Pharmacological Treatment of Olfactory Reference Disorder: A Systematic Review and Historical Analysis of Treatment Patterns Authors: Ligia M Fonseca, Verônica Hühne, Murilo Branco, Inês M Salles-Fonseca, Gabriela B de Menezesa, Leonardo F Fontenelle Published: 2026 ABSTRACT AbstractObjective:olfactory Reference Disorder (ORD) is characterized by a persistent preoccupation with the belief of emitting a bad or unpleasant body odor or breath, which is either imperceptible or only minimally perceptible to others. To systematically review studies investigating the pharmacological treatment of Olfactory Reference Disorder (ORD) and explore clinical characteristics, treatment response, and historical changes in pharmacological prescribing patterns.Method:systematic review conducted in accordance with PRISMA guidelines. Sociodemographic, clinical, phenomenological, and therapeutic variables were extracted.Results:fifty-two studies (47 case reports, 5 case series, and no controlled clinical trials), totaling 87 patients, were included. Treatment included serotonin reuptake inhibitors in 59.8%, antipsychotics in 52.9%, and both in 22% of cases. The overall treatment response rate was 86.2%, although interpretation is limited by uncontrolled study designs and probable publication bias. A historical analysis revealed that antipsychotics were more frequently prescribed in earlier reports, whereas SRIs became widely used only after the 1990s, reflecting a shift in the conceptualization of ORD toward the obsessive–compulsive spectrum. Also, no statistically significant differences were observed between responders and non-responders regarding sex, illness duration, psychiatric comorbidities, treatment modality, or pharmacological class used. Evidence of publication bias was identified.Conclusions:ORD appears to be a potentially treatable condition, although the available evidence is limited by uncontrolled study designs, heterogeneous outcome measures, and probable publication bias. The shift toward serotonergic treatments parallels the evolving understanding of ORD within the obsessive–compulsive spectrum. Prospective studies using standardized outcome measures are needed..Keywords: olfactory reference syndrome, olfactory reference disorder, obsessive–compulsive and related disorders, pharmacotherapy, selective serotonin reuptake inhibitors Introduction The term Olfactory Reference Syndrome (ORS) was introduced in 1971 by the neurologist Pryse-Phillips through the description of 36 cases, with the aim of distinguishing this condition from olfactory symptoms observed in schizophrenia, depression, and temporal lobe epilepsy. Despite being a clinically recognized condition for many years, ORS was only formally acknowledged as Olfactory Reference Disorder (ORD) in diagnostic manuals in 2019, with the publication of the International Classification of Diseases, 11th Revision (ICD-11). Between its initial description and current conceptualizations, ORD fluctuated across diagnostic systems, being included among somatoform disorders in the Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III), and later classified as a delusional disorder in the DSM-IV (Begum & McKenna, 2011). In the DSM-5, ORD is briefly mentioned under “other specified disorders” within the group of Obsessive–Compulsive and Related Disorders (Soman & Narayanaswamy, 2023). ORD is characterized by a persistent preoccupation with the belief that one is emitting an unpleasant or offensive body odor. Affected individuals typically exhibit excessive self-consciousness regarding the perceived odor, ideas of reference, and the belief that others can detect the smell and react negatively. Consequently, patients often avoid social situations and engage in repetitive and excessive behaviors aimed at checking, minimizing or controlling the perceived source of the smell (Tada, 2023). The onset of symptoms is estimated to occur predominantly from late adolescence to the third decade of life, with a slight male predominance (Reuter et al., 2024). Recent community-based studies estimate the prevalence of ORD to range from 0.5% to 2.1%, although these figures are likely underestimated (Soman & Narayanaswamy, 2023). It is common for psychiatric evaluation to be sought only after multiple consultations with other medical specialists—on average four to five—in attempts to identify a physical cause for the perceived odor. (Prazeres et al., 2010; Teraishi et al., 2012) Until the late twentieth century, when ORD was largely conceptualized as part of the spectrum of delusional disorders, antipsychotics were presumably prescribed more frequently for this condition. As the conceptualization of ORD evolved during the 2000s toward its inclusion within the obsessive–compulsive spectrum, serotonergic antidepressants may have become more commonly prescribed (Begum & McKenna, 2011). Despite its increasing recognition, there remains a paucity of studies systematically evaluating the pharmacological management of ORD, with the last review on the pharmacotherapy of ORD being published more than 15 years ago (Begum & McKenna, 2011). The absence of controlled clinical trials and the predominance of case reports significantly limit the establishment of standardized therapeutic strategies. Therefore, investigating which pharmacological classes have been used and in which clinical contexts is relevant to inform and refine more effective treatmentapproaches. Method Objective To review the literature on the pharmacological treatment of ORD, by describing the types of studies conducted; the overall response rate; sociodemographic, clinical (such as psychiatric comorbidities), and therapeutic characteristics (including different medication classes) of responders versus non-responders to treatment; as well as potential publication bias. Protocol and registration The protocol for this review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) under registration number CRD42024534359. This systematic review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews) guidelines. Eligibility criteria Eligibility criteria were defined based on the key components of the PICOS framework (Participants, Intervention, Comparator, Outcomes, and Study Design), as detailed below: Participants: individuals diagnosed with ORD according to ICD-11 or DSM-5- TR criteria, as well as individuals whose clinical presentation was consistent with these definitions, as determined through clinical interviews and/or validated assessment instruments. Studies conducted prior to the formal establishment of these diagnostic criteria were also included. Intervention: any pharmacological treatment. Comparator: no control group was required. Outcome: a broad definition of treatment response, as defined by the authors of the original studies. Study designs: we searched for randomized and non-randomized controlled trials, open-label studies, case series and case reports, case–control studies, and observational studies. Information sources and study selection The electronic search was conducted on May 17, 2025, in the following databases: PubMed, PsycInfo, Web of Science, and EMBASE. The same search strategy was applied across all four databases, with no restrictions on publication date or language: (“olfactory reference” OR “olfactory reference syndrome” OR “olfactory delusion*” OR “olfactory hallucination*” OR “olfactory paranoia*” OR “chronic olfactory” OR bromidrosiphobia OR bromosis OR autodysoma* OR “body smell” OR “bad odor” OR “corporal smell” OR “corporal odor” OR halitophobia OR “imagined halitosis” OR “psychogenic halitosis” OR “delusional halitosis” OR “hallucinatory halitosis” OR “monosymptomatic hypochondri*”) AND (medication OR pharmaco* OR intervention OR therapeutic* OR treatment OR drug OR “drug-based” OR antidepressant* OR serotoninergic OR “serotonin reuptake inhibitor*” OR “serotonin and noradrenaline reuptake inhibitor*” OR antipsychotic* OR neuroleptic* OR “dopamine blocker*” OR “dopamine antagonist*”). Titles and abstracts of the records retrieved through the electronic search were uploaded to the Rayyan platform. After removal of duplicates, studies were screened and selected independently by two reviewers, initially based on title and abstract. Subsequently, both reviewers assessed the full text of potentially eligible studies to determine inclusion. In cases of disagreement, consensus was sought; if disagreement persisted, a third reviewer was consulted to resolve discrepancies. Risk of Bias assessment Studies were assessed for risk of publication bias by comparing clinical outcomes reported in case reports and case series. Data collection process Data extraction was performed by the first author, and the extracted data were independently reviewed by two co-authors. Discrepancies were resolved through weekly team meetings. The following information was extracted from each article: year and country of publication, author name, sex, age, duration of illness (in years), presence of precipitating factors, source of the perceived odor (e.g., breath, sweat, feces), degree of conviction (delusional vs. non-delusional), presence of psychiatric comorbidities, pharmacological treatment used (including serotonergic antidepressants, antipsychotics, benzodiazepines, and other medications), corresponding dosage (adequate vs. subtherapeutic), previous treatments, use of adjunctive cognitive- behavioral therapy (CBT), and clinical outcome. Data Analysis Data were analyzed using the Statistical Package for the Social Sciences (SPSS). Given the heterogeneity of terminology used to describe psychiatric disorders over the decades, a standardization of variables was adopted. For statistical analysis purposes, the terms obsessive–compulsive neurosis, severe obsessive condition, obsessional neurosis, obsessive–compulsive disorder, and obsessive symptoms were grouped under the umbrella term Obsessive–Compulsive Disorder (OCD). When assessing associated psychiatric comorbidities, several articles reported mood symptoms without explicitly diagnosing a depressive disorder. Accordingly, this category was subdivided into ‘depressive symptoms” and major depressive disorder. The psychotic disorders category included primary psychotic conditions, such as schizophrenia and persistent delusional disorder. Additional comorbidities analyzed included social anxiety disorder, bipolar disorders, cognitive impairment, and substance use. Regarding pharmacological treatment, the use of selective serotonin reuptake inhibitors (SSRIs) was analyzed together with clomipramine, under the category serotonin reuptake inhibitors (SRIs). Treatment response was assessed based on the clinical descriptions provided in the articles, given the absence of standardized outcome measures across studies. Categorical variables were tested for associations with the binary outcome using the chi-square test, while continuous variables were analyzed using Student’s t-test. Results General description of the sample The included studies (n = 51) consisted of case reports (n = 46; see table 1) and case series (n = 5; see table 2). The sample comprised a total of 87 patients diagnosed with ORD, of whom 56.3% were male (n = 49) and 42.5% were female (n = 37). The mean age was34.1 ± 15.1 years, ranging from 14 to 75 years. The mean duration of illness was 7.1 ±8.3 years, with a range from 0.01 to 42 years. Table 1.Case reports Main Author Year Main Author Year Main Author Year Gebara2011Boro2020Rodriguez-Quiroga2014Riding1975Stein1998Bizamcer2008Petho1979Raheman2019Balaban2015Boro2025Cheryak2021Albers2018Davidson1982Ahamed2019Teraishi2012Yoshimura2021Roshan2024Gama-Marques2014Raheman2024Ozen2018Lorea2011Bishop1980Michael2014Wernecke2002Alexander2010Gungoren2018Gebara2011Uvais2023Alhadi2018Alvarez2020Brotman1984Gkini2017Misery2009Aftab2017Ferreira2014Li2022Reddy2022Tee2015Lochner2003Domingues1997Takekita2011Singh2006Takenoshita2021Thomas2014Sayar2003Ozen2016Source: Research data; prepared by the author (2025). Case reports Source: Research data; prepared by the author (2025). Table 2.Case series Main Author Year Main Author Year Senkal2015Malasi1990Suzuki2004Prazeres2010Luckhaus2003 Case series The most frequent psychiatric comorbidities were depressive disorder (31%), obsessive–compulsive disorder (OCD) (12.6%), social anxiety disorder (9.2%), psychotic disorders (3.4%), bipolar disorders (4.6%), cognitive impairment (3.4%), and substance use (5.7%). Depressive and anxiety symptoms without a formal diagnosis were also reported in 24.1% and 14.9% of cases, respectively. The mean age at symptom onset was similar between men (24.6 ± 12.5 years) and women (30.1 ± 17.0 years), with no statistically significant difference (t(84) = 1.47; p > 0.05). Patients with psychiatric comorbidities showed a slightly later onset (27.8 ± 14.8 years) compared with those without comorbidities (24.4 ± 13.8 years), although this difference was not statistically significant (t(85) = 0.93; p > 0.05). In 26 of the 87 cases (approximately 30%), precipitating events were described, such as perceived comments or gestures by others regarding environmental or personal odors, including frowning or placing a hand over the nose. These events were reported as triggering the onset of ORD symptoms. Regarding the descriptions of perceived body odors, sweat was reported in 24% of cases, general body odor in 20%, halitosis in 20%, fecal odor in 10%, urine odor in 5%, and flatulence in 5%. In some cases, more than one odor was described, and other reports included less specific perceptions, such as a “smell of fear” or an odor perceived as offensive to others. Concerning the degree of conviction that the individual was indeed emitting an unpleasant or offensive odor, 52% of the analyzed patients exhibited absolute conviction, while 20% showed moderate conviction, characterized by some degree of doubtregarding the reality of the odor. In the remaining cases, the degree of conviction was not described. Pharmacotheraphy of ORD Treatment included serotonin reuptake inhibitors in 59.8%, antipsychotics in 52.9%, and both in 22% of cases. Cognitive Behavioral Therapy (CBT), particularly exposure and response prevention, was used in association with pharmacotherapy in 23%. Because most studies (48) lacked objective measures of treatment response, we were required to rely, albeit leniently, on the authors’ reports of treatment effectiveness to classify individuals as responders (e.g. Fontenelle et al., 2015). Accordingly, the overall treatment response rate was 86.2% (75 of 87 individuals), regardless of other variables or the treatment modality employed. An attempt was performed to identify sociodemographic and/or clinical features associated with treatment response. When directly comparing patients who responded to treatment with those who did not, the distribution between men and women was similar across groups. Among responders (n=75), 42 patients (56.7%) were male and 32 (43.3%) were female, whereas among non-responders, seven (58.3%) were male and five (41.6%) were female. No statistically significant difference was observed between groups (χ2= 0.01; p > 0.05), indicating that sex did not influence the likelihood of treatment response. Regarding duration of illness, non-responders had a mean illness duration of 6.5 years (SD = 5.19), while responders had a mean duration of 7.22 years (SD = 8.72). This difference was not statistically significant (t = 0.24; p > 0.05), suggesting that symptom duration did not differentiate responders from non-responders. With respect to psychiatric comorbidities, responders were predominantly composed of individuals with comorbid conditions (69.3%, 52 of 75), similarly to non- responders (91.6%, 11 of 12). However, this difference was not statistically significant (p> 0.05), indicating that the presence of comorbidities did not distinguish responders from non-responders. In relation to antipsychotic use, most responders had received antipsychotics (n=40, 53.3%), a proportion similar to that observed among non-responders (n=6, 50.0%). The absence of a statistically significant difference (χ2= 0.01; p > 0.05) suggests that antipsychotic use alone was not associated with clinical response. A similar pattern was observed for the use of SRIs, with the proportion of responders not being statistically different between SRI users (57.5%) and non-users 42.5% (χ2= 0.65; p > 0.05). Among patients who used antipsychotics, a marked difference in outcome distribution was observed when comparing typical and atypical agents. All non-responders had received typical antipsychotics (5 of 5; 100%), whereas all responders had used atypical antipsychotics (0 of 24; 0%). Among responders, 36.8% had used typical antipsychotics (14 of 38) and 63.2% had used atypical antipsychotics (24 of 38). However, due to the presence of cells with zero frequency, Fisher’s exact test was applied and did not indicate a statistically significant association between antipsychotic type and therapeutic outcome (p > 0.05). Among patients who responded to treatment, 50% had used adequate doses of SSRIs/clomipramine (≥ 75% of the recommended therapeutic dose), while the remaining 50% had used subtherapeutic doses (18 of 36). Among non-responders, all cases had used doses below 75% (2 of 2). Although this distribution suggests a potential clinical trend, Fisher’s exact test did not reveal a statistically significant difference between responders and non-responders (p > 0.05). The combined use of SRIs and antipsychotics was observed in 18 responders (24.0%) and in only 1 non-responder (8.3%). Although numerically higher, this difference did not reach statistical significance (Fisher’s exact test, p > 0.05), indicating that combination therapy may suggest clinical benefit but did not demonstrate a statistically confirmed effect. Finally, 24.7% of responders received cognitive-behavioral therapy (CBT). Only 10% of non-responders had undergone CBT. Although the proportions suggest a possible clinical benefit of CBT, the difference was not statistically significant (χ2= 0.91; p > 0.05), indicating that CBT may increase the likelihood of response, but no statistically confirmed difference was observed in the present sample. The analysis of clinical outcome distribution according to publication period revealed that the proportion of responders was similar between studies published up to 1990 and those published thereafter, with 1990 chosen as a reference point representing the beginning of the decade in which SRI prescription became widespread. Among studies published prior to 1990 (n = 12), 83.3% of patients showed a clinical response, whereas in studies published after 1990 (n = 75), the response rate was 86.7%. No statistically significant difference was observed between periods regarding the likelihood of therapeutic response (non-significant χ2; p > 0.05). A difference was observed in antipsychotic prescription over time. Among studies published up to 1990, 75% of patients received antipsychotics (n = 9/12), whereas in studies published after 1990 this proportion decreased to 49.3% (n = 37/75). Statistical analysis, however, indicated that this difference did not reach statistical significance (χ2(1) = 2.72; p > 0.05). Nevertheless, earlier reports showing a higher frequency of neuroleptic use are consistent with the historical conceptualization of ORD as a somatic delusional disorder. The use of SRIs also varied according to the historical period. In studies published before 1990, no patients received SRIs, whereas in studies published after 1990, 69.3% of patients were treated with this pharmacological class (n = 52/75). The association between publication period and SRI use was statistically significant (χ2(1) = 20.68; p < 0.001), reflecting the conceptual shift in the understanding of the disorder over recent decades and its increasing alignment with the obsessive–compulsive spectrum. Publication bias A significant difference was observed between study types regarding clinical outcomes. Case reports showed a substantially higher proportion of treatment response (96.2%) compared with case series (70.6%) (p < 0.01; Fisher’s exact test). This finding suggests the presence of publication bias, as isolated case reports tend to describe more favorable clinical courses, whereas case series include greater variability in treatment response. Discussion The present systematic review analyzed 87 patients described in 52 studies (case reports and case series), suggesting that some individuals with ORD may improve with pharmacological interventions; however, the absence of controlled studies, heterogeneous outcome measures, and probable publication bias substantially limit conclusions regarding efficacy. The similarly high response rates reported across different pharmacological classes, including both antipsychotics and SRIs, together with the absence of control groups and the predominance of case reports, raise the possibility that nonspecific treatment effects, placebo response, expectancy effects, and outcome reporting bias may contribute to the optimistic response rates described in the literature. The results are consistent with the review by Begum and McKenna (2011), performed almost 15 years ago, which analyzed 84 cases and found that more than two- thirds of patients improved, particularly those treated with antidepressants and psychotherapeutic interventions, while responses to antipsychotics were more variable. In addition to summarizing the available evidence on pharmacological treatment, the present review examined historical changes in prescribing patterns across decades. This approach allowed us to observe how shifts in the conceptualization of ORD from a somatic delusional condition to a disorder related to the obsessive–compulsive spectrum were paralleled by changes in pharmacological strategies. Historically, ORD was initially described as a somatic delusional condition, which for decades was associated with the predominant use of antipsychotics. The therapeutic shift toward serotonergic strategies parallels the psychopathological reinterpretation of the disorder: whereas Pryse-Phillips (1971) emphasized delusional features in part of the population, contemporary studies highlight obsessive elements, referential ideation, rumination, and checking behaviors, bringing the syndrome closer to obsessive–compulsive disorder (OCD) rather than primary psychosis. This reconceptualization has been accompanied by an increased use of SRIs in clinical practice, which is consistent with the current classification of ORD within the obsessive–compulsive and related disorders in the ICD-11 (Soman & Narayanaswamy, 2023; Tada, 2023). Another relevant finding of this review was the scarcity of standardized instruments to assess clinical response. Although two validated instruments are currently available to assess symptoms of olfactory reference disorder – the Yale–Brown Obsessive–Compulsive Scale modified for ORS (ORS-YBOCS) and the Olfactory Reference Disorder Questionnaire (ORDQ) (Reuter et al., 2024) – only three studies employed standardized rating scales, highlighting the lack of systematic outcome assessment in the existing literature. Most reports described improvement in a subjective and heterogeneous manner, limiting quantitative comparisons. This limitation had already been highlighted by Begum and McKenna and remains relevant, underscoring the need for standardized outcome measures. Additionally, when analyzing prior treatments, many articles merely listed previously used pharmacological classes without reporting dosage or treatment duration, hindering a critical assessment of whether these trials constituted adequate-dose and adequate-duration treatments with true non-response. The augmentation of pharmacotherapy with CBT, used in 23% of cases, is also consistent with contemporary literature emphasizing the utility of exposure and response prevention and insight-oriented cognitive restructuring (Reuter et al., 2024). Although CBT was not more frequently adopted among responders in the present sample, the high improvement rate (95%) among patients who received CBT suggests a potential benefit, reinforcing current clinical recommendations (Soman & Narayanaswamy, 2023). The lack of association between psychiatric comorbidities and treatment response is likewise consistent with the literature, which indicates that comorbidity is common but does not necessarily predict poorer prognosis (Soman & Narayanaswamy, 2023; Begum & McKenna, 2011). Table 3.Association between clinical outcome and clinical and therapeutic variables in patients with Olfactory Reference Disorder Variable Responder: n (%) Non-responder: n (%) Test p value Gender χ2(1) = 0.010.920 Male (n = 49)42 (56.7)7 (58.3) Female (n = 37)32 (43.3)5 (41.7) Comorbidities Fisher's exact0.250 Yes (n = 63)52 (69.3)11 (91.7) No (n = 24)23 (30.7)1 (8.3) Antipsychotics χ2(1) = 0.010.930 Yes (n = 46)40 (53.3)6 (50.0) No (n = 41)35 (46.7)6 (50.0) Type of antipsychotic Fisher's exact– Typical (n = 19)14 (36.8)5 (100.0) Atypical (n = 24)24 (63.2)0 (0.0) SRI χ2(1) = 0.650.420 Yes (n = 52)46 (57.5)6 (46.2) No (n = 35)34 (42.5)7 (53.8) SRI and antipsychotics Fisher's exact0.204 Yes (n = 19)18 (24.0)1 (8.3) No (n = 68)57 (76.0)11 (91.7) Cognitive-behavioral therapy Fisher's exact0.340 Yes (n = 20)19 (24.7)1 (10.0) No (n = 67)58 (75.3)9 (90.0)Source: Research data; analysis performed using SPSS (version 29), prepared by the author (2025). Association between clinical outcome and clinical and therapeutic variables in patients with Olfactory Reference Disorder Source: Research data; analysis performed using SPSS (version 29), prepared by the author (2025). Another important aspect identified in this review concerns the phenomenological characteristics of ORD. Variability was also observed in the degree of conviction regarding odor perception: in 52% of cases, conviction was described as absolute, while approximately 20% exhibited partial conviction, suggesting a fluctuation between fixed belief and doubt. This phenomenological variability is clinically relevant, as contemporary literature suggests that ORD may lie on a continuum between obsessive ideas with poor insight and overvalued ideas. However, in the present review, it was not possible to establish an association between degree of conviction and treatment response due to the lack of systematic measurement of this variable in the included studies. Case descriptions often lacked detailed psychopathological characterization, which would help distinguish whether the odor was truly perceived as a sensory-perceptual alteration or represented a cognitive-perceptual belief, as well as clarify the degree of conviction. Nonetheless, recent empirical studies have reinforced the overlap between ORD, social anxiety disorder, and obsessive–compulsive psychopathology, supporting the view that ORD may be better conceptualized along a spectrum of disorders characterized by heightened self-referential concerns, shame, and repetitive behaviors (Reuter et al., 2024; Croy et al., 2025). Emerging neuropsychological and olfactory findings also suggest that ORD may involve alterations extending beyond purely delusional beliefs, including abnormalities in olfactory processing, cognitive interpretation, and self-referential perception, although these mechanisms remain poorly understood (Skimming & Miller, 2019; Sofko et al., 2020). Finally, regarding publication bias, the findings are consistent with the well- documented phenomenon whereby case reports tend to emphasize positive therapeutic outcomes, whereas case series include greater clinical heterogeneity and, consequently, lower response rates. This pattern reinforces the need for caution when extrapolating findings from isolated case reports to clinical practice and highlights that the available evidence remains limited and subject to publication bias. Conclusion The present systematic review contributes to the clinical and therapeutic understanding of ORD, highlighting that it is a psychiatric condition frequently marked by intense distress, significant functional impairment, and a high degree of social isolation. Available reports suggest that some patients with ORD may benefit from pharmacological and psychotherapeutic interventions; however, the current evidence base remains weak and insufficient to establish treatment efficacy. No statistically significant differences were observed among the main pharmacological classes employed. Nevertheless, the use of SRIs and CBT appears coherent as first-line approaches, given the current classification of ORD within the obsessive–compulsive spectrum in the ICD-11. The discrepancy between response rates reported in case reports and case series indicates a likely publication bias, whereby cases with more favorable clinical outcomes are more likely to be published. This finding underscores the need for prospective, controlled studies with standardized outcome measures. In conclusion, the treatment of ORD should be conducted in an integrated manner, combining serotonergic pharmacotherapy at adequate doses with structured psychotherapeutic interventions. Progress in this field depends on the development of specific diagnostic instruments, greater clarity regarding the role of insight, and comparative studies that allow the establishment of evidence-based therapeutic strategies.