Evidence-Based Reviews

Not all patients who experience hallucinations have a psychotic disorder. Many physical and psychiatric disorders can manifest with hallucinations, and some patients have >1 disorder that could cause different types of hallucinations. To avoid providing unnecessary or ineffective treatments—and to ensure that patients receive proper care for nonpsychiatric conditions—it is important to accurately diagnose the disorder causing a patient’s hallucinations.

In this article we describe common features and psychiatric and nonpsychiatric causes of auditory, visual, olfactory, gustatory, tactile, and somatic hallucinations. Awareness of typical presentations of hallucinations associated with specific disorders can help narrow the diagnosis and provide appropriate treatment.

Auditory hallucinations

Also known as paracusia, auditory hallucinations are perceptions of sounds without identifiable external stimuli. This type of hallucination has various causes (Table 1).¹ A frequent symptom of schizophrenia, auditory hallucinations can cause substantial distress and functional disability.² Approximately 60% to 90% of patients with schizophrenia and up to 80% of those with affective psychoses experience auditory hallucinations.¹

Auditory hallucinations in psychosis usually are formed and complex.³ A common manifestation is hearing ≥1 voices. A patient might experience 2 voices talking about him in the third person. The voices may be perceived as coming from inside or outside the patient’s head. Some might hear their own thoughts spoken aloud. According to DSM-IV-TR, “hearing voices” is sufficient to diagnose schizophrenia if the hallucinations consist of a voice keeping up a running commentary on the person’s behavior or ≥2 voices conversing with each other.⁴ Auditory hallucinations also are seen in mood disorders but tend to be milder than their psychosis-induced counterparts.

Simple (unformed) auditory hallucinations—referred to as tinnitus—can be caused by disease of the middle ear (otosclerosis) or inner ear. These unformed hallucinations consist of buzzing or tones of varying pitch and timbre.¹

Partial seizures may cause auditory hallucinations. Perceptions of music have been associated with partial seizures.⁵ Curie and colleagues found that 17% of 514 patients with temporal lobe epilepsy had auditory hallucinations as a component of their seizures.⁶ These hallucinations typically are brief, stereotyped sensory impressions and, if formed, may be trivial sentences, previously heard phrases, or commands.

Alcoholic hallucinosis is a hallucinatory syndrome caused by alcohol withdrawal. These hallucinations usually are vocal and typically consist of accusatory, threatening, and/or critical voices directed at the patient.¹ Patients with alcohol hallucinosis also may experience musical auditory hallucinations.^7,8

CNS neoplasms can produce auditory hallucinations in 3% to 10% of patients.⁹ Hemorrhages and arteriovenous malformations in the pontine tegmentum and lower midbrain have been associated with acute onset of auditory hallucinations. The sounds typically are unformed mechanical or seashell-like noises or music.¹⁰

Patients with migraines rarely report auditory hallucinations. When they occur, they typically consist of perceived unilateral tinnitus, phonophobia, or hearing loss.

Table 1

Common causes of auditory hallucinations

Visual hallucinations

Visual hallucinations manifest as visual sensory perceptions in the absence of external stimuli.¹¹ These false perceptions may consist of formed images (eg, people) or unformed images (eg, flashes of light).¹² Visual hallucinations occur in numerous ophthalmologic, neurologic, medical, and psychiatric disorders (Table 2).¹³

DSM-IV-TR lists visual hallucinations as a primary diagnostic criterion for several psychotic disorders, including schizophrenia and schizoaffective disorder,⁴ and they occur in 16% to 72% of patients with these conditions.^14,15 Patients with major depressive disorder or bipolar disorder also may experience visual hallucinations. Visual hallucinations in those with schizophrenia tend to involve vivid scenes with family members, religious figures, and/or animals.¹⁶

Delirium is a transient, reversible cause of cerebral dysfunction that often presents with hallucinations. Several studies have shown that visual hallucinations are the most common type among patients with delirium. Webster and Holroyd found visual hallucinations in 27% of 227 delirium patients.¹⁷

Delirium tremens typically is accompanied by visual hallucinations. Visions of small animals and crawling insects are common.¹⁸ Hallucinations due to drug intoxication or withdrawal generally vary in duration from brief to continuous; such experiences often contribute to agitation.¹⁹

Migraines are a well-recognized cause of visual hallucinations. Up to 31% of those with migraines experience an aura, and nearly 99% of those with aura have visual symptoms.^20,21 The classic visual aura starts as an irregular colored crescent of light with multi-colored edges in the center of the visual field that gradually progresses toward the periphery, lasting <60 minutes. These simple visual hallucinations are most common; more complex hallucinations are seen more frequently in migraine coma and familial hemiplegic migraine.

Approximately 5% of patients with epilepsy have occipital seizures, which almost always have visual manifestations. Epileptic visual hallucinations often are simple, brief, stereotyped, and fragmentary. They usually consist of small, brightly colored spots or shapes that flash.²² Complex visual hallucinations in epilepsy are similar to hypnagogic hallucinations but are rare. Intracranial electroencephalography recordings have shown that pathological excitation of visual cortical areas may be responsible for complex visual hallucinations in epilepsy.¹⁹

Dementia with Lewy bodies (DLB) is associated with visual hallucinations.²³ Visual hallucinations occur in >20% of patients with DLB.²⁴ Patients with DLB may see complex scenarios of people and items that are not present. Visual hallucinations have an 83% positive predictive value for distinguishing DLB from dementia of the Alzheimer’s type.²⁵ There is a strong correlation between Lewy bodies located in the amygdala and parahippocampus and well-formed visual hallucinations.²⁶

Visual hallucinations are common in Parkinson’s disease and may occur in up to one-half of patients.²⁷ Patients with Parkinson’s disease may experience hallucinations similar to those observed in DLB, which can range from seeing a person or animal to more complex, formed, and mobile people, animals, or objects.

Table 2

Common causes of visual hallucinations

Olfactory hallucinations

Also known as phantosmia, olfactory hallucinations involve smelling odors that are not derived from any physical stimulus. They can occur with several psychiatric conditions, including schizophrenia, depression, bipolar disorder, eating disorders, and substance abuse.²⁸ Olfactory hallucinations caused by epileptic activity are rare. They constitute approximately 0.9% of all auras and typically are described as unpleasant. Tumors that affect the medial temporal lobe and mesial temporal sclerosis are associated with olfactory hallucinations.²⁹ Olfactory hallucinations also have been reported in patients with multi-infarct dementia, Alzheimer’s disease, and alcoholic psychosyndromes. In patients with schizophrenia, the smell may be perceived as coming from an external source, whereas patients with depression may perceive the source as internal.³⁰ Patients who perceive that they are the source of an offensive odor—a condition known as olfactory reference syndrome—may wash excessively, overuse deodorants and perfumes, or become socially withdrawn.³⁰

Gustatory hallucinations

Patients with gustatory hallucinations may experience salivation, sensation of thirst, or taste alterations. These hallucinations can be observed when the sylvian fissure that extends to the insula is stimulated electrically.³¹ Similar to olfactory hallucinations, gustatory hallucinations are associated with temporal lobe disease and parietal operculum lesions.^31,32 Sinus diseases have been associated with olfactory and gustatory hallucinations.³³ Brief gustatory hallucinations can be elicited with stimulation of the right rolandic operculum, parietal operculum, amygdala, hippocampus, medial temporal gyrus, and anterior part of right temporal gyrus.³⁴

Tactile hallucinations

These hallucinations may include perceptions of insects crawling over or under the skin (formication) or simulation of pressure on skin.³⁵ They have been associated with substance abuse, toxicity, or withdrawal.²⁸ Tactile hallucinations are characteristic of cocaine or amphetamine intoxication.³⁵

Tactile hallucinations are a rare symptom of schizophrenia. Heveling and colleagues reported a case of a woman, age 68, with chronic schizophrenia who experienced touching and being touched by a “shadow man” several times a day in addition to auditory and visual hallucinations.³⁶ Her symptoms disappeared after 4 weeks of antipsychotic and mood stabilizer therapy.

Tactile hallucinations have been associated with obsessive-compulsive disorder (OCD).³⁷ Fontenelle and colleagues suggested that OCD and psychotic disorders may share dysfunctional dopaminergic circuits.³⁷

Somatic hallucinations

Patients who have somatic hallucinations report perceptions of abnormal body sensations or physical experiences. For example, a patient may have sense of not having a stomach while eating.³⁵

This type of hallucination has been associated with activation of postcentral gyrus, parietal operculum, insula, and inferior parietal lobule on stereoelectroencephalography.³⁴ In a study of cerebral blood flow in 20 geriatric patients with delusional disorder, somatic type who were experiencing somatic hallucinations, positron emission testing scan demonstrated increased perfusion in somatic sensory processing regions, particularly the left postcentral gyrus and the right paracentral lobule.³⁸ Other researchers have linked somatic hallucinations with activation in the primary somatosensory and posterior parietal cortex, areas that normally mediate tactile perception.³⁹

Related Resource

• Teeple RC, Caplan JP, Stern TA. Visual hallucinations: differential diagnosis and treatment. Prim Care Companion J Clin Psychiatry. 2009;11(1):26-32.

Disclosures

Drs. Ali, Patel, Avenido, Bailey, and Jabeen report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Riley is on the board of directors for Vertex Pharmaceuticals.

Acknowledgment

The authors would like to thank Marwah Shahid, Research Associate, Vanderbilt University, Nashville, TN.

Not all patients who experience hallucinations have a psychotic disorder. Many physical and psychiatric disorders can manifest with hallucinations, and some patients have >1 disorder that could cause different types of hallucinations. To avoid providing unnecessary or ineffective treatments—and to ensure that patients receive proper care for nonpsychiatric conditions—it is important to accurately diagnose the disorder causing a patient’s hallucinations.

In this article we describe common features and psychiatric and nonpsychiatric causes of auditory, visual, olfactory, gustatory, tactile, and somatic hallucinations. Awareness of typical presentations of hallucinations associated with specific disorders can help narrow the diagnosis and provide appropriate treatment.

Auditory hallucinations

Also known as paracusia, auditory hallucinations are perceptions of sounds without identifiable external stimuli. This type of hallucination has various causes (Table 1).¹ A frequent symptom of schizophrenia, auditory hallucinations can cause substantial distress and functional disability.² Approximately 60% to 90% of patients with schizophrenia and up to 80% of those with affective psychoses experience auditory hallucinations.¹

Auditory hallucinations in psychosis usually are formed and complex.³ A common manifestation is hearing ≥1 voices. A patient might experience 2 voices talking about him in the third person. The voices may be perceived as coming from inside or outside the patient’s head. Some might hear their own thoughts spoken aloud. According to DSM-IV-TR, “hearing voices” is sufficient to diagnose schizophrenia if the hallucinations consist of a voice keeping up a running commentary on the person’s behavior or ≥2 voices conversing with each other.⁴ Auditory hallucinations also are seen in mood disorders but tend to be milder than their psychosis-induced counterparts.

Simple (unformed) auditory hallucinations—referred to as tinnitus—can be caused by disease of the middle ear (otosclerosis) or inner ear. These unformed hallucinations consist of buzzing or tones of varying pitch and timbre.¹

Partial seizures may cause auditory hallucinations. Perceptions of music have been associated with partial seizures.⁵ Curie and colleagues found that 17% of 514 patients with temporal lobe epilepsy had auditory hallucinations as a component of their seizures.⁶ These hallucinations typically are brief, stereotyped sensory impressions and, if formed, may be trivial sentences, previously heard phrases, or commands.

Alcoholic hallucinosis is a hallucinatory syndrome caused by alcohol withdrawal. These hallucinations usually are vocal and typically consist of accusatory, threatening, and/or critical voices directed at the patient.¹ Patients with alcohol hallucinosis also may experience musical auditory hallucinations.^7,8

CNS neoplasms can produce auditory hallucinations in 3% to 10% of patients.⁹ Hemorrhages and arteriovenous malformations in the pontine tegmentum and lower midbrain have been associated with acute onset of auditory hallucinations. The sounds typically are unformed mechanical or seashell-like noises or music.¹⁰

Patients with migraines rarely report auditory hallucinations. When they occur, they typically consist of perceived unilateral tinnitus, phonophobia, or hearing loss.

Table 1

Common causes of auditory hallucinations

Visual hallucinations

Visual hallucinations manifest as visual sensory perceptions in the absence of external stimuli.¹¹ These false perceptions may consist of formed images (eg, people) or unformed images (eg, flashes of light).¹² Visual hallucinations occur in numerous ophthalmologic, neurologic, medical, and psychiatric disorders (Table 2).¹³

DSM-IV-TR lists visual hallucinations as a primary diagnostic criterion for several psychotic disorders, including schizophrenia and schizoaffective disorder,⁴ and they occur in 16% to 72% of patients with these conditions.^14,15 Patients with major depressive disorder or bipolar disorder also may experience visual hallucinations. Visual hallucinations in those with schizophrenia tend to involve vivid scenes with family members, religious figures, and/or animals.¹⁶

Delirium is a transient, reversible cause of cerebral dysfunction that often presents with hallucinations. Several studies have shown that visual hallucinations are the most common type among patients with delirium. Webster and Holroyd found visual hallucinations in 27% of 227 delirium patients.¹⁷

Delirium tremens typically is accompanied by visual hallucinations. Visions of small animals and crawling insects are common.¹⁸ Hallucinations due to drug intoxication or withdrawal generally vary in duration from brief to continuous; such experiences often contribute to agitation.¹⁹

Migraines are a well-recognized cause of visual hallucinations. Up to 31% of those with migraines experience an aura, and nearly 99% of those with aura have visual symptoms.^20,21 The classic visual aura starts as an irregular colored crescent of light with multi-colored edges in the center of the visual field that gradually progresses toward the periphery, lasting <60 minutes. These simple visual hallucinations are most common; more complex hallucinations are seen more frequently in migraine coma and familial hemiplegic migraine.

Approximately 5% of patients with epilepsy have occipital seizures, which almost always have visual manifestations. Epileptic visual hallucinations often are simple, brief, stereotyped, and fragmentary. They usually consist of small, brightly colored spots or shapes that flash.²² Complex visual hallucinations in epilepsy are similar to hypnagogic hallucinations but are rare. Intracranial electroencephalography recordings have shown that pathological excitation of visual cortical areas may be responsible for complex visual hallucinations in epilepsy.¹⁹

Dementia with Lewy bodies (DLB) is associated with visual hallucinations.²³ Visual hallucinations occur in >20% of patients with DLB.²⁴ Patients with DLB may see complex scenarios of people and items that are not present. Visual hallucinations have an 83% positive predictive value for distinguishing DLB from dementia of the Alzheimer’s type.²⁵ There is a strong correlation between Lewy bodies located in the amygdala and parahippocampus and well-formed visual hallucinations.²⁶

Visual hallucinations are common in Parkinson’s disease and may occur in up to one-half of patients.²⁷ Patients with Parkinson’s disease may experience hallucinations similar to those observed in DLB, which can range from seeing a person or animal to more complex, formed, and mobile people, animals, or objects.

Table 2

Common causes of visual hallucinations

Olfactory hallucinations

Also known as phantosmia, olfactory hallucinations involve smelling odors that are not derived from any physical stimulus. They can occur with several psychiatric conditions, including schizophrenia, depression, bipolar disorder, eating disorders, and substance abuse.²⁸ Olfactory hallucinations caused by epileptic activity are rare. They constitute approximately 0.9% of all auras and typically are described as unpleasant. Tumors that affect the medial temporal lobe and mesial temporal sclerosis are associated with olfactory hallucinations.²⁹ Olfactory hallucinations also have been reported in patients with multi-infarct dementia, Alzheimer’s disease, and alcoholic psychosyndromes. In patients with schizophrenia, the smell may be perceived as coming from an external source, whereas patients with depression may perceive the source as internal.³⁰ Patients who perceive that they are the source of an offensive odor—a condition known as olfactory reference syndrome—may wash excessively, overuse deodorants and perfumes, or become socially withdrawn.³⁰

Gustatory hallucinations

Patients with gustatory hallucinations may experience salivation, sensation of thirst, or taste alterations. These hallucinations can be observed when the sylvian fissure that extends to the insula is stimulated electrically.³¹ Similar to olfactory hallucinations, gustatory hallucinations are associated with temporal lobe disease and parietal operculum lesions.^31,32 Sinus diseases have been associated with olfactory and gustatory hallucinations.³³ Brief gustatory hallucinations can be elicited with stimulation of the right rolandic operculum, parietal operculum, amygdala, hippocampus, medial temporal gyrus, and anterior part of right temporal gyrus.³⁴

Tactile hallucinations

These hallucinations may include perceptions of insects crawling over or under the skin (formication) or simulation of pressure on skin.³⁵ They have been associated with substance abuse, toxicity, or withdrawal.²⁸ Tactile hallucinations are characteristic of cocaine or amphetamine intoxication.³⁵

Tactile hallucinations are a rare symptom of schizophrenia. Heveling and colleagues reported a case of a woman, age 68, with chronic schizophrenia who experienced touching and being touched by a “shadow man” several times a day in addition to auditory and visual hallucinations.³⁶ Her symptoms disappeared after 4 weeks of antipsychotic and mood stabilizer therapy.

Tactile hallucinations have been associated with obsessive-compulsive disorder (OCD).³⁷ Fontenelle and colleagues suggested that OCD and psychotic disorders may share dysfunctional dopaminergic circuits.³⁷

Somatic hallucinations

Patients who have somatic hallucinations report perceptions of abnormal body sensations or physical experiences. For example, a patient may have sense of not having a stomach while eating.³⁵

This type of hallucination has been associated with activation of postcentral gyrus, parietal operculum, insula, and inferior parietal lobule on stereoelectroencephalography.³⁴ In a study of cerebral blood flow in 20 geriatric patients with delusional disorder, somatic type who were experiencing somatic hallucinations, positron emission testing scan demonstrated increased perfusion in somatic sensory processing regions, particularly the left postcentral gyrus and the right paracentral lobule.³⁸ Other researchers have linked somatic hallucinations with activation in the primary somatosensory and posterior parietal cortex, areas that normally mediate tactile perception.³⁹

Related Resource

• Teeple RC, Caplan JP, Stern TA. Visual hallucinations: differential diagnosis and treatment. Prim Care Companion J Clin Psychiatry. 2009;11(1):26-32.

Disclosures

Drs. Ali, Patel, Avenido, Bailey, and Jabeen report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Riley is on the board of directors for Vertex Pharmaceuticals.

Acknowledgment

The authors would like to thank Marwah Shahid, Research Associate, Vanderbilt University, Nashville, TN.

Not all patients who experience hallucinations have a psychotic disorder. Many physical and psychiatric disorders can manifest with hallucinations, and some patients have >1 disorder that could cause different types of hallucinations. To avoid providing unnecessary or ineffective treatments—and to ensure that patients receive proper care for nonpsychiatric conditions—it is important to accurately diagnose the disorder causing a patient’s hallucinations.

In this article we describe common features and psychiatric and nonpsychiatric causes of auditory, visual, olfactory, gustatory, tactile, and somatic hallucinations. Awareness of typical presentations of hallucinations associated with specific disorders can help narrow the diagnosis and provide appropriate treatment.

Auditory hallucinations

Also known as paracusia, auditory hallucinations are perceptions of sounds without identifiable external stimuli. This type of hallucination has various causes (Table 1).¹ A frequent symptom of schizophrenia, auditory hallucinations can cause substantial distress and functional disability.² Approximately 60% to 90% of patients with schizophrenia and up to 80% of those with affective psychoses experience auditory hallucinations.¹

Auditory hallucinations in psychosis usually are formed and complex.³ A common manifestation is hearing ≥1 voices. A patient might experience 2 voices talking about him in the third person. The voices may be perceived as coming from inside or outside the patient’s head. Some might hear their own thoughts spoken aloud. According to DSM-IV-TR, “hearing voices” is sufficient to diagnose schizophrenia if the hallucinations consist of a voice keeping up a running commentary on the person’s behavior or ≥2 voices conversing with each other.⁴ Auditory hallucinations also are seen in mood disorders but tend to be milder than their psychosis-induced counterparts.

Simple (unformed) auditory hallucinations—referred to as tinnitus—can be caused by disease of the middle ear (otosclerosis) or inner ear. These unformed hallucinations consist of buzzing or tones of varying pitch and timbre.¹

Partial seizures may cause auditory hallucinations. Perceptions of music have been associated with partial seizures.⁵ Curie and colleagues found that 17% of 514 patients with temporal lobe epilepsy had auditory hallucinations as a component of their seizures.⁶ These hallucinations typically are brief, stereotyped sensory impressions and, if formed, may be trivial sentences, previously heard phrases, or commands.

Alcoholic hallucinosis is a hallucinatory syndrome caused by alcohol withdrawal. These hallucinations usually are vocal and typically consist of accusatory, threatening, and/or critical voices directed at the patient.¹ Patients with alcohol hallucinosis also may experience musical auditory hallucinations.^7,8

CNS neoplasms can produce auditory hallucinations in 3% to 10% of patients.⁹ Hemorrhages and arteriovenous malformations in the pontine tegmentum and lower midbrain have been associated with acute onset of auditory hallucinations. The sounds typically are unformed mechanical or seashell-like noises or music.¹⁰

Patients with migraines rarely report auditory hallucinations. When they occur, they typically consist of perceived unilateral tinnitus, phonophobia, or hearing loss.

Table 1

Common causes of auditory hallucinations

Visual hallucinations

Visual hallucinations manifest as visual sensory perceptions in the absence of external stimuli.¹¹ These false perceptions may consist of formed images (eg, people) or unformed images (eg, flashes of light).¹² Visual hallucinations occur in numerous ophthalmologic, neurologic, medical, and psychiatric disorders (Table 2).¹³

DSM-IV-TR lists visual hallucinations as a primary diagnostic criterion for several psychotic disorders, including schizophrenia and schizoaffective disorder,⁴ and they occur in 16% to 72% of patients with these conditions.^14,15 Patients with major depressive disorder or bipolar disorder also may experience visual hallucinations. Visual hallucinations in those with schizophrenia tend to involve vivid scenes with family members, religious figures, and/or animals.¹⁶

Delirium is a transient, reversible cause of cerebral dysfunction that often presents with hallucinations. Several studies have shown that visual hallucinations are the most common type among patients with delirium. Webster and Holroyd found visual hallucinations in 27% of 227 delirium patients.¹⁷

Delirium tremens typically is accompanied by visual hallucinations. Visions of small animals and crawling insects are common.¹⁸ Hallucinations due to drug intoxication or withdrawal generally vary in duration from brief to continuous; such experiences often contribute to agitation.¹⁹

Migraines are a well-recognized cause of visual hallucinations. Up to 31% of those with migraines experience an aura, and nearly 99% of those with aura have visual symptoms.^20,21 The classic visual aura starts as an irregular colored crescent of light with multi-colored edges in the center of the visual field that gradually progresses toward the periphery, lasting <60 minutes. These simple visual hallucinations are most common; more complex hallucinations are seen more frequently in migraine coma and familial hemiplegic migraine.

Approximately 5% of patients with epilepsy have occipital seizures, which almost always have visual manifestations. Epileptic visual hallucinations often are simple, brief, stereotyped, and fragmentary. They usually consist of small, brightly colored spots or shapes that flash.²² Complex visual hallucinations in epilepsy are similar to hypnagogic hallucinations but are rare. Intracranial electroencephalography recordings have shown that pathological excitation of visual cortical areas may be responsible for complex visual hallucinations in epilepsy.¹⁹

Dementia with Lewy bodies (DLB) is associated with visual hallucinations.²³ Visual hallucinations occur in >20% of patients with DLB.²⁴ Patients with DLB may see complex scenarios of people and items that are not present. Visual hallucinations have an 83% positive predictive value for distinguishing DLB from dementia of the Alzheimer’s type.²⁵ There is a strong correlation between Lewy bodies located in the amygdala and parahippocampus and well-formed visual hallucinations.²⁶

Visual hallucinations are common in Parkinson’s disease and may occur in up to one-half of patients.²⁷ Patients with Parkinson’s disease may experience hallucinations similar to those observed in DLB, which can range from seeing a person or animal to more complex, formed, and mobile people, animals, or objects.

Table 2

Common causes of visual hallucinations

Olfactory hallucinations

Also known as phantosmia, olfactory hallucinations involve smelling odors that are not derived from any physical stimulus. They can occur with several psychiatric conditions, including schizophrenia, depression, bipolar disorder, eating disorders, and substance abuse.²⁸ Olfactory hallucinations caused by epileptic activity are rare. They constitute approximately 0.9% of all auras and typically are described as unpleasant. Tumors that affect the medial temporal lobe and mesial temporal sclerosis are associated with olfactory hallucinations.²⁹ Olfactory hallucinations also have been reported in patients with multi-infarct dementia, Alzheimer’s disease, and alcoholic psychosyndromes. In patients with schizophrenia, the smell may be perceived as coming from an external source, whereas patients with depression may perceive the source as internal.³⁰ Patients who perceive that they are the source of an offensive odor—a condition known as olfactory reference syndrome—may wash excessively, overuse deodorants and perfumes, or become socially withdrawn.³⁰

Gustatory hallucinations

Patients with gustatory hallucinations may experience salivation, sensation of thirst, or taste alterations. These hallucinations can be observed when the sylvian fissure that extends to the insula is stimulated electrically.³¹ Similar to olfactory hallucinations, gustatory hallucinations are associated with temporal lobe disease and parietal operculum lesions.^31,32 Sinus diseases have been associated with olfactory and gustatory hallucinations.³³ Brief gustatory hallucinations can be elicited with stimulation of the right rolandic operculum, parietal operculum, amygdala, hippocampus, medial temporal gyrus, and anterior part of right temporal gyrus.³⁴

Tactile hallucinations

These hallucinations may include perceptions of insects crawling over or under the skin (formication) or simulation of pressure on skin.³⁵ They have been associated with substance abuse, toxicity, or withdrawal.²⁸ Tactile hallucinations are characteristic of cocaine or amphetamine intoxication.³⁵

Tactile hallucinations are a rare symptom of schizophrenia. Heveling and colleagues reported a case of a woman, age 68, with chronic schizophrenia who experienced touching and being touched by a “shadow man” several times a day in addition to auditory and visual hallucinations.³⁶ Her symptoms disappeared after 4 weeks of antipsychotic and mood stabilizer therapy.

Tactile hallucinations have been associated with obsessive-compulsive disorder (OCD).³⁷ Fontenelle and colleagues suggested that OCD and psychotic disorders may share dysfunctional dopaminergic circuits.³⁷

Somatic hallucinations

Patients who have somatic hallucinations report perceptions of abnormal body sensations or physical experiences. For example, a patient may have sense of not having a stomach while eating.³⁵

This type of hallucination has been associated with activation of postcentral gyrus, parietal operculum, insula, and inferior parietal lobule on stereoelectroencephalography.³⁴ In a study of cerebral blood flow in 20 geriatric patients with delusional disorder, somatic type who were experiencing somatic hallucinations, positron emission testing scan demonstrated increased perfusion in somatic sensory processing regions, particularly the left postcentral gyrus and the right paracentral lobule.³⁸ Other researchers have linked somatic hallucinations with activation in the primary somatosensory and posterior parietal cortex, areas that normally mediate tactile perception.³⁹

Related Resource

• Teeple RC, Caplan JP, Stern TA. Visual hallucinations: differential diagnosis and treatment. Prim Care Companion J Clin Psychiatry. 2009;11(1):26-32.

Disclosures

Drs. Ali, Patel, Avenido, Bailey, and Jabeen report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Riley is on the board of directors for Vertex Pharmaceuticals.

Acknowledgment

The authors would like to thank Marwah Shahid, Research Associate, Vanderbilt University, Nashville, TN.

Discuss this article at www.facebook.com/CurrentPsychiatry

Patients with bipolar disorder (BD) have a high risk for suicidal ideation, suicide attempts, and suicide.^1-3 Approximately 25% to 50% of BD patients attempt suicide at least once, and their attempts often are lethal—the ratio of attempts to completed suicides in BD patients is 3:1, compared with 30:1 in the general population.⁴ Lithium has been shown to effectively stabilize BD patients’ mood and significantly reduce the rates of suicide attempts and completed suicides,^5-9 but does not reduce BD patients’ long-term suicide risk to that of the general population.

Literature on psychotherapeutic treatments for patients with BD primarily focuses on improving patients’ adherence to pharmacotherapy and achieving faster recovery and remission.¹⁰ Nonpharmacologic treatments for patients with BD include psychoeducation, family-focused psychoeducation, cognitive therapy, and interpersonal and social rhythm therapy (Table 1).¹¹ Literature on nonpharmacologic treatments to address suicidality in BD patients is limited,^12,13 and additional psychotherapeutic interventions to reduce suicide risk in BD patients are needed.¹⁴

In this article, I describe a novel psychotherapeutic intervention I use that integrates cognitive therapy principles with ideas derived from the psychosynthesis model.^15,16 It consists of teaching patients to “disidentify” from suicidal thoughts, followed by a guided-imagery exercise in which patients experience a future positive life event with all 5 senses and internalize this experience. This creates a “hook into the future” that changes the present to match the future event and acts as an antidote to suicidal thoughts. I have used this strategy successfully in many patients as an adjunct to pharmacotherapy.

Table 1

Nonpharmacologic interventions for bipolar disorder

A theoretical model

Roberto Assagioli, who established the approach to psychology called psychosynthesis, formulated a fundamental psychological principle in controlling one’s behavior: “We are dominated by everything with which our self becomes identified. We can dominate and control everything from which we disidentify ourselves.”¹⁵ According to the psychosynthesis model, it is easier to change thoughts we identify as foreign to “the self” (ego-dystonic) than thoughts we identify as being part of “the self” (ego-syntonic).

Patients whose suicidal thoughts are ego-syntonic identify with the thoughts as representing themselves and take ownership of these thoughts. Such patients are at a greater risk of acting on suicidal thoughts.

Patients whose suicidal thoughts are ego-dystonic consider the suicidal thoughts foreign to their core self and do not believe such thoughts represent them. In essence, they “disown” the thoughts and typically want to control and eliminate them. Examples of patients’ ego-syntonic vs ego-dystonic suicidal thoughts are listed in Table 2.

This construct calls for an intervention to help patients who have ego-syntonic suicidal thoughts restructure them as a manifestation of BD, rather then the patient’s core self belief. The intervention emphasizes the patient is not “a suicidal patient” but suffers from an illness that may manifest with suicidal ideation. Many BD patients overly identify with their disease, stating, “I am bipolar” or “I am suicidal.” The “I am” statement originates from the verb “to be,” which implies the disease is part of the patient’s identity. The goal of this intervention is to help the patient learn to disidentify from the disease and decide that suicidal thoughts do not represent their core self, but are a manifestation of the underlying disease.

The psychosynthesis model of helping patients disidentify and therefore disown suicidal thoughts is compatible with interventions that use mindfulness-based cognitive therapy training to teach patients to experience their thoughts as just passing through their consciousness without taking ownership of them.¹⁷

Table 2

Examples of ego-syntonic vs ego-dystonic suicidal thoughts

The intervention

Assessment of suicidality is a fundamental skill for every mental health clinician.¹⁸ The psychotherapeutic intervention I use integrates the cognitive therapy principles of reframing, relabeling, and restructuring patients’ thoughts with disidentification from dysfunctional thoughts, feelings, and desires, based on psychosynthesis principles.

First, I conduct a comprehensive mental status examination that includes an in-depth exploration of the patient’s suicidal thoughts to determine if they are ego-syntonic or ego-dystonic. I begin by asking patients to clarify and elaborate on their statements referring to suicide, asking questions such as “Is there a part of you that objects to these thoughts?” and “Is there a part of you that wants to live?” If a patient indicates that he or she does experience inner conflict regarding such thoughts, these thoughts are classified as ego-dystonic. If a patient does not have any counter thoughts regarding the suicidal thoughts and fully identifies with them, the thoughts are classified as ego-syntonic.

I follow this with a treatment plan that helps patients change their view of their suicidal thoughts. I ask the patient to change these suicidal thoughts to ego-dystonic by focusing on the following statement: “I, (patient’s name), am a human being and like all human beings, I have thoughts; however, I am not my thoughts, I am much more than that.” I ask my patient to read this out loud and to mindfully meditate on this statement several times a day to reinforce the new understanding that these suicidal thoughts are a manifestation of the chemical imbalance of the mood disorder, and do not represent the patient as a person.

This intervention is paired with a future-focused internalized imagery experience I have described in previous articles.^19,20 In this part of the treatment, the patient and I discuss a specific expected life milestone that is positive and for which he or she would want to be present (eg, children graduating from high school or college, a wedding, birth of a child/grandchild, etc.). Using guided imagery, the patient experiences this event with all 5 senses during the session. I instruct the patient to internalize the experience and bring it back from the future to the present. This creates a “hook into the future” that is coupled with this desired milestone event in the patient’s life.

The following 3 case studies provide examples of the application of this treatment intervention.

CASE 1: Disidentifying family history

Mrs. G, a 42-year-old mother of 2, suffers from bipolar II disorder with recurrent episodes of depression associated with ego-syntonic suicidal thoughts. She states that at times she feels she is a burden to her husband and children and believes they may be better off without her. She says she believes “ending it all” must be her destiny. After further investigation, I learn Mrs. G has a family history of BD and 3 relatives have committed suicide. This family history may partially explain her belief that suicide must be “in her genes.”

I discuss with Mrs. G the strategy of changing her thoughts. I tell her to write in her journal—which she brings to her sessions—the following statements: “I am a human being. I am an adult woman and mother of 2 children. I know I have thoughts but I am not my thoughts, I am much more than that. I know I have genes but I am not my genes, I am much more than that. I know I have feelings, but I am not my feelings, I am much more than that. I know I have cousins, uncles, aunts, and other relatives but I am not my relatives. I am uniquely myself, different from the others.”

I ask Mrs. G to read these statements out loud and repeat them several times a day to reinforce this new way of perceiving the suicidal thoughts and to disidentify from the thoughts and her family history as it relates to suicide.

Mrs. G and I talk about the future and expected family milestones. When I ask if her son would want her to be present at his college graduation, she says yes. We then discuss in detail the date, time of day, and location of this event, followed by a guided imagery exercise focused on the graduation. She is guided to experience this event with all 5 senses and describes the event in detail, including the expression on the faces of her husband and children, their voices, and the scent of their aftershave lotion. She hears her son saying, “Mom, I love you. Thank you for being there with me all these years. I could not have done it without you.” I ask Mrs. G to internalize these experiences and carve them into her memory. She is instructed to come back from this future-focused guided imagery experience. When her eyes open, she looks at me and describes her experience in great detail, at times using the past tense, which confirms that the future-focused event was internalized.

In her next session, Mrs. G reports an improvement in her sleep and a change in her suicidal thoughts, which now are only fleeting.

CASE 2: Experiencing graduation

Ms. J, age 17, was diagnosed with bipolar I disorder when she was 15. She has a family history of BD in her mother, 2 maternal aunts, her grandmother, and an older sister. All these women have a history of suicidal thoughts and suicide attempts requiring hospital treatment, but no completed suicides.

Ms. J has been taking an adequate combination of mood stabilizers. She has recovered from 2 previous depressive episodes and is experiencing a third relapse with suicidal thoughts. At times, she experiences these thoughts as ego-syntonic; at other times, they are ego-dystonic.

I first educate her about the nature of BD, explaining that her suicidal thoughts are a manifestation of a chemical imbalance in her brain as a result of the depressive relapse. I teach her to use guided imagery to focus on her favorite place of peace and serenity, the beach, which produces immediate relief of the intense anxiety she felt.

After we complete the disidentification exercise, I ask her to focus on her high school graduation ceremony, which is scheduled to take place in 1 year. In a state of guided imagery, she experiences her graduation from high school with all 5 senses. As she returns to a state of full alertness with her eyes open, she describes the graduation ceremony experience in detail using the past tense, as if it had already occurred, thereby creating her own hook into the future. I instruct her to write about this experience in her journal and bring it with her to the next session.

The following session, Ms. J reports that her suicide ideations have “disappeared.” She says this was accompanied by improvements in her overall mood and sleep.

CASE 3: Internalizing the future

Mr. C, a 38-year-old married father of 4 children, has bipolar II disorder and is in a depressed state. He has been treated with optimal doses of mood stabilizers and atypical antipsychotics but continues to have suicidal thoughts. These thoughts are at times ego-syntonic; he says, “My family would be better off without me.” When Mr. C’s mood improves, however, the suicidal thoughts become more ego-dystonic; he expresses fear that he might act out on the thoughts and states that he does not want to die, he really wants to live and get better. He has no history of suicide attempts.

During our session, I ask Mr. C to focus on a new perspective to understand his thoughts by repeating the following statements: “I, JC, am a human being. I know I have a bipolar mood disorder; however, I must remember I am not bipolar. I have suicidal thoughts; however, I am not my thoughts, I am much more than that. I know I want to live, to heal, and to get better. I want to be alive and well so I can see and participate in my children’s graduation from high school and be there when they get married and when my grandchildren are born.”

I teach Mr. C to use guided imagery, during which he experiences such future positive images and milestones in his life in all 5 senses and internalizes them by using the “back from the future” technique.¹⁷ By the end of the session, he reports feeling better, more hopeful, and confident in his abilities to control his suicidal thoughts. I instruct him to write in his diary about his experiences with the future-focused positive milestones and to bring this assignment to his next appointment.

At his next appointment, Mr. C reports that his suicidal thoughts have become more fleeting, lasting for 10 to 30 seconds, and then spontaneously change to focus on issues of the “here and now.” When I ask him to read what he’s written, what stands out is the use of past tense verbs to describe future-focused experiences. For me, this confirms that Mr. C has internalized the future, creating the desirable “future hook” that acts as an antidote to the suicidal thoughts.

Related Resources

• Rouget BW, Aubry JM. Efficacy of psychoeducational approaches on bipolar disorders: a review of the literature. J Affect Disord. 2007;98:11-27.
• Weinberg I, Ronningstam E, Goldblatt MJ, et al. Strategies in treatment of suicidality: identification of common and treatment-specific interventions in empirically supported treatment manuals. J Clin Psychiatry. 2010;71:699-706.

Drug Brand Name

• Lithium • Eskalith, Lithobid

Disclosure

Dr. Torem reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Discuss this article at www.facebook.com/CurrentPsychiatry

Patients with bipolar disorder (BD) have a high risk for suicidal ideation, suicide attempts, and suicide.^1-3 Approximately 25% to 50% of BD patients attempt suicide at least once, and their attempts often are lethal—the ratio of attempts to completed suicides in BD patients is 3:1, compared with 30:1 in the general population.⁴ Lithium has been shown to effectively stabilize BD patients’ mood and significantly reduce the rates of suicide attempts and completed suicides,^5-9 but does not reduce BD patients’ long-term suicide risk to that of the general population.

Literature on psychotherapeutic treatments for patients with BD primarily focuses on improving patients’ adherence to pharmacotherapy and achieving faster recovery and remission.¹⁰ Nonpharmacologic treatments for patients with BD include psychoeducation, family-focused psychoeducation, cognitive therapy, and interpersonal and social rhythm therapy (Table 1).¹¹ Literature on nonpharmacologic treatments to address suicidality in BD patients is limited,^12,13 and additional psychotherapeutic interventions to reduce suicide risk in BD patients are needed.¹⁴

In this article, I describe a novel psychotherapeutic intervention I use that integrates cognitive therapy principles with ideas derived from the psychosynthesis model.^15,16 It consists of teaching patients to “disidentify” from suicidal thoughts, followed by a guided-imagery exercise in which patients experience a future positive life event with all 5 senses and internalize this experience. This creates a “hook into the future” that changes the present to match the future event and acts as an antidote to suicidal thoughts. I have used this strategy successfully in many patients as an adjunct to pharmacotherapy.

Table 1

Nonpharmacologic interventions for bipolar disorder

A theoretical model

Roberto Assagioli, who established the approach to psychology called psychosynthesis, formulated a fundamental psychological principle in controlling one’s behavior: “We are dominated by everything with which our self becomes identified. We can dominate and control everything from which we disidentify ourselves.”¹⁵ According to the psychosynthesis model, it is easier to change thoughts we identify as foreign to “the self” (ego-dystonic) than thoughts we identify as being part of “the self” (ego-syntonic).

Patients whose suicidal thoughts are ego-syntonic identify with the thoughts as representing themselves and take ownership of these thoughts. Such patients are at a greater risk of acting on suicidal thoughts.

Patients whose suicidal thoughts are ego-dystonic consider the suicidal thoughts foreign to their core self and do not believe such thoughts represent them. In essence, they “disown” the thoughts and typically want to control and eliminate them. Examples of patients’ ego-syntonic vs ego-dystonic suicidal thoughts are listed in Table 2.

This construct calls for an intervention to help patients who have ego-syntonic suicidal thoughts restructure them as a manifestation of BD, rather then the patient’s core self belief. The intervention emphasizes the patient is not “a suicidal patient” but suffers from an illness that may manifest with suicidal ideation. Many BD patients overly identify with their disease, stating, “I am bipolar” or “I am suicidal.” The “I am” statement originates from the verb “to be,” which implies the disease is part of the patient’s identity. The goal of this intervention is to help the patient learn to disidentify from the disease and decide that suicidal thoughts do not represent their core self, but are a manifestation of the underlying disease.

The psychosynthesis model of helping patients disidentify and therefore disown suicidal thoughts is compatible with interventions that use mindfulness-based cognitive therapy training to teach patients to experience their thoughts as just passing through their consciousness without taking ownership of them.¹⁷

Table 2

Examples of ego-syntonic vs ego-dystonic suicidal thoughts

The intervention

Assessment of suicidality is a fundamental skill for every mental health clinician.¹⁸ The psychotherapeutic intervention I use integrates the cognitive therapy principles of reframing, relabeling, and restructuring patients’ thoughts with disidentification from dysfunctional thoughts, feelings, and desires, based on psychosynthesis principles.

First, I conduct a comprehensive mental status examination that includes an in-depth exploration of the patient’s suicidal thoughts to determine if they are ego-syntonic or ego-dystonic. I begin by asking patients to clarify and elaborate on their statements referring to suicide, asking questions such as “Is there a part of you that objects to these thoughts?” and “Is there a part of you that wants to live?” If a patient indicates that he or she does experience inner conflict regarding such thoughts, these thoughts are classified as ego-dystonic. If a patient does not have any counter thoughts regarding the suicidal thoughts and fully identifies with them, the thoughts are classified as ego-syntonic.

I follow this with a treatment plan that helps patients change their view of their suicidal thoughts. I ask the patient to change these suicidal thoughts to ego-dystonic by focusing on the following statement: “I, (patient’s name), am a human being and like all human beings, I have thoughts; however, I am not my thoughts, I am much more than that.” I ask my patient to read this out loud and to mindfully meditate on this statement several times a day to reinforce the new understanding that these suicidal thoughts are a manifestation of the chemical imbalance of the mood disorder, and do not represent the patient as a person.

This intervention is paired with a future-focused internalized imagery experience I have described in previous articles.^19,20 In this part of the treatment, the patient and I discuss a specific expected life milestone that is positive and for which he or she would want to be present (eg, children graduating from high school or college, a wedding, birth of a child/grandchild, etc.). Using guided imagery, the patient experiences this event with all 5 senses during the session. I instruct the patient to internalize the experience and bring it back from the future to the present. This creates a “hook into the future” that is coupled with this desired milestone event in the patient’s life.

The following 3 case studies provide examples of the application of this treatment intervention.

CASE 1: Disidentifying family history

Mrs. G, a 42-year-old mother of 2, suffers from bipolar II disorder with recurrent episodes of depression associated with ego-syntonic suicidal thoughts. She states that at times she feels she is a burden to her husband and children and believes they may be better off without her. She says she believes “ending it all” must be her destiny. After further investigation, I learn Mrs. G has a family history of BD and 3 relatives have committed suicide. This family history may partially explain her belief that suicide must be “in her genes.”

I discuss with Mrs. G the strategy of changing her thoughts. I tell her to write in her journal—which she brings to her sessions—the following statements: “I am a human being. I am an adult woman and mother of 2 children. I know I have thoughts but I am not my thoughts, I am much more than that. I know I have genes but I am not my genes, I am much more than that. I know I have feelings, but I am not my feelings, I am much more than that. I know I have cousins, uncles, aunts, and other relatives but I am not my relatives. I am uniquely myself, different from the others.”

I ask Mrs. G to read these statements out loud and repeat them several times a day to reinforce this new way of perceiving the suicidal thoughts and to disidentify from the thoughts and her family history as it relates to suicide.

Mrs. G and I talk about the future and expected family milestones. When I ask if her son would want her to be present at his college graduation, she says yes. We then discuss in detail the date, time of day, and location of this event, followed by a guided imagery exercise focused on the graduation. She is guided to experience this event with all 5 senses and describes the event in detail, including the expression on the faces of her husband and children, their voices, and the scent of their aftershave lotion. She hears her son saying, “Mom, I love you. Thank you for being there with me all these years. I could not have done it without you.” I ask Mrs. G to internalize these experiences and carve them into her memory. She is instructed to come back from this future-focused guided imagery experience. When her eyes open, she looks at me and describes her experience in great detail, at times using the past tense, which confirms that the future-focused event was internalized.

In her next session, Mrs. G reports an improvement in her sleep and a change in her suicidal thoughts, which now are only fleeting.

CASE 2: Experiencing graduation

Ms. J, age 17, was diagnosed with bipolar I disorder when she was 15. She has a family history of BD in her mother, 2 maternal aunts, her grandmother, and an older sister. All these women have a history of suicidal thoughts and suicide attempts requiring hospital treatment, but no completed suicides.

Ms. J has been taking an adequate combination of mood stabilizers. She has recovered from 2 previous depressive episodes and is experiencing a third relapse with suicidal thoughts. At times, she experiences these thoughts as ego-syntonic; at other times, they are ego-dystonic.

I first educate her about the nature of BD, explaining that her suicidal thoughts are a manifestation of a chemical imbalance in her brain as a result of the depressive relapse. I teach her to use guided imagery to focus on her favorite place of peace and serenity, the beach, which produces immediate relief of the intense anxiety she felt.

After we complete the disidentification exercise, I ask her to focus on her high school graduation ceremony, which is scheduled to take place in 1 year. In a state of guided imagery, she experiences her graduation from high school with all 5 senses. As she returns to a state of full alertness with her eyes open, she describes the graduation ceremony experience in detail using the past tense, as if it had already occurred, thereby creating her own hook into the future. I instruct her to write about this experience in her journal and bring it with her to the next session.

The following session, Ms. J reports that her suicide ideations have “disappeared.” She says this was accompanied by improvements in her overall mood and sleep.

CASE 3: Internalizing the future

Mr. C, a 38-year-old married father of 4 children, has bipolar II disorder and is in a depressed state. He has been treated with optimal doses of mood stabilizers and atypical antipsychotics but continues to have suicidal thoughts. These thoughts are at times ego-syntonic; he says, “My family would be better off without me.” When Mr. C’s mood improves, however, the suicidal thoughts become more ego-dystonic; he expresses fear that he might act out on the thoughts and states that he does not want to die, he really wants to live and get better. He has no history of suicide attempts.

During our session, I ask Mr. C to focus on a new perspective to understand his thoughts by repeating the following statements: “I, JC, am a human being. I know I have a bipolar mood disorder; however, I must remember I am not bipolar. I have suicidal thoughts; however, I am not my thoughts, I am much more than that. I know I want to live, to heal, and to get better. I want to be alive and well so I can see and participate in my children’s graduation from high school and be there when they get married and when my grandchildren are born.”

I teach Mr. C to use guided imagery, during which he experiences such future positive images and milestones in his life in all 5 senses and internalizes them by using the “back from the future” technique.¹⁷ By the end of the session, he reports feeling better, more hopeful, and confident in his abilities to control his suicidal thoughts. I instruct him to write in his diary about his experiences with the future-focused positive milestones and to bring this assignment to his next appointment.

At his next appointment, Mr. C reports that his suicidal thoughts have become more fleeting, lasting for 10 to 30 seconds, and then spontaneously change to focus on issues of the “here and now.” When I ask him to read what he’s written, what stands out is the use of past tense verbs to describe future-focused experiences. For me, this confirms that Mr. C has internalized the future, creating the desirable “future hook” that acts as an antidote to the suicidal thoughts.

Related Resources

• Rouget BW, Aubry JM. Efficacy of psychoeducational approaches on bipolar disorders: a review of the literature. J Affect Disord. 2007;98:11-27.
• Weinberg I, Ronningstam E, Goldblatt MJ, et al. Strategies in treatment of suicidality: identification of common and treatment-specific interventions in empirically supported treatment manuals. J Clin Psychiatry. 2010;71:699-706.

Drug Brand Name

• Lithium • Eskalith, Lithobid

Disclosure

Dr. Torem reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Discuss this article at www.facebook.com/CurrentPsychiatry

Patients with bipolar disorder (BD) have a high risk for suicidal ideation, suicide attempts, and suicide.^1-3 Approximately 25% to 50% of BD patients attempt suicide at least once, and their attempts often are lethal—the ratio of attempts to completed suicides in BD patients is 3:1, compared with 30:1 in the general population.⁴ Lithium has been shown to effectively stabilize BD patients’ mood and significantly reduce the rates of suicide attempts and completed suicides,^5-9 but does not reduce BD patients’ long-term suicide risk to that of the general population.

Literature on psychotherapeutic treatments for patients with BD primarily focuses on improving patients’ adherence to pharmacotherapy and achieving faster recovery and remission.¹⁰ Nonpharmacologic treatments for patients with BD include psychoeducation, family-focused psychoeducation, cognitive therapy, and interpersonal and social rhythm therapy (Table 1).¹¹ Literature on nonpharmacologic treatments to address suicidality in BD patients is limited,^12,13 and additional psychotherapeutic interventions to reduce suicide risk in BD patients are needed.¹⁴

In this article, I describe a novel psychotherapeutic intervention I use that integrates cognitive therapy principles with ideas derived from the psychosynthesis model.^15,16 It consists of teaching patients to “disidentify” from suicidal thoughts, followed by a guided-imagery exercise in which patients experience a future positive life event with all 5 senses and internalize this experience. This creates a “hook into the future” that changes the present to match the future event and acts as an antidote to suicidal thoughts. I have used this strategy successfully in many patients as an adjunct to pharmacotherapy.

Table 1

Nonpharmacologic interventions for bipolar disorder

A theoretical model

Roberto Assagioli, who established the approach to psychology called psychosynthesis, formulated a fundamental psychological principle in controlling one’s behavior: “We are dominated by everything with which our self becomes identified. We can dominate and control everything from which we disidentify ourselves.”¹⁵ According to the psychosynthesis model, it is easier to change thoughts we identify as foreign to “the self” (ego-dystonic) than thoughts we identify as being part of “the self” (ego-syntonic).

Patients whose suicidal thoughts are ego-syntonic identify with the thoughts as representing themselves and take ownership of these thoughts. Such patients are at a greater risk of acting on suicidal thoughts.

Patients whose suicidal thoughts are ego-dystonic consider the suicidal thoughts foreign to their core self and do not believe such thoughts represent them. In essence, they “disown” the thoughts and typically want to control and eliminate them. Examples of patients’ ego-syntonic vs ego-dystonic suicidal thoughts are listed in Table 2.

This construct calls for an intervention to help patients who have ego-syntonic suicidal thoughts restructure them as a manifestation of BD, rather then the patient’s core self belief. The intervention emphasizes the patient is not “a suicidal patient” but suffers from an illness that may manifest with suicidal ideation. Many BD patients overly identify with their disease, stating, “I am bipolar” or “I am suicidal.” The “I am” statement originates from the verb “to be,” which implies the disease is part of the patient’s identity. The goal of this intervention is to help the patient learn to disidentify from the disease and decide that suicidal thoughts do not represent their core self, but are a manifestation of the underlying disease.

The psychosynthesis model of helping patients disidentify and therefore disown suicidal thoughts is compatible with interventions that use mindfulness-based cognitive therapy training to teach patients to experience their thoughts as just passing through their consciousness without taking ownership of them.¹⁷

Table 2

Examples of ego-syntonic vs ego-dystonic suicidal thoughts

The intervention

Assessment of suicidality is a fundamental skill for every mental health clinician.¹⁸ The psychotherapeutic intervention I use integrates the cognitive therapy principles of reframing, relabeling, and restructuring patients’ thoughts with disidentification from dysfunctional thoughts, feelings, and desires, based on psychosynthesis principles.

First, I conduct a comprehensive mental status examination that includes an in-depth exploration of the patient’s suicidal thoughts to determine if they are ego-syntonic or ego-dystonic. I begin by asking patients to clarify and elaborate on their statements referring to suicide, asking questions such as “Is there a part of you that objects to these thoughts?” and “Is there a part of you that wants to live?” If a patient indicates that he or she does experience inner conflict regarding such thoughts, these thoughts are classified as ego-dystonic. If a patient does not have any counter thoughts regarding the suicidal thoughts and fully identifies with them, the thoughts are classified as ego-syntonic.

I follow this with a treatment plan that helps patients change their view of their suicidal thoughts. I ask the patient to change these suicidal thoughts to ego-dystonic by focusing on the following statement: “I, (patient’s name), am a human being and like all human beings, I have thoughts; however, I am not my thoughts, I am much more than that.” I ask my patient to read this out loud and to mindfully meditate on this statement several times a day to reinforce the new understanding that these suicidal thoughts are a manifestation of the chemical imbalance of the mood disorder, and do not represent the patient as a person.

This intervention is paired with a future-focused internalized imagery experience I have described in previous articles.^19,20 In this part of the treatment, the patient and I discuss a specific expected life milestone that is positive and for which he or she would want to be present (eg, children graduating from high school or college, a wedding, birth of a child/grandchild, etc.). Using guided imagery, the patient experiences this event with all 5 senses during the session. I instruct the patient to internalize the experience and bring it back from the future to the present. This creates a “hook into the future” that is coupled with this desired milestone event in the patient’s life.

The following 3 case studies provide examples of the application of this treatment intervention.

CASE 1: Disidentifying family history

Mrs. G, a 42-year-old mother of 2, suffers from bipolar II disorder with recurrent episodes of depression associated with ego-syntonic suicidal thoughts. She states that at times she feels she is a burden to her husband and children and believes they may be better off without her. She says she believes “ending it all” must be her destiny. After further investigation, I learn Mrs. G has a family history of BD and 3 relatives have committed suicide. This family history may partially explain her belief that suicide must be “in her genes.”

I discuss with Mrs. G the strategy of changing her thoughts. I tell her to write in her journal—which she brings to her sessions—the following statements: “I am a human being. I am an adult woman and mother of 2 children. I know I have thoughts but I am not my thoughts, I am much more than that. I know I have genes but I am not my genes, I am much more than that. I know I have feelings, but I am not my feelings, I am much more than that. I know I have cousins, uncles, aunts, and other relatives but I am not my relatives. I am uniquely myself, different from the others.”

I ask Mrs. G to read these statements out loud and repeat them several times a day to reinforce this new way of perceiving the suicidal thoughts and to disidentify from the thoughts and her family history as it relates to suicide.

Mrs. G and I talk about the future and expected family milestones. When I ask if her son would want her to be present at his college graduation, she says yes. We then discuss in detail the date, time of day, and location of this event, followed by a guided imagery exercise focused on the graduation. She is guided to experience this event with all 5 senses and describes the event in detail, including the expression on the faces of her husband and children, their voices, and the scent of their aftershave lotion. She hears her son saying, “Mom, I love you. Thank you for being there with me all these years. I could not have done it without you.” I ask Mrs. G to internalize these experiences and carve them into her memory. She is instructed to come back from this future-focused guided imagery experience. When her eyes open, she looks at me and describes her experience in great detail, at times using the past tense, which confirms that the future-focused event was internalized.

In her next session, Mrs. G reports an improvement in her sleep and a change in her suicidal thoughts, which now are only fleeting.

CASE 2: Experiencing graduation

Ms. J, age 17, was diagnosed with bipolar I disorder when she was 15. She has a family history of BD in her mother, 2 maternal aunts, her grandmother, and an older sister. All these women have a history of suicidal thoughts and suicide attempts requiring hospital treatment, but no completed suicides.

Ms. J has been taking an adequate combination of mood stabilizers. She has recovered from 2 previous depressive episodes and is experiencing a third relapse with suicidal thoughts. At times, she experiences these thoughts as ego-syntonic; at other times, they are ego-dystonic.

I first educate her about the nature of BD, explaining that her suicidal thoughts are a manifestation of a chemical imbalance in her brain as a result of the depressive relapse. I teach her to use guided imagery to focus on her favorite place of peace and serenity, the beach, which produces immediate relief of the intense anxiety she felt.

After we complete the disidentification exercise, I ask her to focus on her high school graduation ceremony, which is scheduled to take place in 1 year. In a state of guided imagery, she experiences her graduation from high school with all 5 senses. As she returns to a state of full alertness with her eyes open, she describes the graduation ceremony experience in detail using the past tense, as if it had already occurred, thereby creating her own hook into the future. I instruct her to write about this experience in her journal and bring it with her to the next session.

The following session, Ms. J reports that her suicide ideations have “disappeared.” She says this was accompanied by improvements in her overall mood and sleep.

CASE 3: Internalizing the future

Mr. C, a 38-year-old married father of 4 children, has bipolar II disorder and is in a depressed state. He has been treated with optimal doses of mood stabilizers and atypical antipsychotics but continues to have suicidal thoughts. These thoughts are at times ego-syntonic; he says, “My family would be better off without me.” When Mr. C’s mood improves, however, the suicidal thoughts become more ego-dystonic; he expresses fear that he might act out on the thoughts and states that he does not want to die, he really wants to live and get better. He has no history of suicide attempts.

During our session, I ask Mr. C to focus on a new perspective to understand his thoughts by repeating the following statements: “I, JC, am a human being. I know I have a bipolar mood disorder; however, I must remember I am not bipolar. I have suicidal thoughts; however, I am not my thoughts, I am much more than that. I know I want to live, to heal, and to get better. I want to be alive and well so I can see and participate in my children’s graduation from high school and be there when they get married and when my grandchildren are born.”

I teach Mr. C to use guided imagery, during which he experiences such future positive images and milestones in his life in all 5 senses and internalizes them by using the “back from the future” technique.¹⁷ By the end of the session, he reports feeling better, more hopeful, and confident in his abilities to control his suicidal thoughts. I instruct him to write in his diary about his experiences with the future-focused positive milestones and to bring this assignment to his next appointment.

At his next appointment, Mr. C reports that his suicidal thoughts have become more fleeting, lasting for 10 to 30 seconds, and then spontaneously change to focus on issues of the “here and now.” When I ask him to read what he’s written, what stands out is the use of past tense verbs to describe future-focused experiences. For me, this confirms that Mr. C has internalized the future, creating the desirable “future hook” that acts as an antidote to the suicidal thoughts.

Related Resources

• Rouget BW, Aubry JM. Efficacy of psychoeducational approaches on bipolar disorders: a review of the literature. J Affect Disord. 2007;98:11-27.
• Weinberg I, Ronningstam E, Goldblatt MJ, et al. Strategies in treatment of suicidality: identification of common and treatment-specific interventions in empirically supported treatment manuals. J Clin Psychiatry. 2010;71:699-706.

Drug Brand Name

• Lithium • Eskalith, Lithobid

Disclosure

Dr. Torem reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Introduced into clinical practice more than a half century ago, antipsychotics are still the mainstay of schizophrenia treatment. However, from the earliest reports, antipsychotic efficacy was seemingly inseparable from extrapyramidal side effects (EPS) that manifested as acute and chronic involuntary movement disorders. Although acute extrapyramidal side effects could be prevented and treated, the late-arising symptoms of tardive dyskinesia (TD) seemed irreversible in most cases.

Concerns over TD stimulated extensive research and fueled efforts to develop new antipsychotics that spared the extrapyramidal motor system. Numerous industry-sponsored trials found a reduced risk of EPS—including TD—with newer, second-generation antipsychotics (SGAs), although this advantage diminished when modest doses of low- or mid-potency first-generation antipsychotics (FGAs) were used as the comparator.^1-3 Nevertheless, in addition to the continued potential risk of introducing new cases of TD—even with SGAs—several other factors underscore the need to develop a rational strategy for clinical management of TD, including:

• thousands of patients are left with TD as a legacy of past treatment
• the neurophysiologic mechanisms underlying TD are not well understood
• there is no uniformly effective treatment to reverse TD
• TD may be irreversible in most cases.

Prevention

Because there is no “gold standard” treatment for TD, it is important to minimize the risk of TD by taking preventive measures and detecting incipient signs of the disorder. Preventive principles include:

• confirming and documenting the indication for antipsychotics
• using conservative maintenance doses and opting for lower potency or newer agents
• informing patients and caregivers of risk
• assessing for incipient signs of TD using the Abnormal Involuntary Movement Scale (AIMS),⁴ which should be administered at least every 3 to 6 months.

Confirming the diagnosis

TD presents as a polymorphous involuntary movement disorder,^5-8 most often with nonrhythmic, repetitive, purposeless hyperkinetic symptoms. It usually affects orofacial and lingual musculature (“buccolinguomasticatory syndrome”) with chewing; bruxism; protrusion, curling, or twisting of the tongue; lip smacking, puckering, sucking, and pursing; retraction, grimacing or bridling of the mouth; bulging of the cheeks; or eye blinking and blepharospasm. Choreoathetoid movements of the fingers, hands, or upper or lower extremities also are common. Patients may experience axial symptoms affecting the neck, shoulders, spine, or pelvis. When severe, dyskinesias can affect breathing, swallowing, or speech, and interfere with walking and activities of daily living.

TD may present with nonchoreoathetoid symptoms that can be difficult to distinguish from acute EPS. These may co-exist with classic TD symptoms, but may represent separate subtypes with increased risk of progression, persistence, and severe disability. For example, tardive dystonia, which is estimated to occur in 1% to 4% of patients treated with antipsychotics,⁹ may be more generalized and disabling than TD, and may respond to anticholinergic agents. Akathisia and other movement disorders also occur as tardive variants.¹⁰

Multiple diagnostic schemes for TD have been proposed; criteria proposed by Schooler and Kane have been widely accepted (Table 1).¹¹ TD onset occurs insidiously over ≥3 months of antipsychotic treatment and may begin with tic-like movements or increased eye blinking. TD often is suppressed or masked by ongoing antipsychotic treatment and becomes apparent only when the drug is reduced, switched, or discontinued. Dyskinesias increase with emotional arousal, activation, or distraction, and diminish with relaxation, sleep, or volitional effort. As a result, TD symptoms fluctuate over time; therefore, repeated measurements are necessary for reliable assessment of severity and persistence.

The differential diagnosis of TD necessitates conducting a careful medical and neurologic evaluation of all patients with new-onset movement disorders. Clues to neurologic causes include a family history of movement disorders, sudden onset or progressive course, associated medical or neurologic abnormalities, and asymmetry of symptoms. Some of the medical, neurologic, and psychiatric conditions to consider are listed in Table 2.¹²

Table 1

Schooler-Kane diagnostic criteria for TD

Table 2

Differential diagnosis of tardive dyskinesia

Treatment decisions

If a patient develops TD, clinicians need to make several decisions (Algorithm). First, consider tapering any anticholinergic drugs unless acute EPS are prominent or tardive dystonia is present. Anticholinergic agents can worsen TD but not tardive dystonia; 60% of TD cases improve after discontinuing anticholinergics.¹³ Second, decide whether antipsychotics could be safely tapered or discontinued. If antipsychotics cannot be safely tapered, decide whether to maintain the patient’s present antipsychotic or switch to a more or less potent agent. Finally, decide whether a trial of an adjunctive antidyskinetic drug is warranted. All of these decisions require thorough discussion with patients and their families, accompanied by careful documentation.

Discontinuing, continuing, or switching antipsychotics. Discontinuing antipsychotics once TD becomes apparent is an option. However, the natural course of TD after drug withdrawal is unclear. Although drug withdrawal had been recommended to increase the odds of TD resolution, early studies showed withdrawing antipsychotics may lead to an initial worsening of TD in 33% to 53% of patients (unmasking or withdrawal dyskinesia).¹⁴ With long-term follow-up, 36% to 55% of patients eventually improved, which supports recommendations for drug reduction or withdrawal.¹⁴ However, complete and permanent reversibility beyond the withdrawal period is rare; Glazer et al found only 2% of patients showed complete reversal of TD after drug discontinuation.^15,16 In a meta-analysis, Soares and McGrath¹⁷ reported 37% of patients assigned to placebo across studies showed at least some improvement in TD, but concluded insufficient evidence existed to support drug cessation or reduction as effective treatments for TD, especially when contrasted with robust evidence for the risk of psychotic relapse after drug withdrawal in patients with schizophrenia (53% within 9 months).¹⁸

A second option for a stable patient with good control of psychotic symptoms but established or long-term TD is to continue the antipsychotic, try to gradually reduce the dose, inform patients and caregivers of risks, document the decision, and monitor carefully. In most cases, TD may not progress even with continued antipsychotic treatment, although symptoms may worsen in some cases. However, in a patient with new-onset or early signs of TD, the clinician may be obligated to switch to a lower-potency antipsychotic or newer SGA to improve the chance of resolution; switching is discussed below.

Data on the change in prevalence of TD within a population during continued antipsychotic treatment have been inconsistent. Some studies show an increase, while others show a decrease or no change at all.¹⁹ However, prevalence rates obscure the dynamics of TD in individual patients. Roughly 50% of patients with TD have persistent symptoms, 10% to 30% have reduced symptoms, and 10% to 30% experience increased symptoms during treatment.¹³ Long-term studies estimated that up to 23% of patients may show loss of observable TD symptoms during treatment with FGAs in 1 year.^19,20 Similarly, studies of SGAs have shown reduction of TD ratings; some found greater reductions, some found less reductions, and some no difference compared with FGAs.^19,20 In some studies, improved TD outcomes were correlated with younger age, lower antipsychotic doses, reduced duration of drug treatment and dyskinesia, and increased length of follow-up.

In the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study, there was a significant decline in TD severity ratings among 200 patients with TD at baseline who were randomized to receive 1 of 4 SGAs, but there were no significant differences among these SGAs in decline in AIMS scores (Figure).¹⁹ Fifty-five percent of these patients met criteria for TD at 2 consecutive post-baseline visits, 76% met criteria at some or all post-baseline visits, and 24% did not meet criteria at any subsequent visit. In addition, 32% showed ≥50% decrease and 7% showed ≥50% increase in AIMS score. Thus, similar to past evidence on the course of TD during treatment with FGAs or SGAs, most patients in this trial showed either persistence or fluctuation in observable TD symptoms.

Another alternative is to switch antipsychotics, keeping in mind the risk of destabilizing a patient and precipitating psychotic relapse. More potent antipsychotics—such as haloperidol—suppress TD in approximately 67% of patients and may be necessary to consider in patients with severe, disabling symptoms, although the safety of these drugs in relation to their impact on long-term TD outcome is unclear.^13,21,22 On the other hand, lower-potency drugs and SGAs also have been associated with reduced TD symptoms^23,24; this was confirmed by results of the CATIE trial cited above in which SGAs were associated with a significant reduction in TD severity ratings.¹⁹ Clozapine in particular has been recommended for suppressing TD, especially in cases of tardive dystonia.²⁰ Surprisingly, data are limited and inconsistent in addressing whether high-potency FGAs suppress TD symptoms more than low-potency drugs or SGAs, and whether SGAs may suppress TD by mechanisms other than dopamine receptor blockade, which would enhance symptom remission.^19,25,26

Apart from short-term suppression of TD symptoms, the advantage of switching to lower-potency antipsychotics or other SGAs would be to increase the odds of eventual TD resolution. Although there has been speculation that in contrast to high-potency FGAs, SGAs may increase the possibility of remission by actively reversing TD or by passively allowing time for TD to resolve, existing data are inconclusive as to whether treatment with SGAs or FGAs results in true recovery rather than symptom suppression. To distinguish remission from suppression, a few studies discontinued SGAs. Some reported continued absence of TD,^27,28 but others found unmasking and reappearance of TD.^29-31

Adjunctive antidyskinetic drugs. Agents that have been tested off-label for antidyskinetic effects could be considered if symptoms of TD remain problematic despite optimization of antipsychotic treatment, although none have been confirmed as uniformly effective in randomized controlled trials replicated by different investigators.^13,17,22 These include dopamine-depleting agents, dopamine agonists, noradrenergic agonists and antagonists, GABAergic drugs (benzodiazepines, valproate, levetiracetam), lithium, calcium channel blockers, serotonergic drugs, antioxidants (vitamin E and B6), branched-chain amino acids, neuropeptides, cholinergic precursors, and cholinesterase inhibitors. Electroconvulsive therapy and botulinum toxin or surgical intervention (for tardive dystonia) also may be considered.

Hypotheses proposed to explain TD pathophysiology and thereby justify trials of specific antidyskinetic agents include dopamine receptor hypersensitivity, GABA insufficiency, and structural damage resulting from increased catecholamine metabolism and oxidative free radical production.³² Another hypothesis proposes that TD results from damage to striatal cholinergic interneurons due to loss of dopamine-mediated inhibition.³³ If correct, this implies that cholinesterase inhibitors or cholinergic agonists may suppress TD by directly enhancing post-synaptic cholinergic activity, thereby compensating for the loss of pre-synaptic cholinergic neurons. Several preliminary trials that explored the use of cholinesterase inhibitors had mixed results.^34-36 However, this hypothesis is supported by evidence from animal and human studies that correlated antipsychotic-induced changes in cholinergic activity with the delay in onset, irreversibility, and age-related risk of TD, the worsening of symptoms due to anticholinergic drugs, and the reduced liability of SGAs for causing TD. These findings suggest that further investigation of cholinergic mechanisms underlying TD may be worthwhile.³⁵

Algorithm: Proposed treatment algorithm for tardive dyskinesia
Source: Reprinted from Caroff SN, Hurford I, Lybrand J, et al. Movement disorders induced by antipsychotic drugs: implications of the CATIE Schizophrenia Trial. Neurol Clin. 2011;29:127-148 with permission from Elsevier

Figure: Adjusted^a repeated measures model of change in total AIMS scores for patients with tardive dyskinesia at baseline in CATIE
^a Model adjusted for baseline AIMS, baseline PANSS, and duration of illness. Adjusted P value for reduction in total AIMS score from baseline for all patients: P < .001. Treatment differences between the second-generation antipsychotics during the trial: P = .811
AIMS: Abnormal Involuntary Movement Scale; CATIE: Clinical Antipsychotic Trials of Intervention Effectiveness; PANSS: Positive and Negative Syndrome Scale
Source: Reprinted from Caroff SN, Davis VG, Miller DD, et al; for the CATIE Investigators. Treatment outcomes of patients with tardive dyskinesia and chronic schizophrenia. J Clin Psychiatry. 2011;72(3):295-303 with permission from Physician Postgraduate Press, Inc.Related Resources

• National Institute of Neurological Disorders and Stroke. NINDS Tardive Dyskinesia Information Page. www.ninds.nih.gov/disorders/tardive/tardive.htm.
• WE MOVE (Worldwide Education and Awareness for Movement Disorders). www.wemove.org.

Drug Brand Names

• Botulinum toxin • Botox, Dysport, others
• Clonidine • Catapres
• Clozapine • Clozaril
• Haloperidol • Haldol
• Levetiracetam • Keppra
• Levodopa • Dopar, Larodopa
• Lithium • Lithobid, Eskalith, others
• Nifedipine • Adalat, Afeditab CR, others
• Olanzapine • Zyprexa
• Phenytoin • Dilantin
• Quetiapine • Seroquel
• Reserpine • Serpasil
• Risperidone • Risperdal
• Tetrabenazine • Xenazine
• Valproate • Depakote
• Ziprasidone • Geodon

Disclosures

Drs. Caroff, Dhopesh, and Campbell report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Miller receives research/grant support from AstraZeneca, Bristol-Myers Squibb, Eli Lilly and Company, Ortho-McNeil-Janssen, and Pfizer Inc. and is a consultant to GlaxoSmithKline and Otsuka.

Introduced into clinical practice more than a half century ago, antipsychotics are still the mainstay of schizophrenia treatment. However, from the earliest reports, antipsychotic efficacy was seemingly inseparable from extrapyramidal side effects (EPS) that manifested as acute and chronic involuntary movement disorders. Although acute extrapyramidal side effects could be prevented and treated, the late-arising symptoms of tardive dyskinesia (TD) seemed irreversible in most cases.

Concerns over TD stimulated extensive research and fueled efforts to develop new antipsychotics that spared the extrapyramidal motor system. Numerous industry-sponsored trials found a reduced risk of EPS—including TD—with newer, second-generation antipsychotics (SGAs), although this advantage diminished when modest doses of low- or mid-potency first-generation antipsychotics (FGAs) were used as the comparator.^1-3 Nevertheless, in addition to the continued potential risk of introducing new cases of TD—even with SGAs—several other factors underscore the need to develop a rational strategy for clinical management of TD, including:

• thousands of patients are left with TD as a legacy of past treatment
• the neurophysiologic mechanisms underlying TD are not well understood
• there is no uniformly effective treatment to reverse TD
• TD may be irreversible in most cases.

Prevention

Because there is no “gold standard” treatment for TD, it is important to minimize the risk of TD by taking preventive measures and detecting incipient signs of the disorder. Preventive principles include:

• confirming and documenting the indication for antipsychotics
• using conservative maintenance doses and opting for lower potency or newer agents
• informing patients and caregivers of risk
• assessing for incipient signs of TD using the Abnormal Involuntary Movement Scale (AIMS),⁴ which should be administered at least every 3 to 6 months.

Confirming the diagnosis

TD presents as a polymorphous involuntary movement disorder,^5-8 most often with nonrhythmic, repetitive, purposeless hyperkinetic symptoms. It usually affects orofacial and lingual musculature (“buccolinguomasticatory syndrome”) with chewing; bruxism; protrusion, curling, or twisting of the tongue; lip smacking, puckering, sucking, and pursing; retraction, grimacing or bridling of the mouth; bulging of the cheeks; or eye blinking and blepharospasm. Choreoathetoid movements of the fingers, hands, or upper or lower extremities also are common. Patients may experience axial symptoms affecting the neck, shoulders, spine, or pelvis. When severe, dyskinesias can affect breathing, swallowing, or speech, and interfere with walking and activities of daily living.

TD may present with nonchoreoathetoid symptoms that can be difficult to distinguish from acute EPS. These may co-exist with classic TD symptoms, but may represent separate subtypes with increased risk of progression, persistence, and severe disability. For example, tardive dystonia, which is estimated to occur in 1% to 4% of patients treated with antipsychotics,⁹ may be more generalized and disabling than TD, and may respond to anticholinergic agents. Akathisia and other movement disorders also occur as tardive variants.¹⁰

Multiple diagnostic schemes for TD have been proposed; criteria proposed by Schooler and Kane have been widely accepted (Table 1).¹¹ TD onset occurs insidiously over ≥3 months of antipsychotic treatment and may begin with tic-like movements or increased eye blinking. TD often is suppressed or masked by ongoing antipsychotic treatment and becomes apparent only when the drug is reduced, switched, or discontinued. Dyskinesias increase with emotional arousal, activation, or distraction, and diminish with relaxation, sleep, or volitional effort. As a result, TD symptoms fluctuate over time; therefore, repeated measurements are necessary for reliable assessment of severity and persistence.

The differential diagnosis of TD necessitates conducting a careful medical and neurologic evaluation of all patients with new-onset movement disorders. Clues to neurologic causes include a family history of movement disorders, sudden onset or progressive course, associated medical or neurologic abnormalities, and asymmetry of symptoms. Some of the medical, neurologic, and psychiatric conditions to consider are listed in Table 2.¹²

Table 1

Schooler-Kane diagnostic criteria for TD

Table 2

Differential diagnosis of tardive dyskinesia

Treatment decisions

If a patient develops TD, clinicians need to make several decisions (Algorithm). First, consider tapering any anticholinergic drugs unless acute EPS are prominent or tardive dystonia is present. Anticholinergic agents can worsen TD but not tardive dystonia; 60% of TD cases improve after discontinuing anticholinergics.¹³ Second, decide whether antipsychotics could be safely tapered or discontinued. If antipsychotics cannot be safely tapered, decide whether to maintain the patient’s present antipsychotic or switch to a more or less potent agent. Finally, decide whether a trial of an adjunctive antidyskinetic drug is warranted. All of these decisions require thorough discussion with patients and their families, accompanied by careful documentation.

Discontinuing, continuing, or switching antipsychotics. Discontinuing antipsychotics once TD becomes apparent is an option. However, the natural course of TD after drug withdrawal is unclear. Although drug withdrawal had been recommended to increase the odds of TD resolution, early studies showed withdrawing antipsychotics may lead to an initial worsening of TD in 33% to 53% of patients (unmasking or withdrawal dyskinesia).¹⁴ With long-term follow-up, 36% to 55% of patients eventually improved, which supports recommendations for drug reduction or withdrawal.¹⁴ However, complete and permanent reversibility beyond the withdrawal period is rare; Glazer et al found only 2% of patients showed complete reversal of TD after drug discontinuation.^15,16 In a meta-analysis, Soares and McGrath¹⁷ reported 37% of patients assigned to placebo across studies showed at least some improvement in TD, but concluded insufficient evidence existed to support drug cessation or reduction as effective treatments for TD, especially when contrasted with robust evidence for the risk of psychotic relapse after drug withdrawal in patients with schizophrenia (53% within 9 months).¹⁸

A second option for a stable patient with good control of psychotic symptoms but established or long-term TD is to continue the antipsychotic, try to gradually reduce the dose, inform patients and caregivers of risks, document the decision, and monitor carefully. In most cases, TD may not progress even with continued antipsychotic treatment, although symptoms may worsen in some cases. However, in a patient with new-onset or early signs of TD, the clinician may be obligated to switch to a lower-potency antipsychotic or newer SGA to improve the chance of resolution; switching is discussed below.

Data on the change in prevalence of TD within a population during continued antipsychotic treatment have been inconsistent. Some studies show an increase, while others show a decrease or no change at all.¹⁹ However, prevalence rates obscure the dynamics of TD in individual patients. Roughly 50% of patients with TD have persistent symptoms, 10% to 30% have reduced symptoms, and 10% to 30% experience increased symptoms during treatment.¹³ Long-term studies estimated that up to 23% of patients may show loss of observable TD symptoms during treatment with FGAs in 1 year.^19,20 Similarly, studies of SGAs have shown reduction of TD ratings; some found greater reductions, some found less reductions, and some no difference compared with FGAs.^19,20 In some studies, improved TD outcomes were correlated with younger age, lower antipsychotic doses, reduced duration of drug treatment and dyskinesia, and increased length of follow-up.

In the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study, there was a significant decline in TD severity ratings among 200 patients with TD at baseline who were randomized to receive 1 of 4 SGAs, but there were no significant differences among these SGAs in decline in AIMS scores (Figure).¹⁹ Fifty-five percent of these patients met criteria for TD at 2 consecutive post-baseline visits, 76% met criteria at some or all post-baseline visits, and 24% did not meet criteria at any subsequent visit. In addition, 32% showed ≥50% decrease and 7% showed ≥50% increase in AIMS score. Thus, similar to past evidence on the course of TD during treatment with FGAs or SGAs, most patients in this trial showed either persistence or fluctuation in observable TD symptoms.

Another alternative is to switch antipsychotics, keeping in mind the risk of destabilizing a patient and precipitating psychotic relapse. More potent antipsychotics—such as haloperidol—suppress TD in approximately 67% of patients and may be necessary to consider in patients with severe, disabling symptoms, although the safety of these drugs in relation to their impact on long-term TD outcome is unclear.^13,21,22 On the other hand, lower-potency drugs and SGAs also have been associated with reduced TD symptoms^23,24; this was confirmed by results of the CATIE trial cited above in which SGAs were associated with a significant reduction in TD severity ratings.¹⁹ Clozapine in particular has been recommended for suppressing TD, especially in cases of tardive dystonia.²⁰ Surprisingly, data are limited and inconsistent in addressing whether high-potency FGAs suppress TD symptoms more than low-potency drugs or SGAs, and whether SGAs may suppress TD by mechanisms other than dopamine receptor blockade, which would enhance symptom remission.^19,25,26

Apart from short-term suppression of TD symptoms, the advantage of switching to lower-potency antipsychotics or other SGAs would be to increase the odds of eventual TD resolution. Although there has been speculation that in contrast to high-potency FGAs, SGAs may increase the possibility of remission by actively reversing TD or by passively allowing time for TD to resolve, existing data are inconclusive as to whether treatment with SGAs or FGAs results in true recovery rather than symptom suppression. To distinguish remission from suppression, a few studies discontinued SGAs. Some reported continued absence of TD,^27,28 but others found unmasking and reappearance of TD.^29-31

Adjunctive antidyskinetic drugs. Agents that have been tested off-label for antidyskinetic effects could be considered if symptoms of TD remain problematic despite optimization of antipsychotic treatment, although none have been confirmed as uniformly effective in randomized controlled trials replicated by different investigators.^13,17,22 These include dopamine-depleting agents, dopamine agonists, noradrenergic agonists and antagonists, GABAergic drugs (benzodiazepines, valproate, levetiracetam), lithium, calcium channel blockers, serotonergic drugs, antioxidants (vitamin E and B6), branched-chain amino acids, neuropeptides, cholinergic precursors, and cholinesterase inhibitors. Electroconvulsive therapy and botulinum toxin or surgical intervention (for tardive dystonia) also may be considered.

Hypotheses proposed to explain TD pathophysiology and thereby justify trials of specific antidyskinetic agents include dopamine receptor hypersensitivity, GABA insufficiency, and structural damage resulting from increased catecholamine metabolism and oxidative free radical production.³² Another hypothesis proposes that TD results from damage to striatal cholinergic interneurons due to loss of dopamine-mediated inhibition.³³ If correct, this implies that cholinesterase inhibitors or cholinergic agonists may suppress TD by directly enhancing post-synaptic cholinergic activity, thereby compensating for the loss of pre-synaptic cholinergic neurons. Several preliminary trials that explored the use of cholinesterase inhibitors had mixed results.^34-36 However, this hypothesis is supported by evidence from animal and human studies that correlated antipsychotic-induced changes in cholinergic activity with the delay in onset, irreversibility, and age-related risk of TD, the worsening of symptoms due to anticholinergic drugs, and the reduced liability of SGAs for causing TD. These findings suggest that further investigation of cholinergic mechanisms underlying TD may be worthwhile.³⁵

Algorithm: Proposed treatment algorithm for tardive dyskinesia
Source: Reprinted from Caroff SN, Hurford I, Lybrand J, et al. Movement disorders induced by antipsychotic drugs: implications of the CATIE Schizophrenia Trial. Neurol Clin. 2011;29:127-148 with permission from Elsevier

Figure: Adjusted^a repeated measures model of change in total AIMS scores for patients with tardive dyskinesia at baseline in CATIE
^a Model adjusted for baseline AIMS, baseline PANSS, and duration of illness. Adjusted P value for reduction in total AIMS score from baseline for all patients: P < .001. Treatment differences between the second-generation antipsychotics during the trial: P = .811
AIMS: Abnormal Involuntary Movement Scale; CATIE: Clinical Antipsychotic Trials of Intervention Effectiveness; PANSS: Positive and Negative Syndrome Scale
Source: Reprinted from Caroff SN, Davis VG, Miller DD, et al; for the CATIE Investigators. Treatment outcomes of patients with tardive dyskinesia and chronic schizophrenia. J Clin Psychiatry. 2011;72(3):295-303 with permission from Physician Postgraduate Press, Inc.Related Resources

• National Institute of Neurological Disorders and Stroke. NINDS Tardive Dyskinesia Information Page. www.ninds.nih.gov/disorders/tardive/tardive.htm.
• WE MOVE (Worldwide Education and Awareness for Movement Disorders). www.wemove.org.

Drug Brand Names

• Botulinum toxin • Botox, Dysport, others
• Clonidine • Catapres
• Clozapine • Clozaril
• Haloperidol • Haldol
• Levetiracetam • Keppra
• Levodopa • Dopar, Larodopa
• Lithium • Lithobid, Eskalith, others
• Nifedipine • Adalat, Afeditab CR, others
• Olanzapine • Zyprexa
• Phenytoin • Dilantin
• Quetiapine • Seroquel
• Reserpine • Serpasil
• Risperidone • Risperdal
• Tetrabenazine • Xenazine
• Valproate • Depakote
• Ziprasidone • Geodon

Disclosures

Drs. Caroff, Dhopesh, and Campbell report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Miller receives research/grant support from AstraZeneca, Bristol-Myers Squibb, Eli Lilly and Company, Ortho-McNeil-Janssen, and Pfizer Inc. and is a consultant to GlaxoSmithKline and Otsuka.

Introduced into clinical practice more than a half century ago, antipsychotics are still the mainstay of schizophrenia treatment. However, from the earliest reports, antipsychotic efficacy was seemingly inseparable from extrapyramidal side effects (EPS) that manifested as acute and chronic involuntary movement disorders. Although acute extrapyramidal side effects could be prevented and treated, the late-arising symptoms of tardive dyskinesia (TD) seemed irreversible in most cases.

Concerns over TD stimulated extensive research and fueled efforts to develop new antipsychotics that spared the extrapyramidal motor system. Numerous industry-sponsored trials found a reduced risk of EPS—including TD—with newer, second-generation antipsychotics (SGAs), although this advantage diminished when modest doses of low- or mid-potency first-generation antipsychotics (FGAs) were used as the comparator.^1-3 Nevertheless, in addition to the continued potential risk of introducing new cases of TD—even with SGAs—several other factors underscore the need to develop a rational strategy for clinical management of TD, including:

• thousands of patients are left with TD as a legacy of past treatment
• the neurophysiologic mechanisms underlying TD are not well understood
• there is no uniformly effective treatment to reverse TD
• TD may be irreversible in most cases.

Prevention

Because there is no “gold standard” treatment for TD, it is important to minimize the risk of TD by taking preventive measures and detecting incipient signs of the disorder. Preventive principles include:

• confirming and documenting the indication for antipsychotics
• using conservative maintenance doses and opting for lower potency or newer agents
• informing patients and caregivers of risk
• assessing for incipient signs of TD using the Abnormal Involuntary Movement Scale (AIMS),⁴ which should be administered at least every 3 to 6 months.

Confirming the diagnosis

TD presents as a polymorphous involuntary movement disorder,^5-8 most often with nonrhythmic, repetitive, purposeless hyperkinetic symptoms. It usually affects orofacial and lingual musculature (“buccolinguomasticatory syndrome”) with chewing; bruxism; protrusion, curling, or twisting of the tongue; lip smacking, puckering, sucking, and pursing; retraction, grimacing or bridling of the mouth; bulging of the cheeks; or eye blinking and blepharospasm. Choreoathetoid movements of the fingers, hands, or upper or lower extremities also are common. Patients may experience axial symptoms affecting the neck, shoulders, spine, or pelvis. When severe, dyskinesias can affect breathing, swallowing, or speech, and interfere with walking and activities of daily living.

TD may present with nonchoreoathetoid symptoms that can be difficult to distinguish from acute EPS. These may co-exist with classic TD symptoms, but may represent separate subtypes with increased risk of progression, persistence, and severe disability. For example, tardive dystonia, which is estimated to occur in 1% to 4% of patients treated with antipsychotics,⁹ may be more generalized and disabling than TD, and may respond to anticholinergic agents. Akathisia and other movement disorders also occur as tardive variants.¹⁰

Multiple diagnostic schemes for TD have been proposed; criteria proposed by Schooler and Kane have been widely accepted (Table 1).¹¹ TD onset occurs insidiously over ≥3 months of antipsychotic treatment and may begin with tic-like movements or increased eye blinking. TD often is suppressed or masked by ongoing antipsychotic treatment and becomes apparent only when the drug is reduced, switched, or discontinued. Dyskinesias increase with emotional arousal, activation, or distraction, and diminish with relaxation, sleep, or volitional effort. As a result, TD symptoms fluctuate over time; therefore, repeated measurements are necessary for reliable assessment of severity and persistence.

The differential diagnosis of TD necessitates conducting a careful medical and neurologic evaluation of all patients with new-onset movement disorders. Clues to neurologic causes include a family history of movement disorders, sudden onset or progressive course, associated medical or neurologic abnormalities, and asymmetry of symptoms. Some of the medical, neurologic, and psychiatric conditions to consider are listed in Table 2.¹²

Table 1

Schooler-Kane diagnostic criteria for TD

Table 2

Differential diagnosis of tardive dyskinesia

Treatment decisions

If a patient develops TD, clinicians need to make several decisions (Algorithm). First, consider tapering any anticholinergic drugs unless acute EPS are prominent or tardive dystonia is present. Anticholinergic agents can worsen TD but not tardive dystonia; 60% of TD cases improve after discontinuing anticholinergics.¹³ Second, decide whether antipsychotics could be safely tapered or discontinued. If antipsychotics cannot be safely tapered, decide whether to maintain the patient’s present antipsychotic or switch to a more or less potent agent. Finally, decide whether a trial of an adjunctive antidyskinetic drug is warranted. All of these decisions require thorough discussion with patients and their families, accompanied by careful documentation.

Discontinuing, continuing, or switching antipsychotics. Discontinuing antipsychotics once TD becomes apparent is an option. However, the natural course of TD after drug withdrawal is unclear. Although drug withdrawal had been recommended to increase the odds of TD resolution, early studies showed withdrawing antipsychotics may lead to an initial worsening of TD in 33% to 53% of patients (unmasking or withdrawal dyskinesia).¹⁴ With long-term follow-up, 36% to 55% of patients eventually improved, which supports recommendations for drug reduction or withdrawal.¹⁴ However, complete and permanent reversibility beyond the withdrawal period is rare; Glazer et al found only 2% of patients showed complete reversal of TD after drug discontinuation.^15,16 In a meta-analysis, Soares and McGrath¹⁷ reported 37% of patients assigned to placebo across studies showed at least some improvement in TD, but concluded insufficient evidence existed to support drug cessation or reduction as effective treatments for TD, especially when contrasted with robust evidence for the risk of psychotic relapse after drug withdrawal in patients with schizophrenia (53% within 9 months).¹⁸

A second option for a stable patient with good control of psychotic symptoms but established or long-term TD is to continue the antipsychotic, try to gradually reduce the dose, inform patients and caregivers of risks, document the decision, and monitor carefully. In most cases, TD may not progress even with continued antipsychotic treatment, although symptoms may worsen in some cases. However, in a patient with new-onset or early signs of TD, the clinician may be obligated to switch to a lower-potency antipsychotic or newer SGA to improve the chance of resolution; switching is discussed below.

Data on the change in prevalence of TD within a population during continued antipsychotic treatment have been inconsistent. Some studies show an increase, while others show a decrease or no change at all.¹⁹ However, prevalence rates obscure the dynamics of TD in individual patients. Roughly 50% of patients with TD have persistent symptoms, 10% to 30% have reduced symptoms, and 10% to 30% experience increased symptoms during treatment.¹³ Long-term studies estimated that up to 23% of patients may show loss of observable TD symptoms during treatment with FGAs in 1 year.^19,20 Similarly, studies of SGAs have shown reduction of TD ratings; some found greater reductions, some found less reductions, and some no difference compared with FGAs.^19,20 In some studies, improved TD outcomes were correlated with younger age, lower antipsychotic doses, reduced duration of drug treatment and dyskinesia, and increased length of follow-up.

In the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study, there was a significant decline in TD severity ratings among 200 patients with TD at baseline who were randomized to receive 1 of 4 SGAs, but there were no significant differences among these SGAs in decline in AIMS scores (Figure).¹⁹ Fifty-five percent of these patients met criteria for TD at 2 consecutive post-baseline visits, 76% met criteria at some or all post-baseline visits, and 24% did not meet criteria at any subsequent visit. In addition, 32% showed ≥50% decrease and 7% showed ≥50% increase in AIMS score. Thus, similar to past evidence on the course of TD during treatment with FGAs or SGAs, most patients in this trial showed either persistence or fluctuation in observable TD symptoms.

Another alternative is to switch antipsychotics, keeping in mind the risk of destabilizing a patient and precipitating psychotic relapse. More potent antipsychotics—such as haloperidol—suppress TD in approximately 67% of patients and may be necessary to consider in patients with severe, disabling symptoms, although the safety of these drugs in relation to their impact on long-term TD outcome is unclear.^13,21,22 On the other hand, lower-potency drugs and SGAs also have been associated with reduced TD symptoms^23,24; this was confirmed by results of the CATIE trial cited above in which SGAs were associated with a significant reduction in TD severity ratings.¹⁹ Clozapine in particular has been recommended for suppressing TD, especially in cases of tardive dystonia.²⁰ Surprisingly, data are limited and inconsistent in addressing whether high-potency FGAs suppress TD symptoms more than low-potency drugs or SGAs, and whether SGAs may suppress TD by mechanisms other than dopamine receptor blockade, which would enhance symptom remission.^19,25,26

Apart from short-term suppression of TD symptoms, the advantage of switching to lower-potency antipsychotics or other SGAs would be to increase the odds of eventual TD resolution. Although there has been speculation that in contrast to high-potency FGAs, SGAs may increase the possibility of remission by actively reversing TD or by passively allowing time for TD to resolve, existing data are inconclusive as to whether treatment with SGAs or FGAs results in true recovery rather than symptom suppression. To distinguish remission from suppression, a few studies discontinued SGAs. Some reported continued absence of TD,^27,28 but others found unmasking and reappearance of TD.^29-31

Adjunctive antidyskinetic drugs. Agents that have been tested off-label for antidyskinetic effects could be considered if symptoms of TD remain problematic despite optimization of antipsychotic treatment, although none have been confirmed as uniformly effective in randomized controlled trials replicated by different investigators.^13,17,22 These include dopamine-depleting agents, dopamine agonists, noradrenergic agonists and antagonists, GABAergic drugs (benzodiazepines, valproate, levetiracetam), lithium, calcium channel blockers, serotonergic drugs, antioxidants (vitamin E and B6), branched-chain amino acids, neuropeptides, cholinergic precursors, and cholinesterase inhibitors. Electroconvulsive therapy and botulinum toxin or surgical intervention (for tardive dystonia) also may be considered.

Hypotheses proposed to explain TD pathophysiology and thereby justify trials of specific antidyskinetic agents include dopamine receptor hypersensitivity, GABA insufficiency, and structural damage resulting from increased catecholamine metabolism and oxidative free radical production.³² Another hypothesis proposes that TD results from damage to striatal cholinergic interneurons due to loss of dopamine-mediated inhibition.³³ If correct, this implies that cholinesterase inhibitors or cholinergic agonists may suppress TD by directly enhancing post-synaptic cholinergic activity, thereby compensating for the loss of pre-synaptic cholinergic neurons. Several preliminary trials that explored the use of cholinesterase inhibitors had mixed results.^34-36 However, this hypothesis is supported by evidence from animal and human studies that correlated antipsychotic-induced changes in cholinergic activity with the delay in onset, irreversibility, and age-related risk of TD, the worsening of symptoms due to anticholinergic drugs, and the reduced liability of SGAs for causing TD. These findings suggest that further investigation of cholinergic mechanisms underlying TD may be worthwhile.³⁵

Algorithm: Proposed treatment algorithm for tardive dyskinesia
Source: Reprinted from Caroff SN, Hurford I, Lybrand J, et al. Movement disorders induced by antipsychotic drugs: implications of the CATIE Schizophrenia Trial. Neurol Clin. 2011;29:127-148 with permission from Elsevier

Figure: Adjusted^a repeated measures model of change in total AIMS scores for patients with tardive dyskinesia at baseline in CATIE
^a Model adjusted for baseline AIMS, baseline PANSS, and duration of illness. Adjusted P value for reduction in total AIMS score from baseline for all patients: P < .001. Treatment differences between the second-generation antipsychotics during the trial: P = .811
AIMS: Abnormal Involuntary Movement Scale; CATIE: Clinical Antipsychotic Trials of Intervention Effectiveness; PANSS: Positive and Negative Syndrome Scale
Source: Reprinted from Caroff SN, Davis VG, Miller DD, et al; for the CATIE Investigators. Treatment outcomes of patients with tardive dyskinesia and chronic schizophrenia. J Clin Psychiatry. 2011;72(3):295-303 with permission from Physician Postgraduate Press, Inc.Related Resources

• National Institute of Neurological Disorders and Stroke. NINDS Tardive Dyskinesia Information Page. www.ninds.nih.gov/disorders/tardive/tardive.htm.
• WE MOVE (Worldwide Education and Awareness for Movement Disorders). www.wemove.org.

Drug Brand Names

• Botulinum toxin • Botox, Dysport, others
• Clonidine • Catapres
• Clozapine • Clozaril
• Haloperidol • Haldol
• Levetiracetam • Keppra
• Levodopa • Dopar, Larodopa
• Lithium • Lithobid, Eskalith, others
• Nifedipine • Adalat, Afeditab CR, others
• Olanzapine • Zyprexa
• Phenytoin • Dilantin
• Quetiapine • Seroquel
• Reserpine • Serpasil
• Risperidone • Risperdal
• Tetrabenazine • Xenazine
• Valproate • Depakote
• Ziprasidone • Geodon

Disclosures

Drs. Caroff, Dhopesh, and Campbell report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dr. Miller receives research/grant support from AstraZeneca, Bristol-Myers Squibb, Eli Lilly and Company, Ortho-McNeil-Janssen, and Pfizer Inc. and is a consultant to GlaxoSmithKline and Otsuka.

Discuss this article at www.facebook.com/CurrentPsychiatry

Ms. M, age 31, has bipolar I disorder and takes lamotrigine, 200 mg/d, and aripiprazole, 10 mg/d. She was first hospitalized at age 20 for a manic episode and was discharged on lithium, 1,200 mg/d. She was hospitalized again at age 25 for a depressive episode that occurred after she stopped taking lithium because of undesirable side effects. She was switched to lamotrigine, 200 mg/d, which she tolerated well. Aripiprazole, 10 mg/d, was added 1 year later to address emergence of mild mood elevation symptoms.

During a recent follow-up appointment, Ms. M expresses interest in getting pregnant in the next 6 months. Her mood has been stable for 5 years and she asks if she should stop taking her medications in preparation for pregnancy. What would you recommend?

Because the typical age of onset for bipolar disorder (BD) is late adolescence or early adulthood, women are at risk for new onset or recurrence of mood episodes throughout their peak reproductive years. This article updates practitioners on the treatment of BD during pregnancy, including preconception planning and the risks and benefits of medication use during pregnancy. We also cover treatment considerations during the postpartum period, such as prophylaxis of mood episodes and mood stabilizer treatment for women who breast-feed.

Prenatal planning

Ideally, “prenatal planning” should begin long before women with BD prepare to have children. Because one-half of pregnancies in the United States are unplanned¹ and manic episodes may result in impulsivity and increased sexual activity, all women of reproductive age with BD should be counseled about birth control and risks of unplanned pregnancies. Discussions about risks of in utero exposure to psychotropics should occur when medications are first prescribed. Because certain mood stabilizers, (eg, carbamazepine) may decrease efficacy of oral contraceptives by inducing cytochrome P450 (CYP450) enzymes, women taking these medications also should be counseled about additional methods of birth control.²

Oral contraceptives also may affect mood stabilizer levels through similar mechanisms. Because of CYP450 induction, lamotrigine serum levels are lower during the 3 “active” weeks of exposure to exogenous estrogen. During the “pill-free” last week, lamotrigine levels may increase up to 54%.³

Because mood stabilizers such as valproate are associated with teratogenic risks, women with BD should be asked about contraception at every visit.⁴ Valproate also has been associated with an increased risk of menstrual cycle irregularity. Some studies have shown that even before initiating mood stabilizers, women with BD have a higher incidence of menstrual cycle irregularity than women without BD, which suggests the link between polycystic ovarian syndrome (PCOS) and BD may be independent of medications and part of the endophenotype.⁵

The importance of prenatal vitamins should be discussed. The recommended folate dosage for women planning to become pregnant is 0.4 to 1 mg/d and 0.8 to 5 mg/d for women with either a previous pregnancy with neural tube defects or those taking an antiepileptic medication.⁶

Table 1⁷ summarizes recommendations to improve prenatal planning in women with BD. Goals include:

• meeting with the patient at least 3 months before conception to review current menstrual cycle functioning. If your patient exhibits any signs or symptoms of PCOS, consider referral to a gynecologist
• meeting with patient and partner/significant supports to discuss treatment decisions
• optimizing the patient’s mood before conception, preferably for at least 3 to 6 months
• prescribing monotherapy at the lowest therapeutic dose if clinically feasible
• assessing the patient’s personal preferences and beliefs regarding medication use during pregnancy and breast-feeding
• assessing the patient’s capacity to understand the risks and benefits of medication continuation/discontinuation during pregnancy, including risk for relapse, current literature on teratogenicity, perinatal complications, and neurodevelopmental studies. Document that the patient provides informed consent.

Table 1

Pregnancy and BD: Medication management guidelines

CASE CONTINUED: Medication decisions

Ms. M’s first question is, “Should I stop taking my medications?” Ms. M and her psychiatrist review the risks and benefits of medication exposure during pregnancy (Table 2) and decide against discontinuing all medications because of her history of relapse when she stopped lithium. Because Ms. M’s mood has been stable for 5 years, she and her psychiatrist decide to limit her medications to lamotrigine monotherapy at her current dose, and agree to slowly taper aripiprazole. One week later, Ms. M calls and states she has a positive pregnancy test and is wondering if she should stop aripiprazole all at once. Ms. M is advised to continue with the original plan to slowly taper aripiprazole.

Table 2

Potential risks of continuing or discontinuing medications for BD during pregnancy

Medication risks/benefits

Women with BD have a high rate of relapse associated with abrupt discontinuation of pharmacotherapy during pregnancy. As such, patients and their partners and families should be cautioned against rapid discontinuation of medications.⁸ The risk to mother and fetus is particularly high for women with a history of recurrent, severe mood episodes. These patients face not only a high risk of recurrence of mood episodes, but also the inherent danger of impulsivity, poor self-care, and suicidality associated with mania, depression, and mixed states. In these cases, continuing a medication (other than known teratogens such as valproate) that has effectively stabilized mood may be preferred to discontinuation; these decisions are made after careful risk/benefit assessment.

Carefully reviewing the patient’s history is essential to assessing the risks and benefits of tapering medications before pregnancy. Consider the frequency and severity of your patient’s mood episodes, and whether a switch in mood state was rapid or had a prodromal phase. If a patient currently has a stable mood, a history of mild to moderate mood episodes, a history of prodromal symptoms (eg, gradually increasing sleep disturbances and mood deterioration), and no history of rapid switches, gradually discontinuing medications before or during pregnancy may be considered. However, encourage women to enlist their partners and family members to monitor for warning symptoms and advocate for early medication intervention. Because insomnia is a sign of relapse for many patients, educate women and their families about the importance of maintaining a regular sleep/wake cycle and alerting care providers if this cycle changes.

Mood stabilizers with the greatest risk for teratogenicity are valproate, carbamazepine, and lithium.⁹ Valproate is associated with a 6% to 13% risk of congenital malformation, including neural tube defects (1% to 2%) and cardiac or craniofacial defects.³ Risks increase at doses >800 mg/d.¹⁰ Potential perinatal complications associated with valproate include heart rate deceleration, abnormal tone (hypotonia or hypertonia), and growth retardation.¹¹ Neurobehavioral sequelae include lower IQ scores and increased risk of autism.¹²

Carbamazepine is associated with a 2% to 5% risk of congenital malformation, including neural tube defects and cardiac or craniofacial defects.⁴ Perinatal complications associated with carbamazepine include vitamin K deficiency.⁴ The neurobehavioral sequelae of carbamazepine are controversial; most prospective studies do not suggest long-term cognitive deficits.¹³ It is strongly recommended that valproate and carbamazepine be avoided, if possible, in women with BD who plan to become pregnant in the near future.

Prospective studies of lithium have shown a 2.8% rate of congenital malformations, which is much lower than the 11% rate found in retrospective studies.¹⁴ Ebstein’s anomaly—downward displacement of the tricuspid valve—is estimated to occur in .05% to 0.1% of infants exposed to lithium, which is 10 to 20 times the base rate, but a low absolute risk.¹¹

It is recommended women taking lithium during pregnancy complete a fetal high resolution ultrasound and echocardiogram at 16 to 18 weeks.¹¹ Perinatal complications associated with lithium include prematurity, hypotonia, hypothyroidism, hepatic abnormalities, respiratory distress, and nephrogenic diabetes insipidus.¹⁵ When prescribing lithium, divided doses are recommended to maintain a stable serum level. Serum lithium levels should be monitored frequently, and higher doses may be needed because of increased glomerular filtration rate and plasma volume throughout pregnancy.¹⁰ Because of fluid shifts at delivery—including blood loss during delivery and postpartum diuresis and diaphoresis—there is a risk of lithium toxicity at this time. Some researchers have suggested suspending lithium treatment during labor or 24 to 48 hours before planned induction or Caesarean section may lower this risk, with re-administration after delivery when medically stable.¹⁶ Women should be followed closely for signs of lithium toxicity and have lithium levels monitored as clinically indicated.¹⁶ There is insufficient data to support any neurobehavioral sequelae of in utero exposure to lithium; however, there are few long-term follow up studies using standardized measures.¹⁷

Lamotrigine is associated with a 1.9% to 4.6% rate of congenital malformations, including cleft lip/palate (8.9/1,000 vs 0.5 to 1.2/1,000 baseline).⁴ Studies suggest that rates of malformations (cardiac, genitourinary, gastrointestinal, neural tube defect) are dose-dependent: 1.3% at dosages <100 mg/d, 1.9% at 100 to 200 mg/d, and 5.4% at >200 mg/d.¹⁸ Because cleft lip and palate are formed by late second trimester, it is recommended to attempt to keep the lamotrigine dose <200 mg/d during the first and second trimesters. Higher doses of lamotrigine may be needed in the third trimester because of increased renal clearance.¹⁹ There is insufficient data to support any lamotrigine-associated neurobehavioral effects, and unlike studies of valproate, follow-up evaluations of lamotrigine-exposed children have not shown lower IQs.²⁰

Evidence about the reproductive safety of other mood stabilizers used in BD is limited. A recent population-based cohort study did not show increased risk of major malformations in children exposed to topiramate, gabapentin, or oxcarbazepine during the first trimester of pregnancy.²¹ Topiramate often is used in combination with other mood stabilizers for weight control, and studies suggest that polypharmacy with topiramate, especially at higher doses and with valproate, increases the risk of major congenital malformations, especially cleft lip and cleft palate.²² Consequently, topiramate is not recommended for women planning to conceive.

Antipsychotics. Although there is increasing information about outcomes of neonates exposed to atypical antipsychotics during pregnancy, the literature still is limited. The greatest number of studies have evaluated olanzapine, risperidone, and quetiapine and show the rate of congenital malformations is 0.9% to 4.1%, which is consistent with general population rates.^23-26 Perinatal complications associated with these atypical antipsychotics include neonatal extrapyramidal syndrome (EPS), possible neonatal adaptation/withdrawal syndrome, and an increased risk of the infant being either large or small for gestational age. Because atypical antipsychotics may increase the risk of metabolic syndrome, women should be counseled about the possible increased risk for gestational diabetes with these medications. None of these drugs have been associated with neurobehavioral sequelae, but long-term follow-up studies of exposed infants are lacking.

For aripiprazole, asenapine, ziprasidone, iloperidone, and lurasidone there is insufficient data about rate of congenital malformations, obstetric complications, and neurobehavioral sequelae. However, perinatal complications associated with these medications include risk of EPS and withdrawal symptoms.^25,26

CASE CONTINUED: Worsening mood symptoms

During pregnancy, Ms. M’s mood is stable on lamotrigine, 200 mg/d, and she participates in individual interpersonally oriented psychotherapy to address anxieties related to becoming a mother. However, late in her third trimester, Ms. M reports worsening symptoms, including depressed mood, insomnia, fatigue, and poor motivation. She also learns her mother had an episode of postpartum depression. Ms. M and her doctor discuss the risks of postpartum relapse, but she declines additional medication for prophylaxis because she is concerned about its impact on breast-feeding.

Two days after delivery, Ms. M complains of increased insomnia and depressed mood, and her husband reports she is not getting out of bed. She describes thoughts and images of throwing her baby out the window, and feels her thoughts are controlled by something outside of herself. Ms. M suspects her husband is having an affair.

Postpartum risks

All women with BD should be counseled regarding prophylaxis with mood stabilizers during the postpartum period. Women with BD are at high risk of mania and psychosis postpartum, particularly those with a personal or family history of postpartum psychosis. Postpartum psychosis frequently presents with an abrupt onset, shortly after delivery (Table 3). Although it may present with the classic symptoms of mania or psychotic depression, it also may have features of delirium.²⁷

Clinicians should immediately implement treatment with mood stabilizers and antipsychotics to manage acute psychotic symptoms, while also ruling out medical causes or comorbidities. Hospitalization should be considered. Aggressive treatment of insomnia will help stabilize mood. Electroconvulsive therapy can be used in treatment-refractory or urgent cases.¹⁰ Lastly, because approximately 4% of women with postpartum psychosis commit infanticide, all mother/child interactions should be closely supervised.²⁷

In small prospective studies, use of lithium within 48 hours of delivery decreased the risk of relapse of postpartum psychosis within the first 3 months.^28,29 In lower-risk patients who have discontinued pharmacotherapy during pregnancy, restarting medication before or immediately after delivery should be considered. At the same time, it is important to minimize sleep disruption, particularly postpartum. Psychoeducation—ideally begun in the preconception counseling visit—is extremely important for emphasizing the need for postpartum sleep.

Table 3

Consequences of postpartum mood relapse

Breast-feeding concerns

Data on risks of infant exposure to medications through breast milk are largely limited to case reports and case series. All mood-stabilizing medications have been found to pass into breast milk at varying concentrations.²⁸ If a patient chooses to breast-feed, she should inform her pediatrician of this decision, and she and her support system should be educated about signs of neonatal toxicity. Ideally, the psychiatrist should liaise with the patient’s pediatrician, especially when infants are premature, because the child’s liver metabolism may be immature, leading to higher serum drug levels and in some cases drug accumulation. Encourage patients to consider bottle feeding, either their own breast milk, pumped and stored, or formula. This will allow others to assist with feedings and the patient to have more consistent sleep, which could stabilize mood.

Lamotrigine concentrations in breast milk are highly variable.³⁰ Lamotrigine is processed through glucuronidation, a process that is immature in neonates. One study found serum lamotrigine levels in infants were 23% to 33% of maternal serum levels and milk/plasma ratios were highly variable, ranging from 6% to 147%.³⁰ Infants exposed to lamotrigine in breast milk should be monitored for rash and signs of thrombocytosis, and if clinically indicated, lamotrigine levels should be checked.³⁰ Valproate has a low infant serum/maternal serum ratio; there are rare case reports of hepatotoxicity and thrombocytopenia. Although valproate can be reinitiated because of its lower breast milk concentration, it is not a drug of choice in reproductive-age women because of the many issues described above, including risks during pregnancy, PCOS, and effect on oral contraceptives.

Carbamazepine serum levels in breast-feeding infants range from 6% to 65%; hepatic dysfunction, sedation, and poor feeding have been reported in infants in rare instances.³¹

Historically, lithium has been considered incompatible with breast-feeding, but recent reports suggest with careful monitoring it may not be contraindicated. In 10 mother/infant pairs, lithium levels in breast milk and infant serum diminished over time, with no adverse neonatal effects.³² However, if an infant does breast-feed, it may be important to monitor thyroid-stimulating hormone, blood urea nitrogen-to-creatinine ratio, and ECG in both mother and infant, especially if the mother is taking high doses of lithium.

The safety of breast-feeding while treated with atypical antipsychotics is largely unknown. Case reports indicate low transmission of these medications into breast milk.²⁸

CASE CONTINUED

Ms. M is admitted for psychiatric hospitalization because of worsening psychotic symptoms, poor self-care, and persistent thoughts of harming her baby. She agrees to restart aripiprazole, which is titrated to 20 mg/d. Breast-feeding is not pursued. She is discharged in 10 days after she no longer has thoughts of harming her baby, delusions, or psychotic or suicidal ideation. She and her family agree to close supervision by her family and outpatient follow-up.

Related Resources

• The Hospital for Sick Children. Pregnancy and breastfeeding resources. www.motherisk.org/women/pregnancyResources.jsp.
• U.S. National Library of Medicine. TOXNET toxicology data network. http://toxnet.nlm.nih.gov.

Drug Brand Names

• Aripiprazole • Abilify
• Asenapine • Saphris
• Carbamazepine • Equetro, Tegretol
• Gabapentin • Neurontin
• Iloperidone • Fanapt
• Lamotrigine • Lamictal
• Lithium • Eskalith, Lithobid
• Lurasidone • Latuda
• Olanzapine • Zyprexa
• Oxcarbazepine • Trileptal
• Quetiapine • Seroquel
• Risperidone • Risperdal
• Topiramate • Topamax
• Valproate • Depacon
• Ziprasidone • Geodon

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Acknowledgement

The authors would like to thank Natasha Barthel, BS, for her assistance with this article.

Discuss this article at www.facebook.com/CurrentPsychiatry

Ms. M, age 31, has bipolar I disorder and takes lamotrigine, 200 mg/d, and aripiprazole, 10 mg/d. She was first hospitalized at age 20 for a manic episode and was discharged on lithium, 1,200 mg/d. She was hospitalized again at age 25 for a depressive episode that occurred after she stopped taking lithium because of undesirable side effects. She was switched to lamotrigine, 200 mg/d, which she tolerated well. Aripiprazole, 10 mg/d, was added 1 year later to address emergence of mild mood elevation symptoms.

During a recent follow-up appointment, Ms. M expresses interest in getting pregnant in the next 6 months. Her mood has been stable for 5 years and she asks if she should stop taking her medications in preparation for pregnancy. What would you recommend?

Because the typical age of onset for bipolar disorder (BD) is late adolescence or early adulthood, women are at risk for new onset or recurrence of mood episodes throughout their peak reproductive years. This article updates practitioners on the treatment of BD during pregnancy, including preconception planning and the risks and benefits of medication use during pregnancy. We also cover treatment considerations during the postpartum period, such as prophylaxis of mood episodes and mood stabilizer treatment for women who breast-feed.

Prenatal planning

Ideally, “prenatal planning” should begin long before women with BD prepare to have children. Because one-half of pregnancies in the United States are unplanned¹ and manic episodes may result in impulsivity and increased sexual activity, all women of reproductive age with BD should be counseled about birth control and risks of unplanned pregnancies. Discussions about risks of in utero exposure to psychotropics should occur when medications are first prescribed. Because certain mood stabilizers, (eg, carbamazepine) may decrease efficacy of oral contraceptives by inducing cytochrome P450 (CYP450) enzymes, women taking these medications also should be counseled about additional methods of birth control.²

Oral contraceptives also may affect mood stabilizer levels through similar mechanisms. Because of CYP450 induction, lamotrigine serum levels are lower during the 3 “active” weeks of exposure to exogenous estrogen. During the “pill-free” last week, lamotrigine levels may increase up to 54%.³

Because mood stabilizers such as valproate are associated with teratogenic risks, women with BD should be asked about contraception at every visit.⁴ Valproate also has been associated with an increased risk of menstrual cycle irregularity. Some studies have shown that even before initiating mood stabilizers, women with BD have a higher incidence of menstrual cycle irregularity than women without BD, which suggests the link between polycystic ovarian syndrome (PCOS) and BD may be independent of medications and part of the endophenotype.⁵

The importance of prenatal vitamins should be discussed. The recommended folate dosage for women planning to become pregnant is 0.4 to 1 mg/d and 0.8 to 5 mg/d for women with either a previous pregnancy with neural tube defects or those taking an antiepileptic medication.⁶

Table 1⁷ summarizes recommendations to improve prenatal planning in women with BD. Goals include:

• meeting with the patient at least 3 months before conception to review current menstrual cycle functioning. If your patient exhibits any signs or symptoms of PCOS, consider referral to a gynecologist
• meeting with patient and partner/significant supports to discuss treatment decisions
• optimizing the patient’s mood before conception, preferably for at least 3 to 6 months
• prescribing monotherapy at the lowest therapeutic dose if clinically feasible
• assessing the patient’s personal preferences and beliefs regarding medication use during pregnancy and breast-feeding
• assessing the patient’s capacity to understand the risks and benefits of medication continuation/discontinuation during pregnancy, including risk for relapse, current literature on teratogenicity, perinatal complications, and neurodevelopmental studies. Document that the patient provides informed consent.

Table 1

Pregnancy and BD: Medication management guidelines

CASE CONTINUED: Medication decisions

Ms. M’s first question is, “Should I stop taking my medications?” Ms. M and her psychiatrist review the risks and benefits of medication exposure during pregnancy (Table 2) and decide against discontinuing all medications because of her history of relapse when she stopped lithium. Because Ms. M’s mood has been stable for 5 years, she and her psychiatrist decide to limit her medications to lamotrigine monotherapy at her current dose, and agree to slowly taper aripiprazole. One week later, Ms. M calls and states she has a positive pregnancy test and is wondering if she should stop aripiprazole all at once. Ms. M is advised to continue with the original plan to slowly taper aripiprazole.

Table 2

Potential risks of continuing or discontinuing medications for BD during pregnancy

Medication risks/benefits

Women with BD have a high rate of relapse associated with abrupt discontinuation of pharmacotherapy during pregnancy. As such, patients and their partners and families should be cautioned against rapid discontinuation of medications.⁸ The risk to mother and fetus is particularly high for women with a history of recurrent, severe mood episodes. These patients face not only a high risk of recurrence of mood episodes, but also the inherent danger of impulsivity, poor self-care, and suicidality associated with mania, depression, and mixed states. In these cases, continuing a medication (other than known teratogens such as valproate) that has effectively stabilized mood may be preferred to discontinuation; these decisions are made after careful risk/benefit assessment.

Carefully reviewing the patient’s history is essential to assessing the risks and benefits of tapering medications before pregnancy. Consider the frequency and severity of your patient’s mood episodes, and whether a switch in mood state was rapid or had a prodromal phase. If a patient currently has a stable mood, a history of mild to moderate mood episodes, a history of prodromal symptoms (eg, gradually increasing sleep disturbances and mood deterioration), and no history of rapid switches, gradually discontinuing medications before or during pregnancy may be considered. However, encourage women to enlist their partners and family members to monitor for warning symptoms and advocate for early medication intervention. Because insomnia is a sign of relapse for many patients, educate women and their families about the importance of maintaining a regular sleep/wake cycle and alerting care providers if this cycle changes.

Mood stabilizers with the greatest risk for teratogenicity are valproate, carbamazepine, and lithium.⁹ Valproate is associated with a 6% to 13% risk of congenital malformation, including neural tube defects (1% to 2%) and cardiac or craniofacial defects.³ Risks increase at doses >800 mg/d.¹⁰ Potential perinatal complications associated with valproate include heart rate deceleration, abnormal tone (hypotonia or hypertonia), and growth retardation.¹¹ Neurobehavioral sequelae include lower IQ scores and increased risk of autism.¹²

Carbamazepine is associated with a 2% to 5% risk of congenital malformation, including neural tube defects and cardiac or craniofacial defects.⁴ Perinatal complications associated with carbamazepine include vitamin K deficiency.⁴ The neurobehavioral sequelae of carbamazepine are controversial; most prospective studies do not suggest long-term cognitive deficits.¹³ It is strongly recommended that valproate and carbamazepine be avoided, if possible, in women with BD who plan to become pregnant in the near future.

Prospective studies of lithium have shown a 2.8% rate of congenital malformations, which is much lower than the 11% rate found in retrospective studies.¹⁴ Ebstein’s anomaly—downward displacement of the tricuspid valve—is estimated to occur in .05% to 0.1% of infants exposed to lithium, which is 10 to 20 times the base rate, but a low absolute risk.¹¹

It is recommended women taking lithium during pregnancy complete a fetal high resolution ultrasound and echocardiogram at 16 to 18 weeks.¹¹ Perinatal complications associated with lithium include prematurity, hypotonia, hypothyroidism, hepatic abnormalities, respiratory distress, and nephrogenic diabetes insipidus.¹⁵ When prescribing lithium, divided doses are recommended to maintain a stable serum level. Serum lithium levels should be monitored frequently, and higher doses may be needed because of increased glomerular filtration rate and plasma volume throughout pregnancy.¹⁰ Because of fluid shifts at delivery—including blood loss during delivery and postpartum diuresis and diaphoresis—there is a risk of lithium toxicity at this time. Some researchers have suggested suspending lithium treatment during labor or 24 to 48 hours before planned induction or Caesarean section may lower this risk, with re-administration after delivery when medically stable.¹⁶ Women should be followed closely for signs of lithium toxicity and have lithium levels monitored as clinically indicated.¹⁶ There is insufficient data to support any neurobehavioral sequelae of in utero exposure to lithium; however, there are few long-term follow up studies using standardized measures.¹⁷

Lamotrigine is associated with a 1.9% to 4.6% rate of congenital malformations, including cleft lip/palate (8.9/1,000 vs 0.5 to 1.2/1,000 baseline).⁴ Studies suggest that rates of malformations (cardiac, genitourinary, gastrointestinal, neural tube defect) are dose-dependent: 1.3% at dosages <100 mg/d, 1.9% at 100 to 200 mg/d, and 5.4% at >200 mg/d.¹⁸ Because cleft lip and palate are formed by late second trimester, it is recommended to attempt to keep the lamotrigine dose <200 mg/d during the first and second trimesters. Higher doses of lamotrigine may be needed in the third trimester because of increased renal clearance.¹⁹ There is insufficient data to support any lamotrigine-associated neurobehavioral effects, and unlike studies of valproate, follow-up evaluations of lamotrigine-exposed children have not shown lower IQs.²⁰

Evidence about the reproductive safety of other mood stabilizers used in BD is limited. A recent population-based cohort study did not show increased risk of major malformations in children exposed to topiramate, gabapentin, or oxcarbazepine during the first trimester of pregnancy.²¹ Topiramate often is used in combination with other mood stabilizers for weight control, and studies suggest that polypharmacy with topiramate, especially at higher doses and with valproate, increases the risk of major congenital malformations, especially cleft lip and cleft palate.²² Consequently, topiramate is not recommended for women planning to conceive.

Antipsychotics. Although there is increasing information about outcomes of neonates exposed to atypical antipsychotics during pregnancy, the literature still is limited. The greatest number of studies have evaluated olanzapine, risperidone, and quetiapine and show the rate of congenital malformations is 0.9% to 4.1%, which is consistent with general population rates.^23-26 Perinatal complications associated with these atypical antipsychotics include neonatal extrapyramidal syndrome (EPS), possible neonatal adaptation/withdrawal syndrome, and an increased risk of the infant being either large or small for gestational age. Because atypical antipsychotics may increase the risk of metabolic syndrome, women should be counseled about the possible increased risk for gestational diabetes with these medications. None of these drugs have been associated with neurobehavioral sequelae, but long-term follow-up studies of exposed infants are lacking.

For aripiprazole, asenapine, ziprasidone, iloperidone, and lurasidone there is insufficient data about rate of congenital malformations, obstetric complications, and neurobehavioral sequelae. However, perinatal complications associated with these medications include risk of EPS and withdrawal symptoms.^25,26

CASE CONTINUED: Worsening mood symptoms

During pregnancy, Ms. M’s mood is stable on lamotrigine, 200 mg/d, and she participates in individual interpersonally oriented psychotherapy to address anxieties related to becoming a mother. However, late in her third trimester, Ms. M reports worsening symptoms, including depressed mood, insomnia, fatigue, and poor motivation. She also learns her mother had an episode of postpartum depression. Ms. M and her doctor discuss the risks of postpartum relapse, but she declines additional medication for prophylaxis because she is concerned about its impact on breast-feeding.

Two days after delivery, Ms. M complains of increased insomnia and depressed mood, and her husband reports she is not getting out of bed. She describes thoughts and images of throwing her baby out the window, and feels her thoughts are controlled by something outside of herself. Ms. M suspects her husband is having an affair.

Postpartum risks

All women with BD should be counseled regarding prophylaxis with mood stabilizers during the postpartum period. Women with BD are at high risk of mania and psychosis postpartum, particularly those with a personal or family history of postpartum psychosis. Postpartum psychosis frequently presents with an abrupt onset, shortly after delivery (Table 3). Although it may present with the classic symptoms of mania or psychotic depression, it also may have features of delirium.²⁷

Clinicians should immediately implement treatment with mood stabilizers and antipsychotics to manage acute psychotic symptoms, while also ruling out medical causes or comorbidities. Hospitalization should be considered. Aggressive treatment of insomnia will help stabilize mood. Electroconvulsive therapy can be used in treatment-refractory or urgent cases.¹⁰ Lastly, because approximately 4% of women with postpartum psychosis commit infanticide, all mother/child interactions should be closely supervised.²⁷

In small prospective studies, use of lithium within 48 hours of delivery decreased the risk of relapse of postpartum psychosis within the first 3 months.^28,29 In lower-risk patients who have discontinued pharmacotherapy during pregnancy, restarting medication before or immediately after delivery should be considered. At the same time, it is important to minimize sleep disruption, particularly postpartum. Psychoeducation—ideally begun in the preconception counseling visit—is extremely important for emphasizing the need for postpartum sleep.

Table 3

Consequences of postpartum mood relapse

Breast-feeding concerns

Data on risks of infant exposure to medications through breast milk are largely limited to case reports and case series. All mood-stabilizing medications have been found to pass into breast milk at varying concentrations.²⁸ If a patient chooses to breast-feed, she should inform her pediatrician of this decision, and she and her support system should be educated about signs of neonatal toxicity. Ideally, the psychiatrist should liaise with the patient’s pediatrician, especially when infants are premature, because the child’s liver metabolism may be immature, leading to higher serum drug levels and in some cases drug accumulation. Encourage patients to consider bottle feeding, either their own breast milk, pumped and stored, or formula. This will allow others to assist with feedings and the patient to have more consistent sleep, which could stabilize mood.

Lamotrigine concentrations in breast milk are highly variable.³⁰ Lamotrigine is processed through glucuronidation, a process that is immature in neonates. One study found serum lamotrigine levels in infants were 23% to 33% of maternal serum levels and milk/plasma ratios were highly variable, ranging from 6% to 147%.³⁰ Infants exposed to lamotrigine in breast milk should be monitored for rash and signs of thrombocytosis, and if clinically indicated, lamotrigine levels should be checked.³⁰ Valproate has a low infant serum/maternal serum ratio; there are rare case reports of hepatotoxicity and thrombocytopenia. Although valproate can be reinitiated because of its lower breast milk concentration, it is not a drug of choice in reproductive-age women because of the many issues described above, including risks during pregnancy, PCOS, and effect on oral contraceptives.

Carbamazepine serum levels in breast-feeding infants range from 6% to 65%; hepatic dysfunction, sedation, and poor feeding have been reported in infants in rare instances.³¹

Historically, lithium has been considered incompatible with breast-feeding, but recent reports suggest with careful monitoring it may not be contraindicated. In 10 mother/infant pairs, lithium levels in breast milk and infant serum diminished over time, with no adverse neonatal effects.³² However, if an infant does breast-feed, it may be important to monitor thyroid-stimulating hormone, blood urea nitrogen-to-creatinine ratio, and ECG in both mother and infant, especially if the mother is taking high doses of lithium.

The safety of breast-feeding while treated with atypical antipsychotics is largely unknown. Case reports indicate low transmission of these medications into breast milk.²⁸

CASE CONTINUED

Ms. M is admitted for psychiatric hospitalization because of worsening psychotic symptoms, poor self-care, and persistent thoughts of harming her baby. She agrees to restart aripiprazole, which is titrated to 20 mg/d. Breast-feeding is not pursued. She is discharged in 10 days after she no longer has thoughts of harming her baby, delusions, or psychotic or suicidal ideation. She and her family agree to close supervision by her family and outpatient follow-up.

Related Resources

• The Hospital for Sick Children. Pregnancy and breastfeeding resources. www.motherisk.org/women/pregnancyResources.jsp.
• U.S. National Library of Medicine. TOXNET toxicology data network. http://toxnet.nlm.nih.gov.

Drug Brand Names

• Aripiprazole • Abilify
• Asenapine • Saphris
• Carbamazepine • Equetro, Tegretol
• Gabapentin • Neurontin
• Iloperidone • Fanapt
• Lamotrigine • Lamictal
• Lithium • Eskalith, Lithobid
• Lurasidone • Latuda
• Olanzapine • Zyprexa
• Oxcarbazepine • Trileptal
• Quetiapine • Seroquel
• Risperidone • Risperdal
• Topiramate • Topamax
• Valproate • Depacon
• Ziprasidone • Geodon

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Acknowledgement

The authors would like to thank Natasha Barthel, BS, for her assistance with this article.

Discuss this article at www.facebook.com/CurrentPsychiatry

Ms. M, age 31, has bipolar I disorder and takes lamotrigine, 200 mg/d, and aripiprazole, 10 mg/d. She was first hospitalized at age 20 for a manic episode and was discharged on lithium, 1,200 mg/d. She was hospitalized again at age 25 for a depressive episode that occurred after she stopped taking lithium because of undesirable side effects. She was switched to lamotrigine, 200 mg/d, which she tolerated well. Aripiprazole, 10 mg/d, was added 1 year later to address emergence of mild mood elevation symptoms.

During a recent follow-up appointment, Ms. M expresses interest in getting pregnant in the next 6 months. Her mood has been stable for 5 years and she asks if she should stop taking her medications in preparation for pregnancy. What would you recommend?

Because the typical age of onset for bipolar disorder (BD) is late adolescence or early adulthood, women are at risk for new onset or recurrence of mood episodes throughout their peak reproductive years. This article updates practitioners on the treatment of BD during pregnancy, including preconception planning and the risks and benefits of medication use during pregnancy. We also cover treatment considerations during the postpartum period, such as prophylaxis of mood episodes and mood stabilizer treatment for women who breast-feed.

Prenatal planning

Ideally, “prenatal planning” should begin long before women with BD prepare to have children. Because one-half of pregnancies in the United States are unplanned¹ and manic episodes may result in impulsivity and increased sexual activity, all women of reproductive age with BD should be counseled about birth control and risks of unplanned pregnancies. Discussions about risks of in utero exposure to psychotropics should occur when medications are first prescribed. Because certain mood stabilizers, (eg, carbamazepine) may decrease efficacy of oral contraceptives by inducing cytochrome P450 (CYP450) enzymes, women taking these medications also should be counseled about additional methods of birth control.²

Oral contraceptives also may affect mood stabilizer levels through similar mechanisms. Because of CYP450 induction, lamotrigine serum levels are lower during the 3 “active” weeks of exposure to exogenous estrogen. During the “pill-free” last week, lamotrigine levels may increase up to 54%.³

Because mood stabilizers such as valproate are associated with teratogenic risks, women with BD should be asked about contraception at every visit.⁴ Valproate also has been associated with an increased risk of menstrual cycle irregularity. Some studies have shown that even before initiating mood stabilizers, women with BD have a higher incidence of menstrual cycle irregularity than women without BD, which suggests the link between polycystic ovarian syndrome (PCOS) and BD may be independent of medications and part of the endophenotype.⁵

The importance of prenatal vitamins should be discussed. The recommended folate dosage for women planning to become pregnant is 0.4 to 1 mg/d and 0.8 to 5 mg/d for women with either a previous pregnancy with neural tube defects or those taking an antiepileptic medication.⁶

Table 1⁷ summarizes recommendations to improve prenatal planning in women with BD. Goals include:

• meeting with the patient at least 3 months before conception to review current menstrual cycle functioning. If your patient exhibits any signs or symptoms of PCOS, consider referral to a gynecologist
• meeting with patient and partner/significant supports to discuss treatment decisions
• optimizing the patient’s mood before conception, preferably for at least 3 to 6 months
• prescribing monotherapy at the lowest therapeutic dose if clinically feasible
• assessing the patient’s personal preferences and beliefs regarding medication use during pregnancy and breast-feeding
• assessing the patient’s capacity to understand the risks and benefits of medication continuation/discontinuation during pregnancy, including risk for relapse, current literature on teratogenicity, perinatal complications, and neurodevelopmental studies. Document that the patient provides informed consent.

Table 1

Pregnancy and BD: Medication management guidelines

CASE CONTINUED: Medication decisions

Ms. M’s first question is, “Should I stop taking my medications?” Ms. M and her psychiatrist review the risks and benefits of medication exposure during pregnancy (Table 2) and decide against discontinuing all medications because of her history of relapse when she stopped lithium. Because Ms. M’s mood has been stable for 5 years, she and her psychiatrist decide to limit her medications to lamotrigine monotherapy at her current dose, and agree to slowly taper aripiprazole. One week later, Ms. M calls and states she has a positive pregnancy test and is wondering if she should stop aripiprazole all at once. Ms. M is advised to continue with the original plan to slowly taper aripiprazole.

Table 2

Potential risks of continuing or discontinuing medications for BD during pregnancy

Medication risks/benefits

Women with BD have a high rate of relapse associated with abrupt discontinuation of pharmacotherapy during pregnancy. As such, patients and their partners and families should be cautioned against rapid discontinuation of medications.⁸ The risk to mother and fetus is particularly high for women with a history of recurrent, severe mood episodes. These patients face not only a high risk of recurrence of mood episodes, but also the inherent danger of impulsivity, poor self-care, and suicidality associated with mania, depression, and mixed states. In these cases, continuing a medication (other than known teratogens such as valproate) that has effectively stabilized mood may be preferred to discontinuation; these decisions are made after careful risk/benefit assessment.

Carefully reviewing the patient’s history is essential to assessing the risks and benefits of tapering medications before pregnancy. Consider the frequency and severity of your patient’s mood episodes, and whether a switch in mood state was rapid or had a prodromal phase. If a patient currently has a stable mood, a history of mild to moderate mood episodes, a history of prodromal symptoms (eg, gradually increasing sleep disturbances and mood deterioration), and no history of rapid switches, gradually discontinuing medications before or during pregnancy may be considered. However, encourage women to enlist their partners and family members to monitor for warning symptoms and advocate for early medication intervention. Because insomnia is a sign of relapse for many patients, educate women and their families about the importance of maintaining a regular sleep/wake cycle and alerting care providers if this cycle changes.

Mood stabilizers with the greatest risk for teratogenicity are valproate, carbamazepine, and lithium.⁹ Valproate is associated with a 6% to 13% risk of congenital malformation, including neural tube defects (1% to 2%) and cardiac or craniofacial defects.³ Risks increase at doses >800 mg/d.¹⁰ Potential perinatal complications associated with valproate include heart rate deceleration, abnormal tone (hypotonia or hypertonia), and growth retardation.¹¹ Neurobehavioral sequelae include lower IQ scores and increased risk of autism.¹²

Carbamazepine is associated with a 2% to 5% risk of congenital malformation, including neural tube defects and cardiac or craniofacial defects.⁴ Perinatal complications associated with carbamazepine include vitamin K deficiency.⁴ The neurobehavioral sequelae of carbamazepine are controversial; most prospective studies do not suggest long-term cognitive deficits.¹³ It is strongly recommended that valproate and carbamazepine be avoided, if possible, in women with BD who plan to become pregnant in the near future.

Prospective studies of lithium have shown a 2.8% rate of congenital malformations, which is much lower than the 11% rate found in retrospective studies.¹⁴ Ebstein’s anomaly—downward displacement of the tricuspid valve—is estimated to occur in .05% to 0.1% of infants exposed to lithium, which is 10 to 20 times the base rate, but a low absolute risk.¹¹

It is recommended women taking lithium during pregnancy complete a fetal high resolution ultrasound and echocardiogram at 16 to 18 weeks.¹¹ Perinatal complications associated with lithium include prematurity, hypotonia, hypothyroidism, hepatic abnormalities, respiratory distress, and nephrogenic diabetes insipidus.¹⁵ When prescribing lithium, divided doses are recommended to maintain a stable serum level. Serum lithium levels should be monitored frequently, and higher doses may be needed because of increased glomerular filtration rate and plasma volume throughout pregnancy.¹⁰ Because of fluid shifts at delivery—including blood loss during delivery and postpartum diuresis and diaphoresis—there is a risk of lithium toxicity at this time. Some researchers have suggested suspending lithium treatment during labor or 24 to 48 hours before planned induction or Caesarean section may lower this risk, with re-administration after delivery when medically stable.¹⁶ Women should be followed closely for signs of lithium toxicity and have lithium levels monitored as clinically indicated.¹⁶ There is insufficient data to support any neurobehavioral sequelae of in utero exposure to lithium; however, there are few long-term follow up studies using standardized measures.¹⁷

Lamotrigine is associated with a 1.9% to 4.6% rate of congenital malformations, including cleft lip/palate (8.9/1,000 vs 0.5 to 1.2/1,000 baseline).⁴ Studies suggest that rates of malformations (cardiac, genitourinary, gastrointestinal, neural tube defect) are dose-dependent: 1.3% at dosages <100 mg/d, 1.9% at 100 to 200 mg/d, and 5.4% at >200 mg/d.¹⁸ Because cleft lip and palate are formed by late second trimester, it is recommended to attempt to keep the lamotrigine dose <200 mg/d during the first and second trimesters. Higher doses of lamotrigine may be needed in the third trimester because of increased renal clearance.¹⁹ There is insufficient data to support any lamotrigine-associated neurobehavioral effects, and unlike studies of valproate, follow-up evaluations of lamotrigine-exposed children have not shown lower IQs.²⁰

Evidence about the reproductive safety of other mood stabilizers used in BD is limited. A recent population-based cohort study did not show increased risk of major malformations in children exposed to topiramate, gabapentin, or oxcarbazepine during the first trimester of pregnancy.²¹ Topiramate often is used in combination with other mood stabilizers for weight control, and studies suggest that polypharmacy with topiramate, especially at higher doses and with valproate, increases the risk of major congenital malformations, especially cleft lip and cleft palate.²² Consequently, topiramate is not recommended for women planning to conceive.

Antipsychotics. Although there is increasing information about outcomes of neonates exposed to atypical antipsychotics during pregnancy, the literature still is limited. The greatest number of studies have evaluated olanzapine, risperidone, and quetiapine and show the rate of congenital malformations is 0.9% to 4.1%, which is consistent with general population rates.^23-26 Perinatal complications associated with these atypical antipsychotics include neonatal extrapyramidal syndrome (EPS), possible neonatal adaptation/withdrawal syndrome, and an increased risk of the infant being either large or small for gestational age. Because atypical antipsychotics may increase the risk of metabolic syndrome, women should be counseled about the possible increased risk for gestational diabetes with these medications. None of these drugs have been associated with neurobehavioral sequelae, but long-term follow-up studies of exposed infants are lacking.

For aripiprazole, asenapine, ziprasidone, iloperidone, and lurasidone there is insufficient data about rate of congenital malformations, obstetric complications, and neurobehavioral sequelae. However, perinatal complications associated with these medications include risk of EPS and withdrawal symptoms.^25,26

CASE CONTINUED: Worsening mood symptoms

During pregnancy, Ms. M’s mood is stable on lamotrigine, 200 mg/d, and she participates in individual interpersonally oriented psychotherapy to address anxieties related to becoming a mother. However, late in her third trimester, Ms. M reports worsening symptoms, including depressed mood, insomnia, fatigue, and poor motivation. She also learns her mother had an episode of postpartum depression. Ms. M and her doctor discuss the risks of postpartum relapse, but she declines additional medication for prophylaxis because she is concerned about its impact on breast-feeding.

Two days after delivery, Ms. M complains of increased insomnia and depressed mood, and her husband reports she is not getting out of bed. She describes thoughts and images of throwing her baby out the window, and feels her thoughts are controlled by something outside of herself. Ms. M suspects her husband is having an affair.

Postpartum risks

All women with BD should be counseled regarding prophylaxis with mood stabilizers during the postpartum period. Women with BD are at high risk of mania and psychosis postpartum, particularly those with a personal or family history of postpartum psychosis. Postpartum psychosis frequently presents with an abrupt onset, shortly after delivery (Table 3). Although it may present with the classic symptoms of mania or psychotic depression, it also may have features of delirium.²⁷

Clinicians should immediately implement treatment with mood stabilizers and antipsychotics to manage acute psychotic symptoms, while also ruling out medical causes or comorbidities. Hospitalization should be considered. Aggressive treatment of insomnia will help stabilize mood. Electroconvulsive therapy can be used in treatment-refractory or urgent cases.¹⁰ Lastly, because approximately 4% of women with postpartum psychosis commit infanticide, all mother/child interactions should be closely supervised.²⁷

In small prospective studies, use of lithium within 48 hours of delivery decreased the risk of relapse of postpartum psychosis within the first 3 months.^28,29 In lower-risk patients who have discontinued pharmacotherapy during pregnancy, restarting medication before or immediately after delivery should be considered. At the same time, it is important to minimize sleep disruption, particularly postpartum. Psychoeducation—ideally begun in the preconception counseling visit—is extremely important for emphasizing the need for postpartum sleep.

Table 3

Consequences of postpartum mood relapse

Breast-feeding concerns

Data on risks of infant exposure to medications through breast milk are largely limited to case reports and case series. All mood-stabilizing medications have been found to pass into breast milk at varying concentrations.²⁸ If a patient chooses to breast-feed, she should inform her pediatrician of this decision, and she and her support system should be educated about signs of neonatal toxicity. Ideally, the psychiatrist should liaise with the patient’s pediatrician, especially when infants are premature, because the child’s liver metabolism may be immature, leading to higher serum drug levels and in some cases drug accumulation. Encourage patients to consider bottle feeding, either their own breast milk, pumped and stored, or formula. This will allow others to assist with feedings and the patient to have more consistent sleep, which could stabilize mood.

Lamotrigine concentrations in breast milk are highly variable.³⁰ Lamotrigine is processed through glucuronidation, a process that is immature in neonates. One study found serum lamotrigine levels in infants were 23% to 33% of maternal serum levels and milk/plasma ratios were highly variable, ranging from 6% to 147%.³⁰ Infants exposed to lamotrigine in breast milk should be monitored for rash and signs of thrombocytosis, and if clinically indicated, lamotrigine levels should be checked.³⁰ Valproate has a low infant serum/maternal serum ratio; there are rare case reports of hepatotoxicity and thrombocytopenia. Although valproate can be reinitiated because of its lower breast milk concentration, it is not a drug of choice in reproductive-age women because of the many issues described above, including risks during pregnancy, PCOS, and effect on oral contraceptives.

Carbamazepine serum levels in breast-feeding infants range from 6% to 65%; hepatic dysfunction, sedation, and poor feeding have been reported in infants in rare instances.³¹

Historically, lithium has been considered incompatible with breast-feeding, but recent reports suggest with careful monitoring it may not be contraindicated. In 10 mother/infant pairs, lithium levels in breast milk and infant serum diminished over time, with no adverse neonatal effects.³² However, if an infant does breast-feed, it may be important to monitor thyroid-stimulating hormone, blood urea nitrogen-to-creatinine ratio, and ECG in both mother and infant, especially if the mother is taking high doses of lithium.

The safety of breast-feeding while treated with atypical antipsychotics is largely unknown. Case reports indicate low transmission of these medications into breast milk.²⁸

CASE CONTINUED

Ms. M is admitted for psychiatric hospitalization because of worsening psychotic symptoms, poor self-care, and persistent thoughts of harming her baby. She agrees to restart aripiprazole, which is titrated to 20 mg/d. Breast-feeding is not pursued. She is discharged in 10 days after she no longer has thoughts of harming her baby, delusions, or psychotic or suicidal ideation. She and her family agree to close supervision by her family and outpatient follow-up.

Related Resources

• The Hospital for Sick Children. Pregnancy and breastfeeding resources. www.motherisk.org/women/pregnancyResources.jsp.
• U.S. National Library of Medicine. TOXNET toxicology data network. http://toxnet.nlm.nih.gov.

Drug Brand Names

• Aripiprazole • Abilify
• Asenapine • Saphris
• Carbamazepine • Equetro, Tegretol
• Gabapentin • Neurontin
• Iloperidone • Fanapt
• Lamotrigine • Lamictal
• Lithium • Eskalith, Lithobid
• Lurasidone • Latuda
• Olanzapine • Zyprexa
• Oxcarbazepine • Trileptal
• Quetiapine • Seroquel
• Risperidone • Risperdal
• Topiramate • Topamax
• Valproate • Depacon
• Ziprasidone • Geodon

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Acknowledgement

The authors would like to thank Natasha Barthel, BS, for her assistance with this article.

Discuss this article at www.facebook.com/CurrentPsychiatry

Over the past 50 years, anecdotal reports linking cannabis sativa (marijuana) and psychosis have been steadily accumulating, giving rise to the notion of “cannabis psychosis.” Despite this historic connection, marijuana often is regarded as a “soft drug” with few harmful effects. However, this benign view is now being revised, along with mounting research demonstrating a clear association between cannabis and psychosis.

In this article, I review evidence on marijuana’s impact on the risk of developing psychotic disorders, as well as the potential contributions of “medical” marijuana and other legally available products containing synthetic cannabinoids to psychosis risk.

Cannabis use and psychosis

Cannabis use has a largely deleterious effect on patients with psychotic disorders, and typically is associated with relapse, poor treatment adherence, and worsening psychotic symptoms.^1,2 There is, however, evidence that some patients with schizophrenia might benefit from treatment with cannabidiol,^3-5 another constituent of marijuana, as well as delta-9-tetrahydrocannabinol (Δ-9-THC), the principle psychoactive constituent of cannabis.^6,7

The acute psychotic potential of cannabis has been demonstrated by studies that documented psychotic symptoms (eg, hallucinations, paranoid delusions, derealization) in a dose-dependent manner among healthy volunteers administered Δ-9-THC under experimental conditions.^8-10 Various cross-sectional epidemiologic studies also have revealed an association between cannabis use and acute or chronic psychosis.^11,12

In the absence of definitive evidence from randomized, long-term, placebo-controlled trials, the strongest evidence of a connection between cannabis use and development of a psychotic disorder comes from prospective, longitudinal cohort studies. In the past 15 years, new evidence has emerged from 7 such studies that cumulatively provide strong support for an association between cannabis use as an adolescent or young adult and a greater risk for developing a psychotic disorder such as schizophrenia.^13-19 These longitudinal studies surveyed for self-reported cannabis use before psychosis onset and controlled for a variety of potential confounding factors (eg, other drug use and demographic, social, and psychological variables). Three meta-analyses of these and other studies concluded an increased risk of psychosis is associated with cannabis use, with an odds ratio of 1.4 to 2.9 (meaning the risk of developing psychosis with any history of cannabis use is up to 3-fold higher compared with those who did not use cannabis).^11,20,21 In addition, this association appears to be dose-related, with increasing amounts of cannabis use linked to greater risk—1 study found an odds ratio of 7 for psychosis among daily cannabis users.¹⁶

There are several ways to explain the link between cannabis use and psychosis, and a causal relationship has not yet been firmly established (Table 1).^{1-7,11-19,21-25} Current evidence supports that cannabis is a “component cause” of chronic psychosis, meaning although neither necessary nor sufficient, cannabis use at a young age increases the likelihood of developing schizophrenia or other psychotic disorders.²⁶ This risk may be greatest for young persons with some psychosis vulnerability (eg, those with attenuated psychotic symptoms).^16,18

The overall magnitude of risk appears to be modest, and cannabis use is only 1 of myriad factors that increase the risk of psychosis.²⁷ Furthermore, most cannabis users do not develop psychosis. However, the risk associated with cannabis occurs during a vulnerable time of development and is modifiable. Based on conservative estimates, 8% of emergent schizophrenia cases and 14% of more broadly defined emergent psychosis cases could be prevented if it were possible to eliminate cannabis use among young people.^11,26 Therefore, reducing cannabis use among young people vulnerable to psychosis should be a clinical and public health priority.

Table 1

Hypotheses linking cannabis and psychosis

Medical marijuana

Although cannabis extracts were marketed by major pharmaceutical companies and widely used by consumers for various ailments during the late 1800s, medicinal cannabis use in the United States declined significantly during the early 20^th century. In 1937, the Marihuana Tax Act was passed, effectively putting a stop to physicians prescribing cannabis for medical purposes. The FDA currently classifies cannabis as a Schedule I drug (eg, high abuse potential, no currently accepted medical use, lack of safety data) and the use of cannabis and its prescription by physicians are prohibited under federal law.

However, in recognition of the potential medical benefits of cannabis, 16 states have legalized medicinal use (“medical marijuana”) over the past several years. Laws and regulations governing medical marijuana vary from state to state. For example, in California, adults who obtain a recommendation from a physician and register for a Medical Marijuana Identification Card can legally purchase cannabis from a state-recognized dispensary and use it in a non-public setting. The physician’s “recommendation” (not a prescription) is based upon the determination that “the person’s health would benefit from the use of marijuana in the treatment of cancer, anorexia, AIDS, chronic pain, spasticity, glaucoma, arthritis, migraine, or any other illness for which marijuana provides relief”²⁸ (emphasis added). Although no state has yet legalized cannabis use for recreational purposes, with such regulations, an increasing number of jurisdictions have provided a way for consumers to easily obtain marijuana for loosely defined medical purposes.

Medical marijuana dispensaries offer a variety of cannabis strains, each with a different advertised “high” based upon variable proportions of Δ-9-THC and other constituents. The Δ-9-THC content of medical marijuana is about twice that of “street” marijuana, even with the latter’s Δ-9-THC content rising to >10% over the past 15 years.^29,30 Therefore, medical marijuana is not only legal, but generally offers a more potent Δ-9-THC dose than typical street marijuana.

A single case of psychosis emerging in the context of medical marijuana has been reported in the literature.³¹ A 24-year-old man with mild, transient psychotic symptoms switched from street cannabis to medical marijuana for its superior potency and to conform with the law. He obtained a physician’s recommendation based on diagnoses of “posttraumatic stress disorder” and “pain.” After several months of increasingly frequent medical marijuana use, he developed florid and persistent psychotic symptoms necessitating antipsychotic medication, and was diagnosed with schizophrenia.

Although causality cannot be established based on this report, taken together with evidence that higher-potency cannabis is associated with a greater risk of psychotic emergence,³² this case raises concerns about the iatrogenic and psychotoxic liability of medical marijuana use among those vulnerable to psychosis. Policy decisions about medical marijuana and its use among patients with psychiatric illness must be informed by evidence of its psychotic potential.

Synthetic cannabinoids

Synthetic cannabinoids were developed in the 1960s for research purposes and potential clinical applications, but have not been FDA-approved for therapeutic use.³³ Over the past 5 years, however, a variety of “herbal incense” products bearing names such as “Spice,” “K2,” and “Aroma” have emerged in Europe and the United States that contain botanicals laced with synthetic cannabinoids (Table 2).

Although herbal incense products are labeled “not for human consumption,” they are sold by “head shops” and on the Internet without age restrictions and typically are purchased for the sole purpose of ingesting them, usually by smoking. Their desired effects resemble cannabis intoxication, including sedation, relaxation, altered consciousness, and euphoria. The products initially had the added appeal of being legal and undetectable in routine drug screening. Although not listed among the product’s ingredients, chemical analyses confirmed these products typically contained 1 of 3 families of synthetic cannabinoid1 and cannabinoid2 (CB1/CB2) receptor agonists, designated by the prefixes JWH-, CP-, and HU-.³⁴ The compounds most commonly found in these analyses (JWH-018; CP-47,497; HU-210) have significantly greater potency (ie, CB1 receptor affinity) compared with Δ-9-THC.^33,34

The growing popularity of herbal incense products has prompted health concerns based on reports of emergency presentations for adverse effects, including tachycardia, agitation, excess sedation, and loss of consciousness.^33,35,36 In addition, 8 anecdotal reports of psychosis associated with herbal incense (with a total of 33 patients) have emerged since 2010 (Table 3). Among them, a variety of psychotic symptoms are described in patients ranging in age from adolescence to adulthood, both with and without histories of psychosis. For those without a pre-existing psychotic disorder, symptoms were typically self-limited.

In the most recently presented case series of patients without pre-existing psychosis (N = 10), symptoms resolved in 70% of patients within 8 days, but 30% had psychosis that persisted beyond 5-month follow-up.³⁷ Collectively, these reports suggest that synthetic cannabinoid intoxication is associated with acute psychosis as well as exacerbations of previously stable psychotic disorders, and also may have a propensity to trigger a chronic psychotic disorder among vulnerable individuals.

Because of health concerns and the abuse potential of herbal incense products, many have been banned in several European countries, 18 U.S. states, and the U.S. military.^33,38 In March 2011, the FDA placed 5 synthetic cannabinoids (JWH-018, JWH-073, JWH-200, CP-47,497, and cannabicyclohexanol) on Schedule I, making them illegal to possess or sell in the United States.³⁸ However, there are hundreds of synthetic cannabinoid homologues, and herbal incense manufacturers have rapidly adapted by substituting other synthetic cannabinoids not yet banned by existing legislation.³⁴ The effects of these newly arising compounds in humans, including their psychotic potential, are largely unknown.

Table 2

Herbal incense products and synthetic cannabinoids

Table 3

Case reports of psychosis associated with synthetic cannabinoids

Related Resources

• Murray RM, Morrison PD, Henquet C, et al. Cannabis, the mind and society: the hash realities. Nat Rev Neurosci. 2007;8(11):885-895.
• European Monitoring Centre for Drugs and Drug Addiction: Synthetic cannabinoids and “spice.” www.emcdda.europa.eu/publications/drug-profiles/synthetic-cannabinoids.
• U.S. Department of Justice, Drug Enforcement Agency, Office of Diversion Control: Schedules of controlled substances: temporary placement of five synthetic cannabinoids into Schedule I. www.deadiversion.usdoj.gov/fed_regs/rules/2011/fr0301.htm.

Discuss this article at www.facebook.com/CurrentPsychiatry

Over the past 50 years, anecdotal reports linking cannabis sativa (marijuana) and psychosis have been steadily accumulating, giving rise to the notion of “cannabis psychosis.” Despite this historic connection, marijuana often is regarded as a “soft drug” with few harmful effects. However, this benign view is now being revised, along with mounting research demonstrating a clear association between cannabis and psychosis.

In this article, I review evidence on marijuana’s impact on the risk of developing psychotic disorders, as well as the potential contributions of “medical” marijuana and other legally available products containing synthetic cannabinoids to psychosis risk.

Cannabis use and psychosis

Cannabis use has a largely deleterious effect on patients with psychotic disorders, and typically is associated with relapse, poor treatment adherence, and worsening psychotic symptoms.^1,2 There is, however, evidence that some patients with schizophrenia might benefit from treatment with cannabidiol,^3-5 another constituent of marijuana, as well as delta-9-tetrahydrocannabinol (Δ-9-THC), the principle psychoactive constituent of cannabis.^6,7

The acute psychotic potential of cannabis has been demonstrated by studies that documented psychotic symptoms (eg, hallucinations, paranoid delusions, derealization) in a dose-dependent manner among healthy volunteers administered Δ-9-THC under experimental conditions.^8-10 Various cross-sectional epidemiologic studies also have revealed an association between cannabis use and acute or chronic psychosis.^11,12

In the absence of definitive evidence from randomized, long-term, placebo-controlled trials, the strongest evidence of a connection between cannabis use and development of a psychotic disorder comes from prospective, longitudinal cohort studies. In the past 15 years, new evidence has emerged from 7 such studies that cumulatively provide strong support for an association between cannabis use as an adolescent or young adult and a greater risk for developing a psychotic disorder such as schizophrenia.^13-19 These longitudinal studies surveyed for self-reported cannabis use before psychosis onset and controlled for a variety of potential confounding factors (eg, other drug use and demographic, social, and psychological variables). Three meta-analyses of these and other studies concluded an increased risk of psychosis is associated with cannabis use, with an odds ratio of 1.4 to 2.9 (meaning the risk of developing psychosis with any history of cannabis use is up to 3-fold higher compared with those who did not use cannabis).^11,20,21 In addition, this association appears to be dose-related, with increasing amounts of cannabis use linked to greater risk—1 study found an odds ratio of 7 for psychosis among daily cannabis users.¹⁶

There are several ways to explain the link between cannabis use and psychosis, and a causal relationship has not yet been firmly established (Table 1).^{1-7,11-19,21-25} Current evidence supports that cannabis is a “component cause” of chronic psychosis, meaning although neither necessary nor sufficient, cannabis use at a young age increases the likelihood of developing schizophrenia or other psychotic disorders.²⁶ This risk may be greatest for young persons with some psychosis vulnerability (eg, those with attenuated psychotic symptoms).^16,18

The overall magnitude of risk appears to be modest, and cannabis use is only 1 of myriad factors that increase the risk of psychosis.²⁷ Furthermore, most cannabis users do not develop psychosis. However, the risk associated with cannabis occurs during a vulnerable time of development and is modifiable. Based on conservative estimates, 8% of emergent schizophrenia cases and 14% of more broadly defined emergent psychosis cases could be prevented if it were possible to eliminate cannabis use among young people.^11,26 Therefore, reducing cannabis use among young people vulnerable to psychosis should be a clinical and public health priority.

Table 1

Hypotheses linking cannabis and psychosis