Disinhibition as a Side Effect of Treatment with Fluvoxamine in Pediatric Patients with Obsessive-Compulsive Disorder

Introduction

C linically, obsessive-compulsive disorder (OCD) is defined by obsessions and compulsions. Obsessions are characterized by thoughts or images that are unwanted, recurrent, and intrusive, such as a fear of germs or a fear of being hurt. Compulsions are behaviors or mental acts that a person feels compelled to perform to neutralize an obsession or to feel “just right,” such as checking, hand washing, counting, or repeating an action (American Psychiatric Association 1994).

Serotonin-selective reuptake inhibitors (SSRIs) are commonly used in children and adolescents to treat major depression and anxiety disorders, with 60%–76% positive response (Research Units for Pediatric Psychopharmacology [RUPP] Anxiety Study Group 2001; March et al. 2004; Kennard et al. 2006). The current approach to treating OCD includes both pharmacological and behavioral interventions. SSRIs have been widely used among psychiatrists because they are effective and generally are well tolerated by patients, interact minimally with other neurotransmitter receptors (Warrington 1992), have few anticholinergic side effects, and do not require cardiac monitoring (Leonard et al. 1997). Of the SSRIs, fluoxetine (ages 7–17), sertraline (ages 6–17), and fluvoxamine (ages 8–17) have been approved by the Food and Drug Administration for use in children and adolescents (Rigoni 2004). A number of adverse effects of SSRIs have been reported in the literature for this age group. These include behavioral activation, manic symptoms, and an increased risk of reporting of suicidal thoughts (Walkup and Labellarte 2001; Leslie et al. 2005). A case series described five pediatric patients treated with paroxetine or fluoxetine who exhibited an amotivational syndrome defined by apathy, carelessness, lack of motivation, and, in one case, disinhibition (Garland and Baerg 2001). In another study, 12 of 24 children with various psychiatric conditions treated with fluoxetine for 8 weeks developed behavioral changes: 11 displayed motor hyperactivity, 6 became disinhibited, 3 exhibited internal overexcitement, and 11 suffered from sleep difficulties (Riddle et al. 1991). Similar side effects were also described in another study that reviewed 43 pediatric cases with anxiety disorders and found 2 patients who manifested both apathy and disinhibition after being treated with fluvoxamine for 32 weeks (Reinblatt and Riddle 2006). Treatment with fluvoxamine for as little as 8 weeks led to rates of increased activity, disinhibition, or activation as high as 45% among 22 pediatric patients with anxiety disorders (Reinblatt et al. 2009).

To determine whether our clinic patients diagnosed with OCD were more or less prone to side effects with SSRIs, we conducted a retrospective chart review of all pediatric patients seen in the past 7 months in our outpatient Anxiety Disorders Clinic. By describing the details of the individual cases, we hope to share our experience and raise awareness about adverse effects that are potentially life threatening. We believe that this information is timely and relevant, as disinhibition differs from symptoms of agitation, irritability, or mania, and is not a well-known side effect among our colleagues with few clinicians aware of the need to monitor for this side effect. Although the sample is small, the inability to predict the development of disinhibition on an individual level based on a pharmacogenetic profile is a novel finding of this report.

Case 1

W is a 12-year-old Korean boy who had his first outpatient psychiatric evaluation in October 2008 due to a fear of germs. At that time, he washed his hands excessively and resisted going to school due to a fear that he may be exposed to vomit at school. He was distressed whenever someone touched his nose or mouth for fear that he may get sick. He refused to touch public doorknobs or go upstairs at home by himself. He did not like odd numbers. He was very selective about his food and clothing. He also reported hypersensitivity to smell, sound, touch, and taste.

He was treated with cognitive behavior therapy, clonazepam 0.25 mg and fluoxetine 20 mg in November 2008, but there was no change in his OCD symptoms. Within 4 weeks of starting fluoxetine, W had difficulty falling asleep and exhibited bizarre behavior, for example, sawing a piece of plywood at 2 a.m. He also became aggressive—hitting and punching his mother several times per day. Fluoxetine was discontinued in December 2008, and he began treatment with mirtazapine, for unclear reasons. On this medication he developed irritability and weight gain, but there was no benefit recalled by the family. Mirtazapine was discontinued and W was started on fluvoxamine in April 2009 and titrated up to 150 mg over the course of 3 months.

Within 4–6 weeks of starting fluvoxamine, W's behavior changed. He no longer thought of the consequences before engaging in potentially dangerous activities. On two occasions, he left home without shoes or a helmet, bicycling past major intersections to go shopping. His mother followed alongside him in the car and repeatedly requested that he get off his bicycle and come in the car. Before starting fluvoxamine, W had always worn a bicycle helmet and never left home at night without permission. On another occasion, he climbed onto the roof of a neighbor's garage. In June 2009, he wanted to light matches and slammed the phone into his mother's chest when his father put the matches in water. When he heard his parents call 911, he tried to climb out the third floor window to escape to the roof. He was taken to the emergency room where he was referred to an inpatient child and adolescent psychiatry unit. Impulsivity, oppositionality, aggression, and poor judgment resolved once fluvoxamine was discontinued.

We presented to our outpatient clinic in July 2009 after discharge from our partial hospitalization program. At that time his Child Yale Brown Obsessive Compulsive Scale (CY-BOCS) score was 32. A low dose of sertraline was initiated to manage his OCD symptoms. Pharmacogenetic testing revealed that W is an extensive metabolizer for CYP2C19 and CYP 2D6. On the basis of his genotype and the corresponding metabolizing phenotype, the laboratory recommended no modification in dosing based on genotype results. No further disinhibited behaviors have occurred to date while he has been maintained on 25 mg of sertraline. His symptoms are noticeably improving with cognitive behavior therapy and a supportive school atmosphere.

Case 2

K is an 11-year-old Hispanic boy previously diagnosed with OCD, attention-deficit/hyperactivity disorder-combined type, chronic motor tic disorder, and a learning disability, who presented for an outpatient psychiatric evaluation in our clinic April 2009. Before coming to our clinic, he had trials of atomoxetine, methylphenidate, aripiprazole, amphetamine/dextroamphetamine, clonidine, quetiapine, oxcarbazepine, guanfacine, risperidone, bupropion, clomipramine, fluoxetine, and sertraline. Indications, doses, benefits, and adverse reactions could not be recalled for these multiple medications. Neuroleptics were likely used to treat motor tics that included a torso twisting motion that had left a paraspinal bruise. Prior evaluation in the division of developmental and behavioral pediatrics at our hospital had ruled out an autism spectrum disorder, but minor social awkwardness was noted during follow-up appointments.

K's symptoms at the time of presentation included feeling itchy after a shower, discomfort wearing underwear due to a wool-like sensation, and refusal to get water on his face. He also worried excessively about swimming in the ocean, “because everyone gets killed.” Fluvoxamine 25 mg was added to his clomipramine, 75 mg at bedtime, bupropion SR 100 mg twice daily, and risperidone 0.5 mg at bedtime, with the plan to discontinue bupropion and risperidone once his OCD symptoms were under control. Nearly a month after he was prescribed fluvoxamine, the patient's mother reported that he had become angry and disrespectful. He told his teachers things that he had never said before, such as, “Mind your own business,” and uncharacteristically, he received detention in school. Fluvoxamine was discontinued and his impulsive, disrespectful speech resolved.

Pharmacogenetic testing revealed that K's CYP2C19 and CYP2D6 genotype are consistent with a metabolism phenotype known as an extensive metabolizer. On the basis of his genotype and the corresponding metabolizing phenotype, the laboratory recommended no modification in dosing based on genotype results.

Case 3

B is a 10-year-old Caucasian boy with OCD and Tourette's disorder who was initially evaluated our clinic in the spring of 2008. Presenting symptoms included echolalia, eye blinking, mouth stretching, and squeaking sounds. His obsessions and compulsions included a fear of contamination for which he would avoid sitting on or touching things he perceived as dirty, the need to touch things symmetrically, and the urge to find a word that rhymes with a swear word he heard. He was distressed by the latter compulsion due to his religious Christian upbringing and he struggled with low self-esteem and difficulty falling asleep due to his worries. He had a limited diet as he was unable to tolerate most fruits and vegetables due to their texture. Additional symptoms included forgetfulness, poor organizational skills, and stool withholding for which he was treated with Miralax. His CY-BOCS score was 23. He was started on fluvoxamine 50 mg at bedtime, which was titrated up to 50 mg twice daily.

After 286 days on fluvoxamine, B's mother called to report several unusual incidents that were atypical for B. He and a friend had gone down a slide inside a garbage can. He threw a hatchet at a spray can in the garage. He also dropped his drawers at school, which he claimed was an accident. Fluvoxamine was tapered and discontinued.

Sertraline was then started at 25 mg and titrated up to 75 mg daily. Within a month, his mother reported her concern that symptoms of disinhibition were returning—B had put a toy handcuff in his back pocket and told his mother that he would handcuff a peer at school. Sertraline was decreased to 25 mg. His obsessions and compulsions resurfaced with an increase in repeating compulsions, checking his foot daily for peeling skin, and worrying that his mother would die although her only symptom was a slight fever. He was referred to the psychology division for cognitive behavior therapy.

Pharmacogenetic testing revealed that B's CYP2C19 phenotype is an extensive metabolizer and his CYP2D6 phenotype is a poor metabolizer. On the basis of his genotype and the corresponding metabolizing phenotype, the laboratory recommended that B could tolerate dosing aimed at 69% of the standard dose for fluvoxamine.

Discussion

Seventeen of the 51 patients who were found to have OCD in the 7 months under consideration in this study were previously or are currently treated with fluvoxamine. Of these 17 patients, 3 became disinhibited within 9 months of the start of treatment. This infrequent but severe side effect warrants attention as impaired judgment and disrespectful behavior observed in pediatric patients may potentially lead to devastating consequences. These symptoms should alert the prescribing clinician to lower the dose of medication and monitor the patient closely.

How might clinicians predict whether a patient will become disinhibited on an SSRI? Might knowing the individual's unique genetic makeup that determines pharmacokinetic factors regarding drug metabolism be helpful in predicting the potential development of a side effect? As described in the Introduction section, fluvoxamine is not the only SSRI with the potential of causing disinhibition, but it is used here as an illustrative example.

Many psychotropic medications are metabolized by cytochromes P450 (CYPs) 2C19 and 2D6; these are enzyme systems with genetic polymorphisms known to influence substrate metabolism. As part of clinical care at our institution, CYP2C19 and CYP2D6 pharmacogenetic testing is performed for inpatient psychiatric admissions, and it is becoming routine in the outpatient psychiatric division as well. Both CYP2C19 and CYP2D6 genotypes can predict four metabolizer phenotypes, from lowest to highest efficiency: Poor, intermediate, extensive, and ultrarapid. At this time, our institution tests for alleles *1, *3,*4, and *5 (deletion) and duplication in CYP2D6 and the *1 and *2 alleles in CYP2C19. Each combination of two alleles (the genotype) is used to predict the phenotype. Phenotype may then be used to anticipate whether a patient may have higher or lower than expected exposures to medications metabolized by those enzymes. For example, if a patient is a CYP2D6 poor metabolizer and treated with a medication metabolized by CYP2D6, they are at risk for higher concentrations, due to a decreased ability to metabolize the medication, compared with the majority of the population (extensive metabolizers). These higher concentrations may increase the risk for side effects and adverse reactions. Another major drug-metabolizing enzyme in psychiatric medicine is CYP3A4; our institution does not offer CYP3A4 genotyping for clinical purposes, as there are no specific genetic polymorphisms that have consistently demonstrated a significant change in the pharmacokinetics of CYP3A4 substrates.

Fluvoxamine is metabolized to inactive metabolites primarily by CYP2D6 and CYP1A2. In addition, fluvoxamine strongly inhibits the hepatic P450 isoenzymes CYP1A2 (Brosen et al. 1993) and CYP2C19 (Xu et al. 1996), and more weakly inhibits CYP3A4 (Fleishaken and Hulst 1994), CYP2C9 (Schmider et al. 1997), and CYP2D6 (Crewe et al. 1992; Hemeryck and Belpaire 2002), which may lead to drug–drug interactions.

The hope of pharmacogenetic testing is to identify the metabolic phenotypes of the patients and thereby anticipate possible responses to different dosages of medications that are metabolized by or interact with specific enzymes (de Leon et al. 2008). This hope was shown to be a potentially useful tool on inpatient pediatric psychiatric units in a retrospective analysis (Prows et al. 2009). They examined the correlation between the CYP2D6 and CYP2C19 genotype-predicted combined phenotype and a behavioral intervention score to measure psychotropic efficacy as well as the drug's tolerability by reported adverse reactions. As would expected, they found that CYP2D6- or CYP2C19-dependent psychotropic medications had the highest efficacy in poor metabolizers as well as the highest number of adverse drug reactions. Medications were least effective, as indicated by the highest behavioral intervention score, in ultrarapid metabolizers, and the lowest number of adverse reactions was reported in these patients.

Although pharmacogenetic testing might hold future promise in providing optimized patient-specific therapy, the sole identification of a patient's genotype and phenotype may not prospectively predict the predilection for fluvoxamine-induced disinhibition or other adverse reactions as can be seen in the three cases described above. Two of these patients were extensive metabolizers of both CYP2C19 and CYP2D6 and thus theoretically able to tolerate at least 100% of recommended dosing, yet they developed an adverse response on a low dose of medication. One patient was a CYP2D6 poor metabolizer, suggesting a higher than expected exposure to the medication and higher risk for side effects. Drug–drug interactions, as well as CYP1A2 genetic variability, are also expected to influence fluvoxamine exposure and response. Given our current state of knowledge, the clinician who wishes to prescribe in a personalized manner will need to account for the influence of other genetic, environmental, and personal factors in addition to the pharmacokinetics and pharmacodynamics of each drug (De Leon 2009). As can be seen from Table 1, family history of bipolar disorder may in fact be more relevant in predicting the likelihood of disinhibition or the development of agitation in response to low-dose SSRI treatment than the extent of metabolism.

True manic symptoms or episodes were not described by the families of the patients in the three cases above. There was no report of decreased need for sleep, hypersexuality, grandiosity, or pressured speech. Instead, the symptoms were impulsive in nature, providing evidence of poor judgment and frontal lobe disinhibition. Yet, these clustering of symptoms may be related in a basic mechanistic way to elements that underlie the disinhibition seen in mania. As Starkstein and Robinson (1997) reviewed, various disinhibition syndromes may result from head injuries, brain tumors, strokes, or epilepsy that lead to symptoms that range from mildly inappropriate social behavior to mania. Interestingly, a strong association was found between disinhibited behavior and dysfunction in the orbitofrontal and basotemporal cortices of the right hemisphere.

To explain the symptoms of patients with OCD, Rapoport (1991) and Rauch et al. (1994) proposed a model of executive function that views the main problem in these patients as an inability to inhibit nonadaptive responses. Imaging studies support their view and implicate a neural circuit that involves altered activity in the frontal cortical regions, basal ganglia, and thalamus in symptomatic patients (e.g., Baxter et al. 1988; Rubin et al. 1992; Perani et al. 1995). In a task used to examine behavioral disinhibition in subjects undergoing functional magnetic resonance imaging (Strakowski et al. 2008), prefrontal activation was reported in patients experiencing their first manic episode of bipolar disorder but not in healthy control subjects. Since relatives of patients with bipolar disorder may have inherited altered connectivity in prefrontal regions (e.g., Allin et al. 2010; Surguladze et al. 2010), it is plausible that upon exposure to fluvoxamine, our patients exhibited disinhibited behavior that relates in a small measure to the type of disinhibition associated with changes in frontal lobe circuitry. However, we caution that these subjects did not exhibit a full blow manic episode and it is unclear whether the disinhibited behavior can be considered a result of a bipolar diathesis; therefore, we see no need for initiating treatment with a mood stabilizer whenever symptoms of disinhibition are present, as merely lowering the dose of fluvoxamine led to resolution of the symptoms. We intend to continue to closely monitor these patients and treat the symptoms of OCD or other illnesses as they emerge.