Current Psychiatry

I thought Dr. Kowatch and colleagues took an important first step in pointing out that not all mood swings in children and adolescents are symptoms of bipolar disorder (“Not all mood swings are bipolar disorder,“ Current Psychiatry, February 2011, p. 38-52). They reviewed some of the other psychiatric conditions known to cause labile moods. One glaring omission is borderline personality disorder (BPD).

I am the medical director of a specialized unit that uses dialectical behavioral therapy (DBT) to treat children and adolescents with BPD. We have treated approximately 300 young women on the residential unit and many present similarly: multiple hospitalizations, multiple robust yet failed medication trials, severe and recurrent self-injury, suicide attempts, and a large degree of hopelessness. Most arrive with previous diagnoses of mood disorder not otherwise specified, bipolar disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, and others. It is when their outpatient psychiatrists and mental health teams have grown frustrated at the lack of enduring progress and are faced with the treatment demands of the borderline patient that a BPD diagnosis is considered. Even though research suggests that BPD—or at least some of its symptoms—begins in the late latency period of childhood,¹ treatment typically is not sought until late adolescence. This is the case despite the fact that BPD has a better prognosis than other serious mental illnesses, such as bipolar disorder.^2,3 Adult BPD patients almost universally recognize that their inability to regulate their mood started in late childhood and early adolescence. Structural and functional neuroimaging has revealed a dysfunctional network of brain regions that seem to mediate important aspects of BPD symptomatology.^4-6

These children have marked mood swings and great difficulty regulating their moods. The mood swings of BPD are not responsive to current medication unless there is comorbid bipolar disorder, in which case treatment with mood stabilizers helps improve vegetative symptoms such as sleep and energy, and reduce racing thoughts, pressured speech, and irritability. What these medications do not treat is the “reactive” mood swings that are characteristic of BPD. The mood reactivity often is triggered by interpersonal or intrapersonal conflict and rarely is long-lived.

Many children and adolescents are moody and most do not have a major psychiatric disorder. Of those who do, it is a great risk to patients’ health to not consider BPD, especially given new and empirically validated treatments, such as DBT. Astute clinicians should keep this diagnosis in mind when treating adolescents with moodiness, particularly when the mood is predominantly reactive to life’s stressors, when other features of the presentation do not fit neatly into a bipolar picture, and when multiple medications fail. On our unit, we have seen that the cognitive-behavioral strategies of DBT help patients even when BPD is not the diagnosis.

I would like to thank Dr. Kowatch and colleagues for expanding our thinking on mood swings and encourage readers to go one step further.

Blaise Aguirre, MD
Medical Director
Adolescent DBT Residential Program
McLean Hospital
Belmont, MA
Instructor in Psychiatry
Harvard Medical School
Boston, MA

References

1. Zanarini MC, Frankenburg FR, Khera GS, et al. Treatment histories of borderline inpatients. Compr Psychiatry. 2001;42:144-150.

2. Tohen M, Hennen J, Zarate CM Jr, et al. Two-year syndromal and functional recovery in 219 cases of first-episode major affective disorder with psychotic features. Am J Psychiatry. 2000;157:220-228.

3. Coryell W, Endicott J, Maser JD, et al. The likelihood of recurrence in bipolar affective disorder: the importance of episode recency. J Affect Disord. 1995;33:201-206.

4. De La Fuente JM, Goldman S, Stanus E, et al. Brain glucose metabolism in borderline personality disorder. J Psychiatr Res. 1997;31:531-541.

5. Soloff PH, Meltzer CC, Becker C, et al. Impulsivity and prefrontal hypometabolism in borderline personality disorder. Psychiatry Res. 2003;123:153-163.

6. Juengling FD, Schmahl C, Hesslinger B, et al. Positron emission tomography in female patients with borderline personality disorder. J Psychiatr Res. 2003;37:109-115.

The authors respond

We welcome comments about the importance of a thorough diagnostic evaluation to tease out possible etiologies of “mood swings, “ including psychosocial factors, as in personality disorders. Nevertheless, the debate about diagnosing personality disorders in children and adolescents is not settled. Developmentally, children and adolescents have continuous changes in biology and brain function. There is significantly more empirical evidence of reactive attachment disorders in childhood and adolescence that integrate the affective changes seen in children who live in chaotic environments. DSM defines BPD as a pervasive pattern of instability of interpersonal relationships that begins by early adulthood.¹ Many children with diagnoses of posttraumatic stress disorder, mood disorder, bipolar disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, etc. also can have difficulties in relating to others caused by their neurobiologic deficits, which also may limit response to medication. Furthermore, children with learning disorders also can misperceive motives of others and thus have pervasive patterns of relational instability.

The research Dr. Aguirre suggested is based on treatment histories and not rigorous study methodology. Most empirical evidence of personality disorders is strongly influenced by psychoanalytic literature regarding object relations, which is in flux because of emerging attention to attachment theory and progress in neurologic studies in the evaluation of temperamental variations related to the influence of mirror neurons.²

We also take issue with the comment that “BPD has a better prognosis than other serious men tal illn esses, such as bipolar disorder” There have been significant efforts in studying the role family can have in the outcomes of mood disorder treatment.³

Finally, there is evidence that in adults medication can be beneficial in treating the affective deregulation of patients with BPD who do not have comorbid disorders.⁴

Robert A. Kowatch, MD, PhD
Professor of Psychiatry and Pediatrics

Erin Monroe, CNS
Clinical Nurse Specialist
Division of Psychiatry

Sergio V. Delgado, MD
Associate Professor of Psychiatry
and Pediatrics
Cincinnati Children’s Hospital Medical Center
Cincinnati, OH

I thought Dr. Kowatch and colleagues took an important first step in pointing out that not all mood swings in children and adolescents are symptoms of bipolar disorder (“Not all mood swings are bipolar disorder,“ Current Psychiatry, February 2011, p. 38-52). They reviewed some of the other psychiatric conditions known to cause labile moods. One glaring omission is borderline personality disorder (BPD).

I am the medical director of a specialized unit that uses dialectical behavioral therapy (DBT) to treat children and adolescents with BPD. We have treated approximately 300 young women on the residential unit and many present similarly: multiple hospitalizations, multiple robust yet failed medication trials, severe and recurrent self-injury, suicide attempts, and a large degree of hopelessness. Most arrive with previous diagnoses of mood disorder not otherwise specified, bipolar disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, and others. It is when their outpatient psychiatrists and mental health teams have grown frustrated at the lack of enduring progress and are faced with the treatment demands of the borderline patient that a BPD diagnosis is considered. Even though research suggests that BPD—or at least some of its symptoms—begins in the late latency period of childhood,¹ treatment typically is not sought until late adolescence. This is the case despite the fact that BPD has a better prognosis than other serious mental illnesses, such as bipolar disorder.^2,3 Adult BPD patients almost universally recognize that their inability to regulate their mood started in late childhood and early adolescence. Structural and functional neuroimaging has revealed a dysfunctional network of brain regions that seem to mediate important aspects of BPD symptomatology.^4-6

These children have marked mood swings and great difficulty regulating their moods. The mood swings of BPD are not responsive to current medication unless there is comorbid bipolar disorder, in which case treatment with mood stabilizers helps improve vegetative symptoms such as sleep and energy, and reduce racing thoughts, pressured speech, and irritability. What these medications do not treat is the “reactive” mood swings that are characteristic of BPD. The mood reactivity often is triggered by interpersonal or intrapersonal conflict and rarely is long-lived.

Many children and adolescents are moody and most do not have a major psychiatric disorder. Of those who do, it is a great risk to patients’ health to not consider BPD, especially given new and empirically validated treatments, such as DBT. Astute clinicians should keep this diagnosis in mind when treating adolescents with moodiness, particularly when the mood is predominantly reactive to life’s stressors, when other features of the presentation do not fit neatly into a bipolar picture, and when multiple medications fail. On our unit, we have seen that the cognitive-behavioral strategies of DBT help patients even when BPD is not the diagnosis.

I would like to thank Dr. Kowatch and colleagues for expanding our thinking on mood swings and encourage readers to go one step further.

Blaise Aguirre, MD
Medical Director
Adolescent DBT Residential Program
McLean Hospital
Belmont, MA
Instructor in Psychiatry
Harvard Medical School
Boston, MA

References

1. Zanarini MC, Frankenburg FR, Khera GS, et al. Treatment histories of borderline inpatients. Compr Psychiatry. 2001;42:144-150.

2. Tohen M, Hennen J, Zarate CM Jr, et al. Two-year syndromal and functional recovery in 219 cases of first-episode major affective disorder with psychotic features. Am J Psychiatry. 2000;157:220-228.

3. Coryell W, Endicott J, Maser JD, et al. The likelihood of recurrence in bipolar affective disorder: the importance of episode recency. J Affect Disord. 1995;33:201-206.

4. De La Fuente JM, Goldman S, Stanus E, et al. Brain glucose metabolism in borderline personality disorder. J Psychiatr Res. 1997;31:531-541.

5. Soloff PH, Meltzer CC, Becker C, et al. Impulsivity and prefrontal hypometabolism in borderline personality disorder. Psychiatry Res. 2003;123:153-163.

6. Juengling FD, Schmahl C, Hesslinger B, et al. Positron emission tomography in female patients with borderline personality disorder. J Psychiatr Res. 2003;37:109-115.

The authors respond

We welcome comments about the importance of a thorough diagnostic evaluation to tease out possible etiologies of “mood swings, “ including psychosocial factors, as in personality disorders. Nevertheless, the debate about diagnosing personality disorders in children and adolescents is not settled. Developmentally, children and adolescents have continuous changes in biology and brain function. There is significantly more empirical evidence of reactive attachment disorders in childhood and adolescence that integrate the affective changes seen in children who live in chaotic environments. DSM defines BPD as a pervasive pattern of instability of interpersonal relationships that begins by early adulthood.¹ Many children with diagnoses of posttraumatic stress disorder, mood disorder, bipolar disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, etc. also can have difficulties in relating to others caused by their neurobiologic deficits, which also may limit response to medication. Furthermore, children with learning disorders also can misperceive motives of others and thus have pervasive patterns of relational instability.

The research Dr. Aguirre suggested is based on treatment histories and not rigorous study methodology. Most empirical evidence of personality disorders is strongly influenced by psychoanalytic literature regarding object relations, which is in flux because of emerging attention to attachment theory and progress in neurologic studies in the evaluation of temperamental variations related to the influence of mirror neurons.²

We also take issue with the comment that “BPD has a better prognosis than other serious men tal illn esses, such as bipolar disorder” There have been significant efforts in studying the role family can have in the outcomes of mood disorder treatment.³

Finally, there is evidence that in adults medication can be beneficial in treating the affective deregulation of patients with BPD who do not have comorbid disorders.⁴

Robert A. Kowatch, MD, PhD
Professor of Psychiatry and Pediatrics

Erin Monroe, CNS
Clinical Nurse Specialist
Division of Psychiatry

Sergio V. Delgado, MD
Associate Professor of Psychiatry
and Pediatrics
Cincinnati Children’s Hospital Medical Center
Cincinnati, OH

I thought Dr. Kowatch and colleagues took an important first step in pointing out that not all mood swings in children and adolescents are symptoms of bipolar disorder (“Not all mood swings are bipolar disorder,“ Current Psychiatry, February 2011, p. 38-52). They reviewed some of the other psychiatric conditions known to cause labile moods. One glaring omission is borderline personality disorder (BPD).

I am the medical director of a specialized unit that uses dialectical behavioral therapy (DBT) to treat children and adolescents with BPD. We have treated approximately 300 young women on the residential unit and many present similarly: multiple hospitalizations, multiple robust yet failed medication trials, severe and recurrent self-injury, suicide attempts, and a large degree of hopelessness. Most arrive with previous diagnoses of mood disorder not otherwise specified, bipolar disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, and others. It is when their outpatient psychiatrists and mental health teams have grown frustrated at the lack of enduring progress and are faced with the treatment demands of the borderline patient that a BPD diagnosis is considered. Even though research suggests that BPD—or at least some of its symptoms—begins in the late latency period of childhood,¹ treatment typically is not sought until late adolescence. This is the case despite the fact that BPD has a better prognosis than other serious mental illnesses, such as bipolar disorder.^2,3 Adult BPD patients almost universally recognize that their inability to regulate their mood started in late childhood and early adolescence. Structural and functional neuroimaging has revealed a dysfunctional network of brain regions that seem to mediate important aspects of BPD symptomatology.^4-6

These children have marked mood swings and great difficulty regulating their moods. The mood swings of BPD are not responsive to current medication unless there is comorbid bipolar disorder, in which case treatment with mood stabilizers helps improve vegetative symptoms such as sleep and energy, and reduce racing thoughts, pressured speech, and irritability. What these medications do not treat is the “reactive” mood swings that are characteristic of BPD. The mood reactivity often is triggered by interpersonal or intrapersonal conflict and rarely is long-lived.

Many children and adolescents are moody and most do not have a major psychiatric disorder. Of those who do, it is a great risk to patients’ health to not consider BPD, especially given new and empirically validated treatments, such as DBT. Astute clinicians should keep this diagnosis in mind when treating adolescents with moodiness, particularly when the mood is predominantly reactive to life’s stressors, when other features of the presentation do not fit neatly into a bipolar picture, and when multiple medications fail. On our unit, we have seen that the cognitive-behavioral strategies of DBT help patients even when BPD is not the diagnosis.

I would like to thank Dr. Kowatch and colleagues for expanding our thinking on mood swings and encourage readers to go one step further.

Blaise Aguirre, MD
Medical Director
Adolescent DBT Residential Program
McLean Hospital
Belmont, MA
Instructor in Psychiatry
Harvard Medical School
Boston, MA

References

1. Zanarini MC, Frankenburg FR, Khera GS, et al. Treatment histories of borderline inpatients. Compr Psychiatry. 2001;42:144-150.

2. Tohen M, Hennen J, Zarate CM Jr, et al. Two-year syndromal and functional recovery in 219 cases of first-episode major affective disorder with psychotic features. Am J Psychiatry. 2000;157:220-228.

3. Coryell W, Endicott J, Maser JD, et al. The likelihood of recurrence in bipolar affective disorder: the importance of episode recency. J Affect Disord. 1995;33:201-206.

4. De La Fuente JM, Goldman S, Stanus E, et al. Brain glucose metabolism in borderline personality disorder. J Psychiatr Res. 1997;31:531-541.

5. Soloff PH, Meltzer CC, Becker C, et al. Impulsivity and prefrontal hypometabolism in borderline personality disorder. Psychiatry Res. 2003;123:153-163.

6. Juengling FD, Schmahl C, Hesslinger B, et al. Positron emission tomography in female patients with borderline personality disorder. J Psychiatr Res. 2003;37:109-115.

The authors respond

We welcome comments about the importance of a thorough diagnostic evaluation to tease out possible etiologies of “mood swings, “ including psychosocial factors, as in personality disorders. Nevertheless, the debate about diagnosing personality disorders in children and adolescents is not settled. Developmentally, children and adolescents have continuous changes in biology and brain function. There is significantly more empirical evidence of reactive attachment disorders in childhood and adolescence that integrate the affective changes seen in children who live in chaotic environments. DSM defines BPD as a pervasive pattern of instability of interpersonal relationships that begins by early adulthood.¹ Many children with diagnoses of posttraumatic stress disorder, mood disorder, bipolar disorder, oppositional defiant disorder, attention-deficit/hyperactivity disorder, etc. also can have difficulties in relating to others caused by their neurobiologic deficits, which also may limit response to medication. Furthermore, children with learning disorders also can misperceive motives of others and thus have pervasive patterns of relational instability.

The research Dr. Aguirre suggested is based on treatment histories and not rigorous study methodology. Most empirical evidence of personality disorders is strongly influenced by psychoanalytic literature regarding object relations, which is in flux because of emerging attention to attachment theory and progress in neurologic studies in the evaluation of temperamental variations related to the influence of mirror neurons.²

We also take issue with the comment that “BPD has a better prognosis than other serious men tal illn esses, such as bipolar disorder” There have been significant efforts in studying the role family can have in the outcomes of mood disorder treatment.³

Finally, there is evidence that in adults medication can be beneficial in treating the affective deregulation of patients with BPD who do not have comorbid disorders.⁴

Robert A. Kowatch, MD, PhD
Professor of Psychiatry and Pediatrics

Erin Monroe, CNS
Clinical Nurse Specialist
Division of Psychiatry

Sergio V. Delgado, MD
Associate Professor of Psychiatry
and Pediatrics
Cincinnati Children’s Hospital Medical Center
Cincinnati, OH

You’ve heard about the 2 certainties in life: death and taxes. In psychiatric practice with complex and chronic patients, there is a third certainty: polypharmacy. It ranges from thoughtful to indiscriminate and seems to be entrenched in clinical practice, possibly reflecting practitioners’ desperation in trying to manage severely ill, treatment-resistant patients, usually in the absence of evidence-based guidelines.

I never fail to encounter polypharmacy in hospitals or clinics where I consult. I always wondered how the patient’s doctor knew which drug was exerting a therapeutic effect or which drug was causing side effects (parkinsonism, akathisia, sedation, orthostasis, dizziness, headache, blurry vision, etc. ). Over time, I came to categorize polypharmacy into 4 subtypes that span the spectrum from sensible to absurd. Here is my personal classification, which I trust that you, my readers, have witnessed as well.

Necessary polypharmacy. This variant of polypharmacy is evidence-based and proven in double-blind studies to be more effective than monotherapy. The most prominent example is adding an atypical antipsychotic to a mood stabilizer in bipolar mania. In fact, the superior efficacy of combination therapy in bipolar disorder is one of the oldest forms of rational polypharmacy, is supported by FDA trials, and is indicated whenever mood stabilizer monotherapy is not sufficient. For example, combining lithium and valproate is superior to either drug alone. Another example of FDA-approved combinations is combining small doses of an atypical antipsychotic to an antidepressant for treatment-resistant depression.

Reasonable polypharmacy. Although many of the combinations in this category are not FDA-approved, controlled studies support their use for suffering patients. Examples include:

• An atypical antipsychotic added to a selective serotonin reuptake inhibitor (SSRI) for obsessive-compulsive disorder (OCD) patients who do not improve on SSRI monotherapy.
• Modafinil added to clozapine in patients who suffer substantial and persistent daytime sedation or somnolence.
• Combining 2 antidepressants for major depressive disorder patients who partially respond to 1 antidepressant.
• Combining a mood stabilizer with an antidepressant for bipolar depression to prevent mood switching.

Ridiculous polypharmacy. The sky is the limit to the variations and degrees of ridiculous polypharmacy, but the theme is the same: an absurd concoction of psychotropic drugs across several classes, often including multiple agents from 1 or several classes. Here are examples I have seen in patient records:

• Two atypicals, an anticholinergic, a mood stabilizer, an antidepressant, and 2 benzodiazepines.
• Three antipsychotics (2 atypicals and 1 typical), 2 antidepressants, 3 sedative/hypnotics, and an anticonvulsant for weight control.
• This one takes the cake: 6 antipsychotics (2 typicals and 4 atypicals), plus an anticholinergic, 3 mood stabilizers, 2 antidepressants, 2 sleeping pills, a hypoglycemic agent, 2 antihypertensives, and a statin.

Hazardous polypharmacy. In this category, serious medical complications, toxic effects, or death may occur because of careless combinations of drugs that may interact to produce dangerous kinetic interactions or exacerbate a pre-existing medical condition. Examples include:

• Combining 1 psychotropic with another that may inhibit its metabolism (eg, prescribing fluvoxamine to a severely psychotic patient who developed OCD while receiving clozapine). There have been several toxic reactions and even death because fluvoxamine inhibits cytochrome 1A2, which metabolizes clozapine, thus increasing clozapine blood level by 400% to 500%.
• Combining 2 injectable drugs for agitation that may cause a serious medical complication. An example would be injecting a benzodiazepine such as lorazepam in a patient receiving olanzapine IM, which can cause severe respiratory depression or death.
• Combining several drugs, each of which may prolong the QTc interval, resulting in syncope or torsade de pointes.

Psychopharmacology can relieve the terrible anguish of psychosis, depression, or anxiety, but it also can carry iatrogenic risks if it is not based on scientific evidence. The practice of psychopharmacology requires the fully integrated skills of medical and psychiatric training to maximize benefit while avoiding harm. It also requires basic arithmetic skills: to consider subtracting drugs, not only adding them!

You’ve heard about the 2 certainties in life: death and taxes. In psychiatric practice with complex and chronic patients, there is a third certainty: polypharmacy. It ranges from thoughtful to indiscriminate and seems to be entrenched in clinical practice, possibly reflecting practitioners’ desperation in trying to manage severely ill, treatment-resistant patients, usually in the absence of evidence-based guidelines.

I never fail to encounter polypharmacy in hospitals or clinics where I consult. I always wondered how the patient’s doctor knew which drug was exerting a therapeutic effect or which drug was causing side effects (parkinsonism, akathisia, sedation, orthostasis, dizziness, headache, blurry vision, etc. ). Over time, I came to categorize polypharmacy into 4 subtypes that span the spectrum from sensible to absurd. Here is my personal classification, which I trust that you, my readers, have witnessed as well.

Necessary polypharmacy. This variant of polypharmacy is evidence-based and proven in double-blind studies to be more effective than monotherapy. The most prominent example is adding an atypical antipsychotic to a mood stabilizer in bipolar mania. In fact, the superior efficacy of combination therapy in bipolar disorder is one of the oldest forms of rational polypharmacy, is supported by FDA trials, and is indicated whenever mood stabilizer monotherapy is not sufficient. For example, combining lithium and valproate is superior to either drug alone. Another example of FDA-approved combinations is combining small doses of an atypical antipsychotic to an antidepressant for treatment-resistant depression.

Reasonable polypharmacy. Although many of the combinations in this category are not FDA-approved, controlled studies support their use for suffering patients. Examples include:

• An atypical antipsychotic added to a selective serotonin reuptake inhibitor (SSRI) for obsessive-compulsive disorder (OCD) patients who do not improve on SSRI monotherapy.
• Modafinil added to clozapine in patients who suffer substantial and persistent daytime sedation or somnolence.
• Combining 2 antidepressants for major depressive disorder patients who partially respond to 1 antidepressant.
• Combining a mood stabilizer with an antidepressant for bipolar depression to prevent mood switching.

Ridiculous polypharmacy. The sky is the limit to the variations and degrees of ridiculous polypharmacy, but the theme is the same: an absurd concoction of psychotropic drugs across several classes, often including multiple agents from 1 or several classes. Here are examples I have seen in patient records:

• Two atypicals, an anticholinergic, a mood stabilizer, an antidepressant, and 2 benzodiazepines.
• Three antipsychotics (2 atypicals and 1 typical), 2 antidepressants, 3 sedative/hypnotics, and an anticonvulsant for weight control.
• This one takes the cake: 6 antipsychotics (2 typicals and 4 atypicals), plus an anticholinergic, 3 mood stabilizers, 2 antidepressants, 2 sleeping pills, a hypoglycemic agent, 2 antihypertensives, and a statin.

Hazardous polypharmacy. In this category, serious medical complications, toxic effects, or death may occur because of careless combinations of drugs that may interact to produce dangerous kinetic interactions or exacerbate a pre-existing medical condition. Examples include:

• Combining 1 psychotropic with another that may inhibit its metabolism (eg, prescribing fluvoxamine to a severely psychotic patient who developed OCD while receiving clozapine). There have been several toxic reactions and even death because fluvoxamine inhibits cytochrome 1A2, which metabolizes clozapine, thus increasing clozapine blood level by 400% to 500%.
• Combining 2 injectable drugs for agitation that may cause a serious medical complication. An example would be injecting a benzodiazepine such as lorazepam in a patient receiving olanzapine IM, which can cause severe respiratory depression or death.
• Combining several drugs, each of which may prolong the QTc interval, resulting in syncope or torsade de pointes.

Psychopharmacology can relieve the terrible anguish of psychosis, depression, or anxiety, but it also can carry iatrogenic risks if it is not based on scientific evidence. The practice of psychopharmacology requires the fully integrated skills of medical and psychiatric training to maximize benefit while avoiding harm. It also requires basic arithmetic skills: to consider subtracting drugs, not only adding them!

You’ve heard about the 2 certainties in life: death and taxes. In psychiatric practice with complex and chronic patients, there is a third certainty: polypharmacy. It ranges from thoughtful to indiscriminate and seems to be entrenched in clinical practice, possibly reflecting practitioners’ desperation in trying to manage severely ill, treatment-resistant patients, usually in the absence of evidence-based guidelines.

I never fail to encounter polypharmacy in hospitals or clinics where I consult. I always wondered how the patient’s doctor knew which drug was exerting a therapeutic effect or which drug was causing side effects (parkinsonism, akathisia, sedation, orthostasis, dizziness, headache, blurry vision, etc. ). Over time, I came to categorize polypharmacy into 4 subtypes that span the spectrum from sensible to absurd. Here is my personal classification, which I trust that you, my readers, have witnessed as well.

Necessary polypharmacy. This variant of polypharmacy is evidence-based and proven in double-blind studies to be more effective than monotherapy. The most prominent example is adding an atypical antipsychotic to a mood stabilizer in bipolar mania. In fact, the superior efficacy of combination therapy in bipolar disorder is one of the oldest forms of rational polypharmacy, is supported by FDA trials, and is indicated whenever mood stabilizer monotherapy is not sufficient. For example, combining lithium and valproate is superior to either drug alone. Another example of FDA-approved combinations is combining small doses of an atypical antipsychotic to an antidepressant for treatment-resistant depression.

Reasonable polypharmacy. Although many of the combinations in this category are not FDA-approved, controlled studies support their use for suffering patients. Examples include:

• An atypical antipsychotic added to a selective serotonin reuptake inhibitor (SSRI) for obsessive-compulsive disorder (OCD) patients who do not improve on SSRI monotherapy.
• Modafinil added to clozapine in patients who suffer substantial and persistent daytime sedation or somnolence.
• Combining 2 antidepressants for major depressive disorder patients who partially respond to 1 antidepressant.
• Combining a mood stabilizer with an antidepressant for bipolar depression to prevent mood switching.

Ridiculous polypharmacy. The sky is the limit to the variations and degrees of ridiculous polypharmacy, but the theme is the same: an absurd concoction of psychotropic drugs across several classes, often including multiple agents from 1 or several classes. Here are examples I have seen in patient records:

• Two atypicals, an anticholinergic, a mood stabilizer, an antidepressant, and 2 benzodiazepines.
• Three antipsychotics (2 atypicals and 1 typical), 2 antidepressants, 3 sedative/hypnotics, and an anticonvulsant for weight control.
• This one takes the cake: 6 antipsychotics (2 typicals and 4 atypicals), plus an anticholinergic, 3 mood stabilizers, 2 antidepressants, 2 sleeping pills, a hypoglycemic agent, 2 antihypertensives, and a statin.

Hazardous polypharmacy. In this category, serious medical complications, toxic effects, or death may occur because of careless combinations of drugs that may interact to produce dangerous kinetic interactions or exacerbate a pre-existing medical condition. Examples include:

• Combining 1 psychotropic with another that may inhibit its metabolism (eg, prescribing fluvoxamine to a severely psychotic patient who developed OCD while receiving clozapine). There have been several toxic reactions and even death because fluvoxamine inhibits cytochrome 1A2, which metabolizes clozapine, thus increasing clozapine blood level by 400% to 500%.
• Combining 2 injectable drugs for agitation that may cause a serious medical complication. An example would be injecting a benzodiazepine such as lorazepam in a patient receiving olanzapine IM, which can cause severe respiratory depression or death.
• Combining several drugs, each of which may prolong the QTc interval, resulting in syncope or torsade de pointes.

Psychopharmacology can relieve the terrible anguish of psychosis, depression, or anxiety, but it also can carry iatrogenic risks if it is not based on scientific evidence. The practice of psychopharmacology requires the fully integrated skills of medical and psychiatric training to maximize benefit while avoiding harm. It also requires basic arithmetic skills: to consider subtracting drugs, not only adding them!

Discuss this article at www.facebook.com/CurrentPsychiatry

In patients with schizophrenia, positive symptoms typically respond to treatment, while negative and cognitive symptoms often persist and contribute to chronic disability.¹ Schizophrenia also is associated with widespread neurocognitive deficits—including impairments in executive functioning, learning, memory, and processing speed—that are a core feature of the disorder and may precede illness onset.²

Current treatment is based on the dopamine model of schizophrenia, which proposes that dopaminergic dysfunction is the basis for symptoms and cognitive deficits.³ Although this model is effective in guiding treatment for some patients, most show persistent disability despite receiving the best available treatment. Over the last 2 decades, researchers have developed alternative conceptual models of schizophrenia based on the psychotomimetic effects of compounds such as phencyclidine (PCP) and ketamine.⁴ These compounds function primarily by blocking N-methyl-D-aspartate (NMDA)-type glutamate receptors (NMDARs), which has lead researchers to focus on glutamatergic neurotransmission and NMDARs as a basis for new drug development. This article describes the glutamatergic model of schizophrenia and its implications for future treatments.

Dopaminergic models

Since the discovery of chlorpromazine almost 60 years ago, the dopamine model of schizophrenia has been widely accepted. It has gone through several iterations but in general suggests that schizophrenia is caused by dopaminergic system dysfunction, particularly increased dopamine within subcortical brain regions such as the striatum or nucleus accumben.³ The ability of amphetamine or other dopaminergic agents to induce symptoms closely resembling positive symptoms supports this model, as do genetic studies that show dopamine-related genes are associated with schizophrenia.⁵ In addition, all antipsychotics block dopamine type 2 receptors.

Unfortunately, limitations of this model continue to limit treatment:

• Dopaminergic compounds such as amphetamine do not induce negative symptoms or cognitive deficits similar to those observed in schizophrenia.
• Dopamine receptor blockers do not reverse cognitive dysfunction or negative symptoms.
• Dopaminergic instability observed during acute decompensation appears to resolve after stabilization even without symptom remission.
• Although dopaminergic systems preferentially innervate frontal brain regions, cognitive deficits in schizophrenia appear to be widespread, involving sensory as well as frontal brain systems.

Thus, dopaminergic dysfunction appears to account for only a part of schizophrenia’s symptomatic and neurocognitive profile.

Glutamatergic model

Approximately 20 years ago, researchers proposed an alternate schizophrenia model based on the observed clinical actions of “dissociative anesthetics,” including PCP and ketamine. PCP was patented in 1953 as a surgical anesthetic, but serious side effects, such as hallucinations, agitation, and catatonic-like reactions, soon curtailed its clinical use. As early as 1959, some researchers noted similarities between PCP psychosis and schizophrenia.^4,6

The binding site for PCP and other dissociative anesthetics (“PCP receptor”) was first described in 1979 and subsequently localized within the ion channel formed by the NMDAR. Glutamate is the primary excitatory neurotransmitter in the brain, and binds to NMDA and non-NMDA (eg, metabotropic or alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid [AMPA]) receptors. Binding of PCP prevents glutamate from activating NMDARs, which suggests that the pathogenesis of schizophrenia may be caused by dysfunction of NMDARs in particular or of the glutamatergic system in general. Unlike dopamine, the glutamatergic system is distributed throughout the brain and plays a prominent role in sensory processing and higher-level functions such as memory and executive functioning (Figure).⁶ Therefore, glutamatergic theories open new approaches for potential schizophrenia treatments, most of which are now entering clinical evaluation.

Figure: The wide reach of glutamatergic dysfunction
NMDA: N-methyl-D-aspartate; NMDAR: N-methyl-D-aspartate-type glutamate receptors
Source: Reference 6

Effects of NMDAR antagonists

In initial studies with PCP and ketamine in the early 1960s, researchers noted that these agents produced psychotic effects similar to schizophrenia symptoms.⁶ Further confirmation was obtained from retrospective studies of PCP abusers.⁶ It was not until the 1990s, however, that studies using modern operationalized symptom and neuropsychological rating scales were conducted. In those studies, healthy participants developed positive symptoms, negative symptoms, and cognitive dysfunction after receiving ketamine.^7,8 Moreover, in these studies the balance between negative and positive symptoms was similar to that typically observed in schizophrenia, as was the pattern of cognitive dysfunction. Therefore, unlike dopaminergic agents, NMDAR antagonists appear to be able to produce the full constellation of symptoms and cognitive deficits associated with schizophrenia.

Similarly, ketamine worsened positive and negative symptoms in patients diagnosed with schizophrenia.⁹ Although acute challenge with NMDAR antagonists does not produce schizophrenia-like auditory hallucinations in healthy controls, it does induce sensory distortions similar to those seen in individuals with early schizophrenia and does exacerbate pre-existing hallucinations in schizophrenia patients.¹⁰ Thus, acute challenge with NMDAR antagonists appears to re-create a state similar to the earliest stages of schizophrenia.⁶

NMDAR antagonists also reproduce the widespread neuropsychological abnormalities of schizophrenia (Figure).⁶ Ketamine infusion results in the severity and type of disorganized thinking seen in schizophrenia. Given the importance of neurocognitive dysfunction to the conceptualization of schizophrenia, these findings further support a glutamatergic model.

Sensory processing deficits

A key difference between dopaminergic and glutamatergic models is prediction of sensory processing deficits. Traditionally, dopaminergic models have viewed cognitive deficits of schizophrenia as being driven “top down” from higher order brain regions such as the prefrontal cortex, or from local dysfunction within regions such as the striatum.¹¹ In contrast, glutamatergic models predict that deficits also should be observed within sensory brain regions, such as the primary auditory and visual cortex.

Because of the focus on higher-level brain dysfunction, little research on sensory processing deficits was performed until recently. It has become increasingly clear that:

• patients with schizophrenia show severe deficits in early auditory and visual processing
• these deficits significantly contribute to patterns of cognitive dysfunction and psychosocial impairment.^12,13

In the auditory system, patients show deficits in pitch perception and, specifically, the ability to match tones after a brief delay. Schizophrenia patients show dysfunction in a specific part of the visual system called the magnocellular visual system. Deficits in these regions lead to impaired ability to detect emotion based on vocal intonation or facial expression, among other deficits.

In addition, reading ability—which was once thought to be normal in patients with schizophrenia—has been found to be severely disturbed.¹⁴ As in developmental dyslexia, impairments relate to dysfunction of underlying auditory and visual brain regions. Administering NMDAR antagonists to humans or animals causes deficits in the auditory and visual system similar to those seen in schizophrenia, which confirms the importance of NMDA dysfunction.

Glutamate-based treatments

Because NMDAR antagonists can induce schizophrenia symptoms, the most straightforward approach for treatment is to develop compounds that stimulate glutamate or NMDAR function (Table). The NMDAR contains modulatory sites that may be appropriate targets for drug development, including one that binds the amino acids glycine and D-serine and a redox site that is sensitive to brain glutathione levels. Reductions in brain D-serine and glutathione levels have been reported in schizophrenia, which suggests that impaired NMDAR regulation may contribute directly to brain dysfunction.¹⁵ Other treatment approaches being developed include targeting glycine transporters, which indirectly regulate brain levels of glycine, or metabotropic glutamate receptors, which modulate both pre-synaptic glutamate release and post-synaptic NMDAR function.

Table

Glutamatergic drugs in development

Glycine/D-serine site agonists. To date, most studies have used glutamatergic drugs adjunctive to antipsychotics and targeted the glycine/D-serine modulatory site, in part because glycine and D-serine are natural compounds and therefore FDA approval for their use could be obtained without the extensive preclinical development usually required for new chemical entities.¹⁶ Unfortunately, these agents are less potent than traditional pharmaceuticals, and delivering optimal doses may be impossible. Nevertheless, positive studies with these compounds have provided proof-of-concept for development of agents with higher affinity and specificity.

Studies have used glycine administered at doses up to 60 g/d, D-serine up to 8 g/d, or D-alanine approximately 6 g/d. For glycine, 60 g/d is the highest dose that can be given because of concerns about tolerability and replacement of other essential amino acids. D-serine originally was tested at approximately 2 g/d with promising results, but a recent open-label trial suggested that higher doses may be more efficacious.¹⁷ D-serine doses are limited by potential renal toxicity, as demonstrated in rodents studies.

Although not all studies of glycine/D-serine site agonists have been positive, a recent meta-analysis suggests significant improvement in negative symptoms across studies.¹⁸ Variability in statistical results across studies is related primarily to degree of placebo effect within individual trials, with a mean improvement in negative symptoms of approximately 15%. Glycine/D-serine site agonists seem to be less effective when combined with clozapine, possibly because clozapine may already enhance the glutamatergic system and increase synaptic glycine levels.⁶

One study that evaluated effects of open-label glycine in individuals with schizophrenia symptoms observed a large effect-size improvement, including early remission in 3 of 10 patients.¹⁹ These data—if confirmed by double-blind trials—would indicate that glycine/d-serine site agonists might have utility in treating the schizophrenia prodrome.

Glycine transport inhibitors. A potential indirect approach to raising glycine levels in the brain is using GlyT1-type glycine transport inhibitors (GTIs). GlyT1 transporters are co-localized in brain with NMDARs and modulate local glycine levels. Rather than binding directly to the NMDAR glycine binding site, GTIs increase glycine levels in the synapse by preventing its removal by GlyT1 transporters. Their function is analogous to using selective serotonin reuptake inhibitors to increase serotonin levels in patients with depression.⁶

Sarcosine (N-methylglycine) is a naturally occurring GlyT1 inhibitor that has been used in early clinical trials in Taiwan. Initial studies with sarcosine showed efficacy similar to—and in some cases better than—that of direct glycine/D-serine site agonists when added to first-generation or non-clozapine second-generation antipsychotics.¹⁸ Sarcosine also has been found to be effective for acute treatment of schizophrenia.²⁰ At present, however, sarcosine is not available for experimental use in the United States because of toxicity considerations.

Using high-affinity GTIs for schizophrenia was first proposed in the mid-1990s,⁶ but such compounds are only now entering clinical efficacy studies. Most recently, phase II results were presented for RG1678, a compound developed by Hoffman LaRoche.²¹ The study targeted persistent negative symptoms in patients receiving chronic antipsychotic treatment. Adding RG1678, 10 mg and 30 mg, to antipsychotics led to significant improvement in persistent negative symptoms vs placebo. These promising results are being followed up in phase III studies.

Other glutamatergic options. Few compounds are available to modulate NMDARs at sites other than the glycine/D-serine site. One study administered N-acetylcysteine, a glutathione precursor, as a potential treatment for persistent negative symptoms.²² Encouraging clinical results were observed in this double-blind study, along with improvement in electrophysiologic measures, negative symptoms, and overall functioning, but the study was limited by relatively high rates of noncompletion. Preclinical studies have combined D-serine with an inhibitor of D-amino acid oxidase to prevent D-serine breakdown.²³ In rodents, this approach produces a 30-fold increase in D-serine potency.

Tetrahydrobiopterin (BH₄) is a cofactor for enzymes responsible for the synthesis of dopamine and other monoamines, and presynaptic release of dopamine and glutamate. Reductions in BH₄ levels have been reported in schizophrenia, which suggests that this compound may be etiologically important.²⁴ Researchers have initiated a study of this compound in schizophrenia.

Other schizophrenia models propose that the crucial issue is not NMDA blockade but subsequent dysregulation of presynaptic glutamate release. Type 2/3 metabotropic glutamate receptors (mGluR2/3) are located on presynaptic glutamate terminals and inhibit presynaptic glutamate release. mGluR2/3 agonists have been shown to reverse ketamine’s effects in humans and in animal models,^25,26 which suggests a potential role in schizophrenia treatment.

The first mGluR2/3 agonist entered into monotherapy clinical efficacy trials for schizophrenia was LY-2140023. In an initial trial, this compound showed significant efficacy in improving positive and negative symptoms, comparable to that of olanzapine.²⁷ However, a follow-up study failed because of a large placebo effect,²⁸ which leaves the efficacy question unresolved.

In contrast to mGluR2/3, type 5 metabotropic receptors (mGluR5) are co-localized with NMDA receptors and potentiate activation. Thus, mGluR5 agonists also may be effective for treating schizophrenia. These compounds remain in preclinical development. Other approaches, such as stimulating specific types of GABA receptors to overcome glutamatergic deficits, remain promising but have not been tested in definitive clinical trials.

Related Resources

• Kantrowitz JT, Javitt DC. N-methyl-D-aspartate (NMDA) receptor dysfunction or dysregulation: the final common pathway on the road to schizophrenia? Brain Res Bull. 2010; 83(3-4): 108-121.
• Kantrowitz JT, Javitt DC. Glutamatergic approaches to the conceptualization and treatment of schizophrenia. In: Kantrowitz JT, Javitt DC, eds. Handbook of neurochemistry and molecular neurobiology: schizophrenia. 3rd ed. New York, NY: Springer; 2009.

Drug Brand Names

• Chlorpromazine • Thorazine
• Clozapine • Clozaril
• Ketamine • Ketalar
• Olanzapine • Zyprexa

Disclosures

Dr. Javitt receives grant/research support from Jazz Pharmaceuticals, Pfizer Inc., and Roche and is a consultant to AstraZeneca, Cypress, Eli Lilly and Company, NPS Pharmaceuticals, Sepracor, Solvay, Sunovion, and Takeda. He holds intellectual property rights for use of glycine, D-serine, and glycine transport inhibitors in treatment of schizophrenia and related disorders.

Dr. Kantrowitz receives grant/research support from Eli Lilly and Company, Jazz Pharmaceuticals, Pfizer Inc., Roche, and Sepracor.

Preparation of this manuscript was supported in part by National Institute of Health grants R01 DA03383, R37 MH49334, and P50 MH086385.

Discuss this article at www.facebook.com/CurrentPsychiatry

In patients with schizophrenia, positive symptoms typically respond to treatment, while negative and cognitive symptoms often persist and contribute to chronic disability.¹ Schizophrenia also is associated with widespread neurocognitive deficits—including impairments in executive functioning, learning, memory, and processing speed—that are a core feature of the disorder and may precede illness onset.²

Current treatment is based on the dopamine model of schizophrenia, which proposes that dopaminergic dysfunction is the basis for symptoms and cognitive deficits.³ Although this model is effective in guiding treatment for some patients, most show persistent disability despite receiving the best available treatment. Over the last 2 decades, researchers have developed alternative conceptual models of schizophrenia based on the psychotomimetic effects of compounds such as phencyclidine (PCP) and ketamine.⁴ These compounds function primarily by blocking N-methyl-D-aspartate (NMDA)-type glutamate receptors (NMDARs), which has lead researchers to focus on glutamatergic neurotransmission and NMDARs as a basis for new drug development. This article describes the glutamatergic model of schizophrenia and its implications for future treatments.

Dopaminergic models

Since the discovery of chlorpromazine almost 60 years ago, the dopamine model of schizophrenia has been widely accepted. It has gone through several iterations but in general suggests that schizophrenia is caused by dopaminergic system dysfunction, particularly increased dopamine within subcortical brain regions such as the striatum or nucleus accumben.³ The ability of amphetamine or other dopaminergic agents to induce symptoms closely resembling positive symptoms supports this model, as do genetic studies that show dopamine-related genes are associated with schizophrenia.⁵ In addition, all antipsychotics block dopamine type 2 receptors.

Unfortunately, limitations of this model continue to limit treatment:

• Dopaminergic compounds such as amphetamine do not induce negative symptoms or cognitive deficits similar to those observed in schizophrenia.
• Dopamine receptor blockers do not reverse cognitive dysfunction or negative symptoms.
• Dopaminergic instability observed during acute decompensation appears to resolve after stabilization even without symptom remission.
• Although dopaminergic systems preferentially innervate frontal brain regions, cognitive deficits in schizophrenia appear to be widespread, involving sensory as well as frontal brain systems.

Thus, dopaminergic dysfunction appears to account for only a part of schizophrenia’s symptomatic and neurocognitive profile.

Glutamatergic model

Approximately 20 years ago, researchers proposed an alternate schizophrenia model based on the observed clinical actions of “dissociative anesthetics,” including PCP and ketamine. PCP was patented in 1953 as a surgical anesthetic, but serious side effects, such as hallucinations, agitation, and catatonic-like reactions, soon curtailed its clinical use. As early as 1959, some researchers noted similarities between PCP psychosis and schizophrenia.^4,6

The binding site for PCP and other dissociative anesthetics (“PCP receptor”) was first described in 1979 and subsequently localized within the ion channel formed by the NMDAR. Glutamate is the primary excitatory neurotransmitter in the brain, and binds to NMDA and non-NMDA (eg, metabotropic or alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid [AMPA]) receptors. Binding of PCP prevents glutamate from activating NMDARs, which suggests that the pathogenesis of schizophrenia may be caused by dysfunction of NMDARs in particular or of the glutamatergic system in general. Unlike dopamine, the glutamatergic system is distributed throughout the brain and plays a prominent role in sensory processing and higher-level functions such as memory and executive functioning (Figure).⁶ Therefore, glutamatergic theories open new approaches for potential schizophrenia treatments, most of which are now entering clinical evaluation.

Figure: The wide reach of glutamatergic dysfunction
NMDA: N-methyl-D-aspartate; NMDAR: N-methyl-D-aspartate-type glutamate receptors
Source: Reference 6

Effects of NMDAR antagonists

In initial studies with PCP and ketamine in the early 1960s, researchers noted that these agents produced psychotic effects similar to schizophrenia symptoms.⁶ Further confirmation was obtained from retrospective studies of PCP abusers.⁶ It was not until the 1990s, however, that studies using modern operationalized symptom and neuropsychological rating scales were conducted. In those studies, healthy participants developed positive symptoms, negative symptoms, and cognitive dysfunction after receiving ketamine.^7,8 Moreover, in these studies the balance between negative and positive symptoms was similar to that typically observed in schizophrenia, as was the pattern of cognitive dysfunction. Therefore, unlike dopaminergic agents, NMDAR antagonists appear to be able to produce the full constellation of symptoms and cognitive deficits associated with schizophrenia.

Similarly, ketamine worsened positive and negative symptoms in patients diagnosed with schizophrenia.⁹ Although acute challenge with NMDAR antagonists does not produce schizophrenia-like auditory hallucinations in healthy controls, it does induce sensory distortions similar to those seen in individuals with early schizophrenia and does exacerbate pre-existing hallucinations in schizophrenia patients.¹⁰ Thus, acute challenge with NMDAR antagonists appears to re-create a state similar to the earliest stages of schizophrenia.⁶

NMDAR antagonists also reproduce the widespread neuropsychological abnormalities of schizophrenia (Figure).⁶ Ketamine infusion results in the severity and type of disorganized thinking seen in schizophrenia. Given the importance of neurocognitive dysfunction to the conceptualization of schizophrenia, these findings further support a glutamatergic model.

Sensory processing deficits

A key difference between dopaminergic and glutamatergic models is prediction of sensory processing deficits. Traditionally, dopaminergic models have viewed cognitive deficits of schizophrenia as being driven “top down” from higher order brain regions such as the prefrontal cortex, or from local dysfunction within regions such as the striatum.¹¹ In contrast, glutamatergic models predict that deficits also should be observed within sensory brain regions, such as the primary auditory and visual cortex.

Because of the focus on higher-level brain dysfunction, little research on sensory processing deficits was performed until recently. It has become increasingly clear that:

• patients with schizophrenia show severe deficits in early auditory and visual processing
• these deficits significantly contribute to patterns of cognitive dysfunction and psychosocial impairment.^12,13

In the auditory system, patients show deficits in pitch perception and, specifically, the ability to match tones after a brief delay. Schizophrenia patients show dysfunction in a specific part of the visual system called the magnocellular visual system. Deficits in these regions lead to impaired ability to detect emotion based on vocal intonation or facial expression, among other deficits.

In addition, reading ability—which was once thought to be normal in patients with schizophrenia—has been found to be severely disturbed.¹⁴ As in developmental dyslexia, impairments relate to dysfunction of underlying auditory and visual brain regions. Administering NMDAR antagonists to humans or animals causes deficits in the auditory and visual system similar to those seen in schizophrenia, which confirms the importance of NMDA dysfunction.

Glutamate-based treatments

Because NMDAR antagonists can induce schizophrenia symptoms, the most straightforward approach for treatment is to develop compounds that stimulate glutamate or NMDAR function (Table). The NMDAR contains modulatory sites that may be appropriate targets for drug development, including one that binds the amino acids glycine and D-serine and a redox site that is sensitive to brain glutathione levels. Reductions in brain D-serine and glutathione levels have been reported in schizophrenia, which suggests that impaired NMDAR regulation may contribute directly to brain dysfunction.¹⁵ Other treatment approaches being developed include targeting glycine transporters, which indirectly regulate brain levels of glycine, or metabotropic glutamate receptors, which modulate both pre-synaptic glutamate release and post-synaptic NMDAR function.

Table

Glutamatergic drugs in development

Glycine/D-serine site agonists. To date, most studies have used glutamatergic drugs adjunctive to antipsychotics and targeted the glycine/D-serine modulatory site, in part because glycine and D-serine are natural compounds and therefore FDA approval for their use could be obtained without the extensive preclinical development usually required for new chemical entities.¹⁶ Unfortunately, these agents are less potent than traditional pharmaceuticals, and delivering optimal doses may be impossible. Nevertheless, positive studies with these compounds have provided proof-of-concept for development of agents with higher affinity and specificity.

Studies have used glycine administered at doses up to 60 g/d, D-serine up to 8 g/d, or D-alanine approximately 6 g/d. For glycine, 60 g/d is the highest dose that can be given because of concerns about tolerability and replacement of other essential amino acids. D-serine originally was tested at approximately 2 g/d with promising results, but a recent open-label trial suggested that higher doses may be more efficacious.¹⁷ D-serine doses are limited by potential renal toxicity, as demonstrated in rodents studies.

Although not all studies of glycine/D-serine site agonists have been positive, a recent meta-analysis suggests significant improvement in negative symptoms across studies.¹⁸ Variability in statistical results across studies is related primarily to degree of placebo effect within individual trials, with a mean improvement in negative symptoms of approximately 15%. Glycine/D-serine site agonists seem to be less effective when combined with clozapine, possibly because clozapine may already enhance the glutamatergic system and increase synaptic glycine levels.⁶

One study that evaluated effects of open-label glycine in individuals with schizophrenia symptoms observed a large effect-size improvement, including early remission in 3 of 10 patients.¹⁹ These data—if confirmed by double-blind trials—would indicate that glycine/d-serine site agonists might have utility in treating the schizophrenia prodrome.

Glycine transport inhibitors. A potential indirect approach to raising glycine levels in the brain is using GlyT1-type glycine transport inhibitors (GTIs). GlyT1 transporters are co-localized in brain with NMDARs and modulate local glycine levels. Rather than binding directly to the NMDAR glycine binding site, GTIs increase glycine levels in the synapse by preventing its removal by GlyT1 transporters. Their function is analogous to using selective serotonin reuptake inhibitors to increase serotonin levels in patients with depression.⁶

Sarcosine (N-methylglycine) is a naturally occurring GlyT1 inhibitor that has been used in early clinical trials in Taiwan. Initial studies with sarcosine showed efficacy similar to—and in some cases better than—that of direct glycine/D-serine site agonists when added to first-generation or non-clozapine second-generation antipsychotics.¹⁸ Sarcosine also has been found to be effective for acute treatment of schizophrenia.²⁰ At present, however, sarcosine is not available for experimental use in the United States because of toxicity considerations.

Using high-affinity GTIs for schizophrenia was first proposed in the mid-1990s,⁶ but such compounds are only now entering clinical efficacy studies. Most recently, phase II results were presented for RG1678, a compound developed by Hoffman LaRoche.²¹ The study targeted persistent negative symptoms in patients receiving chronic antipsychotic treatment. Adding RG1678, 10 mg and 30 mg, to antipsychotics led to significant improvement in persistent negative symptoms vs placebo. These promising results are being followed up in phase III studies.

Other glutamatergic options. Few compounds are available to modulate NMDARs at sites other than the glycine/D-serine site. One study administered N-acetylcysteine, a glutathione precursor, as a potential treatment for persistent negative symptoms.²² Encouraging clinical results were observed in this double-blind study, along with improvement in electrophysiologic measures, negative symptoms, and overall functioning, but the study was limited by relatively high rates of noncompletion. Preclinical studies have combined D-serine with an inhibitor of D-amino acid oxidase to prevent D-serine breakdown.²³ In rodents, this approach produces a 30-fold increase in D-serine potency.

Tetrahydrobiopterin (BH₄) is a cofactor for enzymes responsible for the synthesis of dopamine and other monoamines, and presynaptic release of dopamine and glutamate. Reductions in BH₄ levels have been reported in schizophrenia, which suggests that this compound may be etiologically important.²⁴ Researchers have initiated a study of this compound in schizophrenia.

Other schizophrenia models propose that the crucial issue is not NMDA blockade but subsequent dysregulation of presynaptic glutamate release. Type 2/3 metabotropic glutamate receptors (mGluR2/3) are located on presynaptic glutamate terminals and inhibit presynaptic glutamate release. mGluR2/3 agonists have been shown to reverse ketamine’s effects in humans and in animal models,^25,26 which suggests a potential role in schizophrenia treatment.

The first mGluR2/3 agonist entered into monotherapy clinical efficacy trials for schizophrenia was LY-2140023. In an initial trial, this compound showed significant efficacy in improving positive and negative symptoms, comparable to that of olanzapine.²⁷ However, a follow-up study failed because of a large placebo effect,²⁸ which leaves the efficacy question unresolved.

In contrast to mGluR2/3, type 5 metabotropic receptors (mGluR5) are co-localized with NMDA receptors and potentiate activation. Thus, mGluR5 agonists also may be effective for treating schizophrenia. These compounds remain in preclinical development. Other approaches, such as stimulating specific types of GABA receptors to overcome glutamatergic deficits, remain promising but have not been tested in definitive clinical trials.

Related Resources

• Kantrowitz JT, Javitt DC. N-methyl-D-aspartate (NMDA) receptor dysfunction or dysregulation: the final common pathway on the road to schizophrenia? Brain Res Bull. 2010; 83(3-4): 108-121.
• Kantrowitz JT, Javitt DC. Glutamatergic approaches to the conceptualization and treatment of schizophrenia. In: Kantrowitz JT, Javitt DC, eds. Handbook of neurochemistry and molecular neurobiology: schizophrenia. 3rd ed. New York, NY: Springer; 2009.

Drug Brand Names

• Chlorpromazine • Thorazine
• Clozapine • Clozaril
• Ketamine • Ketalar
• Olanzapine • Zyprexa

Disclosures

Dr. Javitt receives grant/research support from Jazz Pharmaceuticals, Pfizer Inc., and Roche and is a consultant to AstraZeneca, Cypress, Eli Lilly and Company, NPS Pharmaceuticals, Sepracor, Solvay, Sunovion, and Takeda. He holds intellectual property rights for use of glycine, D-serine, and glycine transport inhibitors in treatment of schizophrenia and related disorders.

Dr. Kantrowitz receives grant/research support from Eli Lilly and Company, Jazz Pharmaceuticals, Pfizer Inc., Roche, and Sepracor.

Preparation of this manuscript was supported in part by National Institute of Health grants R01 DA03383, R37 MH49334, and P50 MH086385.

Discuss this article at www.facebook.com/CurrentPsychiatry

In patients with schizophrenia, positive symptoms typically respond to treatment, while negative and cognitive symptoms often persist and contribute to chronic disability.¹ Schizophrenia also is associated with widespread neurocognitive deficits—including impairments in executive functioning, learning, memory, and processing speed—that are a core feature of the disorder and may precede illness onset.²

Current treatment is based on the dopamine model of schizophrenia, which proposes that dopaminergic dysfunction is the basis for symptoms and cognitive deficits.³ Although this model is effective in guiding treatment for some patients, most show persistent disability despite receiving the best available treatment. Over the last 2 decades, researchers have developed alternative conceptual models of schizophrenia based on the psychotomimetic effects of compounds such as phencyclidine (PCP) and ketamine.⁴ These compounds function primarily by blocking N-methyl-D-aspartate (NMDA)-type glutamate receptors (NMDARs), which has lead researchers to focus on glutamatergic neurotransmission and NMDARs as a basis for new drug development. This article describes the glutamatergic model of schizophrenia and its implications for future treatments.

Dopaminergic models

Since the discovery of chlorpromazine almost 60 years ago, the dopamine model of schizophrenia has been widely accepted. It has gone through several iterations but in general suggests that schizophrenia is caused by dopaminergic system dysfunction, particularly increased dopamine within subcortical brain regions such as the striatum or nucleus accumben.³ The ability of amphetamine or other dopaminergic agents to induce symptoms closely resembling positive symptoms supports this model, as do genetic studies that show dopamine-related genes are associated with schizophrenia.⁵ In addition, all antipsychotics block dopamine type 2 receptors.

Unfortunately, limitations of this model continue to limit treatment:

• Dopaminergic compounds such as amphetamine do not induce negative symptoms or cognitive deficits similar to those observed in schizophrenia.
• Dopamine receptor blockers do not reverse cognitive dysfunction or negative symptoms.
• Dopaminergic instability observed during acute decompensation appears to resolve after stabilization even without symptom remission.
• Although dopaminergic systems preferentially innervate frontal brain regions, cognitive deficits in schizophrenia appear to be widespread, involving sensory as well as frontal brain systems.

Thus, dopaminergic dysfunction appears to account for only a part of schizophrenia’s symptomatic and neurocognitive profile.

Glutamatergic model

Approximately 20 years ago, researchers proposed an alternate schizophrenia model based on the observed clinical actions of “dissociative anesthetics,” including PCP and ketamine. PCP was patented in 1953 as a surgical anesthetic, but serious side effects, such as hallucinations, agitation, and catatonic-like reactions, soon curtailed its clinical use. As early as 1959, some researchers noted similarities between PCP psychosis and schizophrenia.^4,6

The binding site for PCP and other dissociative anesthetics (“PCP receptor”) was first described in 1979 and subsequently localized within the ion channel formed by the NMDAR. Glutamate is the primary excitatory neurotransmitter in the brain, and binds to NMDA and non-NMDA (eg, metabotropic or alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid [AMPA]) receptors. Binding of PCP prevents glutamate from activating NMDARs, which suggests that the pathogenesis of schizophrenia may be caused by dysfunction of NMDARs in particular or of the glutamatergic system in general. Unlike dopamine, the glutamatergic system is distributed throughout the brain and plays a prominent role in sensory processing and higher-level functions such as memory and executive functioning (Figure).⁶ Therefore, glutamatergic theories open new approaches for potential schizophrenia treatments, most of which are now entering clinical evaluation.

Figure: The wide reach of glutamatergic dysfunction
NMDA: N-methyl-D-aspartate; NMDAR: N-methyl-D-aspartate-type glutamate receptors
Source: Reference 6

Effects of NMDAR antagonists

In initial studies with PCP and ketamine in the early 1960s, researchers noted that these agents produced psychotic effects similar to schizophrenia symptoms.⁶ Further confirmation was obtained from retrospective studies of PCP abusers.⁶ It was not until the 1990s, however, that studies using modern operationalized symptom and neuropsychological rating scales were conducted. In those studies, healthy participants developed positive symptoms, negative symptoms, and cognitive dysfunction after receiving ketamine.^7,8 Moreover, in these studies the balance between negative and positive symptoms was similar to that typically observed in schizophrenia, as was the pattern of cognitive dysfunction. Therefore, unlike dopaminergic agents, NMDAR antagonists appear to be able to produce the full constellation of symptoms and cognitive deficits associated with schizophrenia.

Similarly, ketamine worsened positive and negative symptoms in patients diagnosed with schizophrenia.⁹ Although acute challenge with NMDAR antagonists does not produce schizophrenia-like auditory hallucinations in healthy controls, it does induce sensory distortions similar to those seen in individuals with early schizophrenia and does exacerbate pre-existing hallucinations in schizophrenia patients.¹⁰ Thus, acute challenge with NMDAR antagonists appears to re-create a state similar to the earliest stages of schizophrenia.⁶

NMDAR antagonists also reproduce the widespread neuropsychological abnormalities of schizophrenia (Figure).⁶ Ketamine infusion results in the severity and type of disorganized thinking seen in schizophrenia. Given the importance of neurocognitive dysfunction to the conceptualization of schizophrenia, these findings further support a glutamatergic model.

Sensory processing deficits

A key difference between dopaminergic and glutamatergic models is prediction of sensory processing deficits. Traditionally, dopaminergic models have viewed cognitive deficits of schizophrenia as being driven “top down” from higher order brain regions such as the prefrontal cortex, or from local dysfunction within regions such as the striatum.¹¹ In contrast, glutamatergic models predict that deficits also should be observed within sensory brain regions, such as the primary auditory and visual cortex.

Because of the focus on higher-level brain dysfunction, little research on sensory processing deficits was performed until recently. It has become increasingly clear that:

• patients with schizophrenia show severe deficits in early auditory and visual processing
• these deficits significantly contribute to patterns of cognitive dysfunction and psychosocial impairment.^12,13

In the auditory system, patients show deficits in pitch perception and, specifically, the ability to match tones after a brief delay. Schizophrenia patients show dysfunction in a specific part of the visual system called the magnocellular visual system. Deficits in these regions lead to impaired ability to detect emotion based on vocal intonation or facial expression, among other deficits.

In addition, reading ability—which was once thought to be normal in patients with schizophrenia—has been found to be severely disturbed.¹⁴ As in developmental dyslexia, impairments relate to dysfunction of underlying auditory and visual brain regions. Administering NMDAR antagonists to humans or animals causes deficits in the auditory and visual system similar to those seen in schizophrenia, which confirms the importance of NMDA dysfunction.

Glutamate-based treatments

Because NMDAR antagonists can induce schizophrenia symptoms, the most straightforward approach for treatment is to develop compounds that stimulate glutamate or NMDAR function (Table). The NMDAR contains modulatory sites that may be appropriate targets for drug development, including one that binds the amino acids glycine and D-serine and a redox site that is sensitive to brain glutathione levels. Reductions in brain D-serine and glutathione levels have been reported in schizophrenia, which suggests that impaired NMDAR regulation may contribute directly to brain dysfunction.¹⁵ Other treatment approaches being developed include targeting glycine transporters, which indirectly regulate brain levels of glycine, or metabotropic glutamate receptors, which modulate both pre-synaptic glutamate release and post-synaptic NMDAR function.

Table

Glutamatergic drugs in development

Glycine/D-serine site agonists. To date, most studies have used glutamatergic drugs adjunctive to antipsychotics and targeted the glycine/D-serine modulatory site, in part because glycine and D-serine are natural compounds and therefore FDA approval for their use could be obtained without the extensive preclinical development usually required for new chemical entities.¹⁶ Unfortunately, these agents are less potent than traditional pharmaceuticals, and delivering optimal doses may be impossible. Nevertheless, positive studies with these compounds have provided proof-of-concept for development of agents with higher affinity and specificity.

Studies have used glycine administered at doses up to 60 g/d, D-serine up to 8 g/d, or D-alanine approximately 6 g/d. For glycine, 60 g/d is the highest dose that can be given because of concerns about tolerability and replacement of other essential amino acids. D-serine originally was tested at approximately 2 g/d with promising results, but a recent open-label trial suggested that higher doses may be more efficacious.¹⁷ D-serine doses are limited by potential renal toxicity, as demonstrated in rodents studies.

Although not all studies of glycine/D-serine site agonists have been positive, a recent meta-analysis suggests significant improvement in negative symptoms across studies.¹⁸ Variability in statistical results across studies is related primarily to degree of placebo effect within individual trials, with a mean improvement in negative symptoms of approximately 15%. Glycine/D-serine site agonists seem to be less effective when combined with clozapine, possibly because clozapine may already enhance the glutamatergic system and increase synaptic glycine levels.⁶

One study that evaluated effects of open-label glycine in individuals with schizophrenia symptoms observed a large effect-size improvement, including early remission in 3 of 10 patients.¹⁹ These data—if confirmed by double-blind trials—would indicate that glycine/d-serine site agonists might have utility in treating the schizophrenia prodrome.

Glycine transport inhibitors. A potential indirect approach to raising glycine levels in the brain is using GlyT1-type glycine transport inhibitors (GTIs). GlyT1 transporters are co-localized in brain with NMDARs and modulate local glycine levels. Rather than binding directly to the NMDAR glycine binding site, GTIs increase glycine levels in the synapse by preventing its removal by GlyT1 transporters. Their function is analogous to using selective serotonin reuptake inhibitors to increase serotonin levels in patients with depression.⁶

Sarcosine (N-methylglycine) is a naturally occurring GlyT1 inhibitor that has been used in early clinical trials in Taiwan. Initial studies with sarcosine showed efficacy similar to—and in some cases better than—that of direct glycine/D-serine site agonists when added to first-generation or non-clozapine second-generation antipsychotics.¹⁸ Sarcosine also has been found to be effective for acute treatment of schizophrenia.²⁰ At present, however, sarcosine is not available for experimental use in the United States because of toxicity considerations.

Using high-affinity GTIs for schizophrenia was first proposed in the mid-1990s,⁶ but such compounds are only now entering clinical efficacy studies. Most recently, phase II results were presented for RG1678, a compound developed by Hoffman LaRoche.²¹ The study targeted persistent negative symptoms in patients receiving chronic antipsychotic treatment. Adding RG1678, 10 mg and 30 mg, to antipsychotics led to significant improvement in persistent negative symptoms vs placebo. These promising results are being followed up in phase III studies.

Other glutamatergic options. Few compounds are available to modulate NMDARs at sites other than the glycine/D-serine site. One study administered N-acetylcysteine, a glutathione precursor, as a potential treatment for persistent negative symptoms.²² Encouraging clinical results were observed in this double-blind study, along with improvement in electrophysiologic measures, negative symptoms, and overall functioning, but the study was limited by relatively high rates of noncompletion. Preclinical studies have combined D-serine with an inhibitor of D-amino acid oxidase to prevent D-serine breakdown.²³ In rodents, this approach produces a 30-fold increase in D-serine potency.

Tetrahydrobiopterin (BH₄) is a cofactor for enzymes responsible for the synthesis of dopamine and other monoamines, and presynaptic release of dopamine and glutamate. Reductions in BH₄ levels have been reported in schizophrenia, which suggests that this compound may be etiologically important.²⁴ Researchers have initiated a study of this compound in schizophrenia.

Other schizophrenia models propose that the crucial issue is not NMDA blockade but subsequent dysregulation of presynaptic glutamate release. Type 2/3 metabotropic glutamate receptors (mGluR2/3) are located on presynaptic glutamate terminals and inhibit presynaptic glutamate release. mGluR2/3 agonists have been shown to reverse ketamine’s effects in humans and in animal models,^25,26 which suggests a potential role in schizophrenia treatment.

The first mGluR2/3 agonist entered into monotherapy clinical efficacy trials for schizophrenia was LY-2140023. In an initial trial, this compound showed significant efficacy in improving positive and negative symptoms, comparable to that of olanzapine.²⁷ However, a follow-up study failed because of a large placebo effect,²⁸ which leaves the efficacy question unresolved.

In contrast to mGluR2/3, type 5 metabotropic receptors (mGluR5) are co-localized with NMDA receptors and potentiate activation. Thus, mGluR5 agonists also may be effective for treating schizophrenia. These compounds remain in preclinical development. Other approaches, such as stimulating specific types of GABA receptors to overcome glutamatergic deficits, remain promising but have not been tested in definitive clinical trials.

Related Resources

• Kantrowitz JT, Javitt DC. N-methyl-D-aspartate (NMDA) receptor dysfunction or dysregulation: the final common pathway on the road to schizophrenia? Brain Res Bull. 2010; 83(3-4): 108-121.
• Kantrowitz JT, Javitt DC. Glutamatergic approaches to the conceptualization and treatment of schizophrenia. In: Kantrowitz JT, Javitt DC, eds. Handbook of neurochemistry and molecular neurobiology: schizophrenia. 3rd ed. New York, NY: Springer; 2009.

Drug Brand Names

• Chlorpromazine • Thorazine
• Clozapine • Clozaril
• Ketamine • Ketalar
• Olanzapine • Zyprexa

Disclosures

Dr. Javitt receives grant/research support from Jazz Pharmaceuticals, Pfizer Inc., and Roche and is a consultant to AstraZeneca, Cypress, Eli Lilly and Company, NPS Pharmaceuticals, Sepracor, Solvay, Sunovion, and Takeda. He holds intellectual property rights for use of glycine, D-serine, and glycine transport inhibitors in treatment of schizophrenia and related disorders.

Dr. Kantrowitz receives grant/research support from Eli Lilly and Company, Jazz Pharmaceuticals, Pfizer Inc., Roche, and Sepracor.

Preparation of this manuscript was supported in part by National Institute of Health grants R01 DA03383, R37 MH49334, and P50 MH086385.

Discuss this article at www.facebook.com/CurrentPsychiatry

Among DSM axis I diagnoses, bipolar disorder (BD) has the highest rates of comorbid substance use disorders (SUDs).^1-3 Approximately 60% of patients with bipolar I disorder have a lifetime diagnosis of an SUD.¹ Excluding tobacco, alcohol is the substance most often abused by BD patients, followed by cannabis, amphetamines, and cocaine.^1-3

BD patients with comorbid SUD usually exhibit more severe clinical presentations and poorer outcomes than their counterparts without SUDs. Compared with patients with BD alone, those with BD and SUD comorbidity (BD-SUD) experience earlier onset of mood symptoms; higher rates of anxiety disorders, suicide attempts, accidents, hospitalizations, and rapid cycling; more depressive episodes; and lower treatment compliance.^4-9

Several treatment options are available for patients with BD-SUD, including psychotherapy modalities, medications primarily used to treat BD, and medications primarily used to treat SUDs. Evidence-based support for these treatments remains limited, and no treatment of choice has emerged. This article reviews evidence on the longer-term treatment of BD-SUD, including general strategies and specific psychosocial and pharmacologic interventions. Short-term treatment strategies, such as pharmacotherapy for detoxification, are outside the scope of this review.

General strategies

The causes of BD-SUD are complex. Evidence suggests that the presence of affective symptoms is associated with an increased risk for substance misuse. This should be kept in mind when treating a patient with BD-SUD because controlling mood symptoms probably will help control substance abuse. However, evidence also shows that SUDs may be independent of mood episodes. Therefore, treating only mood symptoms in the hope that doing so will control substance abuse may not be enough.

Because the negative impact of SUDs on BD outcome is well documented, inform patients that limiting their use of alcohol and/or drugs is vital to control their mood disorder. Efforts to educate, stimulate, and support patients to moderate or stop their alcohol and/or drug use are likely to result in positive changes.¹⁰ Therefore, treatment for BD-SUD should follow, in part, the same recommendations for treatment of SUDs in patients with no comorbid axis I disorders:

• identify the problem (ie, the existence of a comorbid SUD)
• share your concerns with your patient
• offer appropriate and specific treatments, such as detoxification and/or self-help and counseling programs.¹⁰

Because SUDs usually are chronic and relapsing conditions, periods of drug and/ or alcohol use should be expected and not considered a sign of treatment failure. In addition, integrating treatment for both conditions probably is better than managing each separately. Therefore, targeting BD symptoms with mood-stabilizing medications and substance abuse with nonpharmacologic modalities such as drug counseling likely will bring about the best results.

Compared with BD patients without comorbid SUD, BD-SUD patients have a 7-fold increased risk of antidepressantinduced mania.¹¹ Therefore, antidepressants should be prescribed cautiously for patients with BD-SUD.

Integrated psychosocial therapy

BD-SUD patients may benefit from attending self-help programs such as Alcoholics Anonymous and Narcotics Anonymous, provided their mood is stable enough to allow them to participate. Other forms of psychotherapy for BD-SUD patients include standard group drug counseling and integrated group therapy that simultaneously addresses both conditions.

Integrated group therapy is based on the premise that changing maladaptive mood cognitions and behaviors will facilitate recovery from SUDs, and changing maladaptive substance use cognitions and behaviors will facilitate recovery from mood disorders.¹² In a recent randomized controlled trial, 62 BD-SUD patients were blindly assigned to integrated group therapy or standard group drug counseling and followed for 3 months.¹² Pharmacotherapy was prescribed as usual. Substance use decreased for both groups. However, compared with patients in the drug counseling group, those who participated in integrated group therapy spent fewer days using substances in general and alcohol in particular, fewer days using alcohol to intoxication, and had a shorter time from treatment initiation to the first abstinent month. There were no differences between groups in number of weeks in a mood episode.

Pharmacotherapy options

For a table that summarizes the dosages and indications of the medications used to treat BD-SUD that are described below, visit this article at CurrentPsychiatry.com.

Table

Medications used to treat substance use disorders in bipolar disorder patients*

Lithium. Given its well-documented mood stabilizing effect, lithium would seem to be a reasonable option to treat BD-SUD patients, but scant evidence supports its role as an anti-alcohol or anti-drug medication (Table 1).^13,14 Lithium’s efficacy was evaluated in a study of 25 adolescents suffering from mood disorders (mostly BD) and comorbid SUDs (mostly alcohol and cannabis) randomized to receive lithium or placebo for 6 weeks.¹³ Lithium was well tolerated and improved psychiatric symptoms. At week 3, patients receiving lithium produced fewer positive results on randomly administered urine drug screens than those receiving placebo.

However, lithium seems to have little efficacy in reducing cocaine use in cocaine-dependent patients with bipolar spectrum disorders.¹⁴ In an open-label study, 10 patients with a history of hypomania or cyclothymia received lithium monotherapy for 12 weeks. Although patients experienced improved mood symptoms and decreased cocaine use, the mean decrease was transitory and not statistically significant. Another factor that may limit lithium’s use for BD-SUD patients is that these patients are more likely to comply with valproate treatment than with lithium.¹⁵

Table 1

Lithium for BD patients with substance use disorders

Anticonvulsants. In a double-blind, placebo-controlled study of 59 alcohol-dependent bipolar I disorder patients, lithium plus divalproex sodium was superior to lithium plus placebo in decreasing number of drinking days and number of drinks per day and in increasing periods of abstinence (Table 2).^16-19 Divalproex sodium was well tolerated and liver function improved in the divalproex sodium group compared with the placebo group, which probably was a benefit of decreased alcohol consumption. In addition, there was a strong association between mood symptoms and alcohol use, which suggests that maximizing mood symptom treatment may decrease alcohol use. However, the divalproex sodium and placebo groups did not differ in measures of mood symptoms, which implies that divalproex sodium might exhibit a positive effect on drinking regardless of its mood-stabilizing properties.

Divalproex sodium also has been used to treat BD comorbid with cocaine dependence. In a small open-label study, 15 patients receiving divalproex sodium plus counseling for mood and substance use disorders were followed for 6 weeks.¹⁷ The 7 patients who completed the trial had significantly more cocaine-abstinent days, spent less money on cocaine, and experienced fewer manic and depressive symptoms. However, divalproex sodium’s effect on cocaine use cannot be determined solely from this study because there was no placebo control group.

Despite its widespread use as a mood stabilizer and potential use in alcohol detoxification, carbamazepine scarcely has been studied in BD-SUD patients. A double-blind, placebo-controlled study of 139 cocaine-dependent patients with BD or other affective disorders found that patients taking carbamazepine for 12 weeks experienced modest reductions in positive urine drug screens and increased time to cocaine use.¹⁸ They also reported less cocaine craving than patients taking placebo, and mood symptoms (mostly depressive) improved.

An open-label study used lamotrigine as adjunctive therapy or monotherapy for 62 cocaine-dependent BD patients followed for 36 weeks.¹⁹ There was some decrease in cocaine craving, money spent on cocaine, and rate of depressive and manic symptoms, but no effect on cocaine use. A placebo-controlled trial is necessary to confirm these modest effects.

No studies have evaluated the potential role of topiramate in treating BD-SUD, despite its FDA-approved indication for alcoholism treatment. Topiramate’s well-known safety and tolerability profile in BD patients make it an interesting option for those with co-occurring alcohol dependence.

Table 2

Studies suggest anticonvulsants may reduce alcohol, cocaine use in BD patients

Atypical antipsychotics. In an open-label study, 16 weeks of quetiapine monotherapy effectively decreased alcohol consumption, alcohol craving, and psychotic and affective symptoms in 28 alcoholics with a variety of psychiatric diagnoses, including BD, schizoaffective disorder, and borderline personality disorder (Table 3).^20-24 However, in a double-blind study of augmentation with quetiapine or placebo for 102 alcohol-dependent BD patients, no significant differences in alcohol use were found between groups.²¹

Quetiapine may be effective for treating BD patients with comorbid cocaine dependence. In an open-label study, 12 weeks of quetiapine augmentation in 17 cocaine-dependent BD patients was associated with decreased cocaine craving and improvement in depressive symptoms.²² In another open-label study, 80 BD patients with comorbid cocaine or amphetamine dependence were randomly assigned to receive quetiapine or risperidone as adjunctive therapy or monotherapy for 20 weeks.²³ Both groups showed significantly decreased drug use and drug craving and improved mood. This study suggests that risperidone also may be an option for BD patients with comorbid cocaine or stimulant dependence.

A 20-week, open-label study of 20 BD-SUD patients found that switching patients from their previous antipsychotic to aripiprazole resulted in less cocaine craving, less alcohol craving, and less money spent on alcohol.²⁴

Olanzapine has not been systematically studied in BD-SUD patients. Some case reports suggest that olanzapine may decrease cocaine craving and use in patients with schizoaffective disorder (bipolar type) and alcohol craving and use in BD patients with comorbid alcohol dependence.²⁵

Table 3

Evidence of efficacy for antipsychotics for BD patients with SUDs

SUD medications. Little evidence guides using medications indicated for treating SUDs—such as naltrexone, acamprosate, and disulfiram—as treatment for BD patients (Table 4).^26-29 In an open-label trial of 34 BD patients with alcohol dependence, naltrexone was well tolerated and associated with decreased alcohol craving and use and modest improvement in manic and depressive symptoms.²⁶

In a double-blind, placebo-controlled study, 50 alcohol-dependent BD patients treated with standard mood-stabilizing therapy and cognitive-behavioral therapy were randomized to receive add-on naltrexone, 50 mg/d, or placebo.²⁷ Patients receiving naltrexone showed decreased alcohol consumption, although no measures were statistically significant. Effect sizes of alcohol use decrease and alcohol craving were moderate to large compared with placebo, which suggests that naltrexone may be effective for treating alcoholism in these patients.

Two other studies evaluated naltrexone and disulfiram in patients with BD or other mood disorders.^28,29 Naltrexone was well tolerated, caused no serious adverse side effects, and was significantly more effective than placebo in decreasing drinking rates and increasing the number of abstinent days.^28,29 Disulfiram was as effective as naltrexone, but the combination of both offered no advantage over use of either drug separately.

There are reports of a new-onset manic episode associated with naltrexone use in a patient with opioid dependence, and a manic episode triggered by naltrexone in a patient with BD with comorbid alcohol dependence.^30,31 At both low and high doses, disulfiram is associated with induction of psychotic mania in alcoholic patients without a personal or family history of BD.^32,33

We found no studies that evaluated treating BD patients who abused other substances, such as cannabis or opiates. We recommend that BD patients with these substance use disorders should be referred to treatment modalities that are condition-specific, such as psychotherapy for cannabis use disorders or methadone or naltrexone treatment for opiate dependence. More severe cases of comorbid SUD probably would benefit from a referral to or consultation with a SUD specialist.

Table 4

Naltrexone and disulfiram for BD patients with alcohol dependence

Related Resource

• Tolliver BK. Bipolar disorder and substance abuse: Overcome the challenges of ‘dual diagnosis’ patients. Current Psychiatry. 2010; 9(8): 32-38.

Drug Brand Names

• Acamprosate • Campral
• Aripiprazole • Abilify
• Carbamazepine • Carbatrol, Equetro, others
• Disulfiram • Antabuse
• Divalproex sodium • Depakote,
• Depakote ER Lamotrigine • Lamictal
• Lithium • Eskalith, Lithobid
• Methadone • Dolophine
• Naltrexone • ReVia, Vivitrol
• Quetiapine • Seroquel
• Risperidone • Risperdal
• Topiramate • Topamax
• Valproate • Depacon

Disclosures

Dr. Nery held a temporary work contract as a clinical research physician with Eli Lilly and Company Brazil from May 2009 to November 2009.

Dr. Soares was partly supported by National Institute of Health grants MH 68766, MH 69774, and RR 20571. He receives grant/research support from Bristol-Myers Squibb, Cephalon, GlaxoSmithKline, and Sunovion.

Discuss this article at www.facebook.com/CurrentPsychiatry

Among DSM axis I diagnoses, bipolar disorder (BD) has the highest rates of comorbid substance use disorders (SUDs).^1-3 Approximately 60% of patients with bipolar I disorder have a lifetime diagnosis of an SUD.¹ Excluding tobacco, alcohol is the substance most often abused by BD patients, followed by cannabis, amphetamines, and cocaine.^1-3

BD patients with comorbid SUD usually exhibit more severe clinical presentations and poorer outcomes than their counterparts without SUDs. Compared with patients with BD alone, those with BD and SUD comorbidity (BD-SUD) experience earlier onset of mood symptoms; higher rates of anxiety disorders, suicide attempts, accidents, hospitalizations, and rapid cycling; more depressive episodes; and lower treatment compliance.^4-9

Several treatment options are available for patients with BD-SUD, including psychotherapy modalities, medications primarily used to treat BD, and medications primarily used to treat SUDs. Evidence-based support for these treatments remains limited, and no treatment of choice has emerged. This article reviews evidence on the longer-term treatment of BD-SUD, including general strategies and specific psychosocial and pharmacologic interventions. Short-term treatment strategies, such as pharmacotherapy for detoxification, are outside the scope of this review.

General strategies

The causes of BD-SUD are complex. Evidence suggests that the presence of affective symptoms is associated with an increased risk for substance misuse. This should be kept in mind when treating a patient with BD-SUD because controlling mood symptoms probably will help control substance abuse. However, evidence also shows that SUDs may be independent of mood episodes. Therefore, treating only mood symptoms in the hope that doing so will control substance abuse may not be enough.

Because the negative impact of SUDs on BD outcome is well documented, inform patients that limiting their use of alcohol and/or drugs is vital to control their mood disorder. Efforts to educate, stimulate, and support patients to moderate or stop their alcohol and/or drug use are likely to result in positive changes.¹⁰ Therefore, treatment for BD-SUD should follow, in part, the same recommendations for treatment of SUDs in patients with no comorbid axis I disorders:

• identify the problem (ie, the existence of a comorbid SUD)
• share your concerns with your patient
• offer appropriate and specific treatments, such as detoxification and/or self-help and counseling programs.¹⁰

Because SUDs usually are chronic and relapsing conditions, periods of drug and/ or alcohol use should be expected and not considered a sign of treatment failure. In addition, integrating treatment for both conditions probably is better than managing each separately. Therefore, targeting BD symptoms with mood-stabilizing medications and substance abuse with nonpharmacologic modalities such as drug counseling likely will bring about the best results.

Compared with BD patients without comorbid SUD, BD-SUD patients have a 7-fold increased risk of antidepressantinduced mania.¹¹ Therefore, antidepressants should be prescribed cautiously for patients with BD-SUD.

Integrated psychosocial therapy

BD-SUD patients may benefit from attending self-help programs such as Alcoholics Anonymous and Narcotics Anonymous, provided their mood is stable enough to allow them to participate. Other forms of psychotherapy for BD-SUD patients include standard group drug counseling and integrated group therapy that simultaneously addresses both conditions.

Integrated group therapy is based on the premise that changing maladaptive mood cognitions and behaviors will facilitate recovery from SUDs, and changing maladaptive substance use cognitions and behaviors will facilitate recovery from mood disorders.¹² In a recent randomized controlled trial, 62 BD-SUD patients were blindly assigned to integrated group therapy or standard group drug counseling and followed for 3 months.¹² Pharmacotherapy was prescribed as usual. Substance use decreased for both groups. However, compared with patients in the drug counseling group, those who participated in integrated group therapy spent fewer days using substances in general and alcohol in particular, fewer days using alcohol to intoxication, and had a shorter time from treatment initiation to the first abstinent month. There were no differences between groups in number of weeks in a mood episode.

Pharmacotherapy options

For a table that summarizes the dosages and indications of the medications used to treat BD-SUD that are described below, visit this article at CurrentPsychiatry.com.

Table

Medications used to treat substance use disorders in bipolar disorder patients*

Lithium. Given its well-documented mood stabilizing effect, lithium would seem to be a reasonable option to treat BD-SUD patients, but scant evidence supports its role as an anti-alcohol or anti-drug medication (Table 1).^13,14 Lithium’s efficacy was evaluated in a study of 25 adolescents suffering from mood disorders (mostly BD) and comorbid SUDs (mostly alcohol and cannabis) randomized to receive lithium or placebo for 6 weeks.¹³ Lithium was well tolerated and improved psychiatric symptoms. At week 3, patients receiving lithium produced fewer positive results on randomly administered urine drug screens than those receiving placebo.

However, lithium seems to have little efficacy in reducing cocaine use in cocaine-dependent patients with bipolar spectrum disorders.¹⁴ In an open-label study, 10 patients with a history of hypomania or cyclothymia received lithium monotherapy for 12 weeks. Although patients experienced improved mood symptoms and decreased cocaine use, the mean decrease was transitory and not statistically significant. Another factor that may limit lithium’s use for BD-SUD patients is that these patients are more likely to comply with valproate treatment than with lithium.¹⁵

Table 1

Lithium for BD patients with substance use disorders

Anticonvulsants. In a double-blind, placebo-controlled study of 59 alcohol-dependent bipolar I disorder patients, lithium plus divalproex sodium was superior to lithium plus placebo in decreasing number of drinking days and number of drinks per day and in increasing periods of abstinence (Table 2).^16-19 Divalproex sodium was well tolerated and liver function improved in the divalproex sodium group compared with the placebo group, which probably was a benefit of decreased alcohol consumption. In addition, there was a strong association between mood symptoms and alcohol use, which suggests that maximizing mood symptom treatment may decrease alcohol use. However, the divalproex sodium and placebo groups did not differ in measures of mood symptoms, which implies that divalproex sodium might exhibit a positive effect on drinking regardless of its mood-stabilizing properties.

Divalproex sodium also has been used to treat BD comorbid with cocaine dependence. In a small open-label study, 15 patients receiving divalproex sodium plus counseling for mood and substance use disorders were followed for 6 weeks.¹⁷ The 7 patients who completed the trial had significantly more cocaine-abstinent days, spent less money on cocaine, and experienced fewer manic and depressive symptoms. However, divalproex sodium’s effect on cocaine use cannot be determined solely from this study because there was no placebo control group.

Despite its widespread use as a mood stabilizer and potential use in alcohol detoxification, carbamazepine scarcely has been studied in BD-SUD patients. A double-blind, placebo-controlled study of 139 cocaine-dependent patients with BD or other affective disorders found that patients taking carbamazepine for 12 weeks experienced modest reductions in positive urine drug screens and increased time to cocaine use.¹⁸ They also reported less cocaine craving than patients taking placebo, and mood symptoms (mostly depressive) improved.

An open-label study used lamotrigine as adjunctive therapy or monotherapy for 62 cocaine-dependent BD patients followed for 36 weeks.¹⁹ There was some decrease in cocaine craving, money spent on cocaine, and rate of depressive and manic symptoms, but no effect on cocaine use. A placebo-controlled trial is necessary to confirm these modest effects.

No studies have evaluated the potential role of topiramate in treating BD-SUD, despite its FDA-approved indication for alcoholism treatment. Topiramate’s well-known safety and tolerability profile in BD patients make it an interesting option for those with co-occurring alcohol dependence.

Table 2

Studies suggest anticonvulsants may reduce alcohol, cocaine use in BD patients

Atypical antipsychotics. In an open-label study, 16 weeks of quetiapine monotherapy effectively decreased alcohol consumption, alcohol craving, and psychotic and affective symptoms in 28 alcoholics with a variety of psychiatric diagnoses, including BD, schizoaffective disorder, and borderline personality disorder (Table 3).^20-24 However, in a double-blind study of augmentation with quetiapine or placebo for 102 alcohol-dependent BD patients, no significant differences in alcohol use were found between groups.²¹

Quetiapine may be effective for treating BD patients with comorbid cocaine dependence. In an open-label study, 12 weeks of quetiapine augmentation in 17 cocaine-dependent BD patients was associated with decreased cocaine craving and improvement in depressive symptoms.²² In another open-label study, 80 BD patients with comorbid cocaine or amphetamine dependence were randomly assigned to receive quetiapine or risperidone as adjunctive therapy or monotherapy for 20 weeks.²³ Both groups showed significantly decreased drug use and drug craving and improved mood. This study suggests that risperidone also may be an option for BD patients with comorbid cocaine or stimulant dependence.

A 20-week, open-label study of 20 BD-SUD patients found that switching patients from their previous antipsychotic to aripiprazole resulted in less cocaine craving, less alcohol craving, and less money spent on alcohol.²⁴

Olanzapine has not been systematically studied in BD-SUD patients. Some case reports suggest that olanzapine may decrease cocaine craving and use in patients with schizoaffective disorder (bipolar type) and alcohol craving and use in BD patients with comorbid alcohol dependence.²⁵

Table 3

Evidence of efficacy for antipsychotics for BD patients with SUDs

SUD medications. Little evidence guides using medications indicated for treating SUDs—such as naltrexone, acamprosate, and disulfiram—as treatment for BD patients (Table 4).^26-29 In an open-label trial of 34 BD patients with alcohol dependence, naltrexone was well tolerated and associated with decreased alcohol craving and use and modest improvement in manic and depressive symptoms.²⁶

In a double-blind, placebo-controlled study, 50 alcohol-dependent BD patients treated with standard mood-stabilizing therapy and cognitive-behavioral therapy were randomized to receive add-on naltrexone, 50 mg/d, or placebo.²⁷ Patients receiving naltrexone showed decreased alcohol consumption, although no measures were statistically significant. Effect sizes of alcohol use decrease and alcohol craving were moderate to large compared with placebo, which suggests that naltrexone may be effective for treating alcoholism in these patients.

Two other studies evaluated naltrexone and disulfiram in patients with BD or other mood disorders.^28,29 Naltrexone was well tolerated, caused no serious adverse side effects, and was significantly more effective than placebo in decreasing drinking rates and increasing the number of abstinent days.^28,29 Disulfiram was as effective as naltrexone, but the combination of both offered no advantage over use of either drug separately.

There are reports of a new-onset manic episode associated with naltrexone use in a patient with opioid dependence, and a manic episode triggered by naltrexone in a patient with BD with comorbid alcohol dependence.^30,31 At both low and high doses, disulfiram is associated with induction of psychotic mania in alcoholic patients without a personal or family history of BD.^32,33

We found no studies that evaluated treating BD patients who abused other substances, such as cannabis or opiates. We recommend that BD patients with these substance use disorders should be referred to treatment modalities that are condition-specific, such as psychotherapy for cannabis use disorders or methadone or naltrexone treatment for opiate dependence. More severe cases of comorbid SUD probably would benefit from a referral to or consultation with a SUD specialist.

Table 4

Naltrexone and disulfiram for BD patients with alcohol dependence

Related Resource

• Tolliver BK. Bipolar disorder and substance abuse: Overcome the challenges of ‘dual diagnosis’ patients. Current Psychiatry. 2010; 9(8): 32-38.

Drug Brand Names

• Acamprosate • Campral
• Aripiprazole • Abilify
• Carbamazepine • Carbatrol, Equetro, others
• Disulfiram • Antabuse
• Divalproex sodium • Depakote,
• Depakote ER Lamotrigine • Lamictal
• Lithium • Eskalith, Lithobid
• Methadone • Dolophine
• Naltrexone • ReVia, Vivitrol
• Quetiapine • Seroquel
• Risperidone • Risperdal
• Topiramate • Topamax
• Valproate • Depacon

Disclosures

Dr. Nery held a temporary work contract as a clinical research physician with Eli Lilly and Company Brazil from May 2009 to November 2009.

Dr. Soares was partly supported by National Institute of Health grants MH 68766, MH 69774, and RR 20571. He receives grant/research support from Bristol-Myers Squibb, Cephalon, GlaxoSmithKline, and Sunovion.

Discuss this article at www.facebook.com/CurrentPsychiatry

Among DSM axis I diagnoses, bipolar disorder (BD) has the highest rates of comorbid substance use disorders (SUDs).^1-3 Approximately 60% of patients with bipolar I disorder have a lifetime diagnosis of an SUD.¹ Excluding tobacco, alcohol is the substance most often abused by BD patients, followed by cannabis, amphetamines, and cocaine.^1-3

BD patients with comorbid SUD usually exhibit more severe clinical presentations and poorer outcomes than their counterparts without SUDs. Compared with patients with BD alone, those with BD and SUD comorbidity (BD-SUD) experience earlier onset of mood symptoms; higher rates of anxiety disorders, suicide attempts, accidents, hospitalizations, and rapid cycling; more depressive episodes; and lower treatment compliance.^4-9

Several treatment options are available for patients with BD-SUD, including psychotherapy modalities, medications primarily used to treat BD, and medications primarily used to treat SUDs. Evidence-based support for these treatments remains limited, and no treatment of choice has emerged. This article reviews evidence on the longer-term treatment of BD-SUD, including general strategies and specific psychosocial and pharmacologic interventions. Short-term treatment strategies, such as pharmacotherapy for detoxification, are outside the scope of this review.

General strategies

The causes of BD-SUD are complex. Evidence suggests that the presence of affective symptoms is associated with an increased risk for substance misuse. This should be kept in mind when treating a patient with BD-SUD because controlling mood symptoms probably will help control substance abuse. However, evidence also shows that SUDs may be independent of mood episodes. Therefore, treating only mood symptoms in the hope that doing so will control substance abuse may not be enough.

Because the negative impact of SUDs on BD outcome is well documented, inform patients that limiting their use of alcohol and/or drugs is vital to control their mood disorder. Efforts to educate, stimulate, and support patients to moderate or stop their alcohol and/or drug use are likely to result in positive changes.¹⁰ Therefore, treatment for BD-SUD should follow, in part, the same recommendations for treatment of SUDs in patients with no comorbid axis I disorders:

• identify the problem (ie, the existence of a comorbid SUD)
• share your concerns with your patient
• offer appropriate and specific treatments, such as detoxification and/or self-help and counseling programs.¹⁰

Because SUDs usually are chronic and relapsing conditions, periods of drug and/ or alcohol use should be expected and not considered a sign of treatment failure. In addition, integrating treatment for both conditions probably is better than managing each separately. Therefore, targeting BD symptoms with mood-stabilizing medications and substance abuse with nonpharmacologic modalities such as drug counseling likely will bring about the best results.

Compared with BD patients without comorbid SUD, BD-SUD patients have a 7-fold increased risk of antidepressantinduced mania.¹¹ Therefore, antidepressants should be prescribed cautiously for patients with BD-SUD.

Integrated psychosocial therapy

BD-SUD patients may benefit from attending self-help programs such as Alcoholics Anonymous and Narcotics Anonymous, provided their mood is stable enough to allow them to participate. Other forms of psychotherapy for BD-SUD patients include standard group drug counseling and integrated group therapy that simultaneously addresses both conditions.

Integrated group therapy is based on the premise that changing maladaptive mood cognitions and behaviors will facilitate recovery from SUDs, and changing maladaptive substance use cognitions and behaviors will facilitate recovery from mood disorders.¹² In a recent randomized controlled trial, 62 BD-SUD patients were blindly assigned to integrated group therapy or standard group drug counseling and followed for 3 months.¹² Pharmacotherapy was prescribed as usual. Substance use decreased for both groups. However, compared with patients in the drug counseling group, those who participated in integrated group therapy spent fewer days using substances in general and alcohol in particular, fewer days using alcohol to intoxication, and had a shorter time from treatment initiation to the first abstinent month. There were no differences between groups in number of weeks in a mood episode.

Pharmacotherapy options

For a table that summarizes the dosages and indications of the medications used to treat BD-SUD that are described below, visit this article at CurrentPsychiatry.com.

Table

Medications used to treat substance use disorders in bipolar disorder patients*

Lithium. Given its well-documented mood stabilizing effect, lithium would seem to be a reasonable option to treat BD-SUD patients, but scant evidence supports its role as an anti-alcohol or anti-drug medication (Table 1).^13,14 Lithium’s efficacy was evaluated in a study of 25 adolescents suffering from mood disorders (mostly BD) and comorbid SUDs (mostly alcohol and cannabis) randomized to receive lithium or placebo for 6 weeks.¹³ Lithium was well tolerated and improved psychiatric symptoms. At week 3, patients receiving lithium produced fewer positive results on randomly administered urine drug screens than those receiving placebo.

However, lithium seems to have little efficacy in reducing cocaine use in cocaine-dependent patients with bipolar spectrum disorders.¹⁴ In an open-label study, 10 patients with a history of hypomania or cyclothymia received lithium monotherapy for 12 weeks. Although patients experienced improved mood symptoms and decreased cocaine use, the mean decrease was transitory and not statistically significant. Another factor that may limit lithium’s use for BD-SUD patients is that these patients are more likely to comply with valproate treatment than with lithium.¹⁵

Table 1

Lithium for BD patients with substance use disorders

Anticonvulsants. In a double-blind, placebo-controlled study of 59 alcohol-dependent bipolar I disorder patients, lithium plus divalproex sodium was superior to lithium plus placebo in decreasing number of drinking days and number of drinks per day and in increasing periods of abstinence (Table 2).^16-19 Divalproex sodium was well tolerated and liver function improved in the divalproex sodium group compared with the placebo group, which probably was a benefit of decreased alcohol consumption. In addition, there was a strong association between mood symptoms and alcohol use, which suggests that maximizing mood symptom treatment may decrease alcohol use. However, the divalproex sodium and placebo groups did not differ in measures of mood symptoms, which implies that divalproex sodium might exhibit a positive effect on drinking regardless of its mood-stabilizing properties.

Divalproex sodium also has been used to treat BD comorbid with cocaine dependence. In a small open-label study, 15 patients receiving divalproex sodium plus counseling for mood and substance use disorders were followed for 6 weeks.¹⁷ The 7 patients who completed the trial had significantly more cocaine-abstinent days, spent less money on cocaine, and experienced fewer manic and depressive symptoms. However, divalproex sodium’s effect on cocaine use cannot be determined solely from this study because there was no placebo control group.

Despite its widespread use as a mood stabilizer and potential use in alcohol detoxification, carbamazepine scarcely has been studied in BD-SUD patients. A double-blind, placebo-controlled study of 139 cocaine-dependent patients with BD or other affective disorders found that patients taking carbamazepine for 12 weeks experienced modest reductions in positive urine drug screens and increased time to cocaine use.¹⁸ They also reported less cocaine craving than patients taking placebo, and mood symptoms (mostly depressive) improved.

An open-label study used lamotrigine as adjunctive therapy or monotherapy for 62 cocaine-dependent BD patients followed for 36 weeks.¹⁹ There was some decrease in cocaine craving, money spent on cocaine, and rate of depressive and manic symptoms, but no effect on cocaine use. A placebo-controlled trial is necessary to confirm these modest effects.

No studies have evaluated the potential role of topiramate in treating BD-SUD, despite its FDA-approved indication for alcoholism treatment. Topiramate’s well-known safety and tolerability profile in BD patients make it an interesting option for those with co-occurring alcohol dependence.

Table 2

Studies suggest anticonvulsants may reduce alcohol, cocaine use in BD patients

Atypical antipsychotics. In an open-label study, 16 weeks of quetiapine monotherapy effectively decreased alcohol consumption, alcohol craving, and psychotic and affective symptoms in 28 alcoholics with a variety of psychiatric diagnoses, including BD, schizoaffective disorder, and borderline personality disorder (Table 3).^20-24 However, in a double-blind study of augmentation with quetiapine or placebo for 102 alcohol-dependent BD patients, no significant differences in alcohol use were found between groups.²¹

Quetiapine may be effective for treating BD patients with comorbid cocaine dependence. In an open-label study, 12 weeks of quetiapine augmentation in 17 cocaine-dependent BD patients was associated with decreased cocaine craving and improvement in depressive symptoms.²² In another open-label study, 80 BD patients with comorbid cocaine or amphetamine dependence were randomly assigned to receive quetiapine or risperidone as adjunctive therapy or monotherapy for 20 weeks.²³ Both groups showed significantly decreased drug use and drug craving and improved mood. This study suggests that risperidone also may be an option for BD patients with comorbid cocaine or stimulant dependence.

A 20-week, open-label study of 20 BD-SUD patients found that switching patients from their previous antipsychotic to aripiprazole resulted in less cocaine craving, less alcohol craving, and less money spent on alcohol.²⁴

Olanzapine has not been systematically studied in BD-SUD patients. Some case reports suggest that olanzapine may decrease cocaine craving and use in patients with schizoaffective disorder (bipolar type) and alcohol craving and use in BD patients with comorbid alcohol dependence.²⁵

Table 3

Evidence of efficacy for antipsychotics for BD patients with SUDs

SUD medications. Little evidence guides using medications indicated for treating SUDs—such as naltrexone, acamprosate, and disulfiram—as treatment for BD patients (Table 4).^26-29 In an open-label trial of 34 BD patients with alcohol dependence, naltrexone was well tolerated and associated with decreased alcohol craving and use and modest improvement in manic and depressive symptoms.²⁶

In a double-blind, placebo-controlled study, 50 alcohol-dependent BD patients treated with standard mood-stabilizing therapy and cognitive-behavioral therapy were randomized to receive add-on naltrexone, 50 mg/d, or placebo.²⁷ Patients receiving naltrexone showed decreased alcohol consumption, although no measures were statistically significant. Effect sizes of alcohol use decrease and alcohol craving were moderate to large compared with placebo, which suggests that naltrexone may be effective for treating alcoholism in these patients.

Two other studies evaluated naltrexone and disulfiram in patients with BD or other mood disorders.^28,29 Naltrexone was well tolerated, caused no serious adverse side effects, and was significantly more effective than placebo in decreasing drinking rates and increasing the number of abstinent days.^28,29 Disulfiram was as effective as naltrexone, but the combination of both offered no advantage over use of either drug separately.

There are reports of a new-onset manic episode associated with naltrexone use in a patient with opioid dependence, and a manic episode triggered by naltrexone in a patient with BD with comorbid alcohol dependence.^30,31 At both low and high doses, disulfiram is associated with induction of psychotic mania in alcoholic patients without a personal or family history of BD.^32,33

We found no studies that evaluated treating BD patients who abused other substances, such as cannabis or opiates. We recommend that BD patients with these substance use disorders should be referred to treatment modalities that are condition-specific, such as psychotherapy for cannabis use disorders or methadone or naltrexone treatment for opiate dependence. More severe cases of comorbid SUD probably would benefit from a referral to or consultation with a SUD specialist.

Table 4

Naltrexone and disulfiram for BD patients with alcohol dependence

Related Resource

• Tolliver BK. Bipolar disorder and substance abuse: Overcome the challenges of ‘dual diagnosis’ patients. Current Psychiatry. 2010; 9(8): 32-38.

Drug Brand Names

• Acamprosate • Campral
• Aripiprazole • Abilify
• Carbamazepine • Carbatrol, Equetro, others
• Disulfiram • Antabuse
• Divalproex sodium • Depakote,
• Depakote ER Lamotrigine • Lamictal
• Lithium • Eskalith, Lithobid
• Methadone • Dolophine
• Naltrexone • ReVia, Vivitrol
• Quetiapine • Seroquel
• Risperidone • Risperdal
• Topiramate • Topamax
• Valproate • Depacon

Disclosures

Dr. Nery held a temporary work contract as a clinical research physician with Eli Lilly and Company Brazil from May 2009 to November 2009.

Dr. Soares was partly supported by National Institute of Health grants MH 68766, MH 69774, and RR 20571. He receives grant/research support from Bristol-Myers Squibb, Cephalon, GlaxoSmithKline, and Sunovion.

Discuss this article at www.facebook.com/CurrentPsychiatry

Dr. G, a second-year surgical resident, becomes depressed when his girlfriend abruptly ends their relationship. His phone calls and e-mails seeking an explanation go unanswered. Having long struggled with his self-esteem, Dr. G interprets this rejection as confirmation of his self-criticism.

Because of his work schedule, Dr. G feels that there is no way to see a therapist or psychiatrist and believes that asking for time off to do so would adversely affect his evaluations. He feels too embarrassed and “weak” to disclose his breakup and depression to his colleagues and attending physicians and senses that fellow residents would resent having to “carry his load.” Dr. G has spent the past 2 years moonlighting at the local emergency room and thinks it would be humiliating to go there for psychiatric help. His work performance and attendance decline until eventually his residency director forces him to take a medical leave of absence.

Dr. G feels that his pain will never end. He writes goodbye letters to his family, makes arrangements for his possessions and funeral, and hangs himself from the balcony outside his apartment.

Although the rate of depression among physicians is comparable to that of the general population, physicians’ risk of suicide is markedly higher.¹ Depression and other mood disorders may be under-recognized and inadequately treated in physicians because physicians might:

• be reluctant to seek treatment
• attempt to diagnose and treat themselves
• seek and receive “VIP treatment” from other health care providers.

This article examines physicians’ risk for depression and suicide, licensing concerns and other barriers to effective treatment, and what can be done to overcome such obstacles.

Not immune to depression

Rates of depression are higher in medical students and residents (15% to 30%) than in the general population.^2-4 A longitudinal study of medical students at the University of California, San Francisco showed that students’ rates of depression when they enter medical school are similar to those of the general population, but students’ depression scores rise over time; approximately one-fourth of first- and second-year students were depressed.³ Fahrenkopf et al⁵ reported that 20% of 123 pediatric residents at 3 U.S. children’s hospitals were depressed. These depressed residents made 6.2 times more medication errors than did their non-depressed peers.⁵ For more information on physicians-in-training, see “Treating depression in medical residents“.

After completing residency, the risk of depression persists. The lifetime prevalence of depression among physicians is 13% in men and 20% in women⁶; these rates are comparable to those of the general population. Firth-Cozens⁷ found a range of factors that predict depression among general practitioners; relationships with senior doctors and patients were the main stressors (Table 1).⁷ Although these stressors increase depression risk, Vaillant et al⁸ showed that they did not increase suicide risk in physicians who did not have underlying psychological difficulties when they entered college. Certain personality traits common among physicians, such as self-criticism and perfectionism, may increase risk for depression and substance abuse.⁸

A depressed physician might enter a downward spiral. Feelings of hopelessness and worthlessness frequently lead to declining professional performance. Professional and personal relationships are strained as internal dysphoria manifests as irritability and anger. Spouses and partners can feel overwhelmed and bewildered by changes in the depressed person’s behavior, which may lead to separation or divorce. Patient care and the physician’s professional standing can be endangered. Signs that suggest a physician may be suffering from depression or another mental illness appear in Table 2.⁹

Table 1

Predictors of depression in physicians

Table 2

Manifestations of mental illness in physicians

Increased suicide risk

A review of 14 studies found that the relative risk of suicide in physicians compared with the general population is between 1.1 and 3.4 for men and 2.5 to 5.7 for women.¹ A retrospective study of English and Welsh doctors showed elevated suicide rates in female but not male physicians compared with the general population.¹⁰ There are no recent studies of suicide rates among U.S. physicians. A 1984-1995 study showed that white male physicians have a higher risk for suicide than other white male professionals.¹¹ A survey of 4,500 women physicians found that female doctors are less likely to attempt suicide than the general female population⁶; however, their attempts more often are lethal, perhaps because they have greater knowledge of toxicology and access to lethal drugs.¹²

The relative rate of suicide among medical specialties is unknown. Studies had indicated higher rates of suicide among psychiatrists and anesthesiologists, but these trials were methodologically flawed.¹²

Silverman¹² developed a profile of the physician at high risk for suicide: a workaholic white male age ≥50 or female age ≥45 who is divorced, single, or currently experiencing marital disruption and is suffering from depression. He or she has a substance abuse problem and a history of risk-taking (high-stakes gambling, etc.). Physicians with chronic pain or illness or with a recent change in occupational or financial status also are at risk. Recent increased work demands, personal losses, diminished autonomy, and access to lethal means (medications, firearms) complete the profile.

Protective factors that lower the risk of completed suicide include effective treatment, social and family support, resilience and coping skills, religious faith, and restricted access to lethal means.^13,14

Barriers to treatment

Physicians often are hesitant to seek mental health treatment.¹⁵ They may fear social stigma and could have trouble finding a local provider who they trust but is not a colleague. Physicians might be concerned about confidentiality and fear recrimination by colleagues, facilities where they work, or licensing boards.¹⁶ Givens and Tjia³ found that only 22% of medical students who screened positive for depression sought help and only 42% of students with suicidal ideation received treatment. These students reported that time constraints, confidentiality concerns, stigma, cost, and fear that their illness will be documented on their academic record were major barriers to seeking mental health care.

Licensing concerns. Physicians may be required to disclose a mental health diagnosis or treatment history when applying for or renewing their medical license. Increasingly, medical boards are asking applicants if they have been treated for bipolar disorder, schizophrenia, paranoia, or other disorders.¹⁷ Credentialing bodies, clinics, and hospitals may make similar queries.

In an analysis of 51 medical licensing applications (50 states and the District of Columbia), Schroeder et al¹⁷ determined that 69% contained at least 1 question that was “likely impermissible” or “impermissible” in terms of compliance with the Americans with Disabilities Act (ADA). In 1993, a U.S. District Court found that the New Jersey State Board of Medical Examiners was in violation of the ADA because licensure application questions did not focus on current fitness to practice medicine but rather on information about a candidate’s status as a person with a disability (illness or diagnosis).¹⁸

In Alexander v Margolis,¹⁹ however, the court found that because patient safety is in question, medical licensing boards and credentialing bodies can solicit information about serious mental illness that could lead to impaired performance. Courts have ruled that questions regarding a history of treatment or hospitalization for bipolar disorder or schizophrenia and other psychotic disorders are permissible because they are considered “serious disorders” likely to interfere with a physician’s current ability to practice.²⁰ In a 2008 review of all U.S. -affiliated medical licensing boards (N=54), Polfliet²¹ found that 7 specifically asked applicants about a history of bipolar disorder or schizophrenia, paranoia, and other psychotic disorders. Polfliet²¹ also found that state medical boards’ compliance with ADA guidelines was not uniform and some questions were “just as broad, and potentially discriminatory, as they were before enactment of the ADA.”

Worley²² reported a successful appeal to the Arkansas State Medical Board to revise its licensure questions following a cluster of medical student and physician suicides. The Board changed the question “Have you ever, or are you presently, being treated for a mental health condition?” to “Have you ever been advised or required by any licensing or privileging body to seek treatment for a physical or mental health condition?”

Providing inaccurate information on a medical licensure application may result in denial or revocation,²³ but acknowledging a history of mental health or substance abuse treatment triggers a more in-depth inquiry by the medical board. The lack of distinction between diagnosis and impairment further stigmatizes physicians who seek care and impedes treatment.

Bipolar disorder. The trend in psychiatry toward diagnosing bipolar II disorder and “soft bipolarity” in patients previously diagnosed with and treated for major depression presents a new challenge. Despite no change in their history or functioning, a physician whose diagnosis is changed from depression to bipolar II disorder might be moved from a non-reportable to a board-reportable diagnostic category. With the evolving understanding of bipolar spectrum disorders, medical boards may need to revise their screening questions to ensure that they are seeking information about impairment, not simply the presence of a medical disorder.

Seeking special treatment

Self-treatment. Physicians may attempt to treat their mood disorder with self-prescribed medications before seeking consultation from a psychiatrist. Others use alcohol or illicit drugs to try to alleviate mood disorder symptoms. Self-diagnosis and treatment are not advisable because it is impossible to be objective. Professional boards and state medical boards discourage or prohibit self-prescribing because of the need for ongoing evaluation and monitoring for adverse reactions.

‘VIP’ treatment. When a physician comes to a colleague for help with a mental health issue, both parties might underestimate the severity of the crisis.²⁴ Weintraub²⁵ reported a case series of 12 “VIP” psychiatric inpatients, 10 of whom he described as “therapeutic failures, “including 2 who committed suicide and 3 who left the hospital against medical advice. He observed that improvement occurred only after patients lost their VIP status/treatment.

In a literature review, Groves et al²⁶ found delays in pursuing diagnostic evaluation and treatment for physician patients. He described risks of VIP treatment (Table 3),²⁶ including the physician’s ability to circumscribe the care regimen to obtain “special treatment, “which can create conflict among care providers and other patients. The ailing physician might have trouble relinquishing control. Care providers might not give physician patients necessary information about the illness or treatment because they make assumptions about the physician’s knowledge or fear causing narcissistic injury. Providers’ identification with their peers, deference to their background, and desire to preserve these patients’ autonomy may lead to interventions that are different from those they would provide to other patients.

Treating physicians might underestimate the patient’s suicide risk and tend to not hospitalize a physician patient who faces an imminent risk of self-harm. Similarly, a physician patient might know what key words to use to deny suicidal ideation or avoid hospitalization. Providers assessing physician patients should provide the same interventions they would give to nonphysician patients with the same history and suicide risk factors. To do otherwise is to risk a fatal outcome.

Physician health programs provide confidential treatment and assistance to physicians with mental illness and/or substance abuse problems. Some programs are affiliated with licensing boards, some are branches of the state medical societies, and others are independent of the licensing agencies. Directories of these programs are available from the Federation of State Physician Health Programs and the Federation of State Medical Boards (see Related Resources). Physician health programs aim to help impaired physicians receive treatment and rehabilitation without censure or licensure revocation, provided they comply with treatment and monitoring requirements.

Table 3

Risks of caring for ‘VIP’ patients

Related Resources

• American Foundation for Suicide Prevention. www.afsp.org. 24-hour crisis line: 1-800-273-TALK (8255).
• Center for Patient and Professional Advocacy. www.mc.vanderbilt.edu/root/vumc.php?site=CPPA.
• Depression and Bipolar Support Alliance. www.dbsalliance.org.
• Federation of State Physician Health Programs, Inc. www.fsphp.org.
• National Alliance on Mental Illness. www.nami.org.
• Vanderbilt Center for Professional Health. www.mc.vanderbilt.edu/cph.
• Vanderbilt Comprehensive Assessment Program. www.mc.vanderbilt.edu/root/vcap.

Disclosure

The authors report 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

Dr. G, a second-year surgical resident, becomes depressed when his girlfriend abruptly ends their relationship. His phone calls and e-mails seeking an explanation go unanswered. Having long struggled with his self-esteem, Dr. G interprets this rejection as confirmation of his self-criticism.

Because of his work schedule, Dr. G feels that there is no way to see a therapist or psychiatrist and believes that asking for time off to do so would adversely affect his evaluations. He feels too embarrassed and “weak” to disclose his breakup and depression to his colleagues and attending physicians and senses that fellow residents would resent having to “carry his load.” Dr. G has spent the past 2 years moonlighting at the local emergency room and thinks it would be humiliating to go there for psychiatric help. His work performance and attendance decline until eventually his residency director forces him to take a medical leave of absence.

Dr. G feels that his pain will never end. He writes goodbye letters to his family, makes arrangements for his possessions and funeral, and hangs himself from the balcony outside his apartment.

Although the rate of depression among physicians is comparable to that of the general population, physicians’ risk of suicide is markedly higher.¹ Depression and other mood disorders may be under-recognized and inadequately treated in physicians because physicians might:

• be reluctant to seek treatment
• attempt to diagnose and treat themselves
• seek and receive “VIP treatment” from other health care providers.

This article examines physicians’ risk for depression and suicide, licensing concerns and other barriers to effective treatment, and what can be done to overcome such obstacles.

Not immune to depression

Rates of depression are higher in medical students and residents (15% to 30%) than in the general population.^2-4 A longitudinal study of medical students at the University of California, San Francisco showed that students’ rates of depression when they enter medical school are similar to those of the general population, but students’ depression scores rise over time; approximately one-fourth of first- and second-year students were depressed.³ Fahrenkopf et al⁵ reported that 20% of 123 pediatric residents at 3 U.S. children’s hospitals were depressed. These depressed residents made 6.2 times more medication errors than did their non-depressed peers.⁵ For more information on physicians-in-training, see “Treating depression in medical residents“.

After completing residency, the risk of depression persists. The lifetime prevalence of depression among physicians is 13% in men and 20% in women⁶; these rates are comparable to those of the general population. Firth-Cozens⁷ found a range of factors that predict depression among general practitioners; relationships with senior doctors and patients were the main stressors (Table 1).⁷ Although these stressors increase depression risk, Vaillant et al⁸ showed that they did not increase suicide risk in physicians who did not have underlying psychological difficulties when they entered college. Certain personality traits common among physicians, such as self-criticism and perfectionism, may increase risk for depression and substance abuse.⁸

A depressed physician might enter a downward spiral. Feelings of hopelessness and worthlessness frequently lead to declining professional performance. Professional and personal relationships are strained as internal dysphoria manifests as irritability and anger. Spouses and partners can feel overwhelmed and bewildered by changes in the depressed person’s behavior, which may lead to separation or divorce. Patient care and the physician’s professional standing can be endangered. Signs that suggest a physician may be suffering from depression or another mental illness appear in Table 2.⁹

Table 1

Predictors of depression in physicians

Table 2

Manifestations of mental illness in physicians

Increased suicide risk

A review of 14 studies found that the relative risk of suicide in physicians compared with the general population is between 1.1 and 3.4 for men and 2.5 to 5.7 for women.¹ A retrospective study of English and Welsh doctors showed elevated suicide rates in female but not male physicians compared with the general population.¹⁰ There are no recent studies of suicide rates among U.S. physicians. A 1984-1995 study showed that white male physicians have a higher risk for suicide than other white male professionals.¹¹ A survey of 4,500 women physicians found that female doctors are less likely to attempt suicide than the general female population⁶; however, their attempts more often are lethal, perhaps because they have greater knowledge of toxicology and access to lethal drugs.¹²

The relative rate of suicide among medical specialties is unknown. Studies had indicated higher rates of suicide among psychiatrists and anesthesiologists, but these trials were methodologically flawed.¹²

Silverman¹² developed a profile of the physician at high risk for suicide: a workaholic white male age ≥50 or female age ≥45 who is divorced, single, or currently experiencing marital disruption and is suffering from depression. He or she has a substance abuse problem and a history of risk-taking (high-stakes gambling, etc.). Physicians with chronic pain or illness or with a recent change in occupational or financial status also are at risk. Recent increased work demands, personal losses, diminished autonomy, and access to lethal means (medications, firearms) complete the profile.

Protective factors that lower the risk of completed suicide include effective treatment, social and family support, resilience and coping skills, religious faith, and restricted access to lethal means.^13,14

Barriers to treatment

Physicians often are hesitant to seek mental health treatment.¹⁵ They may fear social stigma and could have trouble finding a local provider who they trust but is not a colleague. Physicians might be concerned about confidentiality and fear recrimination by colleagues, facilities where they work, or licensing boards.¹⁶ Givens and Tjia³ found that only 22% of medical students who screened positive for depression sought help and only 42% of students with suicidal ideation received treatment. These students reported that time constraints, confidentiality concerns, stigma, cost, and fear that their illness will be documented on their academic record were major barriers to seeking mental health care.

Licensing concerns. Physicians may be required to disclose a mental health diagnosis or treatment history when applying for or renewing their medical license. Increasingly, medical boards are asking applicants if they have been treated for bipolar disorder, schizophrenia, paranoia, or other disorders.¹⁷ Credentialing bodies, clinics, and hospitals may make similar queries.

In an analysis of 51 medical licensing applications (50 states and the District of Columbia), Schroeder et al¹⁷ determined that 69% contained at least 1 question that was “likely impermissible” or “impermissible” in terms of compliance with the Americans with Disabilities Act (ADA). In 1993, a U.S. District Court found that the New Jersey State Board of Medical Examiners was in violation of the ADA because licensure application questions did not focus on current fitness to practice medicine but rather on information about a candidate’s status as a person with a disability (illness or diagnosis).¹⁸

In Alexander v Margolis,¹⁹ however, the court found that because patient safety is in question, medical licensing boards and credentialing bodies can solicit information about serious mental illness that could lead to impaired performance. Courts have ruled that questions regarding a history of treatment or hospitalization for bipolar disorder or schizophrenia and other psychotic disorders are permissible because they are considered “serious disorders” likely to interfere with a physician’s current ability to practice.²⁰ In a 2008 review of all U.S. -affiliated medical licensing boards (N=54), Polfliet²¹ found that 7 specifically asked applicants about a history of bipolar disorder or schizophrenia, paranoia, and other psychotic disorders. Polfliet²¹ also found that state medical boards’ compliance with ADA guidelines was not uniform and some questions were “just as broad, and potentially discriminatory, as they were before enactment of the ADA.”

Worley²² reported a successful appeal to the Arkansas State Medical Board to revise its licensure questions following a cluster of medical student and physician suicides. The Board changed the question “Have you ever, or are you presently, being treated for a mental health condition?” to “Have you ever been advised or required by any licensing or privileging body to seek treatment for a physical or mental health condition?”

Providing inaccurate information on a medical licensure application may result in denial or revocation,²³ but acknowledging a history of mental health or substance abuse treatment triggers a more in-depth inquiry by the medical board. The lack of distinction between diagnosis and impairment further stigmatizes physicians who seek care and impedes treatment.

Bipolar disorder. The trend in psychiatry toward diagnosing bipolar II disorder and “soft bipolarity” in patients previously diagnosed with and treated for major depression presents a new challenge. Despite no change in their history or functioning, a physician whose diagnosis is changed from depression to bipolar II disorder might be moved from a non-reportable to a board-reportable diagnostic category. With the evolving understanding of bipolar spectrum disorders, medical boards may need to revise their screening questions to ensure that they are seeking information about impairment, not simply the presence of a medical disorder.

Seeking special treatment

Self-treatment. Physicians may attempt to treat their mood disorder with self-prescribed medications before seeking consultation from a psychiatrist. Others use alcohol or illicit drugs to try to alleviate mood disorder symptoms. Self-diagnosis and treatment are not advisable because it is impossible to be objective. Professional boards and state medical boards discourage or prohibit self-prescribing because of the need for ongoing evaluation and monitoring for adverse reactions.

‘VIP’ treatment. When a physician comes to a colleague for help with a mental health issue, both parties might underestimate the severity of the crisis.²⁴ Weintraub²⁵ reported a case series of 12 “VIP” psychiatric inpatients, 10 of whom he described as “therapeutic failures, “including 2 who committed suicide and 3 who left the hospital against medical advice. He observed that improvement occurred only after patients lost their VIP status/treatment.

In a literature review, Groves et al²⁶ found delays in pursuing diagnostic evaluation and treatment for physician patients. He described risks of VIP treatment (Table 3),²⁶ including the physician’s ability to circumscribe the care regimen to obtain “special treatment, “which can create conflict among care providers and other patients. The ailing physician might have trouble relinquishing control. Care providers might not give physician patients necessary information about the illness or treatment because they make assumptions about the physician’s knowledge or fear causing narcissistic injury. Providers’ identification with their peers, deference to their background, and desire to preserve these patients’ autonomy may lead to interventions that are different from those they would provide to other patients.

Treating physicians might underestimate the patient’s suicide risk and tend to not hospitalize a physician patient who faces an imminent risk of self-harm. Similarly, a physician patient might know what key words to use to deny suicidal ideation or avoid hospitalization. Providers assessing physician patients should provide the same interventions they would give to nonphysician patients with the same history and suicide risk factors. To do otherwise is to risk a fatal outcome.

Physician health programs provide confidential treatment and assistance to physicians with mental illness and/or substance abuse problems. Some programs are affiliated with licensing boards, some are branches of the state medical societies, and others are independent of the licensing agencies. Directories of these programs are available from the Federation of State Physician Health Programs and the Federation of State Medical Boards (see Related Resources). Physician health programs aim to help impaired physicians receive treatment and rehabilitation without censure or licensure revocation, provided they comply with treatment and monitoring requirements.

Table 3

Risks of caring for ‘VIP’ patients

Related Resources

• American Foundation for Suicide Prevention. www.afsp.org. 24-hour crisis line: 1-800-273-TALK (8255).
• Center for Patient and Professional Advocacy. www.mc.vanderbilt.edu/root/vumc.php?site=CPPA.
• Depression and Bipolar Support Alliance. www.dbsalliance.org.
• Federation of State Physician Health Programs, Inc. www.fsphp.org.
• National Alliance on Mental Illness. www.nami.org.
• Vanderbilt Center for Professional Health. www.mc.vanderbilt.edu/cph.
• Vanderbilt Comprehensive Assessment Program. www.mc.vanderbilt.edu/root/vcap.

Disclosure

The authors report 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

Dr. G, a second-year surgical resident, becomes depressed when his girlfriend abruptly ends their relationship. His phone calls and e-mails seeking an explanation go unanswered. Having long struggled with his self-esteem, Dr. G interprets this rejection as confirmation of his self-criticism.

Because of his work schedule, Dr. G feels that there is no way to see a therapist or psychiatrist and believes that asking for time off to do so would adversely affect his evaluations. He feels too embarrassed and “weak” to disclose his breakup and depression to his colleagues and attending physicians and senses that fellow residents would resent having to “carry his load.” Dr. G has spent the past 2 years moonlighting at the local emergency room and thinks it would be humiliating to go there for psychiatric help. His work performance and attendance decline until eventually his residency director forces him to take a medical leave of absence.

Dr. G feels that his pain will never end. He writes goodbye letters to his family, makes arrangements for his possessions and funeral, and hangs himself from the balcony outside his apartment.

Although the rate of depression among physicians is comparable to that of the general population, physicians’ risk of suicide is markedly higher.¹ Depression and other mood disorders may be under-recognized and inadequately treated in physicians because physicians might:

• be reluctant to seek treatment
• attempt to diagnose and treat themselves
• seek and receive “VIP treatment” from other health care providers.

This article examines physicians’ risk for depression and suicide, licensing concerns and other barriers to effective treatment, and what can be done to overcome such obstacles.

Not immune to depression

Rates of depression are higher in medical students and residents (15% to 30%) than in the general population.^2-4 A longitudinal study of medical students at the University of California, San Francisco showed that students’ rates of depression when they enter medical school are similar to those of the general population, but students’ depression scores rise over time; approximately one-fourth of first- and second-year students were depressed.³ Fahrenkopf et al⁵ reported that 20% of 123 pediatric residents at 3 U.S. children’s hospitals were depressed. These depressed residents made 6.2 times more medication errors than did their non-depressed peers.⁵ For more information on physicians-in-training, see “Treating depression in medical residents“.

After completing residency, the risk of depression persists. The lifetime prevalence of depression among physicians is 13% in men and 20% in women⁶; these rates are comparable to those of the general population. Firth-Cozens⁷ found a range of factors that predict depression among general practitioners; relationships with senior doctors and patients were the main stressors (Table 1).⁷ Although these stressors increase depression risk, Vaillant et al⁸ showed that they did not increase suicide risk in physicians who did not have underlying psychological difficulties when they entered college. Certain personality traits common among physicians, such as self-criticism and perfectionism, may increase risk for depression and substance abuse.⁸

A depressed physician might enter a downward spiral. Feelings of hopelessness and worthlessness frequently lead to declining professional performance. Professional and personal relationships are strained as internal dysphoria manifests as irritability and anger. Spouses and partners can feel overwhelmed and bewildered by changes in the depressed person’s behavior, which may lead to separation or divorce. Patient care and the physician’s professional standing can be endangered. Signs that suggest a physician may be suffering from depression or another mental illness appear in Table 2.⁹

Table 1

Predictors of depression in physicians

Table 2

Manifestations of mental illness in physicians

Increased suicide risk

A review of 14 studies found that the relative risk of suicide in physicians compared with the general population is between 1.1 and 3.4 for men and 2.5 to 5.7 for women.¹ A retrospective study of English and Welsh doctors showed elevated suicide rates in female but not male physicians compared with the general population.¹⁰ There are no recent studies of suicide rates among U.S. physicians. A 1984-1995 study showed that white male physicians have a higher risk for suicide than other white male professionals.¹¹ A survey of 4,500 women physicians found that female doctors are less likely to attempt suicide than the general female population⁶; however, their attempts more often are lethal, perhaps because they have greater knowledge of toxicology and access to lethal drugs.¹²

The relative rate of suicide among medical specialties is unknown. Studies had indicated higher rates of suicide among psychiatrists and anesthesiologists, but these trials were methodologically flawed.¹²

Silverman¹² developed a profile of the physician at high risk for suicide: a workaholic white male age ≥50 or female age ≥45 who is divorced, single, or currently experiencing marital disruption and is suffering from depression. He or she has a substance abuse problem and a history of risk-taking (high-stakes gambling, etc.). Physicians with chronic pain or illness or with a recent change in occupational or financial status also are at risk. Recent increased work demands, personal losses, diminished autonomy, and access to lethal means (medications, firearms) complete the profile.

Protective factors that lower the risk of completed suicide include effective treatment, social and family support, resilience and coping skills, religious faith, and restricted access to lethal means.^13,14

Barriers to treatment

Physicians often are hesitant to seek mental health treatment.¹⁵ They may fear social stigma and could have trouble finding a local provider who they trust but is not a colleague. Physicians might be concerned about confidentiality and fear recrimination by colleagues, facilities where they work, or licensing boards.¹⁶ Givens and Tjia³ found that only 22% of medical students who screened positive for depression sought help and only 42% of students with suicidal ideation received treatment. These students reported that time constraints, confidentiality concerns, stigma, cost, and fear that their illness will be documented on their academic record were major barriers to seeking mental health care.

Licensing concerns. Physicians may be required to disclose a mental health diagnosis or treatment history when applying for or renewing their medical license. Increasingly, medical boards are asking applicants if they have been treated for bipolar disorder, schizophrenia, paranoia, or other disorders.¹⁷ Credentialing bodies, clinics, and hospitals may make similar queries.

In an analysis of 51 medical licensing applications (50 states and the District of Columbia), Schroeder et al¹⁷ determined that 69% contained at least 1 question that was “likely impermissible” or “impermissible” in terms of compliance with the Americans with Disabilities Act (ADA). In 1993, a U.S. District Court found that the New Jersey State Board of Medical Examiners was in violation of the ADA because licensure application questions did not focus on current fitness to practice medicine but rather on information about a candidate’s status as a person with a disability (illness or diagnosis).¹⁸

In Alexander v Margolis,¹⁹ however, the court found that because patient safety is in question, medical licensing boards and credentialing bodies can solicit information about serious mental illness that could lead to impaired performance. Courts have ruled that questions regarding a history of treatment or hospitalization for bipolar disorder or schizophrenia and other psychotic disorders are permissible because they are considered “serious disorders” likely to interfere with a physician’s current ability to practice.²⁰ In a 2008 review of all U.S. -affiliated medical licensing boards (N=54), Polfliet²¹ found that 7 specifically asked applicants about a history of bipolar disorder or schizophrenia, paranoia, and other psychotic disorders. Polfliet²¹ also found that state medical boards’ compliance with ADA guidelines was not uniform and some questions were “just as broad, and potentially discriminatory, as they were before enactment of the ADA.”

Worley²² reported a successful appeal to the Arkansas State Medical Board to revise its licensure questions following a cluster of medical student and physician suicides. The Board changed the question “Have you ever, or are you presently, being treated for a mental health condition?” to “Have you ever been advised or required by any licensing or privileging body to seek treatment for a physical or mental health condition?”

Providing inaccurate information on a medical licensure application may result in denial or revocation,²³ but acknowledging a history of mental health or substance abuse treatment triggers a more in-depth inquiry by the medical board. The lack of distinction between diagnosis and impairment further stigmatizes physicians who seek care and impedes treatment.

Bipolar disorder. The trend in psychiatry toward diagnosing bipolar II disorder and “soft bipolarity” in patients previously diagnosed with and treated for major depression presents a new challenge. Despite no change in their history or functioning, a physician whose diagnosis is changed from depression to bipolar II disorder might be moved from a non-reportable to a board-reportable diagnostic category. With the evolving understanding of bipolar spectrum disorders, medical boards may need to revise their screening questions to ensure that they are seeking information about impairment, not simply the presence of a medical disorder.

Seeking special treatment

Self-treatment. Physicians may attempt to treat their mood disorder with self-prescribed medications before seeking consultation from a psychiatrist. Others use alcohol or illicit drugs to try to alleviate mood disorder symptoms. Self-diagnosis and treatment are not advisable because it is impossible to be objective. Professional boards and state medical boards discourage or prohibit self-prescribing because of the need for ongoing evaluation and monitoring for adverse reactions.

‘VIP’ treatment. When a physician comes to a colleague for help with a mental health issue, both parties might underestimate the severity of the crisis.²⁴ Weintraub²⁵ reported a case series of 12 “VIP” psychiatric inpatients, 10 of whom he described as “therapeutic failures, “including 2 who committed suicide and 3 who left the hospital against medical advice. He observed that improvement occurred only after patients lost their VIP status/treatment.

In a literature review, Groves et al²⁶ found delays in pursuing diagnostic evaluation and treatment for physician patients. He described risks of VIP treatment (Table 3),²⁶ including the physician’s ability to circumscribe the care regimen to obtain “special treatment, “which can create conflict among care providers and other patients. The ailing physician might have trouble relinquishing control. Care providers might not give physician patients necessary information about the illness or treatment because they make assumptions about the physician’s knowledge or fear causing narcissistic injury. Providers’ identification with their peers, deference to their background, and desire to preserve these patients’ autonomy may lead to interventions that are different from those they would provide to other patients.

Treating physicians might underestimate the patient’s suicide risk and tend to not hospitalize a physician patient who faces an imminent risk of self-harm. Similarly, a physician patient might know what key words to use to deny suicidal ideation or avoid hospitalization. Providers assessing physician patients should provide the same interventions they would give to nonphysician patients with the same history and suicide risk factors. To do otherwise is to risk a fatal outcome.

Physician health programs provide confidential treatment and assistance to physicians with mental illness and/or substance abuse problems. Some programs are affiliated with licensing boards, some are branches of the state medical societies, and others are independent of the licensing agencies. Directories of these programs are available from the Federation of State Physician Health Programs and the Federation of State Medical Boards (see Related Resources). Physician health programs aim to help impaired physicians receive treatment and rehabilitation without censure or licensure revocation, provided they comply with treatment and monitoring requirements.

Table 3

Risks of caring for ‘VIP’ patients

Related Resources

• American Foundation for Suicide Prevention. www.afsp.org. 24-hour crisis line: 1-800-273-TALK (8255).
• Center for Patient and Professional Advocacy. www.mc.vanderbilt.edu/root/vumc.php?site=CPPA.
• Depression and Bipolar Support Alliance. www.dbsalliance.org.
• Federation of State Physician Health Programs, Inc. www.fsphp.org.
• National Alliance on Mental Illness. www.nami.org.
• Vanderbilt Center for Professional Health. www.mc.vanderbilt.edu/cph.
• Vanderbilt Comprehensive Assessment Program. www.mc.vanderbilt.edu/root/vcap.

Disclosure

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