications can have side effects, the potential for benefit in cases of ISB refractory to non‐pharmacologic treatment may be greater with hormonal therapy. Given the side effects of pharmacotherapies and the effectiveness of combined medication and behavioral approaches, clinicians and researchers should implement and evaluate strategies that use both interventions together.

4.1. Comparison With Other Studies

A recent overview of systematic reviews conducted to inform Canada's BPSD guideline did not identify any systematic reviews published on this topic since 2017 [12]. Our systematic review is the most recent and rigorous synthesis of evidence for treating ISB in people with dementia. All reviews of treatment for ISB suggest an initial attempt to manage symptoms with non‐pharmacologic interventions; however, none synthesize the evidence for different types of such interventions [12, 17, 24, 25]. Suggested initial treatment with non‐pharmacologic interventions is aligned with our findings that these interventions are associated with improvements in ISB in close to 75% of cases. Kindrat et al.'s 2023 narrative summary of pharmacotherapy for managing ISB in nursing home residents suggests doses for different drug classes and identifies potential side effects without indicating which medications had more evidence supporting their use [17]. The authors recommended SSRIs as a first‐line treatment, with antipsychotics suggested if a more urgent or rapid response was necessary [17]. Hormonal agents were recommended in refractory cases [17]. These recommendations were consistent with Sarangi et al.'s review of 31 studies of behavioral and non‐behavioral treatments for ISB in dementia: the authors concluded that SSRIs be trialed as first‐line medications after non‐pharmacologic interventions failed [24]. Ozkan et al. conducted a narrative literature review of pharmacotherapy for ISB and recommended that antipsychotics or anticonvulsant mood stabilizers be first‐line therapies after non‐pharmacologic interventions failed [25]. Each of these reviews is limited by a lack of systematic protocols and by not indicating the count of successful outcomes achieved with pharmacologic and non‐pharmacologic interventions, which may contribute to their differing conclusions.

5. Clinical Application

This systematic review supports the management of people with dementia and ISB, although caution is needed in interpreting our results because they are predominantly informed by case studies and case series as opposed to high‐quality clinical trials. Our systematic review highlights the following clinical implications:

1. Non‐pharmacologic interventions such as patient and caregiver education and patient distraction were effective more than half of the time and should be trialed as a first step in managing ISB, either alone or in combination with pharmacotherapy. Distraction, diversion, and means of otherwise engaging hands seemed to be the most promising approach and should be trialed early.

2. p>Given the side effects of pharmacotherapies and the effectiveness of combined medication and behavioral approaches, clinicians and researchers should implement and evaluate strategies that use both interventions together.

3. Our findings suggest that while both hormonal therapies and antipsychotic medications have side effects, the potential for benefit in cases of ISB refractory to nonpharmacologic therapy may be greater with hormonal therapy.

4. Although extant non‐systematic syntheses conclude that SSRIs, antipsychotics, and anticonvulsants be used as first‐line pharmacotherapy [12, 17, 24, 25], our findings suggest that SSRIs and antipsychotics could be effective in a little more than half of cases. Given the higher risk of side effects with antipsychotics and their associated black box warning [26], SSRIs appear a more reasonable initial treatment for ISB management. ISB may also co‐occur with other behaviors, such as agitation or psychosis. Therefore, a parsimonious approach to prescribing the agents that target ISB and other symptoms (i.e., SSRIs for comorbid agitation and ISB) should be prioritized.

5. Hormonal therapy, such as medroxyprogesterone and anti‐androgen medications, may be more effective among males in situations where there is a very high risk associated with the ISB and the ISB has been refractory to psychosocial interventions and other pharmacological interventions. We posit that publication bias is less likely to affect case studies where multiple interventions are trialed sequentially, as they reveal therapeutic failures that may otherwise go unpublished. In these cases, antipsychotics and SSRIs were more likely to be tried and fail en route to a more effective intervention. In contrast, hormonal therapies were more often the final, effective step in a series of previously ineffective treatments.

We identified only one randomized trial in our systematic review [23]. More high‐quality randomized trial data are needed to inform evidence‐based management of ISB. This finding from our systematic review informed ISB management recommendations in the Canadian Clinical Practice Guidelines for Assessing and Managing Behavioral and Psychiatric Symptoms of Dementia, which indicated an absence of adequate evidence to make recommendations for or against any pharmacotherapies for ISB [27].

6. Limitations

A limitation of this systematic review is that 27 of 161 studies deemed eligible for inclusion based on title and abstract screening were not accessible in full text for further eligibility assessment. Most studies included were case reports, which are susceptible to publication bias [28]. Although sequential trials of interventions helped overcome the trend of exclusively positive effects of interventions being reported, cases in which multiple interventions were trialed were also subject to maturation bias, in which the nature and intensity of BPSD change over months and may become less bothersome independent of treatments offered [29]. Given that ISB likely arises from multiple mechanisms in people with dementia (e.g., disinhibition, perseveration, confusion as to whereabouts) [9], classifications of treatment efficacy by dementia phenotype would help guide clinical practice. However, only 65% of the studies included indicated which type of dementia was being treated. Descriptions of non‐pharmacologic interventions were also limited, frequently only described as “behavioral interventions” that were trialed alongside pharmacotherapy.

7. Conclusions

We synthesized the effectiveness of pharmacologic and non‐pharmacologic interventions for treating ISB in people with dementia. We found that non‐pharmacologic interventions can reduce ISB alone or in combination with medication. Among pharmacologic interventions that were reported more frequently, SSRIs and atypical antipsychotics were effective in just over half of cases, while hormonal interventions often worked as a last resort. To minimize the risk of harm, optimization of nonpharmacologic strategies should always be prioritized, even when pharmacotherapy is required. Higher quality evidence, including randomized trials, of both pharmacologic and nonpharmacologic management options for ISB in people with dementia is needed to inform further evidence‐based recommendations.

Author Contributions

N.E.L. is the corresponding author and attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. N.E.L., J.A.W., and D.P.S. conceived and designed the study. N.E.L. accessed the data and conducted the statistical analysis, and all authors interpreted the data. N.E.L. drafted the manuscript. All authors critically revised the manuscript.

Disclosure

Study sponsors had no role in the study design, methods, data collection, data analysis, results interpretation, or the preparation of the study report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S2.

Lane N. E., Ahuja M., Hatch S., Seitz D. P., McGowan J., and Watt J. A., “Treatment of Inappropriate Sexual Behavior in Persons With Dementia: A Systematic Review,” Journal of the American Geriatrics Society 73, no. 8 (2025): 2589–2597, 10.1111/jgs.19489.